Ts

STATA TIME-SERIES REFERENCE MANUAL
RELEASE 12
A Stata Press Publication StataCorp LP College Station, Texas
Copyright c 19852011 StataCorp LP All rights reserved Version 12
Published by Stata Press, 4905 Lakeway Drive, College Station, Texas 77845 Typeset in TEX Printed in the United States of America 10 9 8 7 6 5 4 3 2 1 ISBN-10: 1-59718-094-7 ISBN-13: 978-1-59718-094-8 This manual is protected by copyright. All rights are reserved. No part of this manual may be reproduced, stored in a retrieval system, or transcribed, in any form or by any meanselectronic, mechanical, photocopy, recording, or otherwisewithout the prior written permission of StataCorp LP unless permitted subject to the terms and conditions of a license granted to you by StataCorp LP to use the software and documentation. No license, express or implied, by estoppel or otherwise, to any intellectual property rights is granted by this document. StataCorp provides this manual as is without warranty of any kind, either expressed or implied, including, but not limited to, the implied warranties of merchantability and tness for a particular purpose. StataCorp may make improvements and/or changes in the product(s) and the program(s) described in this manual at any time and without notice. The software described in this manual is furnished under a license agreement or nondisclosure agreement. The software may be copied only in accordance with the terms of the agreement. It is against the law to copy the software onto DVD, CD, disk, diskette, tape, or any other medium for any purpose other than backup or archival purposes. The automobile dataset appearing on the accompanying media is Copyright c 1979 by Consumers Union of U.S., Inc., Yonkers, NY 10703-1057 and is reproduced by permission from CONSUMER REPORTS, April 1979. Stata, , Stata Press, Mata, , and NetCourse are registered trademarks of StataCorp LP.
Stata and Stata Press are registered trademarks with the World Intellectual Property Organization of the United Nations. NetCourseNow is a trademark of StataCorp LP. Other brand and product names are registered trademarks or trademarks of their respective companies. For copyright information about the software, type help copyright within Stata.
The suggested citation for this software is StataCorp. 2011. Stata: Release 12 . Statistical Software. College Station, TX: StataCorp LP.
Table of contents
intro . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Introduction to time-series manual time series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Introduction to time-series commands arch . . . . . . . . . Autoregressive conditional heteroskedasticity (ARCH) family of estimators arch postestimation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Postestimation tools for arch arma . . . . . . . . . . . . . . . . . . . Autoregressive fractionally integrated moving-average models arma postestimation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Postestimation tools for arma arima . . . . . . . . . . . . . . . . . . . . . . . ARIMA, ARMAX, and other dynamic regression models arima postestimation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Postestimation tools for arima 1 3 11 44 49 67 73 96
corrgram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Tabulate and graph autocorrelations 102 cumsp . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cumulative spectral distribution 109 dfactor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Dynamic-factor models dfactor postestimation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Postestimation tools for dfactor dfgls . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . DF-GLS unit-root test dfuller . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Augmented DickeyFuller unit-root test 112 129 134 140
fcast compute . . . Compute dynamic forecasts of dependent variables after var, svar, or vec 145 fcast graph . . . . . . . . . . . . . . . . . . . Graph forecasts of variables computed by fcast compute 153 haver . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Load data from Haver Analytics database 156 irf irf irf irf irf irf irf irf irf irf irf irf . . . . . . . . . . . . . . . . . Create and analyze IRFs, dynamic-multiplier functions, and FEVDs add . . . . . . . . . . . . . . . . . . . . . . . . . . . Add results from an IRF le to the active IRF le cgraph . . . . . . . . . . . Combine graphs of IRFs, dynamic-multiplier functions, and FEVDs create . . . . . . . . . . . . . . . . . . . . . Obtain IRFs, dynamic-multiplier functions, and FEVDs ctable . . . . . . . . . . . . Combine tables of IRFs, dynamic-multiplier functions, and FEVDs describe . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Describe an IRF le drop . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Drop IRF results from the active IRF le graph . . . . . . . . . . . . . . . . . . . . . . . Graph IRFs, dynamic-multiplier functions, and FEVDs ograph . . . . . . . . . . . . . . . Graph overlaid IRFs, dynamic-multiplier functions, and FEVDs rename . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Rename an IRF result in an IRF le set . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Set the active IRF le table . . . . . . . . . . . . . . . Create tables of IRFs, dynamic-multiplier functions, and FEVDs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Multivariate GARCH models ccc . . . . . . . . . . . . . . Constant conditional correlation multivariate GARCH models ccc postestimation . . . . . . . . . . . . . . . . . . . . . . . . Postestimation tools for mgarch ccc dcc . . . . . . . . . . . . . . Dynamic conditional correlation multivariate GARCH models dcc postestimation . . . . . . . . . . . . . . . . . . . . . . . Postestimation tools for mgarch dcc dvech . . . . . . . . . . . . . . . . . . . . . . . . . . . Diagonal vech multivariate GARCH models dvech postestimation . . . . . . . . . . . . . . . . . . . Postestimation tools for mgarch dvech vcc . . . . . . . . . . . . . . . Varying conditional correlation multivariate GARCH models vcc postestimation . . . . . . . . . . . . . . . . . . . . . . . Postestimation tools for mgarch vcc 161 165 167 172 194 199 202 204 210 216 218 220 225 231 244 247 261 264 276 282 296
mgarch mgarch mgarch mgarch mgarch mgarch mgarch mgarch mgarch
newey . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Regression with NeweyWest standard errors 299 newey postestimation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Postestimation tools for newey 304 pergram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Periodogram pperron . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . PhillipsPerron unit-root test prais . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Prais Winsten and Cochrane Orcutt regression prais postestimation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Postestimation tools for prais 306 314 319 330
ii
psdensity . . . . . . . . . . . . Parametric spectral density estimation after arima, arma, and ucm 332 rolling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Rolling-window and recursive estimation 342 sspace . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . State-space models 350 sspace postestimation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Postestimation tools for sspace 374 tsappend . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Add observations to a time-series dataset tsll . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Fill in gaps in time variable tslter . . . . . . . . . . . . . . . . . . . . . . . . . Filter a time-series, keeping only selected periodicities tslter bk . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . BaxterKing time-series lter tslter bw . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Butterworth time-series lter tslter cf . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ChristianoFitzgerald time-series lter tslter hp . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . HodrickPrescott time-series lter tsline . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Plot time-series data tsreport . . . . . . . . . . . . . . . . . . . Report time-series aspects of a dataset or estimation sample tsrevar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Time-series operator programming command tsset . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Declare data to be time-series data tssmooth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Smooth and forecast univariate time-series data tssmooth dexponential . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Double-exponential smoothing tssmooth exponential . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Single-exponential smoothing tssmooth hwinters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . HoltWinters nonseasonal smoothing tssmooth ma . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Moving-average lter tssmooth nl . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Nonlinear lter tssmooth shwinters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . HoltWinters seasonal smoothing 381 387 391 410 418 427 435 442 447 450 453 469 471 477 485 492 497 499
ucm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Unobserved-components model 508 ucm postestimation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Postestimation tools for ucm 536 var intro . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Introduction to vector autoregressive models var . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Vector autoregressive models var postestimation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Postestimation tools for var var svar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Structural vector autoregressive models var svar postestimation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Postestimation tools for svar varbasic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Fit a simple VAR and graph IRFs or FEVDs varbasic postestimation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Postestimation tools for varbasic vargranger . . . . . . . . . . . . . . . . . . . Perform pairwise Granger causality tests after var or svar varlmar . . . . . . . . . . . . . . . . . . Perform LM test for residual autocorrelation after var or svar varnorm . . . . . . . . . . . . . . . . . . . . Test for normally distributed disturbances after var or svar varsoc . . . . . . . . . . . . . . . . . . . . . . Obtain lag-order selection statistics for VARs and VECMs varstable . . . . . . . . . . . . . . . . . . . . . Check the stability condition of VAR or SVAR estimates varwle . . . . . . . . . . . . . . . . . . . . . . . . . . Obtain Wald lag-exclusion statistics after var or svar vec intro . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Introduction to vector error-correction models vec . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Vector error-correction models vec postestimation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Postestimation tools for vec veclmar . . . . . . . . . . . . . . . . . . . . . . . Perform LM test for residual autocorrelation after vec vecnorm . . . . . . . . . . . . . . . . . . . . . . . . . . Test for normally distributed disturbances after vec vecrank . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Estimate the cointegrating rank of a VECM vecstable . . . . . . . . . . . . . . . . . . . . . . . . . . Check the stability condition of VECM estimates 542 549 561 564 584 587 592 595 599 602 607 613 618 623 642 666 669 672 676 684
wntestb . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Bartletts periodogram-based test for white noise 688 wntestq . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Portmanteau (Q) test for white noise 693 xcorr . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cross-correlogram for bivariate time series 696
iii
Glossary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Subject and author index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
700 709
Cross-referencing the documentation

When reading this manual, you will nd references to other Stata manuals. For example, [U] 26 Overview of Stata estimation commands [R] regress [D] reshape The rst example is a reference to chapter 26, Overview of Stata estimation commands, in the Users Guide; the second is a reference to the regress entry in the Base Reference Manual; and the third is a reference to the reshape entry in the Data-Management Reference Manual. All the manuals in the Stata Documentation have a shorthand notation: [GSM] [GSU] [GSW] [U ] [R] [D ] [G ] [XT] [MI] [MV] [P ] [SEM] [SVY] [ST] [TS] [I] [M ] Getting Started with Stata for Mac Getting Started with Stata for Unix Getting Started with Stata for Windows Stata Users Guide Stata Base Reference Manual Stata Data-Management Reference Manual Stata Graphics Reference Manual Stata Longitudinal-Data/Panel-Data Reference Manual Stata Multiple-Imputation Reference Manual Stata Multivariate Statistics Reference Manual Stata Programming Reference Manual Stata Structural Equation Modeling Reference Manual Stata Survey Data Reference Manual Stata Survival Analysis and Epidemiological Tables Reference Manual Stata Time-Series Reference Manual Stata Quick Reference and Index Mata Reference Manual
Detailed information about each of these manuals may be found online at http://www.stata-press.com/manuals/
Title
intro Introduction to time-series manual
Description
This entry describes this manual and what has changed since Stata 11.
Remarks
This manual documents Statas time-series commands and is referred to as [TS] in cross-references. After this entry, [TS] time series provides an overview of the ts commands. The other parts of this manual are arranged alphabetically. If you are new to Statas time-series features, we recommend that you read the following sections rst: [TS] time series [TS] tsset Introduction to time-series commands Declare a dataset to be time-series data
Stata is continually being updated, and Stata users are always writing new commands. To ensure that you have the latest features, you should install the most recent ofcial update; see [R] update.
Whats new
1. MGARCH, which is to say, multivariate GARCH, which is to say, estimation of multivariate generalized autoregressive conditional heteroskedasticity models of volatility, and this includes constant, dynamic, and varying conditional correlations, also known as the CCC, DCC, and VCC models. Innovations in these models may follow multivariate normal or Students t distributions. See [TS] mgarch. 2. UCM, which is to say, unobserved-components models, also known as structural time-series models that decompose a series into trend, seasonal, and cyclical components, and which were popularized by Harvey (1989). See [TS] ucm. 3. ARFIMA, which is to say, autoregressive fractionally integrated moving-average models, useful for long-memory processes. See [TS] arma. 4. Filters for extracting business and seasonal cycles. Four popular time-series lters are provided: the BaxterKing and the ChristianoFitzgerald band-pass lters, and the Butterworth and the HodrickPrescott high-pass lters. See [TS] tslter. 5. Business dates allow you to dene your own calendars so that they display correctly and lags and leads work as they should. You could create le lse.stbcal that recorded the days the London Stock Exchange is open (or closed) and then Stata would understand format %tblse just as it understands the usual date format %td. Once you dene a calendar, Stata deeply understands it. You can, for instance, easily convert between %tblse and %td values. See [D] datetime business calendars. 6. Improved documentation for date and time variables. Anyone who has ever been puzzled by Statas date and time variables, which is to say, anyone who uses them, should see [D] datetime, [D] datetime translation, and [D] datetime display formats.
1
intro Introduction to time-series manual
7. Contrasts, which is to say, tests of linear hypotheses involving factor variables and their interactions from the most recently t model. Tests include ANOVA-style tests of main effects, simple effects, interactions, and nested effects. Effects can be decomposed into comparisons with reference categories, comparisons of adjacent levels, comparisons with the grand mean, and more. New commands contrast and margins, contrast are available after many time-series estimation commands. See [R] contrast and [R] margins, contrast. 8. Pairwise comparisons available after many time-series estimation commands. See [R] pwcompare and [R] margins, pwcompare. 9. Graphs of margins, marginal effects, contrasts, and pairwise comparisons available after most time-series estimation commands. See [R] marginsplot. 10. Estimation output improved. a. Implied zero coefcients now shown. When a coefcient is omitted, it is now shown as being zero and the reason it was omittedcollinearity, base, emptyis shown in the standard-error column. (The word omitted is shown if the coefcient was omitted because of collinearity.) b. You can set displayed precision for all values in coefcient tables using set cformat, set pformat, and set sformat. Or you may use options cformat(), pformat(), and sformat() now allowed on all estimation commands. See [R] set cformat and [R] estimation options. c. Estimation commands now respect the width of the Results window. This feature may be turned off by new display option nolstretch. See [R] estimation options. d. You can now set whether base levels, empty cells, and omitted are shown using set showbaselevels, set showemptycells, and set showomitted. See [R] set showbaselevels. 11. Spectral densities from parametric models via new postestimation command psdensity lets you estimate using arfima, arima, and ucm and then obtain the implied spectral density. See [TS] psdensity. 12. dvech renamed mgarch dvech. The command for tting the diagonal VECH model is now named mgarch dvech, and innovations may follow multivariate normal or Students t distributions. See [TS] mgarch. 13. Loading data from Haver Analytics supported on all 64-bit Windows. See [TS] haver. 14. Option addplot() now places added graphs above or below. Graph commands that allow option addplot() can now place the added plots above or below the commands plots. Affected by this are the commands corrgram, cumsp, pergram, varstable, vecstable, wntestb, and xcorr. For a complete list of all the new features in Stata 12, see [U] 1.3 Whats new.
Reference
Harvey, A. C. 1989. Forecasting, Structural Time Series Models and the Kalman Filter. Cambridge: Cambridge University Press.
Also see
[U] 1.3 Whats new [R] intro Introduction to base reference manual
Title
time series Introduction to time-series commands
Description
The Time-Series Reference Manual organizes the commands alphabetically, making it easy to nd individual command entries if you know the name of the command. This overview organizes and presents the commands conceptually, that is, according to the similarities in the functions that they perform. The table below lists the manual entries that you should see for additional information.
Data-management tools and time-series operators. These commands help you prepare your data for further analysis. Univariate time series. These commands are grouped together because they are either estimators or lters designed for univariate time series or preestimation or postestimation commands that are conceptually related to one or more univariate time-series estimators. Multivariate time series. These commands are similarly grouped together because they are either estimators designed for use with multivariate time series or preestimation or postestimation commands conceptually related to one or more multivariate time-series estimators.
Within these three broad categories, similar commands have been grouped together.
Data-management tools and time-series operators [TS] tsset Declare data to be time-series data [TS] tsll Fill in gaps in time variable [TS] tsappend Add observations to a time-series dataset [TS] tsreport Report time-series aspects of a dataset or estimation sample [TS] tsrevar Time-series operator programming command [TS] haver Load data from Haver Analytics database [TS] rolling Rolling-window and recursive estimation [D] datetime business calendars User-denable business calendars
Univariate time series Estimators [TS] arma
[TS] [TS] [TS] [TS] [TS] [TS] [TS] [TS] [TS] [TS] [TS]
arma postestimation arima arima postestimation arch arch postestimation newey newey postestimation prais prais postestimation ucm ucm postestimation
Autoregressive fractionally integrated moving-average models Postestimation tools for arma ARIMA, ARMAX, and other dynamic regression models Postestimation tools for arima Autoregressive conditional heteroskedasticity (ARCH) family of estimators Postestimation tools for arch Regression with NeweyWest standard errors Postestimation tools for newey PraisWinsten and CochraneOrcutt regression Postestimation tools for prais Unobserved-components model Postestimation tools for ucm
Time-series smoothers and lters [TS] tslter bk [TS] tslter bw [TS] tslter cf [TS] tslter hp [TS] tssmooth ma [TS] tssmooth dexponential [TS] tssmooth exponential [TS] tssmooth hwinters [TS] tssmooth shwinters [TS] tssmooth nl
BaxterKing time-series lter Butterworth time-series lter ChristianoFitzgerald time-series lter HodrickPrescott time-series lter Moving-average lter Double-exponential smoothing Single-exponential smoothing HoltWinters nonseasonal smoothing HoltWinters seasonal smoothing Nonlinear lter
Diagnostic tools [TS] corrgram [TS] xcorr [TS] cumsp [TS] pergram [TS] psdensity [TS] dfgls [TS] dfuller [TS] pperron [R] regress postestimation time series [TS] wntestb [TS] wntestq
Tabulate and graph autocorrelations Cross-correlogram for bivariate time series Cumulative spectral distribution Periodogram Parametric spectral density estimation DF-GLS unit-root test Augmented DickeyFuller unit-root test PhillipsPerron unit-root test Postestimation tools for regress with time series Bartletts periodogram-based test for white noise Portmanteau (Q) test for white noise
Multivariate time series Estimators [TS] dfactor [TS] dfactor postestimation [TS] mgarch ccc [TS] mgarch ccc postestimation [TS] mgarch dcc [TS] mgarch dcc postestimation [TS] mgarch dvech [TS] mgarch dvech postestimation [TS] mgarch vcc [TS] mgarch vcc postestimation [TS] sspace [TS] sspace postestimation [TS] var [TS] var postestimation [TS] var svar [TS] var svar postestimation [TS] varbasic [TS] varbasic postestimation [TS] vec [TS] vec postestimation Diagnostic tools [TS] varlmar [TS] varnorm [TS] varsoc [TS] varstable [TS] varwle [TS] veclmar [TS] vecnorm [TS] vecrank [TS] vecstable
Dynamic-factor models Postestimation tools for dfactor Constant conditional correlation multivariate GARCH models Postestimation tools for mgarch ccc Dynamic conditional correlation multivariate GARCH models Postestimation tools for mgarch dcc Diagonal vech multivariate GARCH models Postestimation tools for mgarch dvech Varying conditional correlation multivariate GARCH models Postestimation tools for mgarch vcc State-space models Postestimation tools for sspace Vector autoregressive models Postestimation tools for var Structural vector autoregressive models Postestimation tools for svar Fit a simple VAR and graph IRFs or FEVDs Postestimation tools for varbasic Vector error-correction models Postestimation tools for vec
Perform LM test for residual autocorrelation Test for normally distributed disturbances Obtain lag-order selection statistics for VARs and VECMs Check the stability condition of VAR or SVAR estimates Obtain Wald lag-exclusion statistics Perform LM test for residual autocorrelation Test for normally distributed disturbances Estimate the cointegrating rank of a VECM Check the stability condition of VECM estimates
Forecasting, inference, and interpretation [TS] irf create Obtain IRFs, dynamic-multiplier functions, and FEVDs [TS] fcast compute Compute dynamic forecasts of dependent variables [TS] vargranger Perform pairwise Granger causality tests
Graphs and tables [TS] corrgram [TS] xcorr [TS] pergram [TS] irf graph [TS] irf cgraph [TS] irf ograph [TS] irf table [TS] irf ctable [TS] fcast graph [TS] tsline [TS] varstable [TS] vecstable [TS] wntestb
Tabulate and graph autocorrelations Cross-correlogram for bivariate time series Periodogram Graph IRFs, dynamic-multiplier functions, and FEVDs Combine graphs of IRFs, dynamic-multiplier functions, and FEVDs Graph overlaid IRFs, dynamic-multiplier functions, and FEVDs Create tables of IRFs, dynamic-multiplier functions, and FEVDs Combine tables of IRFs, dynamic-multiplier functions, and FEVDs Graph forecasts of variables computed by fcast compute Plot time-series data Check the stability condition of VAR or SVAR estimates Check the stability condition of VECM estimates Bartletts periodogram-based test for white noise
Results management tools [TS] irf add Add results from an IRF le to the active IRF le [TS] irf describe Describe an IRF le [TS] irf drop Drop IRF results from the active IRF le [TS] irf rename Rename an IRF result in an IRF le [TS] irf set Set the active IRF le
Remarks
Remarks are presented under the following headings:
Data-management tools and time-series operators Univariate time series Estimators Time-series smoothers and lters Diagnostic tools Multivariate time series Estimators Diagnostic tools
We also offer a NetCourse on Statas time-series capabilities; see http://www.stata.com/netcourse/nc461.html.
Data-management tools and time-series operators

Because time-series estimators are, by denition, a function of the temporal ordering of the observations in the estimation sample, Statas time-series commands require the data to be sorted and indexed by time, using the tsset command, before they can be used. tsset is simply a way for you to tell Stata which variable in your dataset represents time; tsset then sorts and indexes the data appropriately for use with the time-series commands. Once your dataset has been tsset, you can use Statas time-series operators in data manipulation or programming using that dataset and when specifying the syntax for most time-series commands. Stata has time-series operators for representing the lags, leads, differences, and seasonal differences of a variable. The time-series operators are documented in [TS] tsset.
You can also dene a business-day calendar so that Statas time-series operators respect the structure of missing observations in your data. The most common example is having Monday come after Friday in market data. [D] datetime business calendars provides a discussion and examples. tsset can also be used to declare that your dataset contains cross-sectional time-series data, often referred to as panel data. When you use tsset to declare your dataset to contain panel data, you specify a variable that identies the panels and a variable that identies the time periods. Once your dataset has been tsset as panel data, the time-series operators work appropriately for the data. tsfill, which is documented in [TS] tsll, can be used after tsset to ll in missing times with missing observations. tsset will report any gaps in your data, and tsreport will provide more details about the gaps. tsappend adds observations to a time-series dataset by using the information set by tsset. This function can be particularly useful when you wish to predict out of sample after tting a model with a time-series estimator. tsrevar is a programmers command that provides a way to use varlists that contain time-series operators with commands that do not otherwise support time-series operators. The haver commands documented in [TS] haver allow you to load and describe the contents of a Haver Analytics (http://www.haver.com) le. rolling performs rolling regressions, recursive regressions, and reverse recursive regressions. Any command that saves results in e() or r() can be used with rolling.
Univariate time series

Estimators The six univariate time-series estimators currently available in Stata are arfima, arima, arch, newey, prais, and ucm. newey and prais are really just extensions to ordinary linear regression. When you t a linear regression on time-series data via ordinary least squares (OLS), if the disturbances are autocorrelated, the parameter estimates are usually consistent, but the estimated standard errors tend to be underestimated. Several estimators have been developed to deal with this problem. One strategy is to use OLS for estimating the regression parameters and use a different estimator for the variances, one that is consistent in the presence of autocorrelated disturbances, such as the NeweyWest estimator implemented in newey. Another strategy is to model the dynamics of the disturbances. The estimators found in prais, arima, arch, arfima, and ucm are based on such a strategy. prais implements two such estimators: the PraisWinsten and the CochraneOrcutt generalized least-squares (GLS) estimators. These estimators are GLS estimators, but they are fairly restrictive in that they permit only rst-order autocorrelation in the disturbances. Although they have certain pedagogical and historical value, they are somewhat obsolete. Faster computers with more memory have made it possible to implement full information maximum likelihood (FIML) estimators, such as Statas arima command. These estimators permit much greater exibility when modeling the disturbances and are more efcient estimators. arima provides the means to t linear models with autoregressive moving-average (ARMA) disturbances, or in the absence of linear predictors, autoregressive integrated moving-average (ARIMA) models. This means that, whether you think that your data are best represented as a distributed-lag model, a transfer-function model, or a stochastic difference equation, or you simply wish to apply a BoxJenkins lter to your data, the model can be t using arima. arch, a conditional maximum likelihood estimator, has similar modeling capabilities for the mean of the time series but can also model autoregressive conditional heteroskedasticity in the disturbances with a wide variety of specications for the variance equation.
arfima estimates the parameters of autoregressive fractionally integrated moving-average (ARFIMA) models, which handle higher degrees of dependence than ARIMA models. ARFIMA models allow the autocorrelations to decay at the slower hyperbolic rate, whereas ARIMA models handle processes whose autocorrelations decay at an exponential rate. Unobserved-components models (UCMs) decompose a time series into trend, seasonal, cyclical, and idiosyncratic components and allow for exogenous variables. ucm estimates the parameters of UCMs by maximum likelihood. UCMs can also model the stationary cyclical component using the stochastic-cycle parameterization that has an intuitive frequency-domain interpretation. Time-series smoothers and lters In addition to the estimators mentioned above, Stata also provides time-series lters and smoothers. The BaxterKing and ChristianoFitzgerald band-pass lters and the Butterworth and HodrickPrescott high-pass lters are implemented in tsfilter; see [TS] tslter for an overview. Also included are a simple, uniformly weighted, moving-average lter with unit weights; a weighted moving-average lter in which you can specify the weights; single- and double-exponential smoothers; HoltWinters seasonal and nonseasonal smoothers; and a nonlinear smoother. Most of these smoothers were originally developed as ad hoc procedures and are used for reducing the noise in a time series (smoothing) or forecasting. Although they have limited application for signal extraction, these smoothers have all been found to be optimal for some underlying modern time-series models; see [TS] tssmooth. Diagnostic tools Statas time-series commands also include several preestimation and postestimation diagnostic and interpretation commands. corrgram estimates the autocorrelation function and partial autocorrelation function of a univariate time series, as well as Q statistics. These functions and statistics are often used to determine the appropriate model specication before tting ARIMA models. corrgram can also be used with wntestb and wntestq to examine the residuals after tting a model for evidence of model misspecication. Statas time-series commands also include the commands pergram and cumsp, which provide the log-standardized periodogram and the cumulative-sample spectral distribution, respectively, for time-series analysts who prefer to estimate in the frequency domain rather than the time domain. psdensity computes the spectral density implied by the parameters estimated by arfima, arima, or ucm. The estimated spectral density shows the relative importance of components at different frequencies. xcorr estimates the cross-correlogram for bivariate time series and can similarly be used both for preestimation and postestimation. For example, the cross-correlogram can be used before tting a transfer-function model to produce initial estimates of the IRF. This estimate can then be used to determine the optimal lag length of the input series to include in the model specication. It can also be used as a postestimation tool after tting a transfer function. The cross-correlogram between the residual from a transfer-function model and the prewhitened input series of the model can be examined for evidence of model misspecication. When you t ARMA or ARIMA models, the dependent variable being modeled must be covariance stationary (ARMA models), or the order of integration must be known (ARIMA models). Stata has three commands that can test for the presence of a unit root in a time-series variable: dfuller performs the augmented DickeyFuller test, pperron performs the PhillipsPerron test, and dfgls performs a modied DickeyFuller test. arfima can also be used to investigate the order of integration.
The remaining diagnostic tools for univariate time series are for use after tting a linear model via
OLS with Statas regress command. They are documented collectively in [R] regress postestimation
time series. They include estat dwatson, estat durbinalt, estat bgodfrey, and estat archlm. estat dwatson computes the DurbinWatson d statistic to test for the presence of rstorder autocorrelation in the OLS residuals. estat durbinalt likewise tests for the presence of autocorrelation in the residuals. By comparison, however, Durbins alternative test is more general and easier to use than the DurbinWatson test. With estat durbinalt, you can test for higher orders of autocorrelation, the assumption that the covariates in the model are strictly exogenous is relaxed, and there is no need to consult tables to compute rejection regions, as you must with the DurbinWatson test. estat bgodfrey computes the BreuschGodfrey test for autocorrelation in the residuals, and although the computations are different, the test in estat bgodfrey is asymptotically equivalent to the test in estat durbinalt. Finally, estat archlm performs Engles LM test for the presence of autoregressive conditional heteroskedasticity.
Multivariate time series

Estimators Stata provides commands for tting the most widely applied multivariate time-series models. var and svar t vector autoregressive and structural vector autoregressive models to stationary data. vec ts cointegrating vector error-correction models. dfactor ts dynamic-factor models. mgarch ccc, mgarch dcc, mgarch dvech, and mgarch vcc t multivariate GARCH models. sspace ts state-space models. Many linear time-series models, including vector autoregressive moving-average (VARMA) models and structural time-series models, can be cast as state-space models and t by sspace. Diagnostic tools Before tting a multivariate time-series model, you must specify the number of lags of the dependent variable to include. varsoc produces statistics for determining the order of a VAR or VECM. Several postestimation commands perform the most common specication analysis on a previously tted VAR or SVAR. You can use varlmar to check for serial correlation in the residuals, varnorm to test the null hypothesis that the disturbances come from a multivariate normal distribution, and varstable to see if the tted VAR or SVAR is stable. Two common types of inference about VAR models are whether one variable Granger-causes another and whether a set of lags can be excluded from the model. vargranger reports Wald tests of Granger causation, and varwle reports Wald lag exclusion tests. Similarly, several postestimation commands perform the most common specication analysis on a previously tted VECM. You can use veclmar to check for serial correlation in the residuals, vecnorm to test the null hypothesis that the disturbances come from a multivariate normal distribution, and vecstable to analyze the stability of the previously tted VECM.
VARs and VECMs are often t to produce baseline forecasts. fcast produces dynamic forecasts from previously tted VARs and VECMs.
Many researchers t VARs, SVARs, and VECMs because they want to analyze how unexpected shocks affect the dynamic paths of the variables. Stata has a suite of irf commands for estimating IRF functions and interpreting, presenting, and managing these estimates; see [TS] irf.
10
References
Baum, C. F. 2005. Stata: The language of choice for time-series analysis? Stata Journal 5: 4663. Hamilton, J. D. 1994. Time Series Analysis. Princeton: Princeton University Press. L tkepohl, H. 1993. Introduction to Multiple Time Series Analysis. 2nd ed. New York: Springer. u . 2005. New Introduction to Multiple Time Series Analysis. New York: Springer. Pisati, M. 2001. sg162: Tools for spatial data analysis. Stata Technical Bulletin 60: 2137. Reprinted in Stata Technical Bulletin Reprints, vol. 10, pp. 277298. College Station, TX: Stata Press. Stock, J. H., and M. W. Watson. 2001. Vector autoregressions. Journal of Economic Perspectives 15: 101115.
Also see
[U] 1.3 Whats new [R] intro Introduction to base reference manual
Title
arch Autoregressive conditional heteroskedasticity (ARCH) family of estimators
Syntax
arch depvar options
Model
indepvars
if
in
weight
, options
Description suppress constant term

ARCH terms GARCH terms
noconstant arch(numlist) garch(numlist) saarch(numlist) tarch(numlist) aarch(numlist) narch(numlist) narchk(numlist) abarch(numlist) atarch(numlist) sdgarch(numlist) earch(numlist) egarch(numlist) parch(numlist) tparch(numlist) aparch(numlist) nparch(numlist) nparchk(numlist) pgarch(numlist) constraints(constraints) collinear
Model 2
simple asymmetric ARCH terms threshold ARCH terms asymmetric ARCH terms nonlinear ARCH terms nonlinear ARCH terms with single shift absolute value ARCH terms absolute threshold ARCH terms lags of t news terms in Nelsons (1991) EGARCH model 2 lags of ln(t ) power ARCH terms threshold power ARCH terms asymmetric power ARCH terms nonlinear power ARCH terms nonlinear power ARCH terms with single shift power GARCH terms apply specied linear constraints keep collinear variables include ARCH-in-mean term in the mean-equation specication include specied lags of conditional variance in mean equation apply transformation in exp to any ARCH-in-mean terms specify ARIMA(p, d, q) model for dependent variable autoregressive terms of the structural model disturbance moving-average terms of the structural model disturbances use dist distribution for errors (may be gaussian, normal, t, or ged; default is gaussian) include varlist in the specication of the conditional variance conserve memory during estimation
archm archmlags(numlist) archmexp(exp) arima(# p ,# d ,# q ) ar(numlist) ma(numlist)

Model 3
distribution(dist # ) het(varlist) savespace
11
12
Priming
arch0(xb) arch0(xb0) arch0(xbwt) arch0(xb0wt) arch0(zero) arch0(#) arma0(zero) arma0(p) arma0(q) arma0(pq) arma0(#) condobs(#)
SE/Robust
compute priming values on the basis of the expected unconditional variance; the default compute priming values on the basis of the estimated variance of the residuals from OLS compute priming values on the basis of the weighted sum of squares from OLS residuals compute priming values on the basis of the weighted sum of squares from OLS residuals, with more weight at earlier times set priming values of ARCH terms to zero set priming values of ARCH terms to # set all priming values of ARMA terms to zero; the default begin estimation after observation p, where p is the maximum AR lag in model begin estimation after observation q , where q is the maximum MA lag in model begin estimation after observation (p + q ) set priming values of ARMA terms to # set conditioning observations at the start of the sample to # vcetype may be opg, robust, or oim set condence level; default is level(95) report list of gaps in time series do not display constraints control column formats, row spacing, and line width control the maximization process; seldom used display legend instead of statistics
vce(vcetype)
Reporting
level(#) detail nocnsreport display options

Maximization
maximize options coeflegend
You must tsset your data before using arch; see [TS] tsset. depvar and varlist may contain time-series operators; see [U] 11.4.4 Time-series varlists. by, rolling, statsby, and xi are allowed; see [U] 11.1.10 Prex commands. iweights are allowed; see [U] 11.1.6 weight. coeflegend does not appear in the dialog box. See [U] 20 Estimation and postestimation commands for more capabilities of estimation commands.
To t an ARCH(# m ) model with Gaussian errors, type

. arch depvar
. . . , arch(1/#m )
To t a GARCH(# m , # k ) model assuming that the errors follow Students t distribution with 7 degrees of freedom, type
. arch depvar
. . . , arch(1/#m ) garch(1/#k ) distribution(t 7)
You can also t many other models.
13
Details of syntax
The basic model arch ts is
yt = xt +
2 Var( t ) = t = 0 + A(, ) + B(, )2
(1)
The yt equation may optionally include ARCH-in-mean and ARMA terms:
yt = xt +
i
2 i g(ti ) + ARMA(p, q) +
If no options are specied, A() = B() = 0, and the model collapses to linear regression. The following options add to A() (, , and represent parameters to be estimated): Option arch() garch() saarch() tarch() aarch() narch() narchk() Terms added to A()
A() = A()+ 1,1 A() = A() = A() = A()+ 3,1
2 2 t1 + 1,2 t2 + 2 2 A()+ 2,1 t1 + 2,2 t2 + t1
+ 3,2
t2
+
2 t2 ( t2
A()+ 4,1 2 ( t1 t1
t1 | t1 t1
> 0) + 4,2
t1 ) 2 2
> 0) + + 5,2
2 t2 ) 2
A() = A()+ 5,1 (| A() = A()+ 6,1 ( A() = A()+ 7,1 (
+ 5,1
2
+ 5,2 (|
t2
t2 |
6,1 ) + 6,2 ( 7 ) + 7,2 (
6,2 ) +
t2
7 )2 +
The following options add to B(): Option abarch() atarch() sdgarch() Terms added to B()
B() = B()+ 8,1 | B() = B()+ 9,1 |
t1 |
+ 8,2 |
t2 |
+
t2 |( t2
t1 |( t1
> 0) + 9,2 |
> 0) +
B() = B()+ 10,1 t1 + 10,2 t2 +
Each option requires a numlist argument (see [U] 11.1.8 numlist), which determines the lagged terms included. arch(1) species 1,1 2 , arch(2) species 1,2 2 , arch(1,2) species t2 t1 1,1 2 + 1,2 2 , arch(1/3) species 1,1 2 + 1,2 2 + 1,3 2 , etc. t2 t3 t1 t1 t2 If the earch() or egarch() option is specied, the basic model t is
yt = xt +
i
(2)
2 lnVar( t ) = ln t = 0 + C(ln, z) + A(, ) + B(, )2 2 where zt = t /t . A() and B() are given as above, but A() and B() now add to ln t rather than 2 t . (The options corresponding to A() and B() are rarely specied here.) C() is given by
14
Option earch() egarch()
Terms added to C()
C() = C() +11,1 zt1 + 11,1 (|zt1 | +11,2 zt2 + 11,2 (|zt2 |
2/) 2/) +
2 2 C() = C() +12,1 lnt1 + 12,2 lnt2 +
Instead, if the parch(), tparch(), aparch(), nparch(), nparchk(), or pgarch() options are specied, the basic model t is
yt = xt +
i
(3)
{Var( t )}/2 = t = 0 + D(, ) + A(, ) + B(, )2
where is a parameter to be estimated. A() and B() are given as above, but A() and B() now add to t . (The options corresponding to A() and B() are rarely specied here.) D() is given by Option parch() tparch() aparch() nparch() nparchk() pgarch() Terms added to D()
D() = D()+ 13,1 D() = D() = D() = D() = D() =
t1 + 13,2 t2 + D()+ 14,1 ( t1 > 0) + 14,2 ( t2 > 0) + t1 t2 D()+ 15,1 (| t1 | + 15,1 t1 ) + 15,2 (| t2 | + 15,2 t2 ) D()+ 16,1 | t1 16,1 | + 16,2 | t2 16,2 | + D()+ 17,1 | t1 17 | + 17,2 | t2 17 | + D()+ 18,1 t1 + 18,2 t2 +
Common models
Common term ARCH (Engle 1982) GARCH (Bollerslev 1986) ARCH-in-mean (Engle, Lilien, and Robins 1987) GARCH with ARMA terms EGARCH (Nelson 1991) TARCH, threshold ARCH (Zakoian 1994) GJR, form of threshold ARCH (Glosten, Jagannathan, and Runkle 1993) SAARCH, simple asymmetric ARCH (Engle 1990) PARCH, power ARCH (Higgins and Bera 1992) NARCH, nonlinear ARCH NARCHK, nonlinear ARCH with one shift A-PARCH, asymmetric power ARCH (Ding, Granger, and Engle 1993) NPARCH, nonlinear power ARCH Options to specify arch() arch() garch() archm arch() [garch()] arch() garch() ar() ma() earch() egarch() abarch() atarch() sdgarch() arch() tarch() [garch()] arch() saarch() [garch()] parch() [pgarch()] narch() [garch()] narchk() [garch()] aparch() [pgarch()] nparch() [pgarch()]
15
In all cases, you type arch depvar indepvars , options where options are chosen from the table above. Each option requires that you specify as its argument a numlist that species the lags to be included. For most ARCH models, that value will be 1. For instance, to t the classic rst-order GARCH model on cpi, you would type
. arch cpi, arch(1) garch(1)
If you wanted to t a rst-order GARCH model of cpi on wage, you would type
. arch cpi wage, arch(1) garch(1)
If, for any of the options, you want rst- and second-order terms, specify optionname(1/2). Specifying garch(1) arch(1/2) would t a GARCH model with rst- and second-order ARCH terms. If you specied arch(2), only the lag 2 term would be included.
16
Reading arch output

The regression table reported by arch when using the normal distribution for the errors will appear as
op.depvar depvar x1 x2 L1. L2. _cons ARCHM sigma2 ARMA ar L1. ma L1. HET z1 z2 L1. L2. ARCH arch L1. garch L1. aparch L1. etc. _cons POWER power # # # # # # Coef. Std. Err. z P>|z| [95% Conf. Interval]
... ... ... ... ...
# # #
... ... ... ... ...
# #
... ... ... ... ...
Dividing lines separate equations. The rst one, two, or three equations report the mean model:
yt = xt +
i
The rst equation reports , and the equation will be named [depvar]; if you t a model on d.cpi, the rst equation would be named [cpi]. In Stata, the coefcient on x1 in the above example could be referred to as [depvar] b[x1]. The coefcient on the lag 2 value of x2 would be referred to as [depvar] b[L2.x2]. Such notation would be used, for instance, in a later test command; see [R] test.
17
The [ARCHM] equation reports the coefcients if your model includes ARCH-in-mean terms; see options discussed under the Model 2 tab below. Most ARCH-in-mean models include only a 2 2 contemporaneous variance term, so the term i i g(ti ) becomes t . The coefcient will 2 be [ARCHM] b[sigma2]. If your model includes lags of t , the additional coefcients will be [ARCHM] b[L1.sigma2], and so on. If you specify a transformation g() (option archmexp()), the coefcients will be [ARCHM] b[sigma2ex], [ARCHM] b[L1.sigma2ex], and so on. sigma2ex 2 refers to g(t ), the transformed value of the conditional variance. The [ARMA] equation reports the ARMA coefcients if your model includes them; see options discussed under the Model 2 tab below. This equation includes one or two variables named ar and ma. In later test statements, you could refer to the coefcient on the rst lag of the autoregressive term by typing [ARMA] b[L1.ar] or simply [ARMA] b[L.ar] (the L operator is assumed to be lag 1 if you do not specify otherwise). The second lag on the moving-average term, if there were one, could be referred to by typing [ARMA] b[L2.ma]. The next one, two, or three equations report the variance model. The [HET] equation reports the multiplicative heteroskedasticity if the model includes it. When you t such a model, you specify the variables (and their lags), determining the multiplicative heteroskedasticity; after estimation, their coefcients are simply [HET] b[op.varname]. The [ARCH] equation reports the ARCH, GARCH, etc., terms by referring to variables arch, garch, and so on. For instance, if you specied arch(1) garch(1) when you t the model, the 2 2 conditional variance is given by t = 0 + 1,1 2 + 2,1 t1 . The coefcients would be named t1 [ARCH] b[ cons] (0 ), [ARCH] b[L.arch] (1,1 ), and [ARCH] b[L.garch] (2,1 ). The [POWER] equation appears only if you are tting a variance model in the form of (3) above; the estimated is the coefcient [POWER] b[power]. Also, if you use the distribution() option and specify either Students t or the generalized error distribution but do not specify the degree-of-freedom or shape parameter, then you will see two additional rows in the table. The nal row contains the estimated degree-of-freedom or shape parameter. Immediately preceding the nal row is a transformed version of the parameter that arch used during estimation to ensure that the degree-of-freedom parameter is greater than two or that the shape parameter is positive. The naming convention for estimated ARCH, GARCH, etc., parameters is as follows (denitions for parameters i , i , and i can be found in the tables for A(), B(), C(), and D() above):
18

Option arch() garch() saarch() tarch() aarch() narch() narchk() abarch() atarch() sdgarch() earch() egarch() parch() tparch() aparch() nparch() nparchk() pgarch() 1st parameter 2nd parameter Common parameter
1 2 3 4 5 6 7
= [ARCH] = [ARCH] = [ARCH] = [ARCH] = [ARCH] = [ARCH] = [ARCH]
b[arch] b[garch] b[saarch] b[tarch] b[aarch] b[narch] b[narch]
5 = [ARCH] b[aarch e] 6 = [ARCH] b[narch k] 7 = [ARCH] b[narch k]
8 = [ARCH] b[abarch] 9 = [ARCH] b[atarch] 10 = [ARCH] b[sdgarch] 11 = [ARCH] b[earch] 12 = [ARCH] b[egarch] 13 14 15 16 17 18 = [ARCH] = [ARCH] = [ARCH] = [ARCH] = [ARCH] = [ARCH]
b[parch] b[tparch] b[aparch] b[nparch] b[nparch] b[pgarch]
11 = [ARCH] b[earch a] = [POWER] = [POWER] 15 = [ARCH] b[aparch e] = [POWER] 16 = [ARCH] b[nparch k] = [POWER] 17 = [ARCH] b[nparch k] = [POWER] = [POWER]
b[power] b[power] b[power] b[power] b[power] b[power]
Menu
ARCH/GARCH
Statistics
>
Time series
>
ARCH/GARCH
>
ARCH and GARCH models
EARCH/EGARCH
Statistics
>
Time series
>
ARCH/GARCH
>
Nelsons EGARCH model
ABARCH/ATARCH/SDGARCH
Statistics
>
Time series
>
ARCH/GARCH
>
Threshold ARCH model
ARCH/TARCH/GARCH
Statistics
>
Time series
>
ARCH/GARCH
>
GJR form of threshold ARCH model
ARCH/SAARCH/GARCH
Statistics
>
Time series
>
ARCH/GARCH
>
Simple asymmetric ARCH model
PARCH/PGARCH
Statistics
>
Time series
>
ARCH/GARCH
>
Power ARCH model
NARCH/GARCH
Statistics
>
Time series
>
ARCH/GARCH
>
Nonlinear ARCH model
NARCHK/GARCH
Statistics
>
Time series
>
ARCH/GARCH
>
Nonlinear ARCH model with one shift
APARCH/PGARCH
Statistics
>
Time series
>
ARCH/GARCH
>
Asymmetric power ARCH model
19
NPARCH/PGARCH
Statistics
>
Time series
>
ARCH/GARCH
>
Nonlinear power ARCH model
Description
arch ts regression models in which the volatility of a series varies through time. Usually, periods of high and low volatility are grouped together. ARCH models estimate future volatility as a function of prior volatility. To accomplish this, arch ts models of autoregressive conditional heteroskedasticity (ARCH) by using conditional maximum likelihood. In addition to ARCH terms, models may include multiplicative heteroskedasticity. Gaussian (normal), Students t, and generalized error distributions are supported. Concerning the regression equation itself, models may also contain ARCH-in-mean and ARMA terms.
Options

Model
2 t ).
noconstant; see [R] estimation options. arch(numlist) species the ARCH terms (lags of Specify arch(1) to include rst-order terms, arch(1/2) to specify rst- and second-order terms, arch(1/3) to specify rst-, second-, and third-order terms, etc. Terms may be omitted. Specify arch(1/3 5) to specify terms with lags 1, 2, 3, and 5. All the options work this way. arch() may not be specied with aarch(), narch(), narchk(), nparchk(), or nparch(), as this would result in collinear terms.
2 garch(numlist) species the GARCH terms (lags of t ).
saarch(numlist) species the simple asymmetric ARCH terms. Adding these terms is one way to make the standard ARCH and GARCH models respond asymmetrically to positive and negative innovations. Specifying saarch() with arch() and garch() corresponds to the SAARCH model of Engle (1990). saarch() may not be specied with narch(), narchk(), nparchk(), or nparch(), as this would result in collinear terms. tarch(numlist) species the threshold ARCH terms. Adding these is another way to make the standard ARCH and GARCH models respond asymmetrically to positive and negative innovations. Specifying tarch() with arch() and garch() corresponds to one form of the GJR model (Glosten, Jagannathan, and Runkle 1993). tarch() may not be specied with tparch() or aarch(), as this would result in collinear terms. aarch(numlist) species the lags of the two-parameter term i (| t | + i t )2 . This term provides the same underlying form of asymmetry as including arch() and tarch(), but it is expressed in a different way. aarch() may not be specied with arch() or tarch(), as this would result in collinear terms. narch(numlist) species the lags of the two-parameter term i ( t i )2 . This term allows the minimum conditional variance to occur at a value of lagged innovations other than zero. For any term specied at lag L, the minimum contribution to conditional variance of that lag occurs when 2 tL = L the squared innovations at that lag are equal to the estimated constant L .
20
narch() may not be specied with arch(), saarch(), narchk(), nparchk(), or nparch(), as this would result in collinear terms. narchk(numlist) species the lags of the two-parameter term i ( narch() with held constant for all lags.
t
)2 ; this is a variation of
narchk() may not be specied with arch(), saarch(), narch(), nparchk(), or nparch(), as this would result in collinear terms. abarch(numlist) species lags of the term | t |. atarch(numlist) species lags of | t |( t > 0), where ( t > 0) represents the indicator function returning 1 when true and 0 when false. Like the TARCH terms, these ATARCH terms allow the effect of unanticipated innovations to be asymmetric about zero. sdgarch(numlist) species lags of t . Combining atarch(), abarch(), and sdgarch() produces the model by Zakoian (1994) that the author called the TARCH model. The acronym TARCH, however, refers to any model using thresholding to obtain asymmetry. earch(numlist) species lags of the two-parameter term zt +(|zt | 2/). These terms represent the inuence of newslagged innovationsin Nelsons (1991) EGARCH model. For these terms, zt = t /t , and arch assumes zt N (0, 1). Nelson derived the general form of an EGARCH model for any assumed distribution and performed estimation assuming a generalized error distribution (GED). See Hamilton (1994) for a derivation where zt is assumed normal. The zt terms can be parameterized in either of these two equivalent ways. arch uses Nelsons original parameterization; see Hamilton (1994) for an equivalent alternative.
2 egarch(numlist) species lags of ln(t ).
For the following options, the model is parameterized in terms of h( t ) and t . One is estimated, even when more than one option is specied. parch(numlist) species lags of | t | . parch() combined with pgarch() corresponds to the class of nonlinear models of conditional variance suggested by Higgins and Bera (1992). tparch(numlist) species lags of ( t > 0)| t | , where ( t > 0) represents the indicator function returning 1 when true and 0 when false. As with tarch(), tparch() species terms that allow for a differential impact of good (positive innovations) and bad (negative innovations) news for lags specied by numlist. tparch() may not be specied with tarch(), as this would result in collinear terms. aparch(numlist) species lags of the two-parameter term (| t | + t ) . This asymmetric power ARCH model, A-PARCH, was proposed by Ding, Granger, and Engle (1993) and corresponds to a BoxCox function in the lagged innovations. The authors t the original A-PARCH model on more than 16,000 daily observations of the Standard and Poors 500, and for good reason. As the number of parameters and the exibility of the specication increase, more data are required to estimate the parameters of the conditional heteroskedasticity. See Ding, Granger, and Engle (1993) for a discussion of how seven popular ARCH models nest within the A-PARCH model. When goes to 1, the full term goes to zero for many observations and can then be numerically unstable. nparch(numlist) species lags of the two-parameter term |
t
i | .
nparch() may not be specied with arch(), saarch(), narch(), narchk(), or nparchk(), as this would result in collinear terms. nparchk(numlist) species lags of the two-parameter term | t | ; this is a variation of nparch() with held constant for all lags. This is the direct analog of narchk(), except for the power of . nparchk() corresponds to an extended form of the model of Higgins and Bera (1992) as
21
presented by Bollerslev, Engle, and Nelson (1994). nparchk() would typically be combined with the pgarch() option. nparchk() may not be specied with arch(), saarch(), narch(), narchk(), or nparch(), as this would result in collinear terms. pgarch(numlist) species lags of t . constraints(constraints), collinear; see [R] estimation options.
Model 2
archm species that an ARCH-in-mean term be included in the specication of the mean equation. This term allows the expected value of depvar to depend on the conditional variance. ARCH-in-mean is most commonly used in evaluating nancial time series when a theory supports a tradeoff between asset risk and return. By default, no ARCH-in-mean terms are included in the model. archm species that the contemporaneous expected conditional variance be included in the mean equation. For example, typing
. arch y x, archm arch(1)
species the model

2 yt = 0 + 1 xt + t + 2 t t
= 0 +
2 t1
2 archmlags(numlist) is an expansion of archm that includes lags of the conditional variance t in the mean equation. To specify a contemporaneous and once-lagged variance, specify either archm archmlags(1) or archmlags(0/1).
archmexp(exp) applies the transformation in exp to any ARCH-in-mean terms in the model. The expression should contain an X wherever a value of the conditional variance is to enter the expression. This option can be used to produce the commonly used ARCH-in-mean of the conditional standard deviation. With the example from archm, typing
. arch y x, archm arch(1) archmexp(sqrt(X))
species the mean equation yt = 0 + 1 xt + t +

. arch y x, archm arch(1) archmexp(1/sqrt(X))
t.
Alternatively, typing
species yt = 0 + 1 xt + /t +
t.
arima(# p ,# d ,# q ) is an alternative, shorthand notation for specifying autoregressive models in the dependent variable. The dependent variable and any independent variables are differenced # d times, 1 through # p lags of autocorrelations are included, and 1 through # q lags of moving averages are included. For example, the specication
. arch y, arima(2,1,3)
is equivalent to
. arch D.y, ar(1/2) ma(1/3)
The former is easier to write for classic ARIMA models of the mean equation, but it is not nearly as expressive as the latter. If gaps in the AR or MA lags are to be modeled, or if different operators are to be applied to independent variables, the latter syntax is required.
22
ar(numlist) species the autoregressive terms of the structural model disturbance to be included in the model. For example, ar(1/3) species that lags 1, 2, and 3 of the structural disturbance be included in the model. ar(1,4) species that lags 1 and 4 be included, possibly to account for quarterly effects. If the model does not contain regressors, these terms can also be considered autoregressive terms for the dependent variable; see [TS] arima. ma(numlist) species the moving-average terms to be included in the model. These are the terms for the lagged innovations or white-noise disturbances.
Model 3
distribution(dist # ) species the distribution to assume for the error term. dist may be gaussian, normal, t, or ged. gaussian and normal are synonyms, and # cannot be specied with them. If distribution(t) is specied, arch assumes that the errors follow Students t distribution, and the degree-of-freedom parameter is estimated along with the other parameters of the model. If distribution(t #) is specied, then arch uses Students t distribution with # degrees of freedom. # must be greater than 2. If distribution(ged) is specied, arch assumes that the errors have a generalized error distribution, and the shape parameter is estimated along with the other parameters of the model. If distribution(ged #) is specied, then arch uses the generalized error distribution with shape parameter #. # must be positive. The generalized error distribution is identical to the normal distribution when the shape parameter equals 2. het(varlist) species that varlist be included in the specication of the conditional variance. varlist may contain time-series operators. This varlist enters the variance specication collectively as multiplicative heteroskedasticity; see Judge et al. (1985). If het() is not specied, the model will not contain multiplicative heteroskedasticity. Assume that the conditional variance depends on variables x and w and has an ARCH(1) component. We request this specication by using the het(x w) arch(1) options, and this corresponds to the conditional-variance model
2 t = exp(0 + 1 xt + 2 wt ) + 2 t1
Multiplicative heteroskedasticity enters differently with an EGARCH model because the variance is already specied in logs. For the het(x w) earch(1) egarch(1) options, the variance model is
2 ln(t ) = 0 + 1 xt + 2 wt + zt1 + (|zt1 | 2 2/) + ln(t1 )
savespace conserves memory by retaining only those variables required for estimation. The original dataset is restored after estimation. This option is rarely used and should be specied only if there is insufcient memory to t a model without the option. arch requires considerably more temporary storage during estimation than most estimation commands in Stata.
Priming
arch0(cond method) is a rarely used option that species how to compute the conditioning (presample 2 2 or priming) values for t and 2 . In the presample period, it is assumed that t = 2 and that this t t value is constant. If arch0() is not specied, the priming values are computed as the expected unconditional variance given the current estimates of the coefcients and any ARMA parameters.
23
arch0(xb), the default, species that the priming values are the expected unconditional variance T 2 of the model, which is 1 t /T , where t is computed from the mean equation and any ARMA terms. arch0(xb0) species that the priming values are the estimated variance of the residuals from an OLS estimate of the mean equation. arch0(xbwt) species that the priming values are the weighted sum of the t2 from the current conditional mean equation (and ARMA terms) that places more weight on estimates of 2 at the t beginning of the sample. arch0(xb0wt) species that the priming values are the weighted sum of the t2 from an OLS estimate of the mean equation (and ARMA terms) that places more weight on estimates of 2 t at the beginning of the sample. arch0(zero) species that the priming values are 0. Unlike the priming values for ARIMA models, 0 is generally not a consistent estimate of the presample conditional variance or squared innovations.
2 arch0(#) species that t = to arch0(zero). 2 t
= # for any specied nonnegative #. Thus arch0(0) is equivalent
arma0(cond method) is a rarely used option that species how the t values are initialized at the beginning of the sample for the ARMA component, if the model has one. This option has an effect only when AR or MA terms are included in the model (the ar(), ma(), or arima() options specied). arma0(zero), the default, species that all priming values of t be taken as 0. This ts the model over the entire requested sample and takes t as its expected value of 0 for all lags required by the ARMA terms; see Judge et al. (1985). arma0(p), arma0(q), and arma0(pq) specify that estimation begin after priming the recursions for a certain number of observations. p species that estimation begin after the pth observation in the sample, where p is the maximum AR lag in the model; q species that estimation begin after the q th observation in the sample, where q is the maximum MA lag in the model; and pq species that estimation begin after the (p + q )th observation in the sample. During the priming period, the recursions necessary to generate predicted disturbances are performed, but results are used only to initialize preestimation values of t . To understand the denition of preestimation, say that you t a model in 10/100. If the model is specied with ar(1,2), preestimation refers to observations 10 and 11.
2 The ARCH terms t and 2 are also updated over these observations. Any required lags of t t 2 before the priming period are taken to be their expected value of 0, and 2 and t take the t values specied in arch0().
arma0(#) species that the presample values of t are to be taken as # for all lags required by the ARMA terms. Thus arma0(0) is equivalent to arma0(zero). condobs(#) is a rarely used option that species a xed number of conditioning observations at the start of the sample. Over these priming observations, the recursions necessary to generate 2 predicted disturbances are performed, but only to initialize preestimation values of t , 2 , and t . t Any required lags of t before the initialization period are taken to be their expected value of 0 2 (or the value specied in arma0()), and required values of 2 and t assume the values specied t by arch0(). condobs() can be used if conditioning observations are desired for the lags in the ARCH terms of the model. If arma() is also specied, the maximum number of conditioning observations required by arma() and condobs(#) is used.
24
SE/Robust
vce(vcetype) species the type of standard error reported, which includes types that are robust to some kinds of misspecication and that are derived from asymptotic theory; see [R] vce option. For ARCH models, the robust or quasimaximum likelihood estimates (QMLE) of variance are robust to symmetric nonnormality in the disturbances. The robust variance estimates generally are not robust to functional misspecication of the mean equation; see Bollerslev and Wooldridge (1992). The robust variance estimates computed by arch are based on the full Huber/White/sandwich formulation, as discussed in [P] robust. Many other software packages report robust estimates that set some terms to their expectations of zero (Bollerslev and Wooldridge 1992), which saves them from calculating second derivatives of the log-likelihood function.
Reporting
level(#); see [R] estimation options. detail species that a detailed list of any gaps in the series be reported, including gaps due to missing observations or missing data for the dependent variable or independent variables. nocnsreport; see [R] estimation options. display options: vsquish, cformat(% fmt), pformat(% fmt), sformat(% fmt), and nolstretch; see [R] estimation options.
Maximization
maximize options: difficult, technique(algorithm spec), iterate(#), no log, trace, gradient, showstep, hessian, showtolerance, tolerance(#), ltolerance(#), gtolerance(#), nrtolerance(#), nonrtolerance, and from(init specs); see [R] maximize for all options except gtolerance(), and see below for information on gtolerance(). These options are often more important for ARCH models than for other maximum likelihood models because of convergence problems associated with ARCH models ARCH model likelihoods are notoriously difcult to maximize. Setting technique() to something other than the default or BHHH changes the vcetype to vce(oim). The following options are all related to maximization and are either particularly important in tting ARCH models or not available for most other estimators. gtolerance(#) species the tolerance for the gradient relative to the coefcients. When |gi bi | gtolerance() for all parameters bi and the corresponding elements of the gradient gi , the gradient tolerance criterion is met. The default gradient tolerance for arch is gtolerance(.05). gtolerance(999) may be specied to disable the gradient criterion. If the optimizer becomes stuck with repeated (backed up) messages, the gradient probably still contains substantial values, but an uphill direction cannot be found for the likelihood. With this option, results can often be obtained, but whether the global maximum likelihood has been found is unclear. When the maximization is not going well, it is also possible to set the maximum number of iterations (see [R] maximize) to the point where the optimizer appears to be stuck and to inspect the estimation results at that point. from(init specs) species the initial values of the coefcients. ARCH models may be sensitive to initial values and may have coefcient values that correspond to local maximums. The default starting values are obtained via a series of regressions, producing results that, on
25
the basis of asymptotic theory, are consistent for the and ARMA parameters and generally reasonable for the rest. Nevertheless, these values may not always be feasible in that the likelihood function cannot be evaluated at the initial values arch rst chooses. In such cases, the estimation is restarted with ARCH and ARMA parameters initialized to zero. It is possible, but unlikely, that even these values will be infeasible and that you will have to supply initial values yourself. The standard syntax for from() accepts a matrix, a list of values, or coefcient name value pairs; see [R] maximize. arch also allows the following: from(archb0) sets the starting value for all the ARCH/GARCH/. . . parameters in the conditional-variance equation to 0. from(armab0) sets the starting value for all ARMA parameters in the model to 0. from(archb0 armab0) sets the starting value for all ARCH/GARCH/. . . and ARMA parameters to 0. The following option is available with arch but is not shown in the dialog box: coeflegend; see [R] estimation options.
Remarks
The volatility of a series is not constant through time; periods of relatively low volatility and periods of relatively high volatility tend to be grouped together. This is a commonly observed characteristic of economic time series and is even more pronounced in many frequently sampled nancial series. ARCH models seek to estimate this time-dependent volatility as a function of observed prior volatility. Sometimes the model of volatility is of more interest than the model of the conditional mean. As implemented in arch, the volatility model may also include regressors to account for a structural component in the volatilityusually referred to as multiplicative heteroskedasticity.
ARCH models were introduced by Engle (1982) in a study of ination rates, and there has since been a barrage of proposed parametric and nonparametric specications of autoregressive conditional heteroskedasticity. Overviews of the literature can found in Bollerslev, Engle, and Nelson (1994) and Bollerslev, Chou, and Kroner (1992). Introductions to basic ARCH models appear in many general econometrics texts, including Davidson and MacKinnon (1993, 2004), Greene (2012), Kmenta (1997), Stock and Watson (2011), and Wooldridge (2009). Harvey (1989) and Enders (2004) provide introductions to ARCH in the larger context of econometric time-series modeling, and Hamilton (1994) gives considerably more detail in the same context.
arch ts models of autoregressive conditional heteroskedasticity (ARCH, GARCH, etc.) using conditional maximum likelihood. By conditional, we mean that the likelihood is computed based on 2 an assumed or estimated set of priming values for the squared innovations 2 and variances t prior t to the estimation sample; see Hamilton (1994) or Bollerslev (1986). Sometimes more conditioning is done on the rst a, g , or a + g observations in the sample, where a is the maximum ARCH term lag and g is the maximum GARCH term lag (or the maximum lags from the other ARCH family terms). The original ARCH model proposed by Engle (1982) modeled the variance of a regression models disturbances as a linear function of lagged values of the squared regression disturbances. We can write an ARCH(m) model as
yt = xt + t 2 t = 0 + 1 2 + 2 t1
2 t2
+ + m
2 tm
(conditional mean) (conditional variance)
26
where
2 t
is the squared residuals (or innovations)
i are the ARCH parameters

The ARCH model has a specication for both the conditional mean and the conditional variance, and the variance is a function of the size of prior unanticipated innovations 2 . This model was t generalized by Bollerslev (1986) to include lagged values of the conditional variancea GARCH model. The GARCH(m, k) model is written as
yt = xt +
2 t = 0 + 1
t 2 t1
+ 2
2 t2
+ + m
2 tm
2 2 2 + 1 t1 + 2 t2 + + k tk
where
i are the ARCH parameters i are the GARCH parameters

In his pioneering work, Engle (1982) assumed that the error term, t , followed a Gaussian 2 (normal) distribution: t N (0, t ). However, as Mandelbrot (1963) and many others have noted, the distribution of stock returns appears to be leptokurtotic, meaning that extreme stock returns are more frequent than would be expected if the returns were normally distributed. Researchers have therefore assumed other distributions that can have fatter tails than the normal distribution; arch allows you to t models assuming the errors follow Students t distribution or the generalized error distribution. The t distribution has fatter tails than the normal distribution; as the degree-of-freedom parameter approaches innity, the t distribution converges to the normal distribution. The generalized error distributions tails are fatter than the normal distributions when the shape parameter is less than two and are thinner than the normal distributions when the shape parameter is greater than two. The GARCH model of conditional variance can be considered an ARMA process in the squared innovations, although not in the variances as the equations might seem to suggest; see Hamilton (1994). Specically, the standard GARCH model implies that the squared innovations result from
2 t
= 0 + (1 + 1 )
2 2 t1 + (2 + 2 ) t2 + + (k
+ k )
2 tk
+ wt 1 wt1 2 wt2 3 wt3
where
wt =
2 t
2 t 2 t
wt is a white-noise process that is fundamental for
One of the primary benets of the GARCH specication is its parsimony in identifying the conditional variance. As with ARIMA models, the ARMA specication in GARCH allows the conditional variance to be modeled with fewer parameters than with an ARCH specication alone. Empirically, many series with a conditionally heteroskedastic disturbance have been adequately modeled with a GARCH(1,1) specication. An ARMA process in the disturbances can easily be added to the mean equation. For example, the mean equation can be written with an ARMA(1, 1) disturbance as
yt = xt + (yt1 xt1 ) +
t1
with an obvious generalization to ARMA(p, q) by adding terms; see [TS] arima for more discussion of this specication. This change affects only the conditional-variance specication in that 2 now t results from a different specication of the conditional mean.
27
Much of the literature on ARCH models focuses on alternative specications of the variance equation. arch allows many of these specications to be requested using the saarch() through pgarch() options, which imply that one or more terms may be changed or added to the specication of the variance equation. These alternative specications also address asymmetry. Both the ARCH and GARCH specications imply a symmetric impact of innovations. Whether an innovation 2 is positive or negative makes t 2 no difference to the expected variance t in the ensuing periods; only the size of the innovation mattersgood news and bad news have the same effect. Many theories, however, suggest that positive and negative innovations should vary in their impact. For risk-averse investors, a large unanticipated drop in the market is more likely to lead to higher volatility than a large unanticipated increase (see Black [1976], Nelson [1991]). saarch(), tarch(), aarch(), abarch(), earch(), aparch(), and tparch() allow various specications of asymmetric effects. narch(), narchk(), nparch(), and nparchk() imply an asymmetric impact of a specic form. All the models considered so far have a minimum conditional variance when the lagged innovations are all zero. No news is good news when it comes to keeping the conditional variance small. narch(), narchk(), nparch(), and nparchk() also have a symmetric response to innovations, but they are not centered at zero. The entire news-response function (response to innovations) is shifted horizontally so that minimum variance lies at some specic positive or negative value for prior innovations.
ARCH-in-mean models allow the conditional variance of the series to inuence the conditional mean. This is particularly convenient for modeling the riskreturn relationship in nancial series; the riskier an investment, with all else equal, the lower its expected return. ARCH-in-mean models modify the specication of the conditional mean equation to be
2 yt = xt + t + t
(ARCH-in-mean)
Although this linear form in the current conditional variance has dominated the literature, arch allows the conditional variance to enter the mean equation through a nonlinear transformation g() and for this transformed term to be included contemporaneously or lagged.
2 2 2 yt = xt + 0 g(t ) + 1 g(t1 ) + 2 g(t2 ) + + t
Square root is the most commonly used g() transformation because researchers want to include a linear term for the conditional standard deviation, but any transform g() is allowed.
Example 1: ARCH model

Consider a simple model of the U.S. Wholesale Price Index (WPI) (Enders 2004, 8793), which we also consider in [TS] arima. The data are quarterly over the period 1960q1 through 1990q4. In [TS] arima, we t a model of the continuously compounded rate of change in the WPI, ln(WPIt ) ln(WPIt1 ). The graph of the differenced seriessee [TS] arima clearly shows periods of high volatility and other periods of relative tranquility. This makes the series a good candidate for ARCH modeling. Indeed, price indices have been a common target of ARCH models. Engle (1982) presented the original ARCH formulation in an analysis of U.K. ination rates. First, we t a constant-only model by OLS and test ARCH effects by using Engles Lagrange-multiplier test (estat archlm).
28

. use http://www.stata-press.com/data/r12/wpi1 . regress D.ln_wpi Source Model Residual Total D.ln_wpi _cons SS 0 .02521709 .02521709 Coef. .0108215 df 0 122 122 MS . .000206697 .000206697 t 8.35 P>|t| 0.000 Number of obs F( 0, 122) Prob > F R-squared Adj R-squared Root MSE = = = = = = 123 0.00 . 0.0000 0.0000 .01438
Std. Err. .0012963
[95% Conf. Interval] .0082553 .0133878
. estat archlm, lags(1) LM test for autoregressive conditional heteroskedasticity (ARCH) lags(p) 1 chi2 8.366 H0: no ARCH effects df 1 Prob > chi2 0.0038
vs.
H1: ARCH(p) disturbance
Because the LM test shows a p-value of 0.0038, which is well below 0.05, we reject the null hypothesis of no ARCH(1) effects. Thus we can further estimate the ARCH(1) parameter by specifying arch(1). See [R] regress postestimation time series for more information on Engles LM test. The rst-order generalized ARCH model (GARCH, Bollerslev 1986) is the most commonly used specication for the conditional variance in empirical work and is typically written GARCH(1, 1). We can estimate a GARCH(1, 1) process for the log-differenced series by typing
. arch D.ln_wpi, arch(1) garch(1) (setting optimization to BHHH) Iteration 0: log likelihood = Iteration 1: log likelihood = (output omitted ) Iteration 10: log likelihood = ARCH family regression Sample: 1960q2 - 1990q4 Distribution: Gaussian Log likelihood = 373.234 OPG Std. Err. Number of obs Wald chi2(.) Prob > chi2 = = = 123 . . 355.23458 365.64586 373.1894
D.ln_wpi ln_wpi _cons ARCH arch L1. garch L1. _cons
Coef.
P>|z|
[95% Conf. Interval]
.0061167
.0010616
5.76
0.000
.0040361
.0081974
.4364123
.2437428
1.79
0.073
-.0413147
.9141394
.4544606 .0000269
.1866606 .0000122
2.43 2.20
0.015 0.028
.0886127 2.97e-06
.8203086 .0000508
We have estimated the ARCH(1) parameter to be 0.436 and the GARCH(1) parameter to be 0.454, so our tted GARCH(1, 1) model is
29
yt = 0.0061 +
2 t
t 2 + 0.454 t1
= 0.436
2 t1
where yt = ln(wpit ) ln(wpit1 ). The model Wald test and probability are both reported as missing (.). By convention, Stata reports the model test for the mean equation. Here and fairly often for ARCH models, the mean equation consists only of a constant, and there is nothing to test.
Example 2: ARCH model with ARMA process

We can retain the GARCH(1, 1) specication for the conditional variance and model the mean as an ARMA process with AR(1) and MA(1) terms as well as a fourth-lag MA term to control for quarterly seasonal effects by typing
. arch D.ln_wpi, ar(1) ma(1 4) arch(1) garch(1) (setting optimization to BHHH) Iteration 0: log likelihood = 380.9997 Iteration 1: log likelihood = 388.57823 Iteration 2: log likelihood = 391.34143 Iteration 3: log likelihood = 396.36991 Iteration 4: log likelihood = 398.01098 (switching optimization to BFGS) Iteration 5: log likelihood = 398.23668 BFGS stepping has contracted, resetting BFGS Hessian (0) Iteration 6: log likelihood = 399.21497 Iteration 7: log likelihood = 399.21537 (backed up) (output omitted ) (switching optimization to BHHH) Iteration 15: log likelihood = 399.51441 Iteration 16: log likelihood = 399.51443 Iteration 17: log likelihood = 399.51443 ARCH family regression -- ARMA disturbances Sample: 1960q2 - 1990q4 Number of obs Distribution: Gaussian Wald chi2(3) Log likelihood = 399.5144 Prob > chi2 OPG Std. Err.
= = =
123 153.56 0.0000
D.ln_wpi ln_wpi _cons ARMA ar L1. ma L1. L4. ARCH arch L1. garch L1. _cons
Coef.
P>|z|
.0069541
.0039517
1.76
0.078
-.000791
.0146992
.7922674
.1072225
7.39
0.000
.5821153
1.00242
-.341774 .2451724
.1499943 .1251131
-2.28 1.96
0.023 0.050
-.6357575 -.0000447
-.0477905 .4903896
.2040449
.1244991
1.64
0.101
-.0399688
.4480587
.6949687 .0000119
.1892176 .0000104
3.67 1.14
0.000 0.253
.3241091 -8.52e-06
1.065828 .0000324
30
To clarify exactly what we have estimated, we could write our model as
yt = 0.007 + 0.792 (yt1 0.007) 0.342

2 t
t1
+ 0.245
t4
= 0.204
2 t1
2 .695 t1
where yt = ln(wpit ) ln(wpit1 ). The ARCH(1) coefcient, 0.204, is not signicantly different from zero, but the ARCH(1) and GARCH(1) coefcients are signicant collectively. If you doubt this, you can check with test.
. test [ARCH]L1.arch [ARCH]L1.garch ( 1) ( 2) [ARCH]L.arch = 0 [ARCH]L.garch = 0 chi2( 2) = Prob > chi2 = 84.92 0.0000
(For comparison, we t the model over the same sample used in the example in [TS] arima; Enders ts this GARCH model but over a slightly different sample.)
Technical note
The rather ugly iteration log on the previous result is typical, as difculty in converging is common in ARCH models. This is actually a fairly well-behaved likelihood for an ARCH model. The switching optimization to . . . messages are standard messages from the default optimization method for arch. The backed up messages are typical of BFGS stepping as the BFGS Hessian is often overoptimistic, particularly during early iterations. These messages are nothing to be concerned about. Nevertheless, watch out for the messages BFGS stepping has contracted, resetting BFGS Hessian and backed up, which can ag problems that may result in an iteration log that goes on and on. Stata will never report convergence and will never report nal results. The question is, when do you give up and press Break, and if you do, what then? If the BFGS stepping has contracted message occurs repeatedly (more than, say, ve times), it often indicates that convergence will never be achieved. Literally, it means that the BFGS algorithm was stuck and reset its Hessian and take a steepest-descent step. The backed up message, if it occurs repeatedly, also indicates problems, but only if the likelihood value is simultaneously not changing. If the message occurs repeatedly but the likelihood value is changing, as it did above, all is going well; it is just going slowly. If you have convergence problems, you can specify options to assist the current maximization method or try a different method. Or, your model specication and data may simply lead to a likelihood that is not concave in the allowable region and thus cannot be maximized. If you see the backed up message with no change in the likelihood, you can reset the gradient tolerance to a larger value. Specifying the gtolerance(999) option disables gradient checking, allowing convergence to be declared more easily. This does not guarantee that convergence will be declared, and even if it is, the global maximum likelihood may not have been found. You can also try to specify initial values. Finally, you can try a different maximization method; see options discussed under the Maximization tab above.
31
ARCH models are notorious for having convergence difculties. Unlike in most estimators in Stata, it is common for convergence to require many steps or even to fail. This is particularly true of the explicitly nonlinear terms such as aarch(), narch(), aparch(), or archm (ARCH-in-mean), and of any model with several lags in the ARCH terms. There is not always a solution. You can try other maximization methods or different starting values, but if your data do not support your assumed ARCH structure, convergence simply may not be possible. ARCH models can be susceptible to irrelevant regressors or unnecessary lags, whether in the specication of the conditional mean or in the conditional variance. In these situations, arch will often continue to iterate, making little to no improvement in the likelihood. We view this conservative approach as better than declaring convergence prematurely when the likelihood has not been fully maximized. arch is estimating the conditional form of second sample moments, often with exible functions, and that is asking much of the data.
Technical note
if exp and in range are interpreted differently with commands accepting time-series operators. The time-series operators are resolved before the conditions are tested, which may lead to some confusion. Note the results of the following list commands:
. use http://www.stata-press.com/data/r12/archxmpl . list t y l.y in 5/10 t 5. 6. 7. 8. 9. 10. 1961q1 1961q2 1961q3 1961q4 1962q1 1962q2 y 30.8 30.5 30.5 30.6 30.7 30.6 L.y 30.7 30.8 30.5 30.5 30.6 30.7
. keep in 5/10 (118 observations deleted) . list t y l.y t 1. 2. 3. 4. 5. 6. 1961q1 1961q2 1961q3 1961q4 1962q1 1962q2 y 30.8 30.5 30.5 30.6 30.7 30.6 L.y . 30.8 30.5 30.5 30.6 30.7
We have one more lagged observation for y in the rst case: l.y was resolved before the in restriction was applied. In the second case, the dataset no longer contains the value of y to compute the rst lag. This means that
. use http://www.stata-press.com/data/r12/archxmpl, clear . arch y l.x if twithin(1962q2, 1990q3), arch(1)
is not the same as
32

. keep if twithin(1962q2, 1990q3) . arch y l.x, arch(1)
Example 3: Asymmetric effectsEGARCH model

Continuing with the WPI data, we might be concerned that the economy as a whole responds differently to unanticipated increases in wholesale prices than it does to unanticipated decreases. Perhaps unanticipated increases lead to cash ow issues that affect inventories and lead to more volatility. We can see if the data support this supposition by specifying an ARCH model that allows an asymmetric effect of newsinnovations or unanticipated changes. One of the most popular such models is EGARCH (Nelson 1991). The full rst-order EGARCH model for the WPI can be specied as follows:
. use http://www.stata-press.com/data/r12/wpi1, clear . arch D.ln_wpi, ar(1) ma(1 4) earch(1) egarch(1) (setting optimization to BHHH) Iteration 0: log likelihood = 227.5251 Iteration 1: log likelihood = 381.68426 (output omitted ) Iteration 23: log likelihood = 405.31453 ARCH family regression -- ARMA disturbances Sample: 1960q2 - 1990q4 Number of obs Distribution: Gaussian Wald chi2(3) Log likelihood = 405.3145 Prob > chi2 OPG Std. Err.
= = =
123 156.02 0.0000
D.ln_wpi ln_wpi _cons ARMA ar L1. ma L1. L4. ARCH earch L1. earch_a L1. egarch L1. _cons
Coef.
P>|z|
.0087342
.0034004
2.57
0.010
.0020696
.0153989
.7692121
.0968395
7.94
0.000
.5794101
.9590141
-.3554618 .241463
.1265725 .0863832
-2.81 2.80
0.005 0.005
-.6035393 .072155
-.1073842 .4107711
.4064002
.1163509
3.49
0.000
.1783566
.6344438
.2467349
.1233365
2.00
0.045
.0049999
.4884699
.8417294 -1.4884
.0704079 .6604394
11.96 -2.25
0.000 0.024
.7037325 -2.782837
.9797264 -.1939622
Our result for the variance is

2 ln(t ) = 1.49 + .406 zt1 + .247 ( zt1 2 2/ ) + .842 ln(t1 )
where zt =
t /t ,
which is distributed as N (0, 1).
33
This is a strong indication for a leverage effect. The positive L1.earch coefcient implies that positive innovations (unanticipated price increases) are more destabilizing than negative innovations. The effect appears strong (0.406) and is substantially larger than the symmetric effect (0.247). In fact, the relative scales of the two coefcients imply that the positive leverage completely dominates the symmetric effect. This can readily be seen if we plot what is often referred to as the news-response or news-impact function. This curve shows the resulting conditional variance as a function of unanticipated news, 2 in the form of innovations, that is, the conditional variance t as a function of t . Thus we must 2 evaluate t for various values of t say, 4 to 4and then graph the result.
Example 4: Asymmetric power ARCH model

As an example of a frequently sampled, long-run series, consider the daily closing indices of the Dow Jones Industrial Average, variable dowclose. To avoid the rst half of the century, when the New York Stock Exchange was open for Saturday trading, only data after 1jan1953 are used. The compound return of the series is used as the dependent variable and is graphed below.
DOW, compound return on DJIA
.1 .3 01jan1950 .2 .1 0
01jan1960
01jan1970 date
01jan1980
01jan1990
We formed this difference by referring to D.ln dow, but only after playing a trick. The series is daily, and each observation represents the Dow closing index for the day. Our data included a time variable recorded as a daily date. We wanted, however, to model the log differences in the series, and we wanted the span from Friday to Monday to appear as a single-period difference. That is, the day before Monday is Friday. Because our dataset was tsset with date, the span from Friday to Monday was 3 days. The solution was to create a second variable that sequentially numbered the observations. By tsseting the data with this new variable, we obtained the desired differences.
. generate t = _n . tsset t
34
Now our data look like this:

. use http://www.stata-press.com/data/r12/dow1 . generate dayofwk = dow(date) . list date dayofwk t ln_dow D.ln_dow in 1/8 D. ln_dow . .0058026 -.0054603 -.0048032 -.0013762 -.0100365 -.0080323 .0059433
date 1. 2. 3. 4. 5. 6. 7. 8. 02jan1953 05jan1953 06jan1953 07jan1953 08jan1953 09jan1953 12jan1953 13jan1953
dayofwk 5 1 2 3 4 5 1 2
t 1 2 3 4 5 6 7 8
ln_dow 5.677096 5.682899 5.677439 5.672636 5.671259 5.661223 5.653191 5.659134
. list date dayofwk t ln_dow D.ln_dow in -8/l D. ln_dow .0016198 .0014472 -.011034 .0020638 -.0000877 .009553 -.0052676 -.0148082
date 9334. 9335. 9336. 9337. 9338. 9339. 9340. 9341. 08feb1990 09feb1990 12feb1990 13feb1990 14feb1990 15feb1990 16feb1990 20feb1990
dayofwk 4 5 1 2 3 4 5 2
t 9334 9335 9336 9337 9338 9339 9340 9341
ln_dow 7.880188 7.881635 7.870601 7.872665 7.872577 7.88213 7.876863 7.862054
The difference operator D spans weekends because the specied time variable, t, is not a true date and has a difference of 1 for all observations. We must leave this contrived time variable in place during estimation, or arch will be convinced that our dataset has gaps. If we were using calendar dates, we would indeed have gaps. Ding, Granger, and Engle (1993) t an A-PARCH model of daily returns of the Standard and Poors 500 (S&P 500) for 3jan192830aug1991. We will t the same model for the Dow data shown above. The model includes an AR(1) term as well as the A-PARCH specication of conditional variance.

. arch D.ln_dow, ar(1) aparch(1) pgarch(1) (setting optimization to BHHH) Iteration 0: log likelihood = 31139.547 Iteration 1: log likelihood = 31350.751 (output omitted ) Iteration 68: log likelihood = 32273.555 Iteration 69: log likelihood = 32273.555 ARCH family regression -- AR disturbances Sample: 2 - 9341 Distribution: Gaussian Log likelihood = 32273.56 OPG Std. Err.
35
(backed up)
Number of obs Wald chi2(1) Prob > chi2
= = =
9340 175.46 0.0000
D.ln_dow ln_dow _cons ARMA ar L1. ARCH aparch L1. aparch_e L1. pgarch L1. _cons POWER power
Coef.
P>|z|
.0001786
.0000875
2.04
0.041
7.15e-06
.00035
.1410944
.0106519
13.25
0.000
.1202171
.1619716
.0626323
.0034307
18.26
0.000
.0559082
.0693564
-.3645093
.0378485
-9.63
0.000
-.4386909
-.2903277
.9299015 7.19e-06
.0030998 2.53e-06
299.99 2.84
0.000 0.004
.923826 2.23e-06
.935977 .0000121
1.585187
.0629186
25.19
0.000
1.461869
1.708505
In the iteration log, the nal iteration reports the message backed up. For most estimators, ending on a backed up message would be a cause for great concern, but not with arch or, for that matter, arima, as long as you do not specify the gtolerance() option. arch and arima, by default, monitor the gradient and declare convergence only if, in addition to everything else, the gradient is small enough. The tted model demonstrates substantial asymmetry, with the large negative L1.aparch e coefcient indicating that the market responds with much more volatility to unexpected drops in returns (bad news) than it does to increases in returns (good news).
Example 5: ARCH model with nonnormal errors

Stock returns tend to be leptokurtotic, meaning that large returns (either positive or negative) occur more frequently than one would expect if returns were in fact normally distributed. Here we ret the previous A-PARCH model assuming the errors follow the generalized error distribution, and we let arch estimate the shape parameter of the distribution.
36

. use http://www.stata-press.com/data/r12/dow1, clear . arch D.ln_dow, ar(1) aparch(1) pgarch(1) distribution(ged) (setting optimization to BHHH) Iteration 0: log likelihood = Iteration 1: log likelihood = (output omitted ) Iteration 60: log likelihood = Sample: 2 - 9341 Distribution: GED Log likelihood = 32486.46 OPG Std. Err. 31139.547 31348.13 32486.461 Number of obs Wald chi2(1) Prob > chi2 = = = 9340 178.22 0.0000
ARCH family regression -- AR disturbances
D.ln_dow ln_dow _cons ARMA ar L1. ARCH aparch L1. aparch_e L1. pgarch L1. _cons POWER power /lnshape shape
Coef.
P>|z|
.0002735
.000078
3.51
0.000
.0001207
.0004264
.1337478
.0100188
13.35
0.000
.1141114
.1533842
.0641785
.0049402
12.99
0.000
.0544958
.0738612
-.4051993
.0573039
-7.07
0.000
-.5175128
-.2928857
.9341717 .0000216
.004567 .0000117
204.55 1.84
0.000 0.066
.9252206 -1.39e-06
.9431228 .0000446
1.325311 .3527036 1.422909
.1030753 .009482 .013492
12.86 37.20
0.000 0.000
1.123287 .3341193 1.39671
1.527335 .3712879 1.4496
The ARMA and ARCH coefcients are similar to those we obtained when we assumed normally distributed errors, though we do note that the power term is now closer to 1. The estimated shape parameter for the generalized error distribution is shown at the bottom of the output. Here the shape parameter is 1.42; because it is less than 2, the distribution of the errors has tails that are fatter than they would be if the errors were normally distributed.
Example 6: ARCH model with constraints

Engles (1982) original model, which sparked the interest in ARCH, provides an example requiring constraints. Most current ARCH specications use GARCH terms to provide exible dynamic properties without estimating an excessive number of parameters. The original model was limited to ARCH terms, and to help cope with the collinearity of the terms, a declining lag structure was imposed in the parameters. The conditional variance equation was specied as

2 t = 0 + (.4 t1 t1
37
+ .3
t2 t2
+ .2
t3
+ .1
t4 ) t4
= 0 + .4
+ .3
+ .2
t3
+ .1
From the earlier arch output, we know how the coefcients will be named. In Stata, the formula is
2 t = [ARCH] cons + .4 [ARCH]L1.arch t1 t3
+ .3 [ARCH]L2.arch + .1 [ARCH]L4.arch
t2 t4
+ .2 [ARCH]L3.arch
We could specify these linear constraints many ways, but the following seems fairly intuitive; see [R] constraint for syntax.
. use http://www.stata-press.com/data/r12/wpi1, clear . constraint 1 (3/4)*[ARCH]l1.arch = [ARCH]l2.arch . constraint 2 (2/4)*[ARCH]l1.arch = [ARCH]l3.arch . constraint 3 (1/4)*[ARCH]l1.arch = [ARCH]l4.arch
The original model was t on U.K. ination; we will again use the WPI data and retain our earlier specication of the mean equation, which differs from Engles U.K. ination model. With our constraints, we type
. arch D.ln_wpi, ar(1) ma(1 4) arch(1/4) constraints(1/3) (setting optimization to BHHH) Iteration 0: log likelihood = Iteration 1: log likelihood = (output omitted ) Iteration 9: log likelihood = Sample: 1960q2 - 1990q4 Distribution: Gaussian Log likelihood = 399.4624 ( 1) ( 2) ( 3) .75*[ARCH]L.arch - [ARCH]L2.arch = 0 .5*[ARCH]L.arch - [ARCH]L3.arch = 0 .25*[ARCH]L.arch - [ARCH]L4.arch = 0 OPG Std. Err. 396.80198 399.07809 399.46243 Number of obs Wald chi2(3) Prob > chi2 = = = 123 123.32 0.0000
ARCH family regression -- ARMA disturbances
D.ln_wpi ln_wpi _cons ARMA ar L1. ma L1. L4. ARCH arch L1. L2. L3. L4. _cons
Coef.
P>|z|
.0077204
.0034531
2.24
0.025
.0009525
.0144883
.7388168
.1126811
6.56
0.000
.5179659
.9596676
-.2559691 .2528923
.1442861 .1140185
-1.77 2.22
0.076 0.027
-.5387646 .02942
.0268264 .4763645
.2180138 .1635103 .1090069 .0545034 .0000483
.0737787 .055334 .0368894 .0184447 7.66e-06
2.95 2.95 2.95 2.95 6.30
0.003 0.003 0.003 0.003 0.000
.0734101 .0550576 .0367051 .0183525 .0000333
.3626174 .2719631 .1813087 .0906544 .0000633
38
L1.arch, L2.arch, L3.arch, and L4.arch coefcients have the constrained relative sizes.
Saved results
arch saves the following in e():
Scalars e(N) e(N gaps) e(condobs) e(k) e(k eq) e(k eq model) e(k dv) e(k aux) e(df m) e(ll) e(chi2) e(p) e(archi) e(archany) e(tdf) e(shape) e(tmin) e(tmax) e(power) e(rank) e(ic) e(rc) e(converged) number of observations number of gaps number of conditioning observations number of parameters number of equations in e(b) number of equations in overall model test number of dependent variables number of auxiliary parameters model degrees of freedom log likelihood
2
signicance 2 0 = 2 , priming values 0 1 if model contains ARCH terms, 0 otherwise degrees of freedom for Students t distribution shape parameter for generalized error distribution minimum time maximum time for power ARCH terms rank of e(V) number of iterations return code 1 if converged, 0 otherwise

Macros e(cmd) e(cmdline) e(depvar) e(covariates) e(eqnames) e(wtype) e(wexp) e(title) e(tmins) e(tmaxs) e(dist) e(mhet) e(dfopt) e(chi2type) e(vce) e(vcetype) e(ma) e(ar) e(arch) e(archm) e(archmexp) e(earch) e(egarch) e(aarch) e(narch) e(aparch) e(nparch) e(saarch) e(parch) e(tparch) e(abarch) e(tarch) e(atarch) e(sdgarch) e(pgarch) e(garch) e(opt) e(ml method) e(user) e(technique) e(tech) e(tech steps) e(properties) e(estat cmd) e(predict) e(marginsok) e(marginsnotok)
39
arch command as typed name of dependent variable list of covariates names of equations weight type weight expression title in estimation output formatted minimum time formatted maximum time distribution for error term: gaussian, t, or ged 1 if multiplicative heteroskedasticity yes if degrees of freedom for t distribution or shape parameter for GED distribution was estimated; no otherwise Wald; type of model 2 test vcetype specied in vce() title used to label Std. Err. lags for moving-average terms lags for autoregressive terms lags for ARCH terms ARCH-in-mean lags ARCH-in-mean exp lags for EARCH terms lags for EGARCH terms lags for AARCH terms lags for NARCH terms lags for A-PARCH terms lags for NPARCH terms lags for SAARCH terms lags for PARCH terms lags for TPARCH terms lags for ABARCH terms lags for TARCH terms lags for ATARCH terms lags for SDGARCH terms lags for PGARCH terms lags for GARCH terms type of optimization type of ml method name of likelihood-evaluator program maximization technique maximization technique, including number of iterations number of iterations performed before switching techniques b V program used to implement estat program used to implement predict predictions allowed by margins predictions disallowed by margins
40

Matrices e(b) e(Cns) e(ilog) e(gradient) e(V) e(V modelbased) Functions e(sample)
coefcient vector constraints matrix iteration log (up to 20 iterations) gradient vector variancecovariance matrix of the estimators model-based variance marks estimation sample
Methods and formulas

arch is implemented as an ado-le using the ml commands; see [R] ml. The mean equation for the model t by arch and with ARMA terms can be written as
p p 2 i g(ti ) + i=1 q j=1 p
yt = x t +
ytj xtj
i=1
2 i g(tji )
+
k=1
tk
(conditional mean)
where are the regression parameters, are the ARCH-in-mean parameters, are the autoregression parameters, are the moving-average parameters, and
g() is a general function, see the archmexp() option.

Any of the parameters in this full specication of the conditional mean may be zero. For example, the model need not have moving-average parameters ( = 0) or ARCH-in-mean parameters ( = 0). The variance equation will be one of the following:
2 = 0 + A(, ) + B(, )2
2 ln t = 0 + C( ln, z) + A(, ) + B(, )2 t
(1) (2) (3)

2
= 0 + D(, ) + A(, ) + B(, )
where A(, ), B(, ), C( ln, z), and D(, ) are linear sums of the appropriate ARCH terms; see Details of syntax for more information. Equation (1) is used if no EGARCH or power ARCH terms are included in the model, (2) if EGARCH terms are included, and (3) if any power ARCH terms are included; see Details of syntax. Methods and formulas are presented under the following headings:
Priming values Likelihood from prediction error decomposition Missing data
41
Priming values
The above model is recursive with potentially long memory. It is necessary to assume preestimation 2 sample values for t , 2 , and t to begin the recursions, and the remaining computations are therefore t conditioned on these priming values, which can be controlled using the arch0() and arma0() options. See options discussed under the Priming tab above. The arch0(xb0wt) and arch0(xbwt) options compute a weighted sum of estimated disturbances with more weight on the early observations. With either of these options,
T 1 2 t0 i = 2 t0 i
= (1 .7)
t=0
.7T t1
2 T t
where t0 is the rst observation for which the likelihood is computed; see options discussed under the Priming tab above. The 2 are all computed from the conditional mean equation. If arch0(xb0wt) t is specied, , i , j , and k are taken from initial regression estimates and held constant during optimization. If arch0(xbwt) is specied, the current estimates of , i , j , and k are used to compute 2 on every iteration. If any i is in the mean equation (ARCH-in-mean is specied), the t estimates of 2 from the initial regression estimates are not consistent. t
Likelihood from prediction error decomposition

The likelihood function for ARCH has a particularly simple form. Given priming (or conditioning) 2 values of t , 2 , and t , the mean equation above can be solved recursively for every t (prediction error t decomposition). Likewise, the conditional variance can be computed recursively for each observation by using the variance equation. Using these predicted errors, their associated variances, and the 2 assumption that t N (0, t ), we nd that the log likelihood for each observation t is ln Lt = If we assume that ln Lt = ln
1 2
2 ln(2t ) +
2 t 2 t
t(df), then as given in Hamilton (1994, 662), ln df 2

2 1 t 2 ln (df 2)t + (df + 1) ln 1 + 2 2 (df 2)t
df + 1 2
The likelihood is not dened for df 2, so instead of estimating df directly, we estimate m = ln(df 2). Then df = exp(m) + 2 > 2 for any m. Following Bollerslev, Engle, and Nelson (1994, 2978), the log likelihood for the tth observation, assuming t GED(s), is ln Lt = ln s ln where
s+1 1 t ln 2 ln s1 s 2 t s1 22/s (3s1 )

1/2
To enforce the restriction that s > 0, we estimate r = ln s.
42
This command supports the Huber/White/sandwich estimator of the variance using vce(robust). See [P] robust, particularly Maximum likelihood estimators and Methods and formulas.
Missing data
ARCH allows missing data or missing observations but does not attempt to condition on the 2 surrounding data. If a dynamic component cannot be computed t , 2 , and/or t its priming t value is substituted. If a covariate, the dependent variable, or the entire observation is missing, the observation does not enter the likelihood, and its dynamic components are set to their priming values for that observation. This is acceptable only asymptotically and should not be used with a great deal of missing data.
Robert Fry Engle (1942 ) was born in Syracuse, New York. He earned degrees in physics and economics at Williams College and Cornell and then worked at MIT and the University of California, San Diego, before moving to New York University Stern School of Business in 2000. He was awarded the 2003 Nobel Prize in Economics for research on autoregressive conditional heteroskedasticity and is a leading expert in time-series analysis, especially the analysis of nancial markets.
References
Adkins, L. C., and R. C. Hill. 2008. Using Stata for Principles of Econometrics. 3rd ed. Hoboken, NJ: Wiley. Baum, C. F. 2000. sts15: Tests for stationarity of a time series. Stata Technical Bulletin 57: 3639. Reprinted in Stata Technical Bulletin Reprints, vol. 10, pp. 356360. College Station, TX: Stata Press. Baum, C. F., and R. Sperling. 2000. sts15.1: Tests for stationarity of a time series: Update. Stata Technical Bulletin 58: 3536. Reprinted in Stata Technical Bulletin Reprints, vol. 10, pp. 360362. College Station, TX: Stata Press. Baum, C. F., and V. L. Wiggins. 2000. sts16: Tests for long memory in a time series. Stata Technical Bulletin 57: 3944. Reprinted in Stata Technical Bulletin Reprints, vol. 10, pp. 362368. College Station, TX: Stata Press. Berndt, E. K., B. H. Hall, R. E. Hall, and J. A. Hausman. 1974. Estimation and inference in nonlinear structural models. Annals of Economic and Social Measurement 3/4: 653665. Black, F. 1976. Studies of stock price volatility changes. Proceedings of the American Statistical Association, Business and Economics Statistics 177181. Bollerslev, T. 1986. Generalized autoregressive conditional heteroskedasticity. Journal of Econometrics 31: 307327. Bollerslev, T., R. Y. Chou, and K. F. Kroner. 1992. ARCH modeling in nance. Journal of Econometrics 52: 559. Bollerslev, T., R. F. Engle, and D. B. Nelson. 1994. ARCH models. In Handbook of Econometrics, Volume IV, ed. R. F. Engle and D. L. McFadden. New York: Elsevier. Bollerslev, T., and J. M. Wooldridge. 1992. Quasi-maximum likelihood estimation and inference in dynamic models with time-varying covariances. Econometric Reviews 11: 143172. Davidson, R., and J. G. MacKinnon. 1993. Estimation and Inference in Econometrics. New York: Oxford University Press. . 2004. Econometric Theory and Methods. New York: Oxford University Press. Diebold, F. X. 2003. The ET Interview: Professor Robert F. Engle. Econometric Theory 19: 11591193. Ding, Z., C. W. J. Granger, and R. F. Engle. 1993. A long memory property of stock market returns and a new model. Journal of Empirical Finance 1: 83106. Enders, W. 2004. Applied Econometric Time Series. 2nd ed. New York: Wiley. Engle, R. F. 1982. Autoregressive conditional heteroscedasticity with estimates of the variance of United Kingdom ination. Econometrica 50: 9871007.

. 1990. Discussion: Stock volatility and the crash of 87. Review of Financial Studies 3: 103106.
43
Engle, R. F., D. M. Lilien, and R. P. Robins. 1987. Estimating time varying risk premia in the term structure: The ARCH-M model. Econometrica 55: 391407. Glosten, L. R., R. Jagannathan, and D. E. Runkle. 1993. On the relation between the expected value and the volatility of the nominal excess return on stocks. Journal of Finance 48: 17791801. Greene, W. H. 2012. Econometric Analysis. 7th ed. Upper Saddle River, NJ: Prentice Hall. Hamilton, J. D. 1994. Time Series Analysis. Princeton: Princeton University Press. Harvey, A. C. 1989. Forecasting, Structural Time Series Models and the Kalman Filter. Cambridge: Cambridge University Press. . 1990. The Econometric Analysis of Time Series. 2nd ed. Cambridge, MA: MIT Press. Higgins, M. L., and A. K. Bera. 1992. A class of nonlinear ARCH models. International Economic Review 33: 137158. Hill, R. C., W. E. Grifths, and G. C. Lim. 2011. Principles of Econometrics. 4th ed. Hoboken, NJ: Wiley. Judge, G. G., W. E. Grifths, R. C. Hill, H. L tkepohl, and T.-C. Lee. 1985. The Theory and Practice of Econometrics. u 2nd ed. New York: Wiley. Kmenta, J. 1997. Elements of Econometrics. 2nd ed. Ann Arbor: University of Michigan Press. Mandelbrot, B. 1963. The variation of certain speculative prices. Journal of Business 36: 394419. Nelson, D. B. 1991. Conditional heteroskedasticity in asset returns: A new approach. Econometrica 59: 347370. Press, W. H., S. A. Teukolsky, W. T. Vetterling, and B. P. Flannery. 2007. Numerical Recipes in C: The Art of Scientic Computing. 3rd ed. Cambridge: Cambridge University Press. Stock, J. H., and M. W. Watson. 2011. Introduction to Econometrics. 3rd ed. Boston: AddisonWesley. Wooldridge, J. M. 2009. Introductory Econometrics: A Modern Approach. 4th ed. Cincinnati, OH: South-Western. Zakoian, J. M. 1994. Threshold heteroskedastic models. Journal of Economic Dynamics and Control 18: 931955.
Also see
[TS] arch postestimation Postestimation tools for arch [TS] tsset Declare data to be time-series data [TS] arima ARIMA, ARMAX, and other dynamic regression models [TS] mgarch Multivariate GARCH models [R] regress Linear regression [U] 20 Estimation and postestimation commands
Title
arch postestimation Postestimation tools for arch
Description
The following postestimation commands are available after arch: Command estat estimates lincom Description
AIC, BIC, VCE, and estimation sample summary
cataloging estimation results point estimates, standard errors, testing, and inference for linear combinations of coefcients lrtest likelihood-ratio test margins marginal means, predictive margins, marginal effects, and average marginal effects marginsplot graph the results from margins (prole plots, interaction plots, etc.) nlcom point estimates, standard errors, testing, and inference for nonlinear combinations of coefcients predict predictions, residuals, inuence statistics, and other diagnostic measures predictnl point estimates, standard errors, testing, and inference for generalized predictions test Wald tests of simple and composite linear hypotheses testnl Wald tests of nonlinear hypotheses See the corresponding entries in the Base Reference Manual for details.
Syntax for predict

predict statistic
Main
type
newvar
if
in
, statistic options
Description predicted values for mean equationthe differenced series; the default predicted values for the mean equation in y the undifferenced series predicted values for the conditional variance predicted values of the variance, considering only the multiplicative heteroskedasticity residuals or predicted innovations residuals or predicted innovations in y the undifferenced series
xb y variance het residuals yresiduals
These statistics are available both in and out of sample; type predict for the estimation sample.
. . . if e(sample) . . . if wanted only
44
45
options
Options
Description how to handle the lags of yt make static predictions set starting point for the recursions to time constant calculate considering the structural component only
dynamic(time constant) at(varname | # varname2 | # 2 ) t0(time constant) structural
time constant is a # or a time literal, such as td(1jan1995) or tq(1995q1), etc.; see Conveniently typing SIF values in [D] datetime.
Menu
Statistics
>
Postestimation
>
Predictions, residuals, etc.
Options for predict

Six statistics can be computed by using predict after arch: the predictions of the mean equation (option xb, the default), the undifferenced predictions of the mean equation (option y), the predictions of the conditional variance (option variance), the predictions of the multiplicative heteroskedasticity component of variance (option het), the predictions of residuals or innovations (option residuals), and the predictions of residuals or innovations in terms of y (option yresiduals). Given the dynamic nature of ARCH models and because the dependent variable might be differenced, there are other ways of computing each statistic. We can use all the data on the dependent variable available right up to the time of each prediction (the default, which is often called a one-step prediction), or we can use the data up to a particular time, after which the predicted value of the dependent variable is used recursively to make later predictions (option dynamic()). Either way, we can consider or ignore the ARMA disturbance component, which is considered by default and is ignored if you specify the structural option. We might also be interested in predictions at certain xed points where we 2 specify the prior values of t and t (option at()).
Main
xb, the default, calculates the predictions from the mean equation. If D.depvar is the dependent variable, these predictions are of D.depvar and not of depvar itself. y species that predictions of depvar are to be made even if the model was specied for, say, D.depvar.
2 variance calculates predictions of the conditional variance t .
het calculates predictions of the multiplicative heteroskedasticity component of variance. residuals calculates the residuals. If no other options are specied, these are the predicted innovations t ; that is, they include any ARMA component. If the structural option is specied, these are the residuals from the mean equation, ignoring any ARMA terms; see structural below. The residuals are always from the estimated equation, which may have a differenced dependent variable; if depvar is differenced, they are not the residuals of the undifferenced depvar. yresiduals calculates the residuals for depvar, even if the model was specied for, say, D.depvar. As with residuals, the yresiduals are computed from the model, including any ARMA component. If the structural option is specied, any ARMA component is ignored and yresiduals are the residuals from the structural equation; see structural below.
46
Options
dynamic(time constant) species how lags of yt in the model are to be handled. If dynamic() is not specied, actual values are used everywhere lagged values of yt appear in the model to produce one-step-ahead forecasts. dynamic(time constant) produces dynamic (also known as recursive) forecasts. time constant species when the forecast is to switch from one step ahead to dynamic. In dynamic forecasts, references to yt evaluate to the prediction of yt for all periods at or after time constant; they evaluate to the actual value of yt for all prior periods. dynamic(10), for example, would calculate predictions where any reference to yt with t < 10 evaluates to the actual value of yt and any reference to yt with t 10 evaluates to the prediction of yt . This means that one-step-ahead predictions would be calculated for t < 10 and dynamic predictions would be calculated thereafter. Depending on the lag structure of the model, the dynamic predictions might still refer to some actual values of yt . You may also specify dynamic(.) to have predict automatically switch from one-step-ahead to dynamic predictions at p + q , where p is the maximum AR lag and q is the maximum MA lag. at(varname | # varname2 | # 2 ) makes static predictions. at() and dynamic() may not be specied together. Specifying at() allows static evaluation of results for a given set of disturbances. This is useful, for instance, in generating the news response function. at() species two sets of values to be 2 used for t and t , the dynamic components in the model. These specied values are treated as given. Also, any lagged values of depvar in the model are obtained from the real values of the dependent variable. All computations are based on actual data and the given values. at() requires that you specify two arguments, which can be either a variable name or a number. The rst argument supplies the values to be used for t ; the second supplies the values to be used 2 2 for t . If t plays no role in your model, the second argument may be specied as . to indicate missing. t0(time constant) species the starting point for the recursions to compute the predicted statistics; disturbances are assumed to be 0 for t < t0(). The default is to set t0() to the minimum t observed in the estimation sample, meaning that observations before that are assumed to have disturbances of 0. t0() is irrelevant if structural is specied because then all observations are assumed to have disturbances of 0. t0(5), for example, would begin recursions at t = 5. If your data were quarterly, you might instead type t0(tq(1961q2)) to obtain the same result. Any ARMA component in the mean equation or GARCH term in the conditional-variance equation makes arch recursive and dependent on the starting point of the predictions. This includes one-step-ahead predictions. structural makes the calculation considering the structural component only, ignoring any ARMA terms, and producing the steady-state equilibrium predictions.
47
Remarks
Example 1
Continuing with our EGARCH model example (example 3) in [TS] arch, we can see that predict, 2 at() calculates t given a set of specied innovations ( t , t1 , . . .) and prior conditional variances 2 2 (t1 , t2 , . . .). The syntax is
. predict newvar, variance at(epsilon sigma2)
epsilon and sigma2 are either variables or numbers. Using sigma2 is a little tricky because you specify 2 values of t , which predict is supposed to predict. predict does not simply copy variable sigma2 2 into newvar but uses the lagged values contained in sigma2 to produce the predicted value of t . It does this for all t, and those results are stored in newvar. (If you are interested in dynamic predictions 2 of t , see Options for predict.)
2 2 We will generate predictions for t , assuming that the lagged values of t are 1, and we will vary t from 4 to 4. First, we will create variable et containing t , and then we will create and graph the predictions:
. generate et = (_n-64)/15 . predict sigma2, variance at(et 1) . line sigma2 et in 2/l, m(i) c(l) title(News response function)
News response function

2.5 0 4 conditional variance, onestep .5 1 1.5 2
0 et
The positive asymmetry does indeed dominate the shape of the news response function. In fact, the response is a monotonically increasing function of news. The form of the response function shows that, for our simple model, only positive, unanticipated price increases have the destabilizing effect that we observe as larger conditional variances.
48
Example 2
Continuing with our ARCH model with constraints example (example 6) in [TS] arch, using lincom we can recover the parameter from the original specication.
. lincom [ARCH]l1.arch/.4 ( 1) 2.5 [ARCH]L.arch = 0 D.ln_wpi (1) Coef. .5450344 Std. Err. .1844468 z 2.95 P>|z| 0.003 [95% Conf. Interval] .1835253 .9065436
Any arch parameter could be used to produce an identical estimate.

All postestimation commands listed above are implemented as ado-les.
Also see
[TS] arch Autoregressive conditional heteroskedasticity (ARCH) family of estimators [U] 20 Estimation and postestimation commands
Title
arma Autoregressive fractionally integrated moving-average models
Syntax
arfima depvar options
Model
indepvars
if
in
, options
Description suppress constant term autoregressive terms moving-average terms estimate short-memory model without fractional integration maximum likelihood estimates; the default maximum modied-prole-likelihood estimates apply specied linear constraints do not drop collinear variables vcetype may be oim or robust set condence level; default is level(95) do not display constraints control column formats, row spacing, line width, and display of omitted variables and base and empty cells control the maximization process; seldom used display legend instead of statistics
noconstant ar(numlist) ma(numlist) smemory mle mpl constraints(numlist) collinear

SE/Robust
vce(vcetype)
Reporting
level(#) nocnsreport display options

Maximization
You must tsset your data before using arfima; see [TS] tsset. depvar and indepvars may contain time-series operators; see [U] 11.4.4 Time-series varlists. indepvars may contain factor variables; see [U] 11.4.3 Factor variables. by, rolling, and statsby are allowed; see [U] 11.1.10 Prex commands. coeflegend does not appear in the dialog box. See [U] 20 Estimation and postestimation commands for more capabilities of estimation commands.
Menu
Statistics
>
Time series
>
ARFIMA models
49
50
Description
arfima estimates the parameters of autoregressive fractionally integrated moving-average (ARFIMA) models. Long-memory processes are stationary processes whose autocorrelation functions decay more slowly than short-memory processes. The ARFIMA model provides a parsimonious parameterization of long-memory processes that nests the autoregressive moving-average (ARMA) model, which is widely used for short-memory processes. By allowing for fractional degrees of integration, the ARFIMA model also generalizes the autoregressive integrated moving-average (ARIMA) model with integer degrees of integration. See [TS] arima for ARMA and ARIMA parameter estimation.
Options

Model
noconstant; see [R] estimation options. ar(numlist) species the autoregressive (AR) terms to be included in the model. An AR(p), p 1, specication would be ar(1/p). This model includes all lags from 1 to p, but not all lags need to be included. For example, the specication ar(1 p) would specify an AR(p) with only lags 1 and p included, setting all the other AR lag parameters to 0. ma(numlist) species the moving-average terms to be included in the model. These are the terms for the lagged innovations (white-noise disturbances). ma(1/q ), q 1, species an MA(q ) model, but like the ar() option, not all lags need to be included. smemory causes arfima to t a short-memory model with d = 0. This option causes arfima to estimate the parameters of an ARMA model by a method that is asymptotically equivalent to that produced by arima; see [TS] arima. mle causes arfima to estimate the parameters by maximum likelihood. This method is the default. mpl causes arfima to estimate the parameters by maximum modied prole likelihood (MPL). The MPL estimator of the fractional-difference parameter has less small-sample bias than the maximum likelihood estimator when there are covariates in the model. mpl may only be specied when there is a constant term or indepvars in the model, and it may not be combined with the mle option. constraints(numlist), collinear; see [R] estimation options.
SE/Robust
vce(vcetype) species the type of standard error reported, which includes types that are robust to some kinds of misspecication and that are derived from asymptotic theory; see [R] vce option. Options vce(robust) and mpl may not be combined.
Reporting
level(#), nocnsreport; see [R] estimation options. display options: noomitted, vsquish, noemptycells, baselevels, allbaselevels, cformat(% fmt), pformat(% fmt), sformat(% fmt), and nolstretch; see [R] estimation options.
51
Maximization
maximize options: difficult, technique(algorithm spec), iterate(#), no log, trace, gradient, showstep, hessian, showtolerance, tolerance(#), ltolerance(#), nrtolerance(#), gtolerance(#), nonrtolerance(#), and from(init specs); see [R] maximize for all options. Some special points for arfimas maximize options are listed below. technique(algorithm spec) sets the optimization algorithm. The default algorithm is BFGS and BHHH is not allowed. See [R] maximize for a description of the available optimization algorithms. You can specify multiple optimization methods. For example, technique(bfgs 10 nr) requests that the optimizer perform 10 BFGS iterations and then switch to NewtonRaphson until convergence. iterate(#) sets the maximum number of iterations. When the maximization is not going well, set the maximum number of iterations to the point where the optimizer appears to be stuck and inspect the estimation results at that point. from(matname) allows you to specify starting values for the model parameters in a row vector. We recommend that you use the iterate(0) option, retrieve the initial estimates from e(b), and modify these elements. The following option is available with arfima but is not shown in the dialog box: coeflegend; see [R] estimation options.
Remarks
Long-memory processes are stationary processes whose autocorrelation functions decay more slowly than short-memory processes. Because the autocorrelations die out so slowly, long-memory processes display a type of long-run dependence. The autoregressive fractionally integrated movingaverage (ARFIMA) model provides a parsimonious parameterization of long-memory processes. This parameterization nests the autoregressive moving-average (ARMA) model, which is widely used for short-memory processes. The ARFIMA model also generalizes the autoregressive integrated moving-average (ARIMA) model with integer degrees of integration. ARFIMA models provide a solution for the tendency to overdifference stationary series that exhibit long-run dependence. In the ARIMA approach, a nonstationary time series is differenced d times until the differenced series is stationary, where d is an integer. Such series are said to be integrated of order d, denoted I(d), with not differencing, I(0), being the option for stationary series. Many series exhibit too much dependence to be I(0) but are not I(1), and ARFIMA models are designed to represent these series. The ARFIMA model allows for a continuum of fractional differences, 0.5 < d < 0.5. The generalization to fractional differences allows the ARFIMA model to handle processes that are neither I(0) nor I(1), to test for overdifferencing, and to model long-run effects that only die out at long horizons.
52
Technical note
An ARIMA model for the series yt is given by
(L)(1 L)d yt = (L)
(1)
where (L) = (1 1 L 2 L2 p Lp ) is the autoregressive (AR) polynomial in the lag operator L; Lyt = yt1 ; (L) = (1 + 1 L + 2 L2 + + p Lp ) is the moving-average (MA) lag polynomial; t is the independent and identically distributed innovation term; and d is the integer number of differences required to make the yt stationary. An ARFIMA model is also specied by (1) with the generalization that 0.5 < d < 0.5. Series with d 0.5 are handled by differencing and subsequent ARFIMA modeling.
Because long-memory processes are stationary, one might be tempted to approximate the processes with many terms in an ARMA model. But these approximate models are difcult to t and to interpret because ARMA models with many terms are difcult to estimate and the ARMA parameterization has an inherent short-run nature. In contrast, the ARFIMA model has the d parameter for the long-run dependence and ARMA parameters for short-run dependence. Using different parameters for different types of dependence facilitates estimation and interpretation, as discussed by Sowell (1992a).
Technical note
An ARFIMA model species a fractionally integrated ARMA process. Formally, the ARFIMA model species that
yt = (1 L)d {(L)}1 (L)
The short-run ARMA process (L)1 (L) t captures the short-run effects, and the long-run effects are captured by fractionally integrating the short-run ARMA process. Essentially, the fractional-integration parameter d captures the long-run effects, and the ARMA parameters capture the short-run effects. Having separate parameters for short-run and long-run effects makes the ARFIMA model more exible and easier to interpret than the ARMA model. After estimating the ARFIMA parameters, the short-run effects are obtained by setting d = 0, whereas the long-run effects use the estimated value for d. The short-run effects describe the behavior of the fractionally differenced process (1 L)d yt , whereas the long-run effects describe the behavior of the fractionally integrated yt .
ARFIMA models have been useful in elds as diverse as hydrology and economics. Long-memory processes were rst introduced in hydrology by Hurst (1951). Hosking (1981), in hydrology, and Granger and Joyeux (1980), in economics, independently discovered the ARFIMA representation of long-memory processes. Beran (1994), Baillie (1996), and Palma (2007) provide good introductions to long-memory processes and ARFIMA models.
Example 1: Mount Campito tree ring data

Baillie (1996) discusses a time series of measurements of the widths of the annual rings of a Mount Campito Bristlecone pine. The series contains measurements on rings formed in the tree from 3436 BC to 1969 AD. Essentially, larger widths were good years for the tree and narrower widths were harsh years.
53
We begin by plotting the time series.

. use http://www.stata-press.com/data/r12/campito (Campito Mnt. tree ring data from 3435BC to 1969AD) . tsline width, xlabel(-3435(500)1969) ysize(2)
100 0 3435 2935 2435 1935 1435 20 tree ring width in 0.01mm 40 60 80
935 435 year
65
565
1065
1565
2065
Good years and bad years seem to run together, causing the appearance of local trends. The local trends are evidence of dependence, but they are not as pronounced as those in a nonstationary series. We plot the autocorrelations for another view:
. ac width, ysize(2)
0.80 0.20 Autocorrelations of width 0.00 0.20 0.40 0.60 0
10
20 Lag
30
40
Bartletts formula for MA(q) 95% confidence bands
The autocorrelations do not approach 1 but decay very slowly. Granger and Joyeux (1980) show that the autocorrelations from an ARMA model decay exponentially, whereas the autocorrelations from an ARFIMA process decay at the much slower hyperbolic rate. Box, Jenkins, and Reinsel (2008) dene short-memory processes as those whose autocorrelations decay exponentially fast and long-memory processes as those whose autocorrelations decay at the hyperbolic rate. The above plot of autocorrelations looks closer to hyperbolic than exponential. Together, the above plots make us suspect that the series was generated by a long-memory process. We see evidence that the series is stationary but that the autocorrelations die out much slower than a short-memory process would predict. Given that we believe the data was generated by a stationary process, we begin by tting the data to an ARMA model. We begin by using a short-memory model because a comparison of the results highlights the advantages of using an ARFIMA model for a long-memory process.
54

. arima width, ar(1/2) ma(1) technique(bhhh 4 nr) (setting optimization to BHHH) Iteration 0: log likelihood = -18934.593 Iteration 1: log likelihood = -18914.337 Iteration 2: log likelihood = -18913.407 Iteration 3: log likelihood = -18913.24 (switching optimization to Newton-Raphson) Iteration 4: log likelihood = -18913.214 Iteration 5: log likelihood = -18913.208 Iteration 6: log likelihood = -18913.208 ARIMA regression Sample: -3435 - 1969 Number of obs Wald chi2(3) Log likelihood = -18913.21 Prob > chi2 OIM Std. Err.
= 5405 = 133686.46 = 0.0000
width width _cons ARMA ar L1. L2. ma L1. /sigma
Coef.
P>|z|
42.45055
1.02142
41.56
0.000
40.44861
44.4525
1.264367 -.2848827
.0253199 .0227534
49.94 -12.52
0.000 0.000
1.214741 -.3294785
1.313994 -.2402869
-.8066007 8.005814
.0189699 .0770004
-42.52 103.97
0.000 0.000
-.8437811 7.854896
-.7694204 8.156732
Note: The test of the variance against zero is one sided, and the two-sided confidence interval is truncated at zero.
The roots of the AR polynomial are 0.971 and 0.293, and the root of the MA polynomial is 0.807; all of these are less than one in magnitude, indicating that the series is stationary and invertible but has a high level of persistence. See Hamilton (1994, 59) for how to compute the roots of the polynomials from the estimated coefcients. Below we estimate the parameters of an ARFIMA model with only the fractional difference parameter and a constant.

. arfima width Iteration 0: log likelihood Iteration 1: log likelihood Iteration 2: log likelihood Iteration 3: log likelihood Iteration 4: log likelihood Iteration 5: log likelihood Iteration 6: log likelihood Iteration 7: log likelihood Refining estimates: Iteration 0: log likelihood Iteration 1: log likelihood ARFIMA regression Sample: -3435 - 1969 Log likelihood = -18907.279 OIM Std. Err.
55
= = = = = = = =
-18918.219 -18916.84 -18908.508 -18908.508 -18907.379 -18907.318 -18907.279 -18907.279
(backed up)
= -18907.279 = -18907.279 Number of obs Wald chi2(1) Prob > chi2 = = = 5405 1864.43 0.0000
width width _cons ARFIMA d /sigma2
Coef.
P>|z|
44.01432
9.174318
4.80
0.000
26.03299
61.99565
.4468887 63.92927
.0103497 1.229754
43.18 51.99
0.000 0.000
.4266038 61.519
.4671737 66.33955
The estimate of d is large and statistically signicant. The relative parsimony of the ARFIMA model is illustrated by the fact that the estimates of the standard deviation of the idiosyncratic errors are about the same in the 5-parameter ARMA model and the 3-parameter ARFIMA model.
56
Lets add an AR parameter to the above ARFIMA model:

. arfima width, ar(1) Iteration 0: log likelihood Iteration 1: log likelihood Iteration 2: log likelihood Iteration 3: log likelihood Iteration 4: log likelihood Iteration 5: log likelihood Iteration 6: log likelihood Refining estimates: Iteration 0: log likelihood Iteration 1: log likelihood ARFIMA regression Sample: -3435 - 1969 Log likelihood = -18907.233 OIM Std. Err. = = = = = = = -18910.997 -18910.949 -18908.158 -18907.248 -18907.233 -18907.233 -18907.233
(backed up) (backed up)
= -18907.233 = -18907.233 Number of obs Wald chi2(2) Prob > chi2 = = = 5405 1875.34 0.0000
width width _cons ARFIMA ar L1. d /sigma2
Coef.
P>|z|
43.98774
8.685319
5.06
0.000
26.96482
61.01065
.0063319 .4432473 63.92915
.0210492 .0159172 1.229756
0.30 27.85 51.99
0.764 0.000 0.000
-.0349238 .4120502 61.51887
.0475876 .4744445 66.33943
That the estimated AR term is tiny and statistically insignicant indicates that the d parameter has accounted for all the dependence in the series.
As mentioned above, there is a sense in which the main advantages of an ARFIMA model over an ARMA model for long-memory processes are the relative parsimony of the ARFIMA parameterization and the ability of the ARFIMA parameterization to separate out the long-run effects from the short-run effects. If the true process was generated from an ARFIMA model, an ARMA model with many terms can approximate the process, but the terms make estimation difcult and the lack of separate long-run and short-run parameters complicates interpretation. The previous example highlights the relative parsimony of the ARFIMA model. In the examples below, we illustrate the advantages of having separate parameters for long-run and short-run effects.
Technical note
You may be wondering what long-run effects can be produced by a model for stationary processes. Because the autocorrelations of a long-memory process die out so slowly, the spectral density becomes innite as the frequency goes to 0 and the impulseresponse functions die out at a much slower rate. The spectral density of a process describes the relative contributions of random components at different frequencies to the variance of the process, with the low-frequency components corresponding to long-run effects. See [TS] psdensity for an introduction to estimating and interpreting spectral densities implied by the estimated parameters of parametric models.
57
Granger and Joyeux (1980) motivate ARFIMA models by noting that their implied spectral densities are nite except at frequency 0 with 0 < d < 0.5, whereas stationary ARMA models have nite spectral densities at all frequencies. Granger and Joyeux (1980) argue that the ability of ARFIMA models to capture this long-range dependence, which cannot be captured by stationary ARMA models, is an important advantage of ARFIMA models over ARMA models when modeling long-memory processes. Impulseresponse functions are the coefcients on the innite-order MA representation of a process, and they describe how a shock feeds though the dynamic system. If the process is stationary, the coefcients decay to 0 and they sum to a nite constant. As expected, the coefcients from an ARFIMA model die out at a slower rate than those from an ARMA model. Because the ARMA terms model the short-run effects and the d parameter models the long-run effects, an ARFIMA model species both a short-run impulseresponse function and a long-run impulseresponse function. When an ARMA model is used to approximate a long-memory model, the ARMA impulseresponse-function coefcients confound the two effects.
Example 2
In this example, we model the log of the monthly levels of carbon dioxide above Mauna Loa, Hawaii. To remove the seasonality, we model the twelfth seasonal difference of the log of the series. This example illustrates that the ARFIMA model parameterizes long-run and short-run effects, whereas the ARMA model confounds the two effects. (Sowell [1992a] discusses this point in greater depth.) We begin by tting the series to an ARMA model with an AR(1) term and an MA(2).
. use http://www.stata-press.com/data/r12/mloa . arima S12.log, ar(1) ma(2) (setting optimization to BHHH) Iteration 0: log likelihood = 2000.9262 Iteration 1: log likelihood = 2001.5484 Iteration 2: log likelihood = 2001.5637 Iteration 3: log likelihood = 2001.5641 Iteration 4: log likelihood = 2001.5641 ARIMA regression Sample: 1960m1 - 1990m12 Log likelihood = 2001.564 OPG Std. Err.
= = =
372 500.41 0.0000
S12.log log _cons ARMA ar L1. ma L2. /sigma
Coef.
P>|z|
.0036754
.0002475
14.85
0.000
.0031903
.0041605
.7354346
.0357715
20.56
0.000
.6653237
.8055456
.1353086 .0011129
.0513156 .0000401
2.64 27.77
0.008 0.000
.0347319 .0010344
.2358853 .0011914
All the parameters are statistically signicant, and they indicate a high degree of dependence.
58
Below we nest the previously t ARMA model into an ARFIMA model.

. arfima S12.log, ar(1) ma(2) Iteration 0: log likelihood Iteration 1: log likelihood Iteration 2: log likelihood Iteration 3: log likelihood Iteration 4: log likelihood Refining estimates: Iteration 0: log likelihood Iteration 1: log likelihood ARFIMA regression Sample: 1960m1 - 1990m12 Log likelihood = 2006.0805 OIM Std. Err. Number of obs Wald chi2(3) Prob > chi2 = = = 372 248.88 0.0000 = = = = = = = 2006.0757 2006.0774 2006.0775 2006.0804 2006.0805 2006.0805 2006.0805
S12.log S12.log _cons ARFIMA ar L1. ma L2. d /sigma2
Coef.
P>|z|
.003616
.0012968
2.79
0.005
.0010743
.0061578
.2160894
.101556
2.13
0.033
.0170433
.4151355
.1633916 .4042573 1.20e-06
.0516907 .0805422 8.84e-08
3.16 5.02 13.63
0.002 0.000 0.000
.0620796 .2463974 1.03e-06
.2647035 .5621171 1.38e-06
All the parameters are statistically signicant at the 5% level. That the condence interval for the fractional-difference parameter d includes numbers greater than 0.5 is evidence that the series may be nonstationary. Alternatively, we proceed as if the series is stationary, and the wide condence interval for d reects the difculty of tting a complicated dynamic model with only 372 observations. With the above caveat, we can now proceed to compare the interpretations of the ARMA and ARFIMA estimates. We compare these estimates in terms of their implied spectral densities. The spectral density of a stationary time series describes the relative importance of components at different frequencies. See [TS] psdensity for an introduction to spectral densities. Below we quietly ret the ARMA model and use psdensity to estimate the parametric spectral density implied by the ARMA parameter estimates.
. quietly arima S12.log, ar(1) ma(2) . psdensity d_arma omega1
The psdensity command above put the estimated ARMA spectral density into the new variable d arma at the frequencies stored in the new variable omega1. Below we quietly ret the ARFIMA model and use psdensity to estimate the long-run parametric spectral density and then the short-run parametric spectral density implied by the ARFIMA parameter estimates. The long-run estimates use the estimated d, and the short-run estimates set d to 0 (as is implied by specifying the smemory option). The long-run estimates describe the fractionally integrated series, and the short-run estimates describe the fractionally differenced series.

. quietly arfima S12.log, ar(1) ma(2) . psdensity d_arfima omega2 . psdensity ds_arfima omega3, smemory
59
Now that we have the ARMA estimates, the long-run ARFIMA estimates, and the short-run ARFIMA estimates, we graph them below.
. line d_arma d_arfima omega1, name(lmem) nodraw . line d_arma ds_arfima omega1, name(smem) nodraw . graph combine lmem smem, cols(1) xcommon
0 0
1 Frequency ARMA spectral density
2 ARFIMA longmemory spectral density
0 0
.5
1.5
1 Frequency ARMA spectral density
ARFIMA shortmemory spectral density
The top graph contains a plot of the spectral densities implied by the ARMA parameter estimates and by the long-run ARFIMA parameter estimates. As discussed by Granger and Joyeux (1980), the two models imply different spectral densities for frequencies close to 0 when d > 0. When d > 0, the spectral density implied by the ARFIMA estimates diverges to innity, whereas the spectral density implied by the ARMA estimates remains nite at frequency 0 for stable ARMA processes. This difference reects the ability of ARFIMA models to capture long-run effects that ARMA models only capture as the parameters approach those of an unstable model. The bottom graph contains a plot of the spectral densities implied by the ARMA parameter estimates and by the short-run ARFIMA parameter estimates, which are the ARMA parameters for the fractionally differenced process. Comparing the two plots illustrates the ability of the short-run ARFIMA parameters to capture both low-frequency and high-frequency components in the fractionally differenced series. In contrast, the ARMA parameters captured only low-frequency components in the fractionally integrated series. Comparing the ARFIMA and ARMA spectral densities in the two graphs illustrates that the additional fractional-difference parameter allows the ARFIMA model to identify both long-run and short-run effects, which the ARMA model confounds.
Technical note
As noted above, the spectral density of an ARFIMA process with d > 0 diverges to innity as the frequency goes to 0. In contrast, the spectral density of an ARFIMA process with d < 0 is 0 at frequency 0.
60
The autocorrelation function of an ARFIMA process with d < 0 also decays at the slower hyperbolic rate. ARFIMA processes with d < 0 are sometimes called antipersistent because all the autocorrelations for lags greater than 0 are negative. Hosking (1981), Baillie (1996), and others refer to ARFIMA processes with d < 0 as intermediate memory processes and ARFIMA processes with d > 0 as long-memory processes. Box, Jenkins, and Reinsel (2008, 429) dene long-memory processes as those with the slower hyperbolic rate of decay, which includes ARFIMA processes with d < 0. We follow Box, Jenkins, and Reinsel (2008) and thus call ARFIMA processes for 0.5 < d < 0 and 0 < d < 0.5 long-memory processes. Sowell (1992a) uses the properties of ARFIMA processes with d < 0 to derive tests for whether a series was generated by an I(1) process or an I(d) process with d < 1.
Example 3
In this example, we use arfima to test whether a series is nonstationary. More specically, we test whether the series was generated by an I(1) process by testing whether the rst difference of the series is overdifferenced. We have monthly data on the log of the number of reported cases of mumps in New York City between January 1928 and December 1972. We believe that the series is stationary, after accounting for the monthly seasonal effects. We use an ARFIMA model for differenced series to test the null hypothesis of nonstationarity. We use the condence interval for the d parameter from an ARFIMA model for the rst difference of the log of the series to perform the test. If the right-hand end of the 95% CI is less than 0, we conclude that the differenced series was overdifferenced, which implies that the original series was not nonstationary. More formally, if yt is I(1), then yt = yt yt1 must be I(0). If yt is I(d) with d < 0, then yt is overdifferenced and yt is I(d) with d < 1. We use seasonal indicators to account for the seasonal effects. In the output below, we specify the mpl option to use the MPL estimator that is less biased in the presence of covariates. arfima computes the maximum likelihood estimates (MLE) for the parameters of this stationary and invertible Gaussian process. Alternatively, the maximum MPL estimates may be computed. See Methods and formulas for a description of these two estimation techniques, but sufce it to say that the MLE estimates for d are biased in the presence of exogenous variables, even the constant term, for small samples. The MPL estimator reduces this bias; see Hauser (1999) and Doornik and Ooms (2004).
. use http://www.stata-press.com/data/r12/mumps2, clear (Hipel and Mcleod (1994), http://robjhyndman.com/tsdldata/epi/mumps.dat) . arfima D.log i.month, ma(1 2) mpl Iteration 0: log modified profile Iteration 1: log modified profile Iteration 2: log modified profile Iteration 3: log modified profile Iteration 4: log modified profile Iteration 5: log modified profile Iteration 6: log modified profile Iteration 7: log modified profile Iteration 8: log modified profile Refining estimates: Iteration 0: log modified profile Iteration 1: log modified profile likelihood likelihood likelihood likelihood likelihood likelihood likelihood likelihood likelihood = = = = = = = = = 53.766763 54.38864 54.934726 54.937524 55.002186 55.20462 55.205939 55.205949 55.205949 55.205949 55.205949
likelihood = likelihood =

ARFIMA regression Sample: 1928m2 - 1972m6 Log modified profile likelihood = 55.205949 Number of obs Wald chi2(14) Prob > chi2 = = = 533 1360.28 0.0000
61
D.log D.log month 2 3 4 5 6 7 8 9 10 11 12 _cons ARFIMA ma L1. L2. d
Coef.
OIM Std. Err.
P>|z|
-.220719 .0314683 -.2800296 -.3703179 -.4722035 -.9613239 -1.063042 -.7577301 -.3024251 -.0115317 .0247135 .3656807
.0428112 .0424718 .0460084 .0449932 .0446764 .0448375 .0449272 .0452529 .0462887 .0426911 .0430401 .0303215
-5.16 0.74 -6.09 -8.23 -10.57 -21.44 -23.66 -16.74 -6.53 -0.27 0.57 12.06
0.000 0.459 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.787 0.566 0.000
-.3046275 -.0517749 -.3702043 -.4585029 -.5597676 -1.049204 -1.151098 -.8464242 -.3931494 -.0952046 -.0596435 .3062517
-.1368105 .1147115 -.1898548 -.2821329 -.3846394 -.873444 -.9749868 -.669036 -.2117009 .0721413 .1090705 .4251096
.258056 .1972011 -.2329426
.0684414 .0506439 .067336
3.77 3.89 -3.46
0.000 0.000 0.001
.1239134 .0979409 -.3649187
.3921986 .2964613 -.1009664
We interpret the fact that the estimated 95% CI is strictly less than 0 to mean that the differenced series is overdifferenced, which implies that the original series is stationary.
Saved results
arfima saves the following in e():
Scalars e(N) e(k) e(k eq) e(k dv) e(k aux) e(df m) e(ll) e(chi2) e(p) e(s2) e(tmin) e(tmax) e(ar max) e(ma max) e(rank) e(ic) e(rc) e(converged) e(constant) number of observations number of parameters number of equations in e(b) number of dependent variables number of auxiliary parameters model degrees of freedom log likelihood
2
signicance idiosyncratic error variance estimate, if e(method) = mpl minimum time maximum time maximum AR lag maximum MA lag rank of e(V) number of iterations return code 1 if converged, 0 otherwise 0 if noconstant, 1 otherwise
62

Macros e(cmd) e(cmdline) e(depvar) e(covariates) e(eqnames) e(title) e(tmins) e(tmaxs) e(chi2type) e(vce) e(vcetype) e(ma) e(ar) e(technique) e(tech steps) e(properties) e(estat cmd) e(predict) e(marginsok) e(marginsnotok) Matrices e(b) e(Cns) e(ilog) e(gradient) e(V) e(V modelbased) Functions e(sample)
arfima command as typed name of dependent variable list of covariates names of equations title in estimation output formatted minimum time formatted maximum time Wald; type of model 2 test vcetype specied in vce() title used to label Std. Err. lags for MA terms lags for AR terms maximization technique number of iterations performed before switching techniques b V program used to implement estat program used to implement predict predictions allowed by margins predictions disallowed by margins
coefcient vector constraints matrix iteration log (up to 20 iterations) gradient vector variancecovariance matrix of the estimators model-based variance marks estimation sample

arfima is implemented as an ado-le. Methods and formulas are presented under the following headings:
Introduction The likelihood function The autocovariance function The prole likelihood The MPL
Introduction
We model an observed second-order stationary time-series yt , t = 1, . . . , T , using the
ARFIMA(p, d, q) model dened as
(Lp )(1 L)d (yt xt ) = (Lq ) where
(Lp ) = 1 1 L 2 L2 p Lp (Lq ) = 1 + 1 L + 2 L2 + + q Lq
(1 L)d =
j=0
(1)j
(j + d) Lj (j + 1)(d)
63
and the lag operator is dened as Lj yt = ytj , t = 1, . . . , T and j = 1, . . . , t 1; t N (0, 2 ); () is the gamma function; and 0.5 < d < 0.5, d = 0. The row vector xt contains the exogenous variables specied as indepvars in the arfima syntax. The process is stationary and invertible for 0.5 < d < 0.5; the roots of the AR polynomial, (z) = 1 1 z 2 z 2 p z p = 0, and the MA polynomial, (z) = 1 + 1 z + 2 z 2 + + q z q = 0, lie outside the unit circle and there are no common roots. When 0 < d < 0.5, the process has long memory in that the autocovariance function, h , decays to 0 at a hyperbolic rate, such that h= |h | = . When 0.5 < d < 0, the process also has long memory in that the autocovariance function, h , decays to 0 at a hyperbolic rate such that h= |h | < . (As discussed in the text, some authors refer to ARFIMA processes with 0.5 < d < 0 as having intermediate memory, but we follow Box, Jenkins, and Reinsel [2008] and refer to them as long-memory processes.) Granger and Joyeux (1980), Hosking (1981), Sowell (1992b), Sowell (1992a), Baillie (1996), and Palma (2007) provide overviews of long-memory processes, fractional integration, and introductions to ARFIMA models.
The likelihood function

Estimation of the ARFIMA parameters , , d, and 2 is done by the method of maximum likelihood. The log Gaussian likelihood of y given parameter estimates = ( , , d, , 2 ) is
(y|) =
1 T log(2) + log |V| + (y X) V1 (y X) 2
(2)
where the covariance matrix V has a Toeplitz structure 1 2 0 0 1 1 V= . . . . . . . . .
T 1
T 2
T 3
. . . T 1 . . . T 2 . .. . . . ... 0
Var(yt ) = 0 , Cov(yt , yth ) = h (for h = 1, . . . , t 1), and t = 1, . . . , T (Sowell 1992b).

We use the DurbinLevinson algorithm (Palma 2007; Golub and Van Loan 1996) to factor and invert V. Using only the vector of autocovariances , the DurbinLevinson algorithm will compute = D0.5 L1 (y X), where L is lower triangular and V = LDL and D = Diag(), t = Var(yt ). The algorithm performs these computations without generating the T T matrix L1 . During optimization, we restrict the fractional-integration parameter to (0.5, 0.5) using a logistic transform, d = log {(x + 0.5)/(0.5 x)}, so that the range of d encompasses the real line. During the Rening estimates step, the fractional-integration parameter is transformed back to the restricted space, where we obtain its standard error from the observed information matrix.
The autocovariance function

Computation of the autocovariances h is given by Sowell (1992b) with numerical enhancements by Doornik and Ooms (2003) and is reviewed by Palma (2007, sec. 3.2.4). We reproduce it here. The autocovariance of an ARFIMA(0, d, 0) process is
h = 2
(1 2d) (h + d) (1 d)(d) (1 + h d)
64
where h = 0, 1, . . . . For ARFIMA(p, d, q ), we have

q p
h = 2
i=q j=1
(i)j C(d, p + i h, j )
(3)
where
min(q,q+i)
(i) =
k=max(0,i)
k ki 1
j = j
and
(1 i j )
i=1 m=j
(j m )
C(d, h, ) =
h 2p F (d + h, 1, 1 d + h, ) + F (d h, 1, 1 d h, ) 1 2
F () is the hypergeometric series (Gradshteyn and Ryzhik 2007) F (a, b, c, x) = 1 + ab a(a + 1)b(b + 1) 2 a(a + 1)(a + 2)b(b + 1)(b + 2) 3 x+ x + x + ... c1 c(c + 1) 1 2 c(c + 1)(c + 2) 1 2 3
The series recursions are evaluated backward as Doornik and Ooms (2003) emphasize. Doornik and Ooms (2003) also provide other computational enhancements, such as not dividing by j in (3).
The prole likelihood

Doornik and Ooms (2003) show that the parameters 2 and can be concentrated out of the likelihood. Using (2), the MLE for 2 is
2 =
where R =
1 2 V
1 (y X) R1 (y X) T
(4)
and = (X R1 X)1 X R1 y
(5)
is the weighted least-squares estimates for . Substituting (4) into (2) results in the prole likelihood
p (y|r )
T 2
1 + log(2) +
1 log |R| + log 2 T
We compute the MLEs using the prole likelihood for the reduced parameter set r = ( , , d). Equations (4) and (5) provide MLEs for 2 and to create the full parameter vector = ( , , , d, 2 ). We follow with the Rening estimates step, optimizing on the log likelihood (1). The rening step does not change the estimates; it produces the coefcient variancecovariance matrix from the observed information matrix. Using this prole likelihood prevents the use of the BHHH optimization method because there are no observation-level scores.
65
The MPL
The small-sample MLE for d can be biased when there are exogenous variables in the model. The MPL reduces this bias (Hauser 1999; Doornik and Ooms 2004). The mpl option will direct arfima to use this optimization criterion. The MPL is expressed as
m (y|r )
T {1 + log(2)} 2
1 1 T 2
log |R|
T k2 2
log 2
1 log |X R1 X| 2
where k = rank(X) (An and Bloomeld 1993). There is no MPL estimator for 2 , and you will notice its absence from the coefcient table. However, the unbiased estimate assuming ARFIMA(0, 0, 0),
2 =
(y X) R1 (y X) T k
is stored in e() for postestimation computation of the forecast and residual root mean squared errors.
References
An, S., and P. Bloomeld. 1993. Cox and Reids modication in regression models with correlated errors. Technical report, Department of Statistics, North Carolina State University, Raleigh, NC. Baillie, R. T. 1996. Long memory processes and fractional integration in econometrics. Journal of Econometrics 73: 559. Beran, J. 1994. Statistics for Long-Memory Processes. Boca Raton: Chapman & Hall/CRC. Box, G. E. P., G. M. Jenkins, and G. C. Reinsel. 2008. Time Series Analysis: Forecasting and Control. 4th ed. Hoboken, NJ: Wiley. Doornik, J. A., and M. Ooms. 2003. Computational aspects of maximum likelihood estimation of autoregressive fractionally integrated moving average models. Computational Statistics & Data Analysis 42: 333348. . 2004. Inference and forecasting for ARFIMA models with an application to US and UK ination. Studies in Nonlinear Dynamics & Econometrics 8: 123. Golub, G. H., and C. F. Van Loan. 1996. Matrix Computations. 3rd ed. Baltimore: Johns Hopkins University Press. Gradshteyn, I. S., and I. M. Ryzhik. 2007. Table of Integrals, Series, and Products. 7th ed. San Diego: Elsevier. Granger, C. W. J., and R. Joyeux. 1980. An introduction to long-memory time series models and fractional differencing. Journal of Time Series Analysis 1: 1529. Hamilton, J. D. 1994. Time Series Analysis. Princeton: Princeton University Press. Hauser, M. A. 1999. Maximum likelihood estimators for ARMA and ARFIMA models: a Monte Carlo study. Journal of Statistical Planning and Inference 80: 229255. Hosking, J. R. M. 1981. Fractional differencing. Biometrika 68: 165176. Hurst, H. E. 1951. Long-term storage capacity of reservoirs. Transactions of the American Society of Civil Engineers 116: 770779. Palma, W. 2007. Long-Memory Time Series: Theory and Methods. Hoboken, NJ: Wiley. Sowell, F. 1992a. Modeling long-run behavior with the fractional ARIMA model. Journal of Monetary Economics 29: 277302. . 1992b. Maximum likelihood estimation of stationary univariate fractionally integrated time series models. Journal of Econometrics 53: 165188.
66
Also see
[TS] arma postestimation Postestimation tools for arma [TS] tsset Declare data to be time-series data [TS] arima ARIMA, ARMAX, and other dynamic regression models [TS] sspace State-space models [U] 20 Estimation and postestimation commands
Title
arma postestimation Postestimation tools for arma
Description
The following postestimation command is of special interest after arfima: Command psdensity Description estimate the spectral density
For information about psdensity, see [TS] psdensity. The following standard postestimation commands are also available: Command
Description
contrasts and ANOVA-style joint tests of estimates AIC, BIC, VCE, and estimation sample summary cataloging estimation results point estimates, standard errors, testing, and inference for linear combinations of coefcients lrtest likelihood-ratio test margins marginal means, predictive margins, marginal effects, and average marginal effects marginsplot graph the results from margins (prole plots, interaction plots, etc.) nlcom point estimates, standard errors, testing, and inference for nonlinear combinations of coefcients predict predictions, residuals, inuence statistics, and other diagnostic measures predictnl point estimates, standard errors, testing, and inference for generalized predictions pwcompare pairwise comparisons of estimates test Wald tests of simple and composite linear hypotheses testnl Wald tests of nonlinear hypotheses contrast estat estimates lincom
estat ic, margins, marginsplot, nlcom, and predictnl are not appropriate after arfima, mpl.
See the corresponding entries in the Base Reference Manual for details.
Syntax for predict

predict statistic
Main
type
newvar
if
in
, statistic options
Description predicted values; the default predicted innovations standardized innovations fractionally differenced series
xb residuals rstandard fdifference
These statistics are available both in and out of sample; type predict the estimation sample.
. . . if e(sample) . . . if wanted only for
67
68
options
Options
Description put the estimated root mean squared error of the predicted statistic in a new variable; only permitted with options xb and residuals forecast the time series starting at datetime; only permitted with option xb
rmse( type newvar) dynamic(datetime)
datetime is a # or a time literal, such as td(1jan1995) or tq(1995q1); see [D] datetime.
Menu
Statistics
>
Postestimation
>
Options for predict

Main
xb, the default, calculates the predictions for the level of depvar. residuals calculates the predicted innovations. rstandard calculates the standardized innovations. fdifference calculates the fractionally differenced predictions of depvar.
Options
rmse( type newvar) puts the root mean squared errors of the predicted statistics into the specied new variables. The root mean squared errors measure the variances due to the disturbances but do not account for estimation error. rmse() is only permitted with the xb and residuals options. dynamic(datetime) species when predict starts producing dynamic forecasts. The specied datetime must be in the scale of the time variable specied in tsset, and the datetime must be inside a sample for which observations on the dependent variables are available. For example, dynamic(tq(2008q4)) causes dynamic predictions to begin in the fourth quarter of 2008, assuming that your time variable is quarterly; see [D] datetime. If the model contains exogenous variables, they must be present for the whole predicted sample. dynamic() may only be specied with xb.
Remarks
We assume that you have already read [TS] arma. In this entry, we illustrate some of the features of predict after tting an ARFIMA model using arfima.
Example 1
We have monthly data on the one-year Treasury bill secondary market rate imported from the Federal Reserve Bank (FRED) database using freduse; see Drukker (2006) for an introduction to freduse. Below we t an ARFIMA model with two autoregressive terms and one moving-average term to the data.

. use http://www.stata-press.com/data/r12/tb1yr (FRED, 1-year treasury bill; secondary market rate, monthy 1959-2001) . arfima tb1yr, ar(1/2) ma(1) Iteration 0: log likelihood = -235.31856 Iteration 1: log likelihood = -235.26104 (backed up) Iteration 2: log likelihood = -235.25974 (backed up) Iteration 3: log likelihood = -235.2544 (backed up) Iteration 4: log likelihood = -235.13353 Iteration 5: log likelihood = -235.13065 Iteration 6: log likelihood = -235.12107 Iteration 7: log likelihood = -235.11917 Iteration 8: log likelihood = -235.11869 Iteration 9: log likelihood = -235.11868 Refining estimates: Iteration 0: log likelihood = -235.11868 Iteration 1: log likelihood = -235.11868 ARFIMA regression Sample: 1959m7 - 2001m8 Number of obs = 506 Wald chi2(4) = 1864.15 Log likelihood = -235.11868 Prob > chi2 = 0.0000 OIM Std. Err.
69
tb1yr tb1yr _cons ARFIMA ar L1. L2. ma L1. d /sigma2
Coef.
P>|z|
5.496708
2.92036
1.88
0.060
-.2270925
11.22051
.2326105 .3885212
.1136696 .0835657
2.05 4.65
0.041 0.000
.0098222 .2247354
.4553989 .5523071
.7755849 .460649 .1466495
.0669566 .0646564 .009232
11.58 7.12 15.88
0.000 0.000 0.000
.6443524 .3339247 .1285551
.9068174 .5873733 .1647439
All the parameters are statistically signicant at the 5% level, and they indicate a high degree of dependence in the series. In fact, the condence interval for the fractional-difference parameter d indicates that the series may be nonstationary. We will proceed as if the series is stationary and suppose that it is fractionally integrated of order 0.46. We begin our postestimation analysis by predicting the series in sample:
. predict ptb (option xb assumed)
We continue by using the estimated fractional-difference parameter to fractionally difference the original series and by plotting the original series, the predicted series, and the fractionally differenced series. See [TS] arma for a denition of the fractional-difference operator.
70

. predict fdtb, fdifference . twoway tsline tb1yr ptb fdtb, legend(cols(1))
15 0 1960m1 5 10
1970m1
1980m1 month
1990m1
2000m1
1Year Treasury Bill: Secondary Market Rate xb prediction tb1yr fractionally differenced
The above graph shows that the in-sample predictions appear to track the original series well and that the fractionally differenced series looks much more like a stationary series than does the original.
Example 2
In this example, we use the above estimates to produce a dynamic forecast and a condence interval for the forecast for the one-year treasury bill rate and plot them. We begin by extending the dataset and using predict to put the dynamic forecast in the new ftb variable and the root mean squared error of the forecast in the new rtb variable. (As discussed in Methods and formulas, the root mean squared error of the forecast accounts for the idiosyncratic error but not for the estimation error.)
. tsappend, add(12) . predict ftb, xb dynamic(tm(2001m9)) rmse(rtb)
Now we compute a 90% condence interval around the dynamic forecast and plot the original series, the in-sample forecast, the dynamic forecast, and the condence interval of the dynamic forecast.

. scalar z = invnormal(0.95) . generate lb = ftb - z*rtb if month>=tm(2001m9) (506 missing values generated) . generate ub = ftb + z*rtb if month>=tm(2001m9) (506 missing values generated) . twoway tsline tb1yr ftb if month>tm(1998m12) || > tsrline lb ub if month>=tm(2001m9), > legend(cols(1) label(3 "90% prediction interval"))
7
71
2 1999m1
2000m1
2001m1 month
2002m1
1Year Treasury Bill: Secondary Market Rate xb prediction, dynamic(tm(2001m9)) 90% prediction interval

All postestimation commands listed above are implemented as ado-les. Denote h , h = 1, . . . , t, to be the autocovariance function of the ARFIMA(p, d, q ) process for two observations, yt and yth , h time periods apart. The covariance matrix V of the process of length T has a Toeplitz structure of
0 1 V= . . . T 1
1 0 . . . T 2
2 1 . . . T 3
. . . T 1 . . . T 2 . .. . . . ... 0
where the process variance is 0 = Var(yt ). We factor V = LDL , where L is lower triangular and D = Diag(t ). The structure of L1 is of importance.
L1 =
1 1,1 2,2 . . . T 1,T 1
0 1 2,1 . . . T 1,T 2
0 0 1 . . . T 1,T 2
... ... ... .. .
0 0 0 . . .
. . . T 1,1
0 0 0 . . . 1
72
Let zt = yt xt . The best linear predictor of zt+1 based on z1 , z2 , . . . , zt is zt+1 = t 1 up to, but k=1 t,k ztk+1 . Dene t = (t,t , t,t1 , . . . , t1,1 ) to be the tth row of L 1 not including, the diagonal. Then t = Vt t , where Vt is the t t upper left submatrix of V and t = (1 , 2 , . . . , t ) . Hence, the best linear predictor of the innovations is computed as = L1 z, and the one-step predictions are y = + X. In practice, the computation is
y = L1 y X + X
where L and V are computed from the maximum likelihood estimates. We use the DurbinLevinson algorithm (Palma 2007; Golub and Van Loan 1996) to factor V, invert L, and scale y X using only the vector of estimated autocovariances . The prediction error variances of the one-step predictions are computed recursively in the Durbin Levinson algorithm. They are the t elements in the diagonal matrix D computed from the Cholesky 2 factorization of V. The recursive formula is 0 = 0 , and t = t1 (1 t,t ). Forecasting is carried out as described by Beran (1994, sec. 8.7), zT +k = k V1 z, where k = (T +k1 , T +k2 , . . . , k ). The forecast mean squared error is computed as MSE(zT +k ) = 0 k V1 k . Computation of V1 k is carried out efciently using algorithm 4.7.2 of Golub and Van Loan (1996).
References
Beran, J. 1994. Statistics for Long-Memory Processes. Boca Raton: Chapman & Hall/CRC. Drukker, D. M. 2006. Importing Federal Reserve economic data. Stata Journal 6: 384386. Golub, G. H., and C. F. Van Loan. 1996. Matrix Computations. 3rd ed. Baltimore: Johns Hopkins University Press. Palma, W. 2007. Long-Memory Time Series: Theory and Methods. Hoboken, NJ: Wiley.
Also see
[TS] arma Autoregressive fractionally integrated moving-average models [TS] psdensity Parametric spectral density estimation after arima, arma, and ucm [U] 20 Estimation and postestimation commands
Title
arima ARIMA, ARMAX, and other dynamic regression models
Syntax
Basic syntax for a regression model with ARMA disturbances arima depvar indepvars , ar(numlist) ma(numlist)
Basic syntax for an ARIMA(p, d, q) model arima depvar , arima(# p ,# d ,# q ) Basic syntax for a multiplicative seasonal ARIMA(p, d, q) (P, D, Q)s model arima depvar , arima(# p ,# d ,# q ) sarima(# P ,# D ,# Q ,# s ) Full syntax arima depvar options
Model
indepvars
if
in
weight
, options
Description suppress constant term specify ARIMA(p, d, q ) model for dependent variable autoregressive terms of the structural model disturbance moving-average terms of the structural model disturbance apply specied linear constraints keep collinear variables specify period-#s multiplicative seasonal ARIMA term multiplicative seasonal autoregressive term; may be repeated multiplicative seasonal moving-average term; may be repeated use conditional MLE instead of full MLE conserve memory during estimation use diffuse prior for starting Kalman lter recursions use alternate prior for starting Kalman recursions; seldom used use alternate state vector for starting Kalman lter recursions vcetype may be opg, robust, or oim
noconstant arima(# p ,# d ,# q ) ar(numlist) ma(numlist) constraints(constraints) collinear

Model 2
sarima(# P ,# D ,# Q ,# s ) mar(numlist, #s ) mma(numlist, #s )

Model 3
condition savespace diffuse p0(# | matname) state0(# | matname)

SE/Robust
vce(vcetype)
73
74
Reporting
level(#) detail nocnsreport display options

Maximization
set condence level; default is level(95) report list of gaps in time series do not display constraints control column formats, row spacing, and line width control the maximization process; seldom used display legend instead of statistics
You must tsset your data before using arima; see [TS] tsset. depvar and indepvars may contain time-series operators; see [U] 11.4.4 Time-series varlists. by, rolling, statsby, and xi are allowed; see [U] 11.1.10 Prex commands. iweights are allowed; see [U] 11.1.6 weight. coeflegend does not appear in the dialog box. See [U] 20 Estimation and postestimation commands for more capabilities of estimation commands.
Menu
Statistics
>
Time series
>
ARIMA and ARMAX models
Description
arima ts univariate models with time-dependent disturbances. arima ts a model of depvar on indepvars where the disturbances are allowed to follow a linear autoregressive moving-average (ARMA) specication. The dependent and independent variables may be differenced or seasonally differenced to any degree. When independent variables are included in the specication, such models are often called ARMAX models; and when independent variables are not specied, they reduce to BoxJenkins autoregressive integrated moving-average (ARIMA) models in the dependent variable. Multiplicative seasonal ARMAX and ARIMA models can also be t. Missing data are allowed and are handled using the Kalman lter and methods suggested by Harvey (1989 and 1993); see Methods and formulas. In the full syntax, depvar is the variable being modeled, and the structural or regression part of the model is specied in indepvars. ar() and ma() specify the lags of autoregressive and movingaverage terms, respectively; and mar() and mma() specify the multiplicative seasonal autoregressive and moving-average terms, respectively. arima allows time-series operators in the dependent variable and independent variable lists, and making extensive use of these operators is often convenient; see [U] 11.4.4 Time-series varlists and [U] 13.9 Time-series operators for an extended discussion of time-series operators. arima typed without arguments redisplays the previous estimates.
Options

Model
noconstant; see [R] estimation options. arima(# p ,# d ,# q ) is an alternative, shorthand notation for specifying models with ARMA disturbances. The dependent variable and any independent variables are differenced # d times, and 1 through # p lags of autocorrelations and 1 through # q lags of moving averages are included in the model. For example, the specication

. arima D.y, ar(1/2) ma(1/3)
75
is equivalent to
. arima y, arima(2,1,3)
The latter is easier to write for simple ARMAX and ARIMA models, but if gaps in the AR or MA lags are to be modeled, or if different operators are to be applied to independent variables, the rst syntax is required. ar(numlist) species the autoregressive terms of the structural model disturbance to be included in the model. For example, ar(1/3) species that lags of 1, 2, and 3 of the structural disturbance be included in the model; ar(1 4) species that lags 1 and 4 be included, perhaps to account for additive quarterly effects. If the model does not contain regressors, these terms can also be considered autoregressive terms for the dependent variable. ma(numlist) species the moving-average terms to be included in the model. These are the terms for the lagged innovations (white-noise disturbances). constraints(constraints), collinear; see [R] estimation options. If constraints are placed between structural model parameters and ARMA terms, the rst few iterations may attempt steps into nonstationary areas. This process can be ignored if the nal solution is well within the bounds of stationary solutions.
Model 2
sarima(# P ,# D ,# Q ,#s ) is an alternative, shorthand notation for specifying the multiplicative seasonal components of models with ARMA disturbances. The dependent variable and any independent variables are lag-# s seasonally differenced #D times, and 1 through # P seasonal lags of autoregressive terms and 1 through # Q seasonal lags of moving-average terms are included in the model. For example, the specication
. arima DS12.y, ar(1/2) ma(1/3) mar(1/2,12) mma(1/2,12)
is equivalent to
. arima y, arima(2,1,3) sarima(2,1,2,12)
mar(numlist, # s ) species the lag-# s multiplicative seasonal autoregressive terms. For example, mar(1/2,12) requests that the rst two lag-12 multiplicative seasonal autoregressive terms be included in the model. mma(numlist, # s ) specied the lag-# s multiplicative seasonal moving-average terms. For example, mma(1 3,12) requests that the rst and third (but not the second) lag-12 multiplicative seasonal moving-average terms be included in the model.
Model 3
condition species that conditional, rather than full, maximum likelihood estimates be produced. The presample values for t and t are taken to be their expected value of zero, and the estimate of the variance of t is taken to be constant over the entire sample; see Hamilton (1994, 132). This estimation method is not appropriate for nonstationary series but may be preferable for long series or for models that have one or more long AR or MA lags. diffuse, p0(), and state0() have no meaning for models t from the conditional likelihood and may not be specied with condition.
76
If the series is long and stationary and the underlying data-generating process does not have a long memory, estimates will be similar, whether estimated by unconditional maximum likelihood (the default), conditional maximum likelihood (condition), or maximum likelihood from a diffuse prior (diffuse). In small samples, however, results of conditional and unconditional maximum likelihood may differ substantially; see Ansley and Newbold (1980). Whereas the default unconditional maximum likelihood estimates make the most use of sample information when all the assumptions of the model are met, Harvey (1989) and Ansley and Kohn (1985) argue for diffuse priors often, particularly in ARIMA models corresponding to an underlying structural model. The condition or diffuse options may also be preferred when the model contains one or more long AR or MA lags; this avoids inverting potentially large matrices (see diffuse below). When condition is specied, estimation is performed by the arch command (see [TS] arch), and more control of the estimation process can be obtained using arch directly. condition cannot be specied if the model contains any multiplicative seasonal terms. savespace species that memory use be conserved by retaining only those variables required for estimation. The original dataset is restored after estimation. This option is rarely used and should be used only if there is not enough space to t a model without the option. However, arima requires considerably more temporary storage during estimation than most estimation commands in Stata. diffuse species that a diffuse prior (see Harvey 1989 or 1993) be used as a starting point for the Kalman lter recursions. Using diffuse, nonstationary models may be t with arima (see the p0() option below; diffuse is equivalent to specifying p0(1e9)). By default, arima uses the unconditional expected value of the state vector t (see Methods and formulas) and the mean squared error (MSE) of the state vector to initialize the lter. When the process is stationary, this corresponds to the expected value and expected variance of a random draw from the state vector and produces unconditional maximum likelihood estimates of the parameters. When the process is not stationary, however, this default is not appropriate, and the unconditional MSE cannot be computed. For a nonstationary process, another starting point must be used for the recursions. In the absence of nonsample or presample information, diffuse may be specied to start the recursions from a state vector of zero and a state MSE matrix corresponding to an effectively innite variance on this initial state. This method amounts to an uninformative and improper prior that is updated to a proper MSE as data from the sample become available; see Harvey (1989). Nonstationary models may also correspond to models with innite variance given a particular specication. This and other problems with nonstationary series make convergence difcult and sometimes impossible. diffuse can also be useful if a model contains one or more long AR or MA lags. Computation of the unconditional MSE of the state vector (see Methods and formulas) requires construction and inversion of a square matrix that is of dimension {max(p, q + 1)}2 , where p and q are the maximum AR and MA lags, respectively. If q = 27, for example, we would require a 784-by-784 matrix. Estimation with diffuse does not require this matrix. For large samples, there is little difference between using the default starting point and the diffuse starting point. Unless the series has a long memory, the initial conditions affect the likelihood of only the rst few observations.
77
p0(# | matname) is a rarely specied option that can be used for nonstationary series or when an alternate prior for starting the Kalman recursions is desired (see diffuse above for a discussion of the default starting point and Methods and formulas for background). matname species a matrix to be used as the MSE of the state vector for starting the Kalman lter recursions P1|0 . Instead, one number, #, may be supplied, and the MSE of the initial state vector P1|0 will have this number on its diagonal and all off-diagonal values set to zero. This option may be used with nonstationary series to specify a larger or smaller diagonal for P1|0 than that supplied by diffuse. It may also be used with state0() when you believe that you have a better prior for the initial state vector and its MSE. state0(# | matname) is a rarely used option that species an alternate initial state vector, 1|0 (see Methods and formulas), for starting the Kalman lter recursions. If # is specied, all elements of the vector are taken to be #. The default initial state vector is state0(0).
SE/Robust
vce(vcetype) species the type of standard error reported, which includes types that are robust to some kinds of misspecication and that are derived from asymptotic theory; see [R] vce option. For state-space models in general and ARMAX and ARIMA models in particular, the robust or quasimaximum likelihood estimates (QMLEs) of variance are robust to symmetric nonnormality in the disturbances, including, as a special case, heteroskedasticity. The robust variance estimates are not generally robust to functional misspecication of the structural or ARMA components of the model; see Hamilton (1994, 389) for a brief discussion.
Reporting
level(#); see [R] estimation options. detail species that a detailed list of any gaps in the series be reported, including gaps due to missing observations or missing data for the dependent variable or independent variables. nocnsreport; see [R] estimation options. display options: vsquish, cformat(% fmt), pformat(% fmt), sformat(% fmt), and nolstretch; see [R] estimation options.
Maximization
maximize options: difficult, technique(algorithm spec), iterate(#), no log, trace, gradient, showstep, hessian, showtolerance, tolerance(#), ltolerance(#), nrtolerance(#), gtolerance(#), nonrtolerance(#), and from(init specs); see [R] maximize for all options except gtolerance(), and see below for information on gtolerance(). These options are sometimes more important for ARIMA models than most maximum likelihood models because of potential convergence problems with ARIMA models, particularly if the specied model and the sample data imply a nonstationary model. Several alternate optimization methods, such as BerndtHallHallHausman (BHHH) and Broyden FletcherGoldfarbShanno (BFGS), are provided for ARIMA models. Although ARIMA models are not as difcult to optimize as ARCH models, their likelihoods are nevertheless generally not quadratic and often pose optimization difculties; this is particularly true if a model is nonstationary or nearly nonstationary. Because each method approaches optimization differently, some problems can be successfully optimized by an alternate method when one method fails. Setting technique() to something other than the default or BHHH changes the vcetype to vce(oim). The following options are all related to maximization and are either particularly important in tting ARIMA models or not available for most other estimators.
78
technique(algorithm spec) species the optimization technique to use to maximize the likelihood function. technique(bhhh) species the BerndtHallHallHausman (BHHH) algorithm. technique(dfp) species the DavidonFletcherPowell (DFP) algorithm. technique(bfgs) species the BroydenFletcherGoldfarbShanno (BFGS) algorithm. technique(nr) species Statas modied NewtonRaphson (NR) algorithm. You can specify multiple optimization methods. For example, technique(bhhh 10 nr 20) requests that the optimizer perform 10 BHHH iterations, switch to NewtonRaphson for 20 iterations, switch back to BHHH for 10 more iterations, and so on. The default for arima is technique(bhhh 5 bfgs 10). gtolerance(#) species the tolerance for the gradient relative to the coefcients. When |gi bi | gtolerance() for all parameters bi and the corresponding elements of the gradient gi , the gradient tolerance criterion is met. The default gradient tolerance for arima is gtolerance(.05). gtolerance(999) may be specied to disable the gradient criterion. If the optimizer becomes stuck with repeated (backed up) messages, the gradient probably still contains substantial values, but an uphill direction cannot be found for the likelihood. With this option, results can often be obtained, but whether the global maximum likelihood has been found is unclear. When the maximization is not going well, it is also possible to set the maximum number of iterations (see [R] maximize) to the point where the optimizer appears to be stuck and to inspect the estimation results at that point. from(init specs) allows you to set the starting values of the model coefcients; see [R] maximize for a general discussion and syntax options. The standard syntax for from() accepts a matrix, a list of values, or coefcient name value pairs; see [R] maximize. arima also accepts from(armab0), which sets the starting value for all ARMA parameters in the model to zero prior to optimization.
ARIMA models may be sensitive to initial conditions and may have coefcient values that correspond to local maximums. The default starting values for arima are generally good, particularly in large samples for stationary series.
The following option is available with arima but is not shown in the dialog box: coeflegend; see [R] estimation options.
Remarks
Introduction ARIMA models Multiplicative seasonal ARIMA models ARMAX models Dynamic forecasting
79
Introduction
arima ts both standard ARIMA models that are autoregressive in the dependent variable and structural models with ARMA disturbances. Good introductions to the former models can be found in Box, Jenkins, and Reinsel (2008); Hamilton (1994); Harvey (1993); Newton (1988); Diggle (1990); and many others. The latter models are developed fully in Hamilton (1994) and Harvey (1989), both of which provide extensive treatment of the Kalman lter (Kalman 1960) and the state-space form used by arima to t the models. Consider a rst-order autoregressive moving-average process. Then arima estimates all the parameters in the model
yt = xt + t t = t1 +
where
t1 +
structural equation disturbance, ARMA(1, 1)
is the rst-order autocorrelation parameter is the rst-order moving-average parameter i.i.d. N (0, 2 ), meaning that t is a white-noise disturbance
You can combine the two equations and write a general ARMA(p, q) in the disturbances process as
yt = xt + 1 (yt1 xt1 ) + 2 (yt2 xt2 ) + + p (ytp xtp ) + 1

t1
+ 2
t2
+ + q
tq
It is also common to write the general form of the ARMA model more succinctly using lag operator notation as (Lp )(yt xt ) = (Lq ) t ARMA(p, q) where (Lp ) = 1 1 L 2 L2 p Lp (Lq ) = 1 + 1 L + 2 L2 + + q Lq and Lj yt = ytj . For stationary series, full or unconditional maximum likelihood estimates are obtained via the Kalman lter. For nonstationary series, if some prior information is available, you can specify initial values for the lter by using state0() and p0() as suggested by Hamilton (1994) or assume an uninformative prior by using the diffuse option as suggested by Harvey (1989).
ARIMA models
Pure ARIMA models without a structural component do not have regressors and are often written as autoregressions in the dependent variable, rather than autoregressions in the disturbances from a structural equation. For example, an ARMA(1, 1) model can be written as
yt = + yt1 +
t1
(1a)
Other than a scale factor for the constant term , these models are equivalent to the ARMA in the disturbances formulation estimated by arima, though the latter are more exible and allow a wider class of models.
80
To see this effect, replace xt in the structural equation above with a constant term 0 so that
yt = 0 + t = 0 + t1 + t1 + = 0 + (yt1 0 ) + = (1 )0 + yt1 +
t t1
+ t1 +
t t
(1b)
Equations (1a) and (1b) are equivalent, with = (1 )0 , so whether we consider an ARIMA model as autoregressive in the dependent variable or disturbances is immaterial. Our illustration can easily be extended from the ARMA(1, 1) case to the general ARIMA(p, d, q) case.
Example 1: ARIMA model

Enders (2004, 8793) considers an ARIMA model of the U.S. Wholesale Price Index (WPI) using quarterly data over the period 1960q1 through 1990q4. The simplest ARIMA model that includes differencing and both autoregressive and moving-average components is the ARIMA(1,1,1) specication. We can t this model with arima by typing
. use http://www.stata-press.com/data/r12/wpi1 . arima wpi, arima(1,1,1) (setting optimization to BHHH) Iteration 0: log likelihood = -139.80133 Iteration 1: log likelihood = -135.6278 Iteration 2: log likelihood = -135.41838 Iteration 3: log likelihood = -135.36691 Iteration 4: log likelihood = -135.35892 (switching optimization to BFGS) Iteration 5: log likelihood = -135.35471 Iteration 6: log likelihood = -135.35135 Iteration 7: log likelihood = -135.35132 Iteration 8: log likelihood = -135.35131 ARIMA regression Sample: 1960q2 - 1990q4 Log likelihood = -135.3513 OPG Std. Err.
= = =
123 310.64 0.0000
D.wpi wpi _cons ARMA ar L1. ma L1. /sigma
Coef.
P>|z|
.7498197
.3340968
2.24
0.025
.0950019
1.404637
.8742288
.0545435
16.03
0.000
.7673256
.981132
-.4120458 .7250436
.1000284 .0368065
-4.12 19.70
0.000 0.000
-.6080979 .6529042
-.2159938 .7971829
Examining the estimation results, we see that the AR(1) coefcient is 0.874, the MA(1) coefcient is 0.412, and both are highly signicant. The estimated standard deviation of the white-noise disturbance is 0.725.
81
This model also could have been t by typing

. arima D.wpi, ar(1) ma(1)
The D. placed in front of the dependent variable wpi is the Stata time-series operator for differencing. Thus we would be modeling the rst difference in WPI from the second quarter of 1960 through the fourth quarter of 1990 because the rst observation is lost because of differencing. This second syntax allows a richer choice of models.
Example 2: ARIMA model with additive seasonal effects

After examining rst-differences of WPI, Enders chose a model of differences in the natural logarithms to stabilize the variance in the differenced series. The raw data and rst-difference of the logarithms are graphed below.
US Wholesale Price Index
125 .08
US Wholesale Price Index difference of logs
100
50
1960q1
1970q1 t
1980q1
1990q1
.04 1960q1
25
.02
.02
75
.04
.06
1970q1 t
1980q1
1990q1
On the basis of the autocorrelations, partial autocorrelations (see graphs below), and the results of preliminary estimations, Enders identied an ARMA model in the log-differenced series.
. ac D.ln_wpi, ylabels(-.4(.2).6) . pac D.ln_wpi, ylabels(-.4(.2).6)
0.60 0.60 0.40 Partial autocorrelations of D.ln_wpi 0.20 0.00 0.20 0.40 0
0.40
Autocorrelations of D.ln_wpi 0.20 0.00 0.20 0.40
10
20 Lag
30
40
10
20 Lag
30
40
95% Confidence bands [se = 1/sqrt(n)]
In addition to an autoregressive term and an MA(1) term, an MA(4) term is included to account for a remaining quarterly effect. Thus the model to be t is
82
ln(wpit ) = 0 + 1 { ln(wpit1 ) 0 } + 1
t1
+ 4
t4
We can t this model with arima and Statas standard difference operator:
. arima D.ln_wpi, ar(1) ma(1 4) (setting optimization to BHHH) Iteration 0: log likelihood = Iteration 1: log likelihood = Iteration 2: log likelihood = Iteration 3: log likelihood = Iteration 4: log likelihood = (switching optimization to BFGS) Iteration 5: log likelihood = Iteration 6: log likelihood = Iteration 7: log likelihood = Iteration 8: log likelihood = Iteration 9: log likelihood = Iteration 10: log likelihood = ARIMA regression Sample: 1960q2 - 1990q4 386.0336 OPG Std. Err. Number of obs Wald chi2(3) Prob > chi2 = = = 123 333.60 0.0000 382.67447 384.80754 384.84749 385.39213 385.40983 385.9021 385.95646 386.02979 386.03326 386.03354 386.03357
Log likelihood =
D.ln_wpi ln_wpi _cons ARMA ar L1. ma L1. L4. /sigma
Coef.
P>|z|
.0110493
.0048349
2.29
0.022
.0015731
.0205255
.7806991
.0944946
8.26
0.000
.5954931
.965905
-.3990039 .3090813 .0104394
.1258753 .1200945 .0004702
-3.17 2.57 22.20
0.002 0.010 0.000
-.6457149 .0737003 .0095178
-.1522928 .5444622 .0113609
In this nal specication, the log-differenced series is still highly autocorrelated at a level of 0.781, though innovations have a negative impact in the ensuing quarter (0.399) and a positive seasonal impact of 0.309 in the following year.
Technical note
In one way, the results differ from most of Statas estimation commands: the standard error of the coefcients is reported as OPG Std. Err. As noted in Options, the default standard errors and covariance matrix for arima estimates are derived from the outer product of gradients (OPG). This is one of three asymptotically equivalent methods of estimating the covariance matrix of the coefcients (only two of which are usually tractable to derive). Discussions and derivations of all three estimates can be found in Davidson and MacKinnon (1993), Greene (2012), and Hamilton (1994). Bollerslev, Engle, and Nelson (1994) suggest that the OPG estimates are more numerically stable in time-series regressions when the likelihood and its derivatives depend on recursive computations, which is certainly
83
the case for the Kalman lter. To date, we have found no numerical instabilities in either estimate of the covariance matrixsubject to the stability and convergence of the overall model. Most of Statas estimation commands provide covariance estimates derived from the Hessian of the likelihood function. These alternate estimates can also be obtained from arima by specifying the vce(oim) option.
Multiplicative seasonal ARIMA models

Many time series exhibit a periodic seasonal component, and a seasonal ARIMA model, often abbreviated SARIMA, can then be used. For example, monthly sales data for air conditioners have a strong seasonal component, with sales high in the summer months and low in the winter months. In the previous example, we accounted for quarterly effects by tting the model
(1 1 L){ ln(wpit ) 0 } = (1 + 1 L + 4 L4 )
This is an additive seasonal ARIMA model, in the sense that the rst- and fourth-order MA terms work additively: (1 + 1 L + 4 L4 ). Another way to handle the quarterly effect would be to t a multiplicative seasonal ARIMA model. A multiplicative SARIMA model of order (1, 1, 1) (0, 0, 1)4 for the ln(wpit ) series is
(1 1 L){ ln(wpit ) 0 } = (1 + 1 L)(1 + 4,1 L4 )

or, upon expanding terms,
ln(wpit ) = 0 + 1 { ln(wpit ) 0 } + 1
t1
+ 4,1
t4
+ 1 4,1
t5
(2)
In the notation (1, 1, 1) (0, 0, 1)4 , the (1, 1, 1) means that there is one nonseasonal autoregressive term (1 1 L) and one nonseasonal moving-average term (1 + 1 L) and that the time series is rst-differenced one time. The (0, 0, 1)4 indicates that there is no lag-4 seasonal autoregressive term, that there is one lag-4 seasonal moving-average term (1 + 4,1 L4 ), and that the series is seasonally differenced zero times. This is known as a multiplicative SARIMA model because the nonseasonal and seasonal factors work multiplicatively: (1 + 1 L)(1 + 4,1 L4 ). Multiplying the terms imposes nonlinear constraints on the parameters of the fth-order lagged values; arima imposes these constraints automatically. To further clarify the notation, consider a (2, 1, 1) (1, 1, 2)4 multiplicative SARIMA model:
(1 1 L 2 L2 )(1 4,1 L4 )4 zt = (1 + 1 L)(1 + 4,1 L4 + 4,2 L8 )
(3)
where denotes the difference operator yt = yt yt1 and s denotes the lag-s seasonal difference operator s yt = yt yts . Expanding (3), we have
zt = 1 zt1 + 2 zt2 + 4,1 zt4 1 4,1 zt5 2 4,1 zt6 + 1

where
t1
+ 4,1
t4
+ 1 4,1
t5
+ 4,2
t8
+ 1 4,2
t9
zt = 4 zt = (zt zt4 ) = zt zt1 (zt4 zt5 )
84
and zt = yt xt if regressors are included in the model, zt = yt 0 if just a constant term is included, and zt = yt otherwise. More generally, a (p, d, q) (P, D, Q)s multiplicative SARIMA model is (Lp )s (LP )d D zt = (Lq )s (LQ ) s where s (LP ) = (1 s,1 Ls s,2 L2s s,P LP s ) s (LQ ) = (1 + s,1 Ls + s,2 L2s + + s,Q LQs ) (Lp ) and (Lq ) were dened previously, d means apply the operator d times, and similarly for D . Typically, d and D will be 0 or 1; and p, q , P , and Q will seldom be more than 2 or 3. s s will typically be 4 for quarterly data and 12 for monthly data. In fact, the model can be extended to include both monthly and quarterly seasonal factors, as we explain below. If a plot of the data suggests that the seasonal effect is proportional to the mean of the series, then the seasonal effect is probably multiplicative and a multiplicative SARIMA model may be appropriate. Box, Jenkins, and Reinsel (2008, sec. 9.3.1) suggest starting with a multiplicative SARIMA model with any data that exhibit seasonal patterns and then exploring nonmultiplicative SARIMA models if the multiplicative models do not t the data well. On the other hand, Chateld (2004, 14) suggests that taking the logarithm of the series will make the seasonal effect additive, in which case an additive SARIMA model as t in the previous example would be appropriate. In short, the analyst should probably try both additive and multiplicative SARIMA models to see which provides better ts and forecasts. Unless the diffuse option is used, arima must create square matrices of dimension {max(p, q + 1)}2 , where p and q are the maximum AR and MA lags, respectively; and the inclusion of long seasonal terms can make this dimension rather large. For example, with monthly data, you might t a (0, 1, 1) (0, 1, 2)12 SARIMA model. The maximum MA lag is 2 12 + 1 = 25, requiring a matrix with 262 = 676 rows and columns.
Example 3: Multiplicative SARIMA model

One of the most common multiplicative SARIMA specications is the (0, 1, 1) (0, 1, 1)12 airline model of Box, Jenkins, and Reinsel (2008, sec. 9.2). The dataset airline.dta contains monthly international airline passenger data from January 1949 through December 1960. After rst- and seasonally differencing the data, we do not suspect the presence of a trend component, so we use the noconstant option with arima:

. use http://www.stata-press.com/data/r12/air2 (TIMESLAB: Airline passengers) . generate lnair = ln(air) . arima lnair, arima(0,1,1) sarima(0,1,1,12) noconstant (setting optimization to BHHH) Iteration 0: log likelihood = 223.8437 Iteration 1: log likelihood = 239.80405 (output omitted ) Iteration 8: log likelihood = 244.69651 ARIMA regression Sample: 14 - 144 Number of obs Wald chi2(2) Log likelihood = 244.6965 Prob > chi2 OPG Std. Err.
85
= = =
131 84.53 0.0000
DS12.lnair ARMA ma L1. ARMA12 ma L1. /sigma
Coef.
P>|z|
-.4018324
.0730307
-5.50
0.000
-.5449698
-.2586949
-.5569342 .0367167
.0963129 .0020132
-5.78 18.24
0.000 0.000
-.745704 .0327708
-.3681644 .0406625
Thus our model of the monthly number of international airline passengers is
12 lnairt = 0.402 = 0.037
t1
0.557
t12
+ 0.224
t13
In (2), for example, the coefcient on t13 is the product of the coefcients on the t1 and t12 terms (0.224 0.402 0.557). arima labeled the dependent variable DS12.lnair to indicate that it has applied the difference operator and the lag-12 seasonal difference operator 12 to lnair; see [U] 11.4.4 Time-series varlists for more information. We could have t this model by typing
. arima DS12.lnair, ma(1) mma(1, 12) noconstant
For simple multiplicative models, using the sarima() option is easier, though this second syntax allows us to incorporate more complicated seasonal terms.
The mar() and mma() options can be repeated, allowing us to control for multiple seasonal patterns. For example, we may have monthly sales data that exhibit a quarterly pattern as businesses purchase our product at the beginning of calendar quarters when new funds are budgeted, and our product is purchased more frequently in a few months of the year than in most others, even after we control for quarterly uctuations. Thus we might choose to t the model
(1L)(14,1 L4 )(112,1 L12 )(4 12 salest 0 ) = (1+L)(1+4,1 L4 )(1+12,1 L12 )
86
Although this model looks rather complicated, estimating it using arima is straightforward:
. arima DS4S12.sales, ar(1) mar(1, 4) mar(1, 12) ma(1) mma(1, 4) mma(1, 12)
If we instead wanted to include two lags in the lag-4 seasonal AR term and the rst and third (but not the second) term in the lag-12 seasonal MA term, we would type
. arima DS4S12.sales, ar(1) mar(1 2, 4) mar(1, 12) ma(1) mma(1, 4) mma(1 3, 12)
However, models with multiple seasonal terms can be difcult to t. Usually, one seasonal factor with just one or two AR or MA terms is adequate.
ARMAX models
Thus far all our examples have been pure ARIMA models in which the dependent variable was modeled solely as a function of its past values and disturbances. Also, arima can t ARMAX models, which model the dependent variable in terms of a linear combination of independent variables, as well as an ARMA disturbance process. The prais command, for example, allows you to control for only AR(1) disturbances, whereas arima allows you to control for a much richer dynamic error structure. arima allows for both nonseasonal and seasonal ARMA components in the disturbances.
Example 4: ARMAX model

For a simple example of a model including covariates, we can estimate an update of Friedman and Meiselmans (1963) equation representing the quantity theory of money. They postulate a straightforward relationship between personal-consumption expenditures (consump) and the money supply as measured by M2 (m2). consumpt = 0 + 1 m2t + t Friedman and Meiselman t the model over a period ending in 1956; we will ret the model over the period 1959q1 through 1981q4. We restrict our attention to the period prior to 1982 because the Federal Reserve manipulated the money supply extensively in the later 1980s to control ination, and the relationship between consumption and the money supply becomes much more complex during the later part of the decade. To demonstrate arima, we will include both an autoregressive term and a moving-average term for the disturbances in the model; the original estimates included neither. Thus we model the disturbance of the structural equation as t = t1 + t1 + t As per the original authors, the relationship is estimated on seasonally adjusted data, so there is no need to include seasonal effects explicitly. Obtaining seasonally unadjusted data and simultaneously modeling the structural and seasonal effects might be preferable. We will restrict the estimation to the desired sample by using the tin() function in an if expression; see [D] functions. By leaving the rst argument of tin() blank, we are including all available data through the second date (1981q4). We t the model by typing

. use http://www.stata-press.com/data/r12/friedman2, clear . arima consump m2 if tin(, 1981q4), ar(1) ma(1) (setting optimization to BHHH) Iteration 0: log likelihood = -344.67575 Iteration 1: log likelihood = -341.57248 (output omitted ) Iteration 10: log likelihood = -340.50774 ARIMA regression Sample: 1959q1 - 1981q4 Number of obs Wald chi2(3) Log likelihood = -340.5077 Prob > chi2 OPG Std. Err.
87
= = =
92 4394.80 0.0000
consump consump m2 _cons ARMA ar L1. ma L1. /sigma
Coef.
P>|z|
1.122029 -36.09872
.0363563 56.56703
30.86 -0.64
0.000 0.523
1.050772 -146.9681
1.193286 74.77062
.9348486
.0411323
22.73
0.000
.8542308
1.015467
.3090592 9.655308
.0885883 .5635157
3.49 17.13
0.000 0.000
.1354293 8.550837
.4826891 10.75978
We nd a relatively small money velocity with respect to consumption (1.122) over this period, although consumption is only one facet of the income velocity. We also note a very large rst-order autocorrelation in the disturbances, as well as a statistically signicant rst-order moving average. We might be concerned that our specication has led to disturbances that are heteroskedastic or non-Gaussian. We ret the model by using the vce(robust) option.
88

. arima consump m2 if tin(, 1981q4), ar(1) ma(1) vce(robust) (setting optimization to BHHH) Iteration 0: log pseudolikelihood = -344.67575 Iteration 1: log pseudolikelihood = -341.57248 (output omitted ) Iteration 10: log pseudolikelihood = -340.50774 ARIMA regression Sample: 1959q1 - 1981q4 Number of obs Wald chi2(3) Log pseudolikelihood = -340.5077 Prob > chi2 Semirobust Std. Err.
= = =
92 1176.26 0.0000
consump consump m2 _cons ARMA ar L1. ma L1. /sigma
Coef.
P>|z|
1.122029 -36.09872
.0433302 28.10478
25.89 -1.28
0.000 0.199
1.037103 -91.18308
1.206954 18.98563
.9348486
.0493428
18.95
0.000
.8381385
1.031559
.3090592 9.655308
.1605359 1.082639
1.93 8.92
0.054 0.000
-.0055854 7.533375
.6237038 11.77724
We note a substantial increase in the estimated standard errors, and our once clearly signicant moving-average term is now only marginally signicant.
Dynamic forecasting
Another feature of the arima command is the ability to use predict afterward to make dynamic forecasts. Suppose that we wish to t the regression model
yt = 0 + 1 xt + yt1 +
by using a sample of data from t = 1 . . . T and make forecasts beginning at time f . If we use regress or prais to t the model, then we can use predict to make one-step-ahead forecasts. That is, predict will compute
yf = 0 + 1 xf + yf 1
Most importantly, here predict will use the actual value of y at period f 1 in computing the forecast for time f . Thus, if we use regress or prais, we cannot make forecasts for any periods beyond f = T + 1 unless we have observed values for y for those periods. If we instead t our model with arima, then predict can produce dynamic forecasts by using the Kalman lter. If we use the dynamic(f ) option, then for period f predict will compute
yf = 0 + 1 xf + yf 1
89
by using the observed value of yf 1 just as predict after regress or prais. However, for period f + 1 predict newvar, dynamic(f ) will compute
yf +1 = 0 + 1 xf +1 + yf
using the predicted value of yf instead of the observed value. Similarly, the period f + 2 forecast will be yf +2 = 0 + 1 xf +2 + yf +1 Of course, because our model includes the regressor xt , we can make forecasts only through periods for which we have observations on xt . However, for pure ARIMA models, we can compute dynamic forecasts as far beyond the nal period of our dataset as desired. For more information on predict after arima, see [TS] arima postestimation.
Saved results
arima saves the following in e():
Scalars e(N) e(N gaps) e(k) e(k eq) e(k eq model) e(k dv) e(k1) e(df m) e(ll) e(sigma) e(chi2) e(p) e(tmin) e(tmax) e(ar max) e(ma max) e(rank) e(ic) e(rc) e(converged) number of observations number of gaps number of parameters number of equations in e(b) number of equations in overall model test number of dependent variables number of variables in rst equation model degrees of freedom log likelihood sigma
2
signicance minimum time maximum time maximum AR lag maximum MA lag rank of e(V) number of iterations return code 1 if converged, 0 otherwise
90

Macros e(cmd) e(cmdline) e(depvar) e(covariates) e(eqnames) e(wtype) e(wexp) e(title) e(tmins) e(tmaxs) e(chi2type) e(vce) e(vcetype) e(ma) e(ar) e(mari) e(mmai) e(seasons) e(unsta) e(opt) e(ml method) e(user) e(technique) e(tech steps) e(properties) e(estat cmd) e(predict) e(marginsok) e(marginsnotok) Matrices e(b) e(Cns) e(ilog) e(gradient) e(V) e(V modelbased) Functions e(sample)
arima command as typed name of dependent variable list of covariates names of equations weight type weight expression title in estimation output formatted minimum time formatted maximum time Wald; type of model 2 test vcetype specied in vce() title used to label Std. Err. lags for moving-average terms lags for autoregressive terms multiplicative AR terms and lag i=1... (# seasonal AR terms) multiplicative MA terms and lag i=1... (# seasonal MA terms) seasonal lags in model unstationary or blank type of optimization type of ml method name of likelihood-evaluator program maximization technique number of iterations performed before switching techniques b V program used to implement estat program used to implement predict predictions allowed by margins predictions disallowed by margins coefcient vector constraints matrix iteration log (up to 20 iterations) gradient vector variancecovariance matrix of the estimators model-based variance marks estimation sample

arima is implemented as an ado-le. Estimation is by maximum likelihood using the Kalman lter via the prediction error decomposition; see Hamilton (1994), Gourieroux and Monfort (1997), or, in particular, Harvey (1989). Any of these sources will serve as excellent background for the tting of these models with the state-space form; each source also provides considerable detail on the method outlined below.
91
Methods and formulas are presented under the following headings:

ARIMA model Kalman lter equations Kalman lter or state-space representation of the ARIMA model Kalman lter recursions Kalman lter initial conditions Likelihood from prediction error decomposition Missing data
ARIMA model
The model to be t is
yt = xt + t
p q
t =
i=1
i ti +
j=1
tj
which can be written as the single equation

p q
yt = x t +
i=1
i (yti xti ) +
j=1
tj
Some of the s and s may be constrained to zero or, for multiplicative seasonal models, the products of other parameters.
Kalman lter equations

We will roughly follow Hamiltons (1994) notation and write the Kalman lter t = Ft1 + vt
(state equation) (observation equation)
y t = A x t + H t + wt
and
vt wt
0,
Q 0 0 R
We maintain the standard Kalman lter matrix and vector notation, although for univariate models yt , wt , and R are scalars.
Kalman lter or state-space representation of the ARIMA model

A univariate ARIMA model can be cast in state-space form by dening the Kalman lter matrices as follows (see Hamilton [1994], or Gourieroux and Monfort [1997], for details):
92
1 1 F = 0 0
t1
2 0 1 0
. . . p1 ... 0 ... 0 ... 1
p 0 0 0
0 ... vt = ... ... 0 A = H = [ 1 1 wt = 0

The Kalman lter representation does not require the moving-average terms to be invertible.
. . . q ]
Kalman lter recursions

To demonstrate how missing data are handled, the updating recursions for the Kalman lter will be left in two steps. Writing the updating equations as one step using the gain matrix K is common. We will provide the updating equations with little justication; see the sources listed above for details. As a linear combination of a vector of random variables, the state t can be updated to its expected value on the basis of the prior state as t|t1 = Ft1 + vt1 This state is a quadratic form that has the covariance matrix
(1)
Pt|t1 = FPt1 F + Q
The estimator of yt is
(2)
yt|t1 = xt + H t|t1
which implies an innovation or prediction error t = yt yt|t1 This value or vector has mean squared error (MSE)
Mt = H Pt|t1 H + R
Now the expected value of t conditional on a realization of yt is t = t|t1 + Pt|t1 HM1 t t with MSE
(3) (4)
Pt = Pt|t1 Pt|t1 HM1 H Pt|t1 t
This expression gives the full set of Kalman lter recursions.
93
Kalman lter initial conditions

When the series is stationary, conditional on xt , the initial conditions for the lter can be considered a random draw from the stationary distribution of the state equation. The initial values of the state and the state MSE are the expected values from this stationary distribution. For an ARIMA model, these can be written as
1|0 = 0
and vec(P1|0 ) = (Ir2 F F)1 vec(Q) where vec() is an operator representing the column matrix resulting from stacking each successive column of the target matrix. If the series is not stationary, the initial state conditions do not constitute a random draw from a stationary distribution, and some other values must be chosen. Hamilton (1994) suggests that they be chosen based on prior expectations, whereas Harvey suggests a diffuse and improper prior having a state vector of 0 and an innite variance. This method corresponds to P1|0 with diagonal elements of . Stata allows either approach to be taken for nonstationary seriesinitial priors may be specied with state0() and p0(), and a diffuse prior may be specied with diffuse.
Likelihood from prediction error decomposition

Given the outputs from the Kalman lter recursions and assuming that the state and observation vectors are Gaussian, the likelihood for the state-space model follows directly from the resulting multivariate normal in the predicted innovations. The log likelihood for observation t is
lnLt =
1 ln(2) + ln(|Mt |) t M1 t t 2
This command supports the Huber/White/sandwich estimator of the variance using vce(robust). See [P] robust, particularly Maximum likelihood estimators and Methods and formulas.
Missing data
Missing data, whether a missing dependent variable yt , one or more missing covariates xt , or completely missing observations, are handled by continuing the state-updating equations without any contribution from the data; see Harvey (1989 and 1993). That is, (1) and (2) are iterated for every missing observation, whereas (3) and (4) are ignored. Thus, for observations with missing data, t = t|t1 and Pt = Pt|t1 . Without any information from the sample, this effectively assumes that the prediction error for the missing observations is 0. Other methods of handling missing data on the basis of the EM algorithm have been suggested, for example, Shumway (1984, 1988).
94
George Edward Pelham Box (1919 ) was born in Kent, England, and earned degrees in statistics at the University of London. After work in the chemical industry, he taught and researched at Princeton and the University of Wisconsin. His many major contributions to statistics include papers and books in Bayesian inference, robustness (a term he introduced to statistics), modeling strategy, experimental design and response surfaces, time-series analysis, distribution theory, transformations, and nonlinear estimation. Gwilym Meirion Jenkins (19331982) was a British mathematician and statistician who spent his career in industry and academia, working for extended periods at Imperial College London and the University of Lancaster before running his own company. His interests were centered on time series and he collaborated with G. E. P. Box on what are often called BoxJenkins models. The last years of Jenkins life were marked by a slowly losing battle against Hodgkins disease.
References
Ansley, C. F., and R. Kohn. 1985. Estimation, ltering, and smoothing in state space models with incompletely specied initial conditions. Annals of Statistics 13: 12861316. Ansley, C. F., and P. Newbold. 1980. Finite sample properties of estimators for autoregressive moving average models. Journal of Econometrics 13: 159183. Baum, C. F. 2000. sts15: Tests for stationarity of a time series. Stata Technical Bulletin 57: 3639. Reprinted in Stata Technical Bulletin Reprints, vol. 10, pp. 356360. College Station, TX: Stata Press. Baum, C. F., and T. Room. 2001. sts18: A test for long-range dependence in a time series. Stata Technical Bulletin 60: 3739. Reprinted in Stata Technical Bulletin Reprints, vol. 10, pp. 370373. College Station, TX: Stata Press. Baum, C. F., and R. Sperling. 2000. sts15.1: Tests for stationarity of a time series: Update. Stata Technical Bulletin 58: 3536. Reprinted in Stata Technical Bulletin Reprints, vol. 10, pp. 360362. College Station, TX: Stata Press. Baum, C. F., and V. L. Wiggins. 2000. sts16: Tests for long memory in a time series. Stata Technical Bulletin 57: 3944. Reprinted in Stata Technical Bulletin Reprints, vol. 10, pp. 362368. College Station, TX: Stata Press. Berndt, E. K., B. H. Hall, R. E. Hall, and J. A. Hausman. 1974. Estimation and inference in nonlinear structural models. Annals of Economic and Social Measurement 3/4: 653665. Bollerslev, T., R. F. Engle, and D. B. Nelson. 1994. ARCH models. In Handbook of Econometrics, Volume IV, ed. R. F. Engle and D. L. McFadden. New York: Elsevier. Box, G. E. P. 1983. Obituary: G. M. Jenkins, 19331982. Journal of the Royal Statistical Society, Series A 146: 205206. Box, G. E. P., G. M. Jenkins, and G. C. Reinsel. 2008. Time Series Analysis: Forecasting and Control. 4th ed. Hoboken, NJ: Wiley. Chateld, C. 2004. The Analysis of Time Series: An Introduction. 6th ed. Boca Raton, FL: Chapman & Hall/CRC. David, J. S. 1999. sts14: Bivariate Granger causality test. Stata Technical Bulletin 51: 4041. Reprinted in Stata Technical Bulletin Reprints, vol. 9, pp. 350351. College Station, TX: Stata Press. Davidson, R., and J. G. MacKinnon. 1993. Estimation and Inference in Econometrics. New York: Oxford University Press. DeGroot, M. H. 1987. A conversation with George Box. Statistical Science 2: 239258. Diggle, P. J. 1990. Time Series: A Biostatistical Introduction. Oxford: Oxford University Press. Enders, W. 2004. Applied Econometric Time Series. 2nd ed. New York: Wiley. Friedman, M., and D. Meiselman. 1963. The relative stability of monetary velocity and the investment multiplier in the United States, 18971958. In Stabilization Policies, Commission on Money and Credit, 123126. Englewood Cliffs, NJ: Prentice Hall. Gourieroux, C., and A. Monfort. 1997. Time Series and Dynamic Models. Trans. ed. G. M. Gallo. Cambridge: Cambridge University Press. Greene, W. H. 2012. Econometric Analysis. 7th ed. Upper Saddle River, NJ: Prentice Hall.

Hamilton, J. D. 1994. Time Series Analysis. Princeton: Princeton University Press.
95
Harvey, A. C. 1989. Forecasting, Structural Time Series Models and the Kalman Filter. Cambridge: Cambridge University Press. . 1993. Time Series Models. 2nd ed. Cambridge, MA: MIT Press. Hipel, K. W., and A. I. McLeod. 1994. Time Series Modelling of Water Resources and Environmental Systems. Amsterdam: Elsevier. Kalman, R. E. 1960. A new approach to linear ltering and prediction problems. Transactions of the ASMEJournal of Basic Engineering, Series D 82: 3545. McDowell, A. W. 2002. From the help desk: Transfer functions. Stata Journal 2: 7185. . 2004. From the help desk: Polynomial distributed lag models. Stata Journal 4: 180189. Newton, H. J. 1988. TIMESLAB: A Time Series Analysis Laboratory. Belmont, CA: Wadsworth. Press, W. H., S. A. Teukolsky, W. T. Vetterling, and B. P. Flannery. 2007. Numerical Recipes in C: The Art of Scientic Computing. 3rd ed. Cambridge: Cambridge University Press. Shumway, R. H. 1984. Some applications of the EM algorithm to analyzing incomplete time series data. In Time Series Analysis of Irregularly Observed Data, ed. E. Parzen, 290324. New York: Springer. . 1988. Applied Statistical Time Series Analysis. Upper Saddle River, NJ: Prentice Hall.
Also see
[TS] arima postestimation Postestimation tools for arima [TS] tsset Declare data to be time-series data [TS] arch Autoregressive conditional heteroskedasticity (ARCH) family of estimators [TS] dfactor Dynamic-factor models [TS] mgarch Multivariate GARCH models [TS] prais Prais Winsten and Cochrane Orcutt regression [TS] sspace State-space models [TS] ucm Unobserved-components model [R] regress Linear regression [U] 20 Estimation and postestimation commands
Title
arima postestimation Postestimation tools for arima
Description
The following postestimation command is of special interest after arima: Command psdensity Description estimate the spectral density
For information about psdensity, see [TS] psdensity. The following standard postestimation commands are also available: Command estat estimates lincom Description
Syntax for predict

predict statistic
Main
type
newvar
if
in
, statistic options
Description predicted values for mean equationthe differenced series; the default standard error of the linear prediction predicted values for the mean equation in y the undifferenced series mean squared error of the predicted values residuals or predicted innovations residuals or predicted innovations in y , reversing any time-series operators
xb stdp y mse residuals yresiduals
These statistics are available both in and out of sample; type predict . . . if e(sample) the estimation sample. Predictions are not available for conditional ARIMA models t to panel data.
. . . if wanted only for
96
97
options
Options
Description how to handle the lags of yt set starting point for the recursions to time constant calculate considering the structural component only
dynamic(time constant) t0(time constant) structural
time constant is a # or a time literal, such as td(1jan1995) or tq(1995q1); see Conveniently typing SIF values in [D] datetime.
Menu
Statistics
>
Postestimation
>
Options for predict

Five statistics can be computed using predict after arima: the predictions from the model (the default also given by xb), the predictions after reversing any time-series operators applied to the dependent variable (y), the MSE of xb (mse), the predictions of residuals or innovations (residual), and the predicted residuals or innovations in terms of y (yresiduals). Given the dynamic nature of the ARMA component and because the dependent variable might be differenced, there are other ways of computing each. We can use all the data on the dependent variable that is available right up to the time of each prediction (the default, which is often called a one-step prediction), or we can use the data up to a particular time, after which the predicted value of the dependent variable is used recursively to make later predictions (dynamic()). Either way, we can consider or ignore the ARMA disturbance component (the component is considered by default and is ignored if you specify structural). All calculations can be made in or out of sample.
Main
xb, the default, calculates the predictions from the model. If D.depvar is the dependent variable, these predictions are of D.depvar and not of depvar itself. stdp calculates the standard error of the linear prediction xb. stdp does not include the variation arising from the disturbance equation; use mse to calculate standard errors and condence bands around the predicted values. y species that predictions of depvar be made, even if the model was specied in terms of, say, D.depvar. mse calculates the MSE of the predictions. residuals calculates the residuals. If no other options are specied, these are the predicted innovations t ; that is, they include the ARMA component. If structural is specied, these are the residuals t from the structural equation; see structural below. yresiduals calculates the residuals in terms of depvar, even if the model was specied in terms of, say, D.depvar. As with residuals, the yresiduals are computed from the model, including any ARMA component. If structural is specied, any ARMA component is ignored, and yresiduals are the residuals from the structural equation; see structural below.
98
Options
dynamic(time constant) species how lags of yt in the model are to be handled. If dynamic() is not specied, actual values are used everywhere that lagged values of yt appear in the model to produce one-step-ahead forecasts. dynamic(time constant) produces dynamic (also known as recursive) forecasts. time constant species when the forecast is to switch from one step ahead to dynamic. In dynamic forecasts, references to yt evaluate to the prediction of yt for all periods at or after time constant; they evaluate to the actual value of yt for all prior periods. For example, dynamic(10) would calculate predictions in which any reference to yt with t < 10 evaluates to the actual value of yt and any reference to yt with t 10 evaluates to the prediction of yt . This means that one-step-ahead predictions are calculated for t < 10 and dynamic predictions thereafter. Depending on the lag structure of the model, the dynamic predictions might still refer some actual values of yt . You may also specify dynamic(.) to have predict automatically switch from one-step-ahead to dynamic predictions at p + q , where p is the maximum AR lag and q is the maximum MA lag. t0(time constant) species the starting point for the recursions to compute the predicted statistics; disturbances are assumed to be 0 for t < t0(). The default is to set t0() to the minimum t observed in the estimation sample, meaning that observations before that are assumed to have disturbances of 0. t0() is irrelevant if structural is specied because then all observations are assumed to have disturbances of 0. t0(5) would begin recursions at t = 5. If the data were quarterly, you might instead type t0(tq(1961q2)) to obtain the same result. The ARMA component of ARIMA models is recursive and depends on the starting point of the predictions. This includes one-step-ahead predictions. structural species that the calculation be made considering the structural component only, ignoring the ARMA terms, producing the steady-state equilibrium predictions.
Remarks
We assume that you have already read [TS] arima. See [TS] psdensity for an introduction to estimating spectral densities using the parameters estimated by arima. In this entry, we illustrate some of the features of predict after tting ARIMA, ARMAX, and other dynamic models by using arima. In example 2 of [TS] arima, we t the model
ln(wpit ) = 0 + 1 { ln(wpit1 ) 0 } + 1
by typing
. use http://www.stata-press.com/data/r12/wpi1 . arima D.ln_wpi, ar(1) ma(1 4) (output omitted )
t1
+ 4
t4
If we use the command

. predict xb, xb
then Stata computes xbt as
xbt = 0 + 1 { ln(wpit1 ) 0 } + 1
t1
+ 4
t4
99
where
tj
ln(wpitj ) xbtj 0
tj >0 otherwise
meaning that predict newvar, xb calculates predictions by using the metric of the dependent variable. In this example, the dependent variable represented changes in ln(wpit ), and so the predictions are likewise for changes in that variable. If we instead use
. predict y, y
Stata computes yt as yt = xbt + ln(wpit1 ) so that yt represents the predicted levels of ln(wpit ). In general, predict newvar, y will reverse any time-series operators applied to the dependent variable during estimation. If we want to ignore the ARMA error components when making predictions, we use the structural option,
. predict xbs, xb structural
which generates xbst = 0 because there are no regressors in this model, and
. predict ys, y structural
generates yst = 0 + ln(wpit1 )
Example 1: Dynamic forecasts

An attractive feature of the arima command is the ability to make dynamic forecasts. In example 4 of [TS] arima, we t the model
consumpt = 0 + 1 m2t + t t = t1 +
t1
First, we ret the model by using data up through the rst quarter of 1978, and then we will evaluate the one-step-ahead and dynamic forecasts.
. use http://www.stata-press.com/data/r12/friedman2 . keep if time<=tq(1981q4) (67 observations deleted) . arima consump m2 if tin(, 1978q1), ar(1) ma(1) (output omitted )
To make one-step-ahead forecasts, we type

. predict chat, y (52 missing values generated)
(Because our dependent variable contained no time-series operators, we could have instead used predict chat, xb and accomplished the same thing.) We will also make dynamic forecasts, switching from observed values of consump to forecasted values at the rst quarter of 1978:
. predict chatdy, dynamic(tq(1978q1)) y (52 missing values generated)
The following graph compares the forecasted values to the observed values for the rst few years following the estimation sample:
100
1200
Billions of dollars 1400 1600 1800
2000
Personal consumption
1977q1
1978q1
1979q1 1980q1 Quarter
1981q1
1982q1
Observed Dynamic forecast (1978q1)
Onestepahead forecast
The one-step-ahead forecasts never deviate far from the observed values, though over time the dynamic forecasts have larger errors. To understand why that is the case, rewrite the model as
consumpt = 0 + 1 m2t + t1 +
t1
t t1
= 0 + 1 m2t + consumpt1 0 1 m2t1 +
This form shows that the forecasted value of consumption at time t depends on the value of consumption at time t 1. When making the one-step-ahead forecast for period t, we know the actual value of consumption at time t 1. On the other hand, with the dynamic(tq(1978q1)) option, the forecasted value of consumption for period 1978q1 is based on the observed value of consumption in period 1977q4, but the forecast for 1978q2 is based on the forecast value for 1978q1, the forecast for 1978q3 is based on the forecast value for 1978q2, and so on. Thus, with dynamic forecasts, prior forecast errors accumulate over time. The following graph illustrates this effect.
Forecast error
0 200 1978q1 Forecast Actual 150 100 50
1979q1
1980q1 Quarter
1981q1
1982q1
Onestepahead forecast
Dynamic forecast (1978q1)
101

Also see
[TS] arima ARIMA, ARMAX, and other dynamic regression models [TS] psdensity Parametric spectral density estimation after arima, arma, and ucm [U] 20 Estimation and postestimation commands
Title
corrgram Tabulate and graph autocorrelations
Syntax
Autocorrelations, partial autocorrelations, and portmanteau (Q) statistics corrgram varname if in , corrgram options
Graph autocorrelations with condence intervals ac varname if in , ac options
Graph partial autocorrelations with condence intervals pac varname corrgram options
Main
if
in
, pac options Description calculate # autocorrelations suppress character-based plots calculate partial autocorrelations by using YuleWalker equations
lags(#) noplot yw
ac options
Main
Description calculate # autocorrelations generate a variable to hold the autocorrelations set condence level; default is level(95) calculate autocorrelation by using Fourier transforms change look of dropped lines change look of markers (color, size, etc.) add marker labels; change look or position affect rendition of the condence bands add other plots to the generated graph any options other than by() documented in [G-3] twoway options
lags(#) generate(newvar) level(#) fft

Plot
line options marker options marker label options

CI plot
ciopts(area options)
Add plots
addplot(plot) twoway options
Y axis, X axis, Titles, Legend, Overall
102
103
pac options
Main
Description calculate # partial autocorrelations generate a variable to hold the partial autocorrelations calculate partial autocorrelations by using YuleWalker equations set condence level; default is level(95) change look of dropped lines change look of markers (color, size, etc.) add marker labels; change look or position affect rendition of the condence bands include standardized residual variances in graph affect rendition of the plotted standardized residual variances (SRVs) add other plots to the generated graph any options other than by() documented in [G-3] twoway options
lags(#) generate(newvar) yw level(#)

Plot
line options marker options marker label options

CI plot
SRV plot
srv srvopts(marker options)

Add plots
You must tsset your data before using corrgram, ac, or pac; see [TS] tsset. Also, the time series must be dense (nonmissing and no gaps in the time variable) in the sample if you specify the fft option. varname may contain time-series operators; see [U] 11.4.4 Time-series varlists.
Menu
corrgram
Statistics
>
Time series
>
Graphs
>
Autocorrelations & partial autocorrelations
ac
Statistics
>
Time series
>
Graphs
>
Correlogram (ac)
pac
Statistics
>
Time series
>
Graphs
>
Partial correlogram (pac)
Description
corrgram produces a table of the autocorrelations, partial autocorrelations, and portmanteau (Q) statistics. It also displays a character-based plot of the autocorrelations and partial autocorrelations. See [TS] wntestq for more information on the Q statistic. ac produces a correlogram (a graph of autocorrelations) with pointwise condence intervals that is based on Bartletts formula for MA(q) processes.
104
pac produces a partial correlogram (a graph of partial autocorrelations) with condence intervals calculated using a standard error of 1/ n. The residual variances for each lag may optionally be included on the graph.
Options for corrgram

Main
lags(#) species the number of autocorrelations to calculate. The default is to use min( n/2 2, 40), where n/2 is the greatest integer less than or equal to n/2. noplot prevents the character-based plots from being in the listed table of autocorrelations and partial autocorrelations. yw species that the partial autocorrelations be calculated using the YuleWalker equations instead of using the default regression-based technique. yw cannot be used if srv is used.
Options for ac and pac

Main
lags(#) species the number of autocorrelations to calculate. The default is to use min( n/2 2, 40), where n/2 is the greatest integer less than or equal to n/2. generate(newvar) species a new variable to contain the autocorrelation (ac command) or partial autocorrelation (pac command) values. This option is required if the nograph option is used. nograph (implied when using generate() in the dialog box) prevents ac and pac from constructing a graph. This option requires the generate() option. yw (pac only) species that the partial autocorrelations be calculated using the YuleWalker equations instead of using the default regression-based technique. yw cannot be used if srv is used. level(#) species the condence level, as a percentage, for the condence bands in the ac or pac graph. The default is level(95) or as set by set level; see [R] level. fft (ac only) species that the autocorrelations be calculated using two Fourier transforms. This technique can be faster than simply iterating over the requested number of lags.
Plot
line options, marker options, and marker label options affect the rendition of the plotted autocorrelations (with ac) or partial autocorrelations (with pac). line options specify the look of the dropped lines, including pattern, width, and color; see [G-3] line options. marker options specify the look of markers. This look includes the marker symbol, the marker size, and its color and outline; see [G-3] marker options. marker label options specify if and how the markers are to be labeled; see [G-3] marker label options.
CI plot
ciopts(area options) affects the rendition of the condence bands; see [G-3] area options.
105
SRV plot
srv (pac only) species that the standardized residual variances be plotted with the partial autocorrelations. srv cannot be used if yw is used. srvopts(marker options) (pac only) affects the rendition of the plotted standardized residual variances; see [G-3] marker options. This option implies the srv option.
Add plots
addplot(plot) adds specied plots to the generated graph; see [G-3] addplot option.
twoway options are any of the options documented in [G-3] twoway options, excluding by(). These include options for titling the graph (see [G-3] title options) and for saving the graph to disk (see [G-3] saving option).
Remarks
corrgram tabulates autocorrelations, partial autocorrelations, and portmanteau (Q) statistics and plots the autocorrelations and partial autocorrelations. The Q statistics are the same as those produced by [TS] wntestq. ac produces graphs of the autocorrelations, and pac produces graphs of the partial autocorrelations.
Example 1
Here we use the international airline passengers dataset (Box, Jenkins, and Reinsel 2008, Series G). This dataset has 144 observations on the monthly number of international airline passengers from 1949 through 1960. We can list the autocorrelations and partial autocorrelations by using corrgram.
. use http://www.stata-press.com/data/r12/air2 (TIMESLAB: Airline passengers) . corrgram air, lags(20) -1 0 1 -1 0 1 LAG AC PAC Q Prob>Q [Autocorrelation] [Partial Autocor] 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 0.9480 0.8756 0.8067 0.7526 0.7138 0.6817 0.6629 0.6556 0.6709 0.7027 0.7432 0.7604 0.7127 0.6463 0.5859 0.5380 0.4997 0.4687 0.4499 0.4416 0.9589 -0.3298 0.2018 0.1450 0.2585 -0.0269 0.2043 0.1561 0.5686 0.2926 0.8402 0.6127 -0.6660 -0.3846 0.0787 -0.0266 -0.0581 -0.0435 0.2773 -0.0405 132.14 245.65 342.67 427.74 504.8 575.6 643.04 709.48 779.59 857.07 944.39 1036.5 1118 1185.6 1241.5 1289 1330.4 1367 1401.1 1434.1 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000
106
We can use ac to produce a graph of the autocorrelations.

. ac air, lags(20)
1.00 1.00 Autocorrelations of air 0.50 0.00 0.50 0
10 Lag
15
20
The data probably have a trend component as well as a seasonal component. First-differencing will mitigate the effects of the trend, and seasonal differencing will help control for seasonality. To accomplish this goal, we can use Statas time-series operators. Here we graph the partial autocorrelations after controlling for trends and seasonality. We also use srv to include the standardized residual variances.
. pac DS12.air, lags(20) srv
Partial autocorrelations of DS12.air 0.50 0.00 0.50 1.00
10 Lag
15
20
95% CI Partial autocorrelations of DS12.air Standardized variances

95% Confidence bands [se = 1/sqrt(n)]
See [U] 11.4.4 Time-series varlists for more information about time-series operators.
107
Saved results
corrgram saves the following in r():
Scalars r(lags) r(ac#) r(pac#) r(q#) Matrices r(AC) r(PAC) r(Q) number of lags AC for lag # PAC for lag # Q for lag # vector of autocorrelations vector of partial autocorrelations vector of Q statistics

corrgram, ac, and pac are implemented as ado-les. Box, Jenkins, and Reinsel (2008, sec. 2.1.4); Newton (1988); Chateld (2004); and Hamilton (1994) provide excellent descriptions of correlograms. Newton (1988) also discusses the calculation of the various quantities. The autocovariance function for a time series x1 , x2 , . . . , xn is dened for |v| < n as
R(v) =
1 n
n|v|
(xi x)(xi+v x)
i=1
where x is the sample mean, and the autocorrelation function is then dened as
v =
R(v) R(0)
The variance of v is given by Bartletts formula for MA(q) processes. From Brockwell and Davis (2002, 94), we have v=1 1/n v1 Var(v ) = 1 2 v>1 (i) n 1+2
i=1
The partial autocorrelation at lag v measures the correlation between xt and xt+v after the effects of xt+1 , . . . , xt+v1 have been removed. By default, corrgram and pac use a regression-based method to estimate it. We run an OLS regression of xt on xt1 , . . . , xtv and a constant term. The estimated coefcient on xtv is our estimate of the v th partial autocorrelation. The residual variance is the estimated variance of that regression, which we then standardize by dividing by R(0). If the yw option is specied, corrgram and pac use the YuleWalker equations to estimate the partial autocorrelations. Per Enders (2010, 6667), let vv denote the v th partial autocorrelation coefcient. We then have 11 = 1 and for v > 1
v1
v vv =
j=1
v1,j vj
v1
1
j=1
v1,j j
108
and
vj = v1,j vv v1,vj
j = 1, 2, . . . , v 1
Unlike the regression-based method, the YuleWalker equations-based method ensures that the rstsample partial autocorrelation equal the rst-sample autocorrelation coefcient, as must be true in the population; see Greene (2008, 725). McCullough (1998) discusses other methods of estimating vv ; he nds that relative to other methods, such as linear regression, the YuleWalker equations-based method performs poorly, in part because it is susceptible to numerical error. Box, Jenkins, and Reinsel (2008, 69) also caution against using the YuleWalker equations-based method, especially with data that are nearly nonstationary.
Acknowledgment
The ac and pac commands are based on the ac and pac commands written by Sean Becketti (1992), a past editor of the Stata Technical Bulletin.
References
Becketti, S. 1992. sts1: Autocorrelation and partial autocorrelation graphs. Stata Technical Bulletin 5: 2728. Reprinted in Stata Technical Bulletin Reprints, vol. 1, pp. 221223. College Station, TX: Stata Press. Box, G. E. P., G. M. Jenkins, and G. C. Reinsel. 2008. Time Series Analysis: Forecasting and Control. 4th ed. Hoboken, NJ: Wiley. Brockwell, P. J., and R. A. Davis. 2002. Introduction to Times Series and Forecasting. 2nd ed. New York: Springer. Chateld, C. 2004. The Analysis of Time Series: An Introduction. 6th ed. Boca Raton, FL: Chapman & Hall/CRC. Enders, W. 2010. Applied Econometric Time Series. 3rd ed. New York: Wiley. Greene, W. H. 2008. Econometric Analysis. 6th ed. Upper Saddle River, NJ: Prentice Hall. Hamilton, J. D. 1994. Time Series Analysis. Princeton: Princeton University Press. McCullough, B. D. 1998. Algorithm choice for (partial) autocorrelation functions. Journal of Economic and Social Measurement 24: 265278. Newton, H. J. 1988. TIMESLAB: A Time Series Analysis Laboratory. Belmont, CA: Wadsworth.
Also see
[TS] tsset Declare data to be time-series data [TS] wntestq Portmanteau (Q) test for white noise [TS] pergram Periodogram
Title
cumsp Cumulative spectral distribution
Syntax
cumsp varname options
Main
if
in
, options
Description create newvar holding distribution values affect rendition of the plotted points connected by lines change look of markers (color, size, etc.) add marker labels; change look or position add other plots to the generated graph any options other than by() documented in [G-3] twoway options
generate(newvar)
Plot
cline options marker options marker label options

Add plots
You must tsset your data before using cumsp; see [TS] tsset. Also, the time series must be dense (nonmissing with no gaps in the time variable) in the sample specied. varname may contain time-series operators; see [U] 11.4.4 Time-series varlists.
Menu
Statistics
>
Time series
>
Graphs
>
Cumulative spectral distribution
Description
cumsp plots the cumulative sample spectral-distribution function evaluated at the natural frequencies for a (dense) time series.
Options

Main
generate(newvar) species a new variable to contain the estimated cumulative spectral-distribution values.
Plot
cline options affect the rendition of the plotted points connected by lines; see [G-3] cline options. marker options specify the look of markers. This look includes the marker symbol, the marker size, and its color and outline; see [G-3] marker options. marker label options specify if and how the markers are to be labeled; see [G-3] marker label options.
109
110
Add plots
addplot(plot) provides a way to add other plots to the generated graph; see [G-3] addplot option.
Remarks
Example 1
Here we use the international airline passengers dataset (Box, Jenkins, and Reinsel 2008, Series G). This dataset has 144 observations on the monthly number of international airline passengers from 1949 through 1960. In the cumulative sample spectral distribution function for these data, we also request a vertical line at frequency 1/12. Because the data are monthly, there will be a pronounced jump in the cumulative sample spectral-distribution plot at the 1/12 value if there is an annual cycle in the data.
. use http://www.stata-press.com/data/r12/air2 (TIMESLAB: Airline passengers) . cumsp air, xline(.083333333)
1.00
Airline Passengers (19491960) Cumulative spectral distribution 0.20 0.40 0.60 0.80
0.00
0.00
0.10
0.20 0.30 Frequency
0.40
0.50
Points evaluated at the natural frequencies
The cumulative sample spectral-distribution function clearly illustrates the annual cycle.

cumsp is implemented as an ado-le. A time series of interest is decomposed into a unique set of sinusoids of various frequencies and amplitudes.
0.00
0.20
0.40
0.60
0.80
1.00
Sample spectral distribution function
111
A plot of the sinusoidal amplitudes versus the frequencies for the sinusoidal decomposition of a time series gives us the spectral density of the time series. If we calculate the sinusoidal amplitudes for a discrete set of natural frequencies (1/n, 2/n, . . . , q/n), we obtain the periodogram. Let x(1), . . . , x(n) be a time series, and let k = (k 1)/n denote the natural frequencies for k = 1, . . . , n/2 + 1 where indicates the greatest integer function. Dene
2 Ck
1 = 2 n
x(t)e
t=1
2i(t1)k
2 A plot of nCk versus k is then called the periodogram. 2 The sample spectral density may then be dened as f (k ) = nCk .
If we let f (1 ), . . . , f (Q ) be the sample spectral density function of the time series evaluated at the frequencies j = (j 1)/Q for j = 1, . . . , Q and we let q = Q/2 + 1, then
k
f (j ) F (k ) =
i=1 q
f (j )
i=1
is the sample spectral-distribution function of the time series.
References
Box, G. E. P., G. M. Jenkins, and G. C. Reinsel. 2008. Time Series Analysis: Forecasting and Control. 4th ed. Hoboken, NJ: Wiley. Newton, H. J. 1988. TIMESLAB: A Time Series Analysis Laboratory. Belmont, CA: Wadsworth.
Also see
[TS] tsset Declare data to be time-series data [TS] corrgram Tabulate and graph autocorrelations [TS] pergram Periodogram
Title
dfactor Dynamic-factor models
Syntax
dfactor obs eq fac eq if in , options
obs eq species the equation for the observed dependent variables, and it has the form (depvars = (facvars = exog d exog f , sopts ) , sopts ) fac eq species the equation for the unobserved factors, and it has the form depvars are the observed dependent variables. exog d are the exogenous variables that enter into the equations for the observed dependent variables. (All factors are automatically entered into the equations for the observed dependent variables.) facvars are the names for the unobserved factors in the model. You may specify the names of existing variables in facvars, but dfactor treats them only as names and takes no notice that they are also variables. exog f are the exogenous variables that enter into the equations for the factors. options
Model
Description apply specied linear constraints vcetype may be oim or robust set condence level; default is level(95) do not display constraints control column formats, row spacing, and display of omitted variables and base and empty cells control the maximization process; seldom used specify initial values for the maximization process; seldom used specify the method for calculating the log likelihood; seldom used display legend instead of statistics
constraints(constraints)
SE/Robust
vce(vcetype)
Reporting

Maximization
maximize options from(matname)

Advanced
method(method) coeflegend
112
113
sopts
Model
Description
suppress constant term from the equation; allowed only in obs eq ar(numlist) autoregressive terms structure of autoregressive coefcient matrices arstructure(arstructure) covstructure(covstructure) covariance structure noconstant arstructure diagonal ltriangular general covstructure identity dscalar diagonal unstructured Description diagonal matrix; the default lower triangular matrix general matrix Description identity matrix diagonal scalar matrix diagonal matrix symmetric, positive-denite matrix
method hybrid dejong
Description use the stationary Kalman lter and the De Jong diffuse Kalman lter; the default use the stationary De Jong method and the De Jong diffuse Kalman lter
You must tsset your data before using dfactor; see [TS] tsset. exog d and exog f may contain factor variables; see [U] 11.4.3 Factor variables. depvars, exog d, and exog f may contain time-series operators; see [U] 11.4.4 Time-series varlists. by, statsby, and rolling are allowed; see [U] 11.1.10 Prex commands. coeflegend does not appear in the dialog box. See [U] 20 Estimation and postestimation commands for more capabilities of estimation commands.
Menu
Statistics
>
>
Dynamic-factor models
Description
dfactor estimates the parameters of dynamic-factor models by maximum likelihood. Dynamicfactor models are exible models for multivariate time series in which unobserved factors have a vector autoregressive structure, exogenous covariates are permitted in both the equations for the latent factors and the equations for observable dependent variables, and the disturbances in the equations for the dependent variables may be autocorrelated.
114
Options

Model
constraints(constraints) apply linear constraints. Some specications require linear constraints for parameter identication. noconstant suppresses the constant term. ar(numlist) species the vector autoregressive lag structure in the equation. By default, no lags are included in either the observable or the factor equations. arstructure(diagonal|ltriangular|general) species the structure of the matrices in the vector autoregressive lag structure. arstructure(diagonal) species the matrices to be diagonalseparate parameters for each lag, but no cross-equation autocorrelations. arstructure(diagonal) is the default for both the observable and the factor equations. arstructure(ltriangular) species the matrices to be lower triangularparameterizes a recursive, or Wold causal, structure. arstructure(general) species the matrices to be general matricesseparate parameters for each possible autocorrelation and cross-correlation. covstructure(identity | dscalar | diagonal | unstructured) species the covariance structure of the errors. covstructure(identity) species a covariance matrix equal to an identity matrix, and it is the default for the errors in the factor equations. covstructure(dscalar) species a covariance matrix equal to 2 times an identity matrix. covstructure(diagonal) species a diagonal covariance matrix, and it is the default for the errors in the observable variables. covstructure(unstructured) species a symmetric, positive-denite covariance matrix with parameters for all variances and covariances.
SE/Robust
vce(vcetype) species the estimator for the variancecovariance matrix of the estimator. vce(oim), the default, causes dfactor to use the observed information matrix estimator. vce(robust) causes dfactor to use the Huber/White/sandwich estimator.
Reporting
level(#); see [R] estimation options. nocnsreport; see [R] estimation options. display options: noomitted, vsquish, noemptycells, baselevels, allbaselevels, cformat(% fmt), pformat(% fmt), and sformat(% fmt); see [R] estimation options.
Maximization
maximize options: difficult, technique(algorithm spec), iterate(#), no log, trace, gradient, showstep, hessian, showtolerance, tolerance(#), ltolerance(#), nrtolerance(#), and from(matname); see [R] maximize for all options except from(), and see below for information on from(). These options are seldom used.
115
from(matname) species initial values for the maximization process. from(b0) causes dfactor to begin the maximization algorithm with the values in b0. b0 must be a row vector; the number of columns must equal the number of parameters in the model; and the values in b0 must be in the same order as the parameters in e(b). This option is seldom used.
Advanced
method(method) species how to compute the log likelihood. dfactor writes the model in statespace form and uses sspace to estimate the parameters; see [TS] sspace. method() offers two methods for dealing with some of the technical aspects of the state-space likelihood. This option is seldom used. method(hybrid), the default, uses the Kalman lter with model-based initial values when the model is stationary and uses the De Jong (1988, 1991) diffuse Kalman lter when the model is nonstationary. method(dejong) uses the De Jong (1988) method for estimating the initial values for the Kalman lter when the model is stationary and uses the De Jong (1988, 1991) diffuse Kalman lter when the model is nonstationary. The following option is available with dfactor but is not shown in the dialog box: coeflegend; see [R] estimation options.
Remarks
An introduction to dynamic-factor models Some examples
An introduction to dynamic-factor models

dfactor estimates the parameters of dynamic-factor models by maximum likelihood (ML). Dynamicfactor models represent a vector of k endogenous variables as linear functions of nf < k unobserved factors and some exogenous covariates. The unobserved factors and the disturbances in the equations for the observed variables may follow vector autoregressive structures. Dynamic-factor models have been developed and applied in macroeconomics; see Geweke (1977), Sargent and Sims (1977), Stock and Watson (1989, 1991), and Watson and Engle (1983). Dynamic-factor models are very exible; in a sense, they are too exible. Constraints must be imposed to identify the parameters of dynamic-factor and static-factor models. The parameters in the default specications in dfactor are identied, but other specications require additional restrictions. The factors are identied only up to a sign, which means that the coefcients on the unobserved factors can ip signs and still produce the same predictions and the same log likelihood. The exibility of the model sometimes produces convergence problems. dfactor is designed to handle cases in which the number of modeled endogenous variables, k , is small. The ML estimator is implemented by writing the model in state-space form and by using the Kalman lter to derive and implement the log likelihood. As k grows, the number of parameters quickly exceeds the number that can be estimated.
116
A dynamic-factor model has the form
yt = Pft + Qxt + ut ft = Rwt + A1 ft1 + A2 ft2 + + Atp ftp + t ut = C1 ut1 + C2 ut2 + + Ctq utq +
where the denitions are given in the following table: Item yt P ft Q xt ut R wt Ai t Ci
t t
Dimension k1 k nf nf 1 k nx nx 1 k1 nf nw nw 1 nf nf nf 1 kk k1
Denition vector of dependent variables matrix of parameters vector of unobservable factors matrix of parameters vector of exogenous variables vector of disturbances matrix of parameters vector of exogenous variables matrix of autocorrelation parameters for i {1, 2, . . . , p} vector of disturbances matrix of autocorrelation parameters for i {1, 2, . . . , q} vector of disturbances
By selecting different numbers of factors and lags, the dynamic-factor model encompasses the six models in the table below: Dynamic factors with vector autoregressive errors Dynamic factors Static factors with vector autoregressive errors Static factors Vector autoregressive errors Seemingly unrelated regression (DFAR) (DF) (SFAR) (SF) (VAR) (SUR)
nf nf nf nf nf nf
>0 >0 >0 >0 =0 =0
p>0 p>0 p=0 p=0 p=0 p=0
q q q q q q
>0 =0 >0 =0 >0 =0
In addition to the time-series models, dfactor can estimate the parameters of SF models and SUR models. dfactor can place equality constraints on the disturbance covariances, which sureg and var do not allow.
Some examples Example 1

Stock and Watson (1989, 1991) wrote a simple macroeconomic model as a DF model, estimated the parameters by ML, and extracted an economic indicator. In this example, we estimate the parameters of a DF model. In [TS] dfactor postestimation, we extend this example and extract an economic indicator for the differenced series. We have data on an industrial-production index, ipman; real disposable income, income; an aggregate weakly hours index, hours; and aggregate unemployment, unemp. We believe that these variables are rst-difference stationary. We model their rst-differences as linear functions of an unobserved factor that follows a second-order autoregressive process.

. use http://www.stata-press.com/data/r12/dfex (St. Louis Fed (FRED) macro data) . dfactor (D.(ipman income hours unemp) = , noconstant) (f = , ar(1/2)) searching for initial values .................. (setting technique to bhhh) Iteration 0: log likelihood = -675.18934 Iteration 1: log likelihood = -667.47825 (output omitted ) Refining estimates: Iteration 0: log likelihood = -662.09507 Iteration 1: log likelihood = -662.09507 Dynamic-factor model Sample: 1972m2 - 2008m11 Log likelihood = -662.09507 OIM Std. Err. Number of obs Wald chi2(6) Prob > chi2 = = = 442 751.95 0.0000
117
Coef. f f L1. L2. D.ipman f D.income f D.hours f D.unemp f Variance De.ipman De.income De.hours De.unemp -.0676016 .2177469 .0746338 .3502249 .2651932 .4820398
P>|z|
.0568663 .0624635
4.66 7.72
0.000 0.000
.1537372 .3596136
.3766491 .604466
.0287389
12.19
0.000
.2938976
.4065522
.0217319
3.43
0.001
.0320401
.1172276
.0186769
11.66
0.000
.1811407
.254353
.0071022
-9.52
0.000
-.0815217
-.0536816
.1383158 .2773808 .0911446 .0237232
.0167086 .0188302 .0080847 .0017932
8.28 14.73 11.27 13.23
0.000 0.000 0.000 0.000
.1055675 .2404743 .0752988 .0202086
.1710641 .3142873 .1069903 .0272378
Note: Tests of variances against zero are one sided, and the two-sided confidence intervals are truncated at zero.
For a discussion of the atypical iteration log, see example 1 in [TS] sspace. The header in the output describes the estimation sample, reports the log-likelihood function at the maximum, and gives the results of a Wald test against the null hypothesis that the coefcients on the independent variables, the factors, and the autoregressive components are all zero. In this example, the null hypothesis that all parameters except for the variance parameters are zero is rejected at all conventional levels. The results in the estimation table indicate that the unobserved factor is quite persistent and that it is a signicant predictor for each of the observed variables.
118
dfactor writes the DF model as a state-space model and uses the same methods as sspace to estimate the parameters. Example 5 in [TS] sspace writes the model considered here in state-space form and uses sspace to estimate the parameters.
Technical note
The signs of the coefcients on the unobserved factors are not identied. They are not identied because we can multiply the unobserved factors and the coefcients on the unobserved factors by negative one without changing the log likelihood or any of the model predictions. Altering either the starting values for the maximization process, the maximization technique() used, or the platform on which the command is run can cause the signs of the estimated coefcients on the unobserved factors to change. Changes in the signs of the estimated coefcients on the unobserved factors do not alter the implications of the model or the model predictions.
Example 2
Here we extend the previous example by allowing the errors in the equations for the observables to be autocorrelated. This extension yields a constrained VAR model with an unobserved autocorrelated factor. We estimate the parameters by typing

. dfactor (D.(ipman income hours unemp) = , noconstant ar(1)) (f = , ar(1/2)) searching for initial values .............. (setting technique to bhhh) Iteration 0: log likelihood = -654.19377 Iteration 1: log likelihood = -627.46986 (output omitted ) Refining estimates: Iteration 0: log likelihood = -610.28846 Iteration 1: log likelihood = -610.28846 Dynamic-factor model Sample: 1972m2 - 2008m11 Number of obs = 442 Wald chi2(10) = 990.91 Log likelihood = -610.28846 Prob > chi2 = 0.0000 OIM Std. Err.
119
Coef. f f L1. L2. De.ipman e.ipman LD. De.income e.income LD. De.hours e.hours LD. De.unemp e.unemp LD. D.ipman f D.income f D.hours f D.unemp f Variance De.ipman De.income De.hours De.unemp -.0711994 .1933165 .0760412 .3214972 .4058457 .3663499
P>|z|
.0906183 .0849584
4.48 4.31
0.000 0.000
.2282371 .1998344
.5834544 .5328654
-.2772149
.068808
-4.03
0.000
-.4120761
-.1423538
-.2213824
.0470578
-4.70
0.000
-.3136141
-.1291508
-.3969317
.0504256
-7.87
0.000
-.495764
-.2980994
-.1736835
.0532071
-3.26
0.001
-.2779675
-.0693995
.027982
11.49
0.000
.2666535
.3763408
.0173844
4.37
0.000
.0419684
.110114
.0172969
11.18
0.000
.1594151
.2272179
.0066553
-10.70
0.000
-.0842435
-.0581553
.1387909 .2636239 .0822919 .0218056
.0154558 .0179043 .0071096 .0016658
8.98 14.72 11.57 13.09
0.000 0.000 0.000 0.000
.1084981 .2285322 .0683574 .0185407
.1690837 .2987157 .0962265 .0250704
120
The autoregressive (AR) terms are displayed in error notation. e.varname stands for the error in the equation for varname. The estimate of the pth AR term from y1 on y2 is reported as Lpe.y1 in equation e.y2. In the above output, the estimated rst-order AR term of D.ipman on D.ipman is 0.277 and is labeled as LDe.ipman in equation De.ipman.
The previous two examples illustrate how to use dfactor to estimate the parameters of DF models. Although the previous example indicates that the more general DFAR model ts the data well, we use these data to illustrate how to estimate the parameters of more restrictive models.
Example 3
In this example, we use dfactor to estimate the parameters of a SUR model with constraints on the error-covariance matrix. The model is also a constrained VAR with constraints on the error-covariance matrix, because we include the lags of two dependent variables as exogenous variables to model the dynamic structure of the data. Previous exploratory work suggested that we should drop the lag of D.unemp from the model.

. constraint 1 [cov(De.unemp,De.income)]_cons = 0
121
. dfactor (D.(ipman income unemp) = LD.(ipman income), noconstant > covstructure(unstructured)), constraints(1) searching for initial values ............ (setting technique to bhhh) Iteration 0: log likelihood = -569.3512 Iteration 1: log likelihood = -548.76963 (output omitted ) Refining estimates: Iteration 0: log likelihood = -535.12973 Iteration 1: log likelihood = -535.12973 Dynamic-factor model Sample: 1972m3 - 2008m11 Log likelihood = -535.12973 ( 1) [cov(De.income,De.unemp)]_cons = 0 OIM Std. Err. Number of obs Wald chi2(6) Prob > chi2 = = = 441 88.32 0.0000
Coef. D.ipman ipman LD. income LD. D.income ipman LD. income LD. D.unemp ipman LD. income LD. Variance De.ipman Covariance De.ipman De.income De.unemp Variance De.income Covariance De.income De.unemp Variance De.unemp
P>|z|
.206276
.0471654
4.37
0.000
.1138335
.2987185
.1867384
.0512139
3.65
0.000
.086361
.2871158
.1043733
.0434048
2.40
0.016
.0193015
.1894451
-.1957893
.0471305
-4.15
0.000
-.2881634
-.1034153
-.0865823
.0140747
-6.15
0.000
-.1141681
-.0589964
-.0200749
.0152828
-1.31
0.189
-.0500285
.0098788
.3243902
.0218533
14.84
0.000
.2815584
.3672219
.0445794 -.0298076
.013696 .0047755
3.25 -6.24
0.001 0.000
.0177358 -.0391674
.071423 -.0204478
.2747234
.0185008
14.85
0.000
.2384624
.3109844
(omitted)
.0288866
.0019453
14.85
0.000
.0250738
.0326994
122
The output indicates that the model ts well, except that the lag of rst-differenced income is not a signicant predictor of rst-differenced unemployment.
Technical note
The previous example shows how to use dfactor to estimate the parameters of a SUR model with constraints on the error-covariance matrix. Neither sureg nor var allow for constraints on the error-covariance matrix. Without the constraints on the error-covariance matrix and including the lag of D.unemp,
. dfactor (D.(ipman income unemp) = LD.(ipman income unemp), > noconstant covstructure(unstructured)) (output omitted ) . var D.(ipman income unemp), lags(1) noconstant (output omitted )
and
. sureg (D.ipman LD.(ipman income unemp), noconstant) > (D.income LD.(ipman income unemp), noconstant) > (D.unemp LD.(ipman income unemp), noconstant) (output omitted )
produce the same estimates after allowing for small numerical differences.
Example 4
The previous example estimated the parameters of a constrained VAR model with a constraint on the error-covariance matrix. This example makes two renements on the previous one: we use an unconditional estimator instead of a conditional estimator, and we constrain the AR parameters to have a lower triangular structure. (See the next technical note for a discussion of conditional and unconditional estimators.) The results are

. constraint 1 [cov(De.unemp,De.income)]_cons = 0
123
. dfactor (D.(ipman income unemp) = , ar(1) arstructure(ltriangular) noconstant > covstructure(unstructured)), constraints(1) searching for initial values ............ (setting technique to bhhh) Iteration 0: log likelihood = -543.89836 Iteration 1: log likelihood = -541.47455 (output omitted ) Refining estimates: Iteration 0: log likelihood = -540.36159 Iteration 1: log likelihood = -540.36159 Dynamic-factor model Sample: 1972m2 - 2008m11 Log likelihood = -540.36159 ( 1) [cov(De.income,De.unemp)]_cons = 0 OIM Std. Err. Number of obs Wald chi2(6) Prob > chi2 = = = 442 75.48 0.0000
Coef. De.ipman e.ipman LD. De.income e.ipman LD. e.income LD. De.unemp e.ipman LD. e.income LD. e.unemp LD. Variance De.ipman Covariance De.ipman De.income De.unemp Variance De.income Covariance De.income De.unemp Variance De.unemp
P>|z|
.2297308
.0473147
4.86
0.000
.1369957
.3224659
.1075441
.0433357
2.48
0.013
.0226077
.1924805
-.2209485
.047116
-4.69
0.000
-.3132943
-.1286028
-.0975759
.0151301
-6.45
0.000
-.1272304
-.0679215
-.0000467
.0147848
-0.00
0.997
-.0290244
.0289309
-.0795348
.0482213
-1.65
0.099
-.1740469
.0149773
.3335286
.0224282
14.87
0.000
.2895702
.377487
.0457804 -.0329438
.0139123 .0051423
3.29 -6.41
0.001 0.000
.0185127 -.0430226
.0730481 -.022865
.2743375
.0184657
14.86
0.000
.2381454
.3105296
(omitted)
.0292088
.00199
14.68
0.000
.0253083
.0331092
124
The estimated AR terms of D.income and D.unemp on D.unemp are 0.000047 and 0.079535, and they are not signicant at the 1% or 5% levels. The estimated AR term of D.ipman on D.income is 0.107544 and is signicant at the 5% level but not at the 1% level.
Technical note
We obtained the unconditional estimator in example 4 by specifying the ar() option instead of including the lags of the endogenous variables as exogenous variables, as we did in example 3. The unconditional estimator has an additional observation and is more efcient. This change is analogous to estimating an AR coefcient by arima instead of using regress on the lagged endogenous variable. For example, to obtain the unconditional estimator in a univariate model, typing
. arima D.ipman, ar(1) noconstant technique(nr) (output omitted )
will produce the same estimated AR coefcient as

. dfactor (D.ipman, ar(1) noconstant) (output omitted )
We obtain the conditional estimator by typing either

. regress D.ipman LD.ipman, noconstant (output omitted )
or
. dfactor (D.ipman = LD.ipman, noconstant) (output omitted )
Example 5
In this example, we t regional unemployment data to an SF model. We have data on the unemployment levels for the four regions in the U.S. census: west for the West, south for the South, ne for the Northeast, and midwest for the Midwest. We treat the variables as rst-difference stationary and model the rst-differences of these variables. Using dfactor yields

. use http://www.stata-press.com/data/r12/urate (Monthly unemployment rates in US Census regions) . dfactor (D.(west south ne midwest) = , noconstant) (z = ) searching for initial values ............. (setting technique to bhhh) Iteration 0: log likelihood = 872.72029 Iteration 1: log likelihood = 873.04781 (output omitted ) Refining estimates: Iteration 0: log likelihood = 873.0755 Iteration 1: log likelihood = 873.0755 Dynamic-factor model Sample: 1990m2 - 2008m12 Number of obs Wald chi2(4) Log likelihood = 873.0755 Prob > chi2 OIM Std. Err.
125
= = =
227 342.56 0.0000
Coef. D.west z D.south z D.ne z D.midwest z Variance De.west De.south De.ne De.midwest .0861102 .0918607 .0859494 .0978324
P>|z|
.0065644
14.90
0.000
.0849664
.1106983
.0061762
13.92
0.000
.0738442
.0980546
.0072814
12.62
0.000
.0775893
.106132
.0074652
11.53
0.000
.0714787
.1007417
.0036887 .0038902 .0064074 .0074749
.0005834 .0005228 .0007558 .0008271
6.32 7.44 8.48 9.04
0.000 0.000 0.000 0.000
.0025453 .0028656 .0049261 .0058538
.0048322 .0049149 .0078887 .009096
The estimates indicate that we could reasonably suppose that the unobserved factor has the same effect on the changes in unemployment in all four regions. The output below shows that we cannot reject the null hypothesis that these coefcients are the same.
. test ( 1) ( 2) ( 3) [D.west]z = [D.south]z = [D.ne]z = [D.midwest]z [D.west]z - [D.south]z = 0 [D.west]z - [D.ne]z = 0 [D.west]z - [D.midwest]z = 0 chi2( 3) = 3.58 Prob > chi2 = 0.3109
Example 6
In this example, we impose the constraint that the unobserved factor has the same impact on changes in unemployment in all four regions. This constraint was suggested by the results of the previous example. The previous example did not allow for any dynamics in the variables, a problem we alleviate by allowing the disturbances in the equation for each observable to follow an AR(1) process.
126

. constraint 2 [D.west]z = [D.south]z . constraint 3 [D.west]z = [D.ne]z . constraint 4 [D.west]z = [D.midwest]z . dfactor (D.(west south ne midwest) = , noconstant ar(1)) (z = ), > constraints(2/4) searching for initial values ............. (setting technique to bhhh) Iteration 0: log likelihood = 828.22533 Iteration 1: log likelihood = 874.84221 (output omitted ) Refining estimates: Iteration 0: log likelihood = 880.97488 Iteration 1: log likelihood = 880.97488 Dynamic-factor model Sample: 1990m2 - 2008m12 Number of obs = Wald chi2(5) = Log likelihood = 880.97488 Prob > chi2 = ( 1) [D.west]z - [D.south]z = 0 ( 2) [D.west]z - [D.ne]z = 0 ( 3) [D.west]z - [D.midwest]z = 0 OIM Std. Err.
227 363.34 0.0000
Coef. De.west e.west LD. De.south e.south LD. De.ne e.ne LD. De.midwest e.midwest LD. D.west z D.south z D.ne z D.midwest z Variance De.west De.south De.ne De.midwest .0904724 .0904724 .0904724 .0904724 .2866958
P>|z|
.1297198
.0992663
1.31
0.191
-.0648386
.3242781
-.2829014
.0909205
-3.11
0.002
-.4611023
-.1047004
.0847851
3.38
0.001
.12052
.4528715
.0049427
.0782188
0.06
0.950
-.1483634
.1582488
.0049326
18.34
0.000
.0808047
.1001401
.0049326
18.34
0.000
.0808047
.1001401
.0049326
18.34
0.000
.0808047
.1001401
.0049326
18.34
0.000
.0808047
.1001401
.0038959 .0035518 .0058173 .0075444
.0005111 .0005097 .0006983 .0008268
7.62 6.97 8.33 9.12
0.000 0.000 0.000 0.000
.0028941 .0025528 .0044488 .0059239
.0048977 .0045507 .0071859 .009165
127
The results indicate that the model might not t well. Two of the four AR coefcients are statistically insignicant, while the two signicant coefcients have opposite signs and sum to about zero. We suspect that a DF model might t these data better than an SF model with autocorrelated disturbances.
Saved results
dfactor saves the following in e():
Scalars e(N) e(k) e(k aux) e(k eq) e(k eq model) e(k dv) e(k obser) e(k factor) e(o ar max) e(f ar max) e(df m) e(ll) e(chi2) e(p) e(tmin) e(tmax) e(stationary) e(rank) e(ic) e(rc) e(converged) Macros e(cmd) e(cmdline) e(depvar) e(obser deps) e(covariates) e(indeps) e(factor deps) e(tvar) e(eqnames) e(model) e(title) e(tmins) e(tmaxs) e(o ar) e(f ar) e(observ cov) e(factor cov) e(chi2type) e(vce) e(vcetype) e(opt) e(method) e(initial values) e(technique) e(tech steps) e(datasignature) e(datasignaturevars) e(properties) number of observations number of parameters number of auxiliary parameters number of equations in e(b) number of equations in overall model test number of dependent variables number of observation equations number of factors specied number of AR terms for the disturbances number of AR terms for the factors model degrees of freedom log likelihood
2
signicance minimum time in sample maximum time in sample 1 if the estimated parameters indicate a stationary model, 0 otherwise rank of VCE number of iterations return code 1 if converged, 0 otherwise dfactor command as typed unoperated names of dependent variables in observation equations names of dependent variables in observation equations list of covariates independent variables names of unobserved factors in model variable denoting time within groups names of equations type of dynamic-factor model specied title in estimation output formatted minimum time formatted maximum time list of AR terms for disturbances list of AR terms for factors structure of observation-error covariance matrix structure of factor-error covariance matrix Wald; type of model 2 test vcetype specied in vce() title used to label Std. Err. type of optimization likelihood method type of initial values maximization technique iterations taken in maximization technique(s) the checksum variables used in calculation of checksum b V
128

program used to implement estat program used to implement predict predictions allowed by margins predictions disallowed by margins coefcient vector constraints matrix iteration log (up to 20 iterations) gradient vector variancecovariance matrix of the estimators model-based variance marks estimation sample
e(estat cmd) e(predict) e(marginsok) e(marginsnotok) Matrices e(b) e(Cns) e(ilog) e(gradient) e(V) e(V modelbased) Functions e(sample)

dfactor is implemented as an ado-le. dfactor writes the specied model as a state-space model and uses sspace to estimate the parameters by maximum likelihood. See L tkepohl (2005, 619621) for how to write the DF model u in state-space form. See [TS] sspace for the technical details.
References
De Jong, P. 1988. The likelihood for a state space model. Biometrika 75: 165169. . 1991. The diffuse Kalman lter. Annals of Statistics 19: 10731083. Geweke, J. 1977. The dynamic factor analysis of economic time series models. In Latent Variables in Socioeconomic Models, ed. D. J. Aigner and A. S. Goldberger, 365383. Amsterdam: North-Holland. L tkepohl, H. 2005. New Introduction to Multiple Time Series Analysis. New York: Springer. u Sargent, T. J., and C. A. Sims. 1977. Business cycle modeling without pretending to have too much a priori economic theory. In New Methods in Business Cycle Research: Proceedings from a Conference, ed. C. A. Sims, 45109. Minneapolis: Federal Reserve Bank of Minneapolis. Stock, J. H., and M. W. Watson. 1989. New indexes of coincident and leading economic indicators. In NBER Macroeconomics Annual 1989, ed. O. J. Blanchard and S. Fischer, vol. 4, 351394. Cambridge, MA: MIT Press. . 1991. A probability model of the coincident economic indicators. In Leading Economic Indicators: New Approaches and Forecasting Records, ed. K. Lahiri and G. H. Moore, 6389. Cambridge: Cambridge University Press. Watson, M. W., and R. F. Engle. 1983. Alternative algorithms for the estimation of dymanic factor, MIMIC and varying coefcient regression models. Journal of Econometrics 23: 385400.
Also see
[TS] dfactor postestimation Postestimation tools for dfactor [TS] tsset Declare data to be time-series data [TS] sspace State-space models [TS] var Vector autoregressive models [TS] arima ARIMA, ARMAX, and other dynamic regression models [R] regress Linear regression [R] sureg Zellners seemingly unrelated regression [U] 20 Estimation and postestimation commands
Title
dfactor postestimation Postestimation tools for dfactor
Description
The following standard postestimation commands are available after dfactor: Command estat estimates lincom lrtest nlcom predict predictnl test testnl Description
cataloging estimation results point estimates, standard errors, testing, and inference for linear combinations of coefcients likelihood-ratio test point estimates, standard errors, testing, and inference for nonlinear combinations of coefcients predictions, residuals, inuence statistics, and other diagnostic measures point estimates, standard errors, testing, and inference for generalized predictions Wald tests of simple and composite linear hypotheses Wald tests of nonlinear hypotheses
Syntax for predict

predict statistic
Main
type
{ stub* | newvarlist }
if
in
, statistic options
Description dependent variable, which is xbf + residuals linear predictions using the observable independent variables linear predictions using the observable independent variables plus the factor contributions unobserved factor variables autocorrelated disturbances innovations, the observed dependent variable minus the predicted y
y xb xbf factors residuals innovations
129
130
options
Options
Description specify name(s) of equation(s) for which predictions are to be made put estimated root mean squared errors of predicted objects in new variables begin dynamic forecast at specied time method for predicting unobserved states Description predict using past information predict using all sample information predict using past and contemporaneous information
equation(eqnames) rmse(stub* | newvarlist) dynamic(time constant)

Advanced
smethod(method) method onestep smooth filter
Menu
Statistics
>
Postestimation
>
Options for predict

The mathematical notation used in this section is dened in Description of [TS] dfactor.
Main
y, xb, xbf, factors, residuals, and innovations specify the statistic to be predicted. y, the default, predicts the dependent variables. The predictions include the contributions of the unobserved factors, the linear predictions by using the observable independent variables, and any autocorrelation, Pft + Qxt + ut . xb calculates the linear prediction by using the observable independent variables, Qxt . xbf calculates the contributions of the unobserved factors plus the linear prediction by using the observable independent variables, Pft + Qxt . factors estimates the unobserved factors, ft = Rwt + A1 ft1 + A2 ft2 + + Atp ftp . residuals calculates the autocorrelated residuals, ut = C1 ut1 + C2 ut2 + + Ctq utq . innovations calculates the innovations,
t
= yt Pft + Qxt ut .
Options
equation(eqnames) species the equation(s) for which the predictions are to be calculated. You specify equation names, such as equation(income consumption) or equation(factor1 factor2), to identify the equations. For the factors statistic, you must specify names of equations for factors; for all other statistics, you must specify names of equations for observable variables.
131
If you do not specify equation() and do not specify stub*, the results are the same as if you had specied the name of the rst equation for the predicted statistic. equation() may not be specied with stub*. rmse(stub* | newvarlist) puts the root mean squared errors of the predicted objects into the specied new variables. The root mean squared errors measure the variances due to the disturbances but do not account for estimation error. dynamic(time constant) species when predict starts producing dynamic forecasts. The specied time constant must be in the scale of the time variable specied in tsset, and the time constant must be inside a sample for which observations on the dependent variables are available. For example, dynamic(tq(2008q4)) causes dynamic predictions to begin in the fourth quarter of 2008, assuming that your time variable is quarterly, see [D] datetime. If the model contains exogenous variables, they must be present for the whole predicted sample. dynamic() may not be specied with xb, xbf, innovations, smethod(filter), or smethod(smooth).
Advanced
smethod(method) species the method used to predict the unobserved states in the model. smethod() may not be specied with xb. smethod(onestep), the default, causes predict to use previous information on the dependent variables. The Kalman lter is performed on previous periods, but only the one-step predictions are made for the current period. smethod(smooth) causes predict to estimate the states at each time period using all the sample data by the Kalman smoother. smethod(filter) causes predict to estimate the states at each time period using previous and contemporaneous data by the Kalman lter. The Kalman lter is performed on previous periods and the current period. smethod(filter) may be specied only with factors and residuals.
Remarks
We assume that you have already read [TS] dfactor. In this entry, we illustrate some of the features of predict after using dfactor. dfactor writes the specied model as a state-space model and estimates the parameters by maximum likelihood. The unobserved factors and the residuals are states in the state-space form of the model, and they are estimated by the Kalman lter or the Kalman smoother. The smethod() option controls how these states are estimated. The Kalman lter or Kalman smoother is run over the specied sample. Changing the sample can alter the predicted value for a given observation, because the Kalman lter and Kalman smoother are recursive algorithms. After estimating the parameters of a dynamic-factor model, there are many quantities of potential interest. Here we will discuss several of these statistics and illustrate how to use predict to compute them.
Example 1
Lets begin by estimating the parameters of the dynamic-factor model considered in example 2 in [TS] dfactor.
132

. use http://www.stata-press.com/data/r12/dfex (St. Louis Fed (FRED) macro data) . dfactor (D.(ipman income hours unemp) = , noconstant ar(1)) (f = , ar(1/2)) (output omitted )
While several of the six statistics computed by predict might be of interest, we will look only at a few of these statistics for D.ipman. We begin by obtaining one-step predictions in the estimation sample and a six-month dynamic forecast for D.ipman. The graph of the in-sample predictions indicates that our model accounts only for a small fraction of the variability in D.ipman.
. tsappend, add(6) . predict Dipman_f, dynamic(tm(2008m12)) equation(D.ipman) (option y assumed; fitted values) . tsline D.ipman Dipman_f if month<=tm(2008m11), lcolor(gs13) xtitle("") > legend(rows(2))
4 1970m1
1980m1
1990m1
2000m1
2010m1
Dipman y prediction, Dipman, dynamic(tm(2008m12))
Graphing the last year of the sample and the six-month out-of-sample forecast yields
. tsline D.ipman Dipman_f if month>=tm(2008m1), xtitle("") legend(rows(2))
2 6 2008m1 4 2 0
2008m4
2008m7
2008m10
2009m1
2009m4
Dipman y prediction, Dipman, dynamic(tm(2008m12))
133
Example 2
Another common task is to estimate an unobserved factor. We can estimate the unobserved factor at each time period by using only previous information (the smethod(onestep) option), previous and contemporaneous information (the smethod(filter) option), or all the sample information (the smethod(smooth) option). We are interested in the one-step predictive power of the unobserved factor, so we use the default, smethod(onestep).
. predict fac if e(sample), factor . tsline D.ipman fac, lcolor(gs10) xtitle("") legend(rows(2))
4 1970m1
1980m1
1990m1 Dipman factors, f, onestep
2000m1
2010m1

All postestimation commands listed above are implemented as ado-les. dfactor estimates the parameters by writing the model in state-space form and using sspace. Analogously, predict after dfactor uses the methods described in [TS] sspace postestimation. The unobserved factors and the residuals are states in the state-space form of the model. See Methods and formulas of [TS] sspace postestimation for how predictions are made after estimating the parameters of a state-space model.
Also see
[TS] dfactor Dynamic-factor models [TS] sspace State-space models [TS] sspace postestimation Postestimation tools for sspace [U] 20 Estimation and postestimation commands
Title
dfgls DF-GLS unit-root test
Syntax
dfgls varname options
Main
if
in
, options
Description use # as the highest lag order for DickeyFuller GLS regressions series is stationary around a mean instead of around a linear time trend present interpolated critical values from Elliott, Rothenberg, and Stock (1996)
maxlag(#) notrend ers
You must tsset your data before using dfgls; see [TS] tsset. varname may contain time-series operators; see [U] 11.4.4 Time-series varlists.
Menu
Statistics
>
Time series
>
Tests
>
DF-GLS test for a unit root
Description
dfgls performs a modied DickeyFuller t test for a unit root in which the series has been transformed by a generalized least-squares regression.
Options

Main
maxlag(#) sets the value of k , the highest lag order for the rst-differenced, detrended variable in the DickeyFuller regression. By default, dfgls sets k according to the method proposed by Schwert (1989); that is, dfgls sets kmax = int[12{(T + 1)/100}0.25 ]. notrend species that the alternative hypothesis be that the series is stationary around a mean instead of around a linear time trend. By default, a trend is included. ers species that dfgls should present interpolated critical values from tables presented by Elliott, Rothenberg, and Stock (1996), which they obtained from simulations. See Critical values under Methods and formulas for details.
Remarks
dfgls tests for a unit root in a time series. It performs the modied DickeyFuller t test (known as the DF-GLS test) proposed by Elliott, Rothenberg, and Stock (1996). Essentially, the test is an augmented DickeyFuller test, similar to the test performed by Statas dfuller command, except that the time series is transformed via a generalized least squares (GLS) regression before performing the test. Elliott, Rothenberg, and Stock and later studies have shown that this test has signicantly greater power than the previous versions of the augmented DickeyFuller test.
134
135
dfgls performs the DF-GLS test for the series of models that include 1 to k lags of the rstdifferenced, detrended variable, where k can be set by the user or by the method described in Schwert (1989). Stock and Watson (2011, 644649) provide an excellent discussion of the approach. As discussed in [TS] dfuller, the augmented DickeyFuller test involves tting a regression of the form yt = + yt1 + t + 1 yt1 + 2 yt2 + + k ytk + t and then testing the null hypothesis H0 : = 0. The DG-GLS test is performed analogously but on GLS-detrended data. The null hypothesis of the test is that yt is a random walk, possibly with drift. There are two possible alternative hypotheses: yt is stationary about a linear time trend or yt is stationary with a possibly nonzero mean but with no linear time trend. The default is to use the former. To specify the latter alternative, use the notrend option.
Example 1
Here we use the German macroeconomic dataset and test whether the natural log of income exhibits a unit root. We use the default options with dfgls.
. use http://www.stata-press.com/data/r12/lutkepohl2 (Quarterly SA West German macro data, Bil DM, from Lutkepohl 1993 Table E.1) . dfgls ln_inv DF-GLS for ln_inv Number of obs = 80 Maxlag = 11 chosen by Schwert criterion DF-GLS tau 1% Critical 5% Critical 10% Critical [lags] Test Statistic Value Value Value 11 -2.925 10 -2.671 9 -2.766 8 -3.259 7 -3.536 6 -3.115 5 -3.054 4 -3.016 3 -2.071 2 -1.675 1 -1.752 Opt Lag (Ng-Perron seq t) = Min SC = -6.169137 at lag Min MAIC = -6.136371 at lag -3.610 -3.610 -3.610 -3.610 -3.610 -3.610 -3.610 -3.610 -3.610 -3.610 -3.610 7 with RMSE 4 with RMSE 1 with RMSE -2.763 -2.798 -2.832 -2.865 -2.898 -2.929 -2.958 -2.986 -3.012 -3.035 -3.055 .0388771 .0398949 .0440319 -2.489 -2.523 -2.555 -2.587 -2.617 -2.646 -2.674 -2.699 -2.723 -2.744 -2.762
The null hypothesis of a unit root is not rejected for lags 13, it is rejected at the 10% level for lags 910, and it is rejected at the 5% level for lags 48 and 11. For comparison, we also test for a unit root in log income by using dfuller with two different lag specications. We need to use the trend option with dfuller because it is not included by default.
. dfuller ln_inv, lag(4) trend Augmented Dickey-Fuller test for unit root Test Statistic Z(t) -3.133 Number of obs = 87
Interpolated Dickey-Fuller 1% Critical 5% Critical 10% Critical Value Value Value -4.069 -3.463 -3.158
MacKinnon approximate p-value for Z(t) = 0.0987
136

. dfuller ln_inv, lag(7) trend Augmented Dickey-Fuller test for unit root Number of obs = 84 Interpolated Dickey-Fuller Test 1% Critical 5% Critical 10% Critical Statistic Value Value Value Z(t) -3.994 -4.075 -3.466 -3.160
The critical values and the test statistic produced by dfuller with 4 lags do not support rejecting the null hypothesis, although the approximate p-value is less than 0.1. With 7 lags, the critical values and the test statistic reject the null hypothesis at the 5% level, and the approximate p-value is less than 0.01. That the dfuller results are not as strong as those produced by dfgls is not surprising because the DF-GLS test with a trend has been shown to be more powerful than the standard augmented DickeyFuller test.
Saved results
If maxlag(0) is specied, dfgls saves the following in r():
Scalars r(rmse0) r(dft0) RMSE DF-GLS statistic
Otherwise, dfgls saves the following in r():

Scalars r(maxlag) r(N) r(sclag) r(maiclag) r(optlag) Matrices r(results) highest lag order k number of observations lag chosen by Schwarz criterion lag chosen by modied AIC method lag chosen by sequential-t method
k, MAIC, SIC, RMSE, and DF-GLS statistics

dfgls is implemented as an ado-le. dfgls tests for a unit root. There are two possible alternative hypotheses: yt is stationary around a linear trend or yt is stationary with no linear time trend. Under the rst alternative hypothesis, the DF-GLS test is performed by rst estimating the intercept and trend via GLS. The GLS estimation is performed by generating the new variables, yt , xt , and zt , where
137
y1 = y1 yt = yt yt1 , x1 = 1 xt = 1 , z1 = 1 zt = t (t 1)
and = 1 (13.5/T ). An OLS regression is then estimated for the equation
t = 2, . . . , T t = 2, . . . , T
yt = 0 xt + 1 zt +
The OLS estimators 0 and 1 are then used to remove the trend from yt ; that is, we generate
y = yt (0 + 1 t)
Finally, we perform an augmented DickeyFuller test on the transformed variable by tting the OLS regression
k yt = + yt1 + j=1 j ytj + t
and then test the null hypothesis H0: = 0 by using tabulated critical values. To perform the DF-GLS test under the second alternative hypothesis, we proceed as before but dene = 1 (7/T ), eliminate z from the GLS regression, compute y = yt 0 , t the augmented DickeyFuller regression by using the newly transformed variable, and perform a test of the null hypothesis that = 0 by using the tabulated critical values. dfgls reports the DF-GLS statistic and its critical values obtained from the regression in (1) for k {1, 2, . . . , kmax }. By default, dfgls sets kmax = int[12{(T + 1)/100}0.25 ] as proposed by Schwert (1989), although you can override this choice with another value. The sample size available with kmax lags is used in all the regressions. Because there are kmax lags of the rst-differenced series, kmax + 1 observations are lost, leaving T kmax observations. dfgls requires that the sample of T + 1 observations on yt = (y0 , y1 , . . . , yT ) have no gaps. dfgls reports the results of three different methods for choosing which value of k to use. These are method 1 the NgPerron sequential t, method 2 the minimum Schwarz information criterion (SIC), and method 3 the NgPerron modied Akaike information criterion (MAIC). Although the SIC has a long history in time-series modeling, the NgPerron sequential t was developed by Ng and Perron (1995), and the MAIC was developed by Ng and Perron (2000). The SIC can be calculated using either the log likelihood or the sum-of-squared errors from a regression; dfgls uses the latter denition. Specically, for each k
SIC
= ln(rmse ) + (k + 1)
ln(T kmax ) (T kmax )
138
where
rmse =
1 (T kmax )
et2
t=kmax +1
dfgls reports the value of the smallest SIC and the k that produced it. Ng and Perron (1995) derived a sequential-t algorithm for choosing k : i. Set n = 0 and run the regression in method 2 with all kmax n lags. If the coefcient on kmax is signicantly different from zero at level , choose k to kmax . Otherwise, continue to ii. ii. If n < kmax , set n = n + 1 and continue to iii. Otherwise, set k = 0 and stop. iii. Run the regression in method 2 with kmax n lags. If the coefcient on kmax n is signicantly different from zero at level , choose k to kmax n. Otherwise, return to ii. Per Ng and Perron (1995), dfgls uses = 10%. dfgls reports the k selected by this sequential-t algorithm and the rmse from the regression. Method (3) is based on choosing k to minimize the MAIC. The MAIC is calculated as
MAIC(k) = ln(rmse ) +
where
2{ (k) + k} T kmax
T
(k) =
and y was dened previously. Critical values
1 rmse
2 2 0 t=kmax +1
yt2
By default, dfgls uses the 5% and 10% critical values computed from the response surface analysis of Cheung and Lai (1995). Because Cheung and Lai (1995) did not present results for the 1% case, the 1% critical values are always interpolated from the critical values presented by ERS.
ERS presented critical values, obtained from simulations, for the DF-GLS test with a linear trend and showed that the critical values for the mean-only DF-GLS test were the same as those for the ADF test. If dfgls is run with the ers option, dfgls will present interpolated critical values from these tables. The method of interpolation is standard. For the trend case, below 50 observations and above 200 there is no interpolation; the values for 50 and are reported from the tables. For a value N that lies between two values in the table, say, N1 and N2 , with corresponding critical values CV1 and CV2 , the critical value
cv = CV1 +
N N1 (CV2 CV1 ) N1
is presented. The same method is used for the mean-only case, except that interpolation is possible for values between 50 and 500.
139
Acknowledgments
We thank Christopher F. Baum of Boston College and Richard Sperling for a previous version of dfgls.
References
Cheung, Y.-W., and K. S. Lai. 1995. Lag order and critical values of a modied DickeyFuller test. Oxford Bulletin of Economics and Statistics 57: 411419. Dickey, D. A., and W. A. Fuller. 1979. Distribution of the estimators for autoregressive time series with a unit root. Journal of the American Statistical Association 74: 427431. Elliott, G., T. J. Rothenberg, and J. H. Stock. 1996. Efcient tests for an autoregressive unit root. Econometrica 64: 813836. Ng, S., and P. Perron. 1995. Unit root tests in ARMA models with data-dependent methods for the selection of the truncation lag. Journal of the American Statistical Association 90: 268281. . 2000. Lag length selection and the construction of unit root tests with good size and power. Econometrica 69: 15191554. Schwert, G. W. 1989. Tests for unit roots: A Monte Carlo investigation. Journal of Business and Economic Statistics 2: 147159. Stock, J. H., and M. W. Watson. 2011. Introduction to Econometrics. 3rd ed. Boston: AddisonWesley.
Also see
[TS] dfuller Augmented DickeyFuller unit-root test [TS] pperron PhillipsPerron unit-root test [TS] tsset Declare data to be time-series data [XT] xtunitroot Panel-data unit-root tests
Title
dfuller Augmented DickeyFuller unit-root test
Syntax
dfuller varname options
Main
if
in
, options
Description suppress constant term in regression include trend term in regression include drift term in regression display regression table include # lagged differences
noconstant trend drift regress lags(#)
You must tsset your data before using dfuller; see [TS] tsset. varname may contain time-series operators; see [U] 11.4.4 Time-series varlists.
Menu
Statistics
>
Time series
>
Tests
>
Augmented Dickey-Fuller unit-root test
Description
dfuller performs the augmented DickeyFuller test that a variable follows a unit-root process. The null hypothesis is that the variable contains a unit root, and the alternative is that the variable was generated by a stationary process. You may optionally exclude the constant, include a trend term, and include lagged values of the difference of the variable in the regression.
Options

Main
noconstant suppresses the constant term (intercept) in the model and indicates that the process under the null hypothesis is a random walk without drift. noconstant cannot be used with the trend or drift option. trend species that a trend term be included in the associated regression and that the process under the null hypothesis is a random walk, perhaps with drift. This option may not be used with the noconstant or drift option. drift indicates that the process under the null hypothesis is a random walk with nonzero drift. This option may not be used with the noconstant or trend option. regress species that the associated regression table appear in the output. By default, the regression table is not produced. lags(#) species the number of lagged difference terms to include in the covariate list.
140
141
Remarks
Dickey and Fuller (1979) developed a procedure for testing whether a variable has a unit root or, equivalently, that the variable follows a random walk. Hamilton (1994, 528529) describes the four different cases to which the augmented DickeyFuller test can be applied. The null hypothesis is always that the variable has a unit root. They differ in whether the null hypothesis includes a drift term and whether the regression used to obtain the test statistic includes a constant term and time trend. The true model is assumed to be
yt = + yt1 + ut
where ut is an independently and identically distributed zero-mean error term. In cases one and two, presumably = 0, which is a random walk without drift. In cases three and four, we allow for a drift term by letting be unrestricted. The DickeyFuller test involves tting the model
yt = + yt1 + t + ut
by ordinary least squares (OLS), perhaps setting = 0 or = 0. However, such a regression is likely to be plagued by serial correlation. To control for that, the augmented DickeyFuller test instead ts a model of the form
yt = + yt1 + t + 1 yt1 + 2 yt2 + + k ytk +
(1)
where k is the number of lags specied in the lags() option. The noconstant option removes the constant term from this regression, and the trend option includes the time trend t, which by default is not included. Testing = 0 is equivalent to testing = 1, or, equivalently, that yt follows a unit root process. In the rst case, the null hypothesis is that yt follows a random walk without drift, and (1) is t without the constant term and the time trend t. The second case has the same null hypothesis as the rst, except that we include in the regression. In both cases, the population value of is zero under the null hypothesis. In the third case, we hypothesize that yt follows a unit root with drift, so that the population value of is nonzero; we do not include the time trend in the regression. Finally, in the fourth case, the null hypothesis is that yt follows a unit root with or without drift so that is unrestricted, and we include a time trend in the regression. The following table summarizes the four cases. Process under null hypothesis Random walk without drift Random walk without drift Random walk with drift Random walk with or without drift Regression restrictions dfuller option noconstant (default) drift trend
Case
1 2 3 4
= 0, = 0 =0 =0 (none)
Except in the third case, the t-statistic used to test H0: = 0 does not have a standard distribution. Hamilton (1994, chap. 17) derives the limiting distributions, which are different for each of the three other cases. The critical values reported by dfuller are interpolated based on the tables in Fuller (1996). MacKinnon (1994) shows how to approximate the p-values on the basis of a regression surface, and dfuller also reports that p-value. In the third case, where the regression includes a constant term and under the null hypothesis the series has a nonzero drift parameter , the t statistic has the usual t distribution; dfuller reports the one-sided critical values and p-value for the test of H0 against the alternative Ha: < 0, which is equivalent to < 1.
142
Deciding which case to use involves a combination of theory and visual inspection of the data. If economic theory favors a particular null hypothesis, the appropriate case can be chosen based on that. If a graph of the data shows an upward trend over time, then case four may be preferred. If the data do not show a trend but do have a nonzero mean, then case two would be a valid alternative.
Example 1
In this example, we examine the international airline passengers dataset from Box, Jenkins, and Reinsel (2008, Series G). This dataset has 144 observations on the monthly number of international airline passengers from 1949 through 1960. Because the data show a clear upward trend, we use the trend option with dfuller to include a constant and time trend in the augmented DickeyFuller regression.
. use http://www.stata-press.com/data/r12/air2 (TIMESLAB: Airline passengers) . dfuller air, lags(3) trend regress Augmented Dickey-Fuller test for unit root Number of obs = 140 Interpolated Dickey-Fuller Test 1% Critical 5% Critical 10% Critical Statistic Value Value Value Z(t) -6.936 -4.027 -3.445 -3.145
MacKinnon approximate p-value for Z(t) = 0.0000 D.air air L1. LD. L2D. L3D. _trend _cons Coef. Std. Err. t P>|t| [95% Conf. Interval]
-.5217089 .5572871 .095912 .14511 1.407534 44.49164
.0752195 .0799894 .0876692 .0879922 .2098378 7.78335
-6.94 6.97 1.09 1.65 6.71 5.72
0.000 0.000 0.276 0.101 0.000 0.000
-.67048 .399082 -.0774825 -.0289232 .9925118 29.09753
-.3729379 .7154923 .2693065 .3191433 1.822557 59.88575
Here we can overwhelmingly reject the null hypothesis of a unit root at all common signicance levels. From the regression output, the estimated of 0.522 implies that = (1 0.522) = 0.478. Experiments with fewer or more lags in the augmented regression yield the same conclusion.
Example 2
In this example, we use the German macroeconomic dataset to determine whether the log of consumption follows a unit root. We will again use the trend option, because consumption grows over time.

. use http://www.stata-press.com/data/r12/lutkepohl2 (Quarterly SA West German macro data, Bil DM, from Lutkepohl 1993 Table E.1) . tsset qtr time variable: qtr, 1960q1 to 1982q4 delta: 1 quarter . dfuller ln_consump, lags(4) trend Augmented Dickey-Fuller test for unit root Number of obs = 87 Interpolated Dickey-Fuller Test 1% Critical 5% Critical 10% Critical Statistic Value Value Value Z(t) -1.318 -4.069 -3.463 -3.158
143
As we might expect from economic theory, here we cannot reject the null hypothesis that log consumption exhibits a unit root. Again using different numbers of lag terms yield the same conclusion.
Saved results
dfuller saves the following in r():
Scalars r(N) r(lags) number of observations number of lagged differences r(Zt) r(p) DickeyFuller test statistic MacKinnon approximate p-value (if there is a constant or trend in associated regression)

dfuller is implemented as an ado-le. In the OLS estimation of an AR(1) process with Gaussian errors,
yt = yt1 +
where t are independently and identically distributed as N (0, 2 ) and y0 = 0, the OLS estimate (based on an n-observation time series) of the autocorrelation parameter is given by
n
yt1 yt n =
t=1 n 2 yt t=1
If || < 1, then
n(n ) N (0, 1 2 )
If this result were valid when = 1, the resulting distribution would have a variance of zero. When = 1, the OLS estimate still converges in probability to one, though we need to nd a suitable nondegenerate distribution so that we can perform hypothesis tests of H0 : = 1. Hamilton (1994, chap. 17) provides a superb exposition of the requisite theory.
144
To compute the test statistics, we t the augmented DickeyFuller regression

k
yt = + yt1 + t +
j=1
j ytj + et
via OLS where, depending on the options specied, the constant term or time trend t is omitted and k is the number of lags specied in the lags() option. The test statistic for H0 : = 0 is Zt = / , where is the standard error of . The critical values included in the output are linearly interpolated from the table of values that appears in Fuller (1996), and the MacKinnon approximate p-values use the regression surface published in MacKinnon (1994).
David Alan Dickey (1945 ) was born in Ohio and obtained degrees in mathematics at Miami University and a PhD in statistics at Iowa State University in 1976 as a student of Wayne Fuller. He works at North Carolina State University and specializes in time-series analysis. Wayne Arthur Fuller (1931 ) was born in Iowa, obtained three degrees at Iowa State University and then served on the faculty between 1959 and 2001. He has made many distinguished contributions to time series, measurement-error models, survey sampling, and econometrics.
References
Box, G. E. P., G. M. Jenkins, and G. C. Reinsel. 2008. Time Series Analysis: Forecasting and Control. 4th ed. Hoboken, NJ: Wiley. Dickey, D. A., and W. A. Fuller. 1979. Distribution of the estimators for autoregressive time series with a unit root. Journal of the American Statistical Association 74: 427431. Fuller, W. A. 1996. Introduction to Statistical Time Series. 2nd ed. New York: Wiley. Hamilton, J. D. 1994. Time Series Analysis. Princeton: Princeton University Press. MacKinnon, J. G. 1994. Approximate asymptotic distribution functions for unit-root and cointegration tests. Journal of Business and Economic Statistics 12: 167176. Newton, H. J. 1988. TIMESLAB: A Time Series Analysis Laboratory. Belmont, CA: Wadsworth.
Also see
[TS] tsset Declare data to be time-series data [TS] dfgls DF-GLS unit-root test [TS] pperron PhillipsPerron unit-root test [XT] xtunitroot Panel-data unit-root tests
Title
fcast compute Compute dynamic forecasts of dependent variables after var, svar, or vec
Syntax
After var and svar fcast compute prex After vec fcast compute prex , options2 , options1
prex is the prex appended to the names of the dependent variables to create the names of the variables holding the dynamic forecasts. options1
Main
Description set # periods to forecast; default is step(1) begin dynamic forecasts at time constant use previously saved results estname; default is to use active results replace existing forecast variables that have the same prex suppress asymptotic standard errors obtain standard errors from bootstrapped residuals obtain standard errors from parametric bootstrap estimate bounds by using centiles of bootstrapped dataset perform # bootstrap replications; default is reps(200) suppress the usual dot after each bootstrap replication save bootstrap results as lename; use replace to overwrite existing lename set condence level; default is level(95)
step(#) dynamic(time constant) estimates(estname) replace

Std. Errors
nose bs bsp bscentile reps(#) nodots saving(lename , replace )

Reporting
level(#)
145
146
options2
Main
Description set # periods to forecast; default is step(1) begin dynamic forecasts at time constant use previously saved results estname; default is to use active results replace existing forecast variables that have the same prex save dynamic predictions of the rst-differenced variables suppress asymptotic standard errors set condence level; default is level(95)
step(#) dynamic(time constant) estimates(estname) replace differences

Std. Errors
nose
Reporting
level(#)
Default is to use asymptotic standard errors if no options are specied. fcast compute can be used only after var, svar, and vec; see [TS] var, [TS] var svar, and [TS] vec. You must tsset your data before using fcast compute; see [TS] tsset.
Menu
Statistics
>
>
Dynamic forecasts
>
Compute forecasts (required for graph)
Description
fcast compute produces dynamic forecasts of the dependent variables in a model previously t by var, svar, or vec. fcast compute creates new variables and, if necessary, extends the time frame of the dataset to contain the prediction horizon.
Options

Main
step(#) species the number of periods to be forecast. The default is step(1). dynamic(time constant) species the period to begin the dynamic forecasts. The default is the period after the last observation in the estimation sample. The dynamic() option accepts either a Stata date function that returns an integer or an integer that corresponds to a date using the current tsset format. dynamic() must specify a date in the range of two or more periods into the estimation sample to one period after the estimation sample. estimates(estname) species that fcast compute use the estimation results stored as estname. By default, fcast compute uses the active estimation results. See [R] estimates for more information about saving and restoring previously obtained estimates. replace causes fcast compute to replace the variables in memory with the specied predictions. differences species that fcast compute also save dynamic predictions of the rst-differenced variables. differences can be specied only with vec estimation results.
147
Std. Errors
nose species that the asymptotic standard errors of the forecasted levels and, thus the asymptotic condence intervals for the levels, not be calculated. By default, the asymptotic standard errors and the asymptotic condence intervals of the forecasted levels are calculated. bs species that fcast compute use condence bounds estimated by a simulation method based on bootstrapping the residuals. bsp species that fcast compute use condence bounds estimated via simulation in which the innovations are drawn from a multivariate normal distribution. bscentile species that fcast compute use centiles of the bootstrapped dataset to estimate the bounds of the condence intervals. By default, fcast compute uses the estimated standard errors and the quantiles of the standard normal distribution determined by level(). reps(#) gives the number of repetitions used in the simulations. The default is 200. nodots species that no dots be displayed while obtaining the simulation-based standard errors. By default, for each replication, a dot is displayed. saving(lename , replace ) species the name of the le to hold the dataset that contains the bootstrap replications. The replace option overwrites any le with this name. replace species that lename be overwritten if it exists. This option is not shown in the dialog box.
Reporting
level(#) species the condence level, as a percentage, for condence intervals. The default is level(95) or as set by set level; see [U] 20.7 Specifying the width of condence intervals.
Remarks
Researchers often use VARs and VECMs to construct dynamic forecasts. fcast compute computes dynamic forecasts of the dependent variables in a VAR or VECM previously t by var, svar, or vec. If you are interested in conditional, one-step-ahead predictions, use predict (see [TS] var, [TS] var svar, and [TS] vec). To obtain and analyze dynamic forecasts, you t a model, use fcast compute to compute the dynamic forecasts, and use fcast graph to graph the results.
Example 1
Typing
. . . . use http://www.stata-press.com/data/r12/lutkepohl2 var dln_inc dln_consump dln_inv if qtr<tq(1979q1) fcast compute m2_, step(8) fcast graph m2_dln_inc m2_dln_inv m2_dln_consump, observed
ts a VAR with two lags, computes eight-step dynamic predictions for each endogenous variable, and produces the graph
148
Forecast for dln_inc

.04 .1
Forecast for dln_inv
.02
Forecast for dln_consump

.04
.02
1978q3 1979q1 1979q3 1980q1 1980q3
95% CI observed
.1 1978q3 1979q1 1979q3 1980q1 1980q3
forecast
The graph shows that the model is better at predicting changes in income and investment than in consumption. The graph also shows how quickly the predictions from the two-lag model settle down to their mean values.
fcast compute creates new variables in the dataset. If there are K dependent variables in the previously tted model, fcast compute generates 4K new variables:
K new variables that hold the forecasted levels, named by appending the specied prex to the name of the original variable K estimated lower bounds for the forecast interval, named by appending the specied prex and the sufx LB to the name of the original variable K estimated upper bounds for the forecast interval, named by appending the specied prex and the sufx UB to the name of the original variable K estimated standard errors of the forecast, named by appending the specied prex and the sufx SE to the name of the original variable
If you specify options so that fcast compute does not calculate standard errors, the 3K variables that hold them and the bounds of the condence intervals are not generated. If the model previously t is a VECM, specifying differences generates another K variables that hold the forecasts of the rst differences of the dependent variables, named by appending the prex prexD to the name of the original variable.
Example 2
Plots of the forecasts from different models along with the observations from a holdout sample can provide insights to their relative forecasting performance. Continuing the previous example,
. var dln_inc dln_consump dln_inv if qtr<tq(1979q1), lags(1/6) (output omitted ) . fcast compute m6_, step(8) . graph twoway line m6_dln_inv m2_dln_inv dln_inv qtr > if m6_dln_inv < ., legend(cols(1))
.1
149
.05 1978q4
.05
1979q2
1979q4 quarter
1980q2
1980q4
m6_dln_inv, dyn(1979q2) m2_dln_inv, dyn(1979q2) firstdifference of ln_inv
The model with six lags predicts changes in investment better than the two-lag model in some periods but markedly worse in other periods.

fcast compute is implemented as an ado-le. Predictions after var and svar A VAR with endogenous variables yt and exogenous variables xt can be written as
yt = v + A1 yt1 + + Ap ytp + Bxt + ut

where
t = 1, . . . , T yt = (y1t , . . . , yKt ) is a K 1 random vector, the Ai are xed (K K) matrices of parameters, xt is an (M 1) vector of exogenous variables, B is a (K M ) matrix of coefcients, v is a (K 1) vector of xed parameters, and ut is assumed to be white noise; that is, E(ut ) = 0K E(ut ut ) = E(ut us ) = 0K for t = s
150
fcast compute will dynamically predict the variables in the vector yt conditional on p initial values of the endogenous variables and any exogenous xt . Adopting the notation from L tkepohl (2005, u 402) to t the case at hand, the optimal h-step-ahead forecast of yt+h conditional on xt is
yt (h) = v + A1 yt (h 1) + + Ap yt (h p) + Bxt
If there are no exogenous variables, (1) becomes
(1)
yt (h) = v + A1 yt (h 1) + + Ap yt (h p)
When there are no exogenous variables, fcast compute can compute the asymptotic condence bounds. As shown by L tkepohl (2005, 204205), the asymptotic estimator of the covariance matrix of u the prediction error is given by
y (h) = y (h) +
where
h1
1 (h) T
(2)
y (h) =
i=0
i i 1 T
T h1 h1i h1 h1i
(h) =
Zt B
t=0 i=0
i=0
Zt B
(4)
1 0 0 ... 0 v A1 A2 . . . Ap1 0 0 IK 0 . . . B = 0 0 IK 0 . . .. . . . . . 0 0 0 ... IK Zt = (1, yt , . . . , ytp1 ) 0 = IK

i
0 Ap 0 0 . . . 0
i =
j=1
ij Aj
i = 1, 2, . . .
Aj = 0 for j > p is the estimate of the covariance matrix of the innovations, and is the estimated VCE of the coefcients in the VAR. The formula in (3) is general enough to handle the case in which constraints are placed on the coefcients in the VAR(p).
Equation (2) is made up of two terms. y (h) is the estimated mean squared error (MSE) of the forecast. y (h) estimates the error in the forecast arising from the unseen innovations. T 1 (h) estimates the error in the forecast that is due to using estimated coefcients instead of the true coefcients. As the sample size grows, uncertainty with respect to the coefcient estimates decreases, and T 1 (h) goes to zero.
151
If yt is normally distributed, the bounds for the asymptotic (1 )100% interval around the forecast for the k th component of yt , h periods ahead, are
yk,t (h) z( ) k (h) 2

where k (h) is the k th diagonal element of y (h).
(3)
Specifying the bs option causes the standard errors to be computed via simulation, using bootstrapped residuals. Both var and svar contain estimators for the coefcients of a VAR that are conditional on the rst p observations on the endogenous variables in the data. Similarly, these algorithms are conditional on the rst p observations of the endogenous variables in the data. However, the simulation-based estimates of the standard errors are also conditional on the estimated coefcients. The asymptotic standard errors are not conditional on the coefcient estimates because the second term on the right-hand side of (2) accounts for the uncertainty arising from using estimated parameters. For a simulation with R repetitions, this method uses the following algorithm: 1. Fit the model and save the estimated coefcients. 2. Use the estimated coefcients to calculate the residuals. 3. Repeat steps 3a3c R times. 3a. Draw a simple random sample with replacement of size T + h from the residuals. When the tth observation is drawn, all K residuals are selected, preserving any contemporaneous correlation among the residuals. 3b. Use the sampled residuals, p initial values of the endogenous variables, any exogenous variables, and the estimated coefcients to construct a new sample dataset. 3c. Save the simulated endogenous variables for the h forecast periods in the bootstrapped dataset. 4. For each endogenous variable and each forecast period, the simulated standard error is the estimated standard error of the R simulated forecasts. By default, the upper and lower bounds of the (1 )100% are estimated using the simulation-based estimates of the standard errors and the normality assumption, as in (3). If the bscentile option is specied, the sample centiles for the upper and lower bounds of the R simulated forecasts are used for the upper and lower bounds of the condence intervals. If the bsp option is specied, a parametric simulation algorithm is used. Specically, everything is as above except that 3a is replaced by 3a(bsp) as follows: 3a(bsp). Draw T + h observations from a multivariate normal distribution with covariance matrix . The algorithm above assumes that h forecast periods come after the original sample of T observations. If the h forecast periods lie within the original sample, smaller simulated datasets are sufcient. Dynamic forecasts after vec Methods and formulas of [TS] vec discusses how to obtain the one-step predicted differences and levels. fcast compute uses the previous dynamic predictions as inputs for later dynamic predictions.
152
Per L tkepohl (2005, sec. 6.5), fcast compute uses u
y (h) =
T T d
h1
i i
i=0
where the i are the estimated matrices of impulseresponse functions, T is the number of observations in the sample, d is the number of degrees of freedom, and is the estimated cross-equation variance matrix. The formulas for d and are given in Methods and formulas of [TS] vec. The estimated standard errors at step h are the square roots of the diagonal elements of y (h). Per L tkepohl (2005), the estimated forecast-error variance does not consider parameter uncertainty. u As the sample size gets innitely large, the importance of parameter uncertainty diminishes to zero.
References
Hamilton, J. D. 1994. Time Series Analysis. Princeton: Princeton University Press. L tkepohl, H. 2005. New Introduction to Multiple Time Series Analysis. New York: Springer. u
Also see
[TS] fcast graph Graph forecasts of variables computed by fcast compute [TS] var intro Introduction to vector autoregressive models [TS] vec intro Introduction to vector error-correction models
Title
fcast graph Graph forecasts of variables computed by fcast compute
Syntax
fcast graph varlist if in , options where varlist contains one or more forecasted variables generated by fcast compute. options
Main
Description graph forecasts of the rst-differenced variables (vec only) suppress condence bands include observed values of the predicted variables affect rendition of the forecast lines affect rendition of the condence bands affect rendition of the observed values any options other than by() documented in [G-3] twoway options affect appearance of the combined graph; see [G-3] by option
differences noci observed

Forecast plot
cline options
CI plot
Observed plot
obopts(cline options) twoway options byopts(by option)
Y axis, Time axis, Titles, Legend, Overall
Menu
Statistics
>
>
Dynamic forecasts
>
Graph forecasts
Description
fcast graph graphs dynamic forecasts of the endogenous variables from a VAR(p) or VECM that has already been obtained from fcast compute; see [TS] fcast compute.
Options

Main
differences species that the forecasts of the rst-differenced variables be graphed. This option is available only with forecasts computed by fcast compute after vec. The differences option implies noci. noci species that the condence intervals be suppressed. By default, the condence intervals are included. observed species that observed values of the predicted variables be included in the graph. By default, observed values are not graphed.
153
154
Forecast plot
see
cline options affect the rendition of the plotted lines corresponding to the forecast; [G-3] cline options.
CI plot
ciopts(area options) affects the rendition of the condence bands for the forecasts; see [G-3] area options.
Observed plot
obopts(cline options) affects the rendition of the observed values of the predicted variables; see [G-3] cline options. This option implies the observed option.
Y axis, Time axis, Titles, Legend, Overall
twoway options are any of the options documented in [G-3] twoway options, excluding by(). byopts(by option) are documented in [G-3] by option. These options affect the appearance of the combined graph.
Remarks
fcast graph graphs dynamic forecasts created by fcast compute.
Example 1
In this example, we use a cointegrating VECM to model the state-level unemployment rates in Missouri, Indiana, Kentucky, and Illinois, and we graph the forecasts against a 6-month holdout sample.
. use http://www.stata-press.com/data/r12/urates . vec missouri indiana kentucky illinois if t < tm(2003m7), trend(rconstant) > rank(2) lags(4) (output omitted ) . fcast compute m1_, step(6)

. fcast graph m1_missouri m1_indiana m1_kentucky m1_illinois, observed
155
Forecast for missouri

7 6.5 6
Forecast for indiana
5.5
Forecast for kentucky

6.5 7.5 7
Forecast for illinois
5.5
2003m6
2003m8
2003m10
2003m122003m6
5.5
6.5
2003m8
2003m10
2003m12
95% CI observed
forecast
Because the 95% condence bands for the predicted unemployment rates in Missouri and Indiana do not contain all their observed values, the model does not reliably predict these unemployment rates.

fcast graph is implemented as an ado-le.
Also see
[TS] fcast compute Compute dynamic forecasts of dependent variables after var, svar, or vec [TS] var intro Introduction to vector autoregressive models [TS] vec intro Introduction to vector error-correction models
Title
haver Load data from Haver Analytics database
Syntax
Describe contents of a Haver dataset haver describe lename , detail
Describe specied variables in a Haver dataset haver describe varlist using lename Load Haver dataset haver use lename , use options , detail
Load specied variables from a Haver dataset haver use varlist using lename use options tin( constant , constant ) twithin( constant , constant ) tvar(varname) hmissing(misval) fill clear , use options
Description load data within specied date range same as tin(), except exclude the endpoints of range create time variable varname record missing values as misval include observations with missing data in resulting dataset and record missing values for them clear data in memory before loading the Haver dataset
Menu
File
>
Import
>
Haver Analytics database
Description
Haver Analytics (http://www.haver.com) provides economic and nancial databases in the form of .dat les to which you can purchase access. The haver command allows you to use those datasets with Stata. The haver command is provided only with Stata for Windows. haver describe describes the contents of a Haver le. haver use loads the specied variables from a Haver le into Statas memory. If lename is specied without a sufx, .dat is assumed.
156
157
Option for use with haver describe

detail species that a detailed report of all information available on the variables be presented. By default, the Haver concepts data type, number of observations, aggregate type, difference type, magnitude, and date modied are not reported.
Options for use with haver use

tin( constant , constant ) species the date range of the data to be loaded. constant refers to a date constant specied in the usual way, for example, tin(1jan1999, 31dec1999), which would mean from and including 1 January 1999 through 31 December 1999. twithin( constant , constant ) functions the same as tin(), except that the endpoints of the range will be excluded in the data loaded. tvar(varname) species the name of the time variable Stata will create; the default is tvar(time). The tvar() variable is the name of the variable that you would use to tsset the data after loading, although doing is unnecessary because haver use automatically tssets the data for you. hmissing(misval) species which of Statas 27 missing values (., .a, . . . , .z) is to be recorded when there are missing values in the Haver dataset. Two kinds of missing values can occur when loading a Haver dataset. One is created when a variables data does not span the entire time series, and these missing values are always recorded as . by Stata. The other corresponds to an actual missing value recorded in the Haver format. hmissing() sets what is recorded when this second kind of missing value is encountered. You could specify hmissing(.h) to cause Statas .h code to be recorded here. The default is to store . for both kinds of missing values. See [U] 12.2.1 Missing values. fill species that observations with no data be left in the resulting dataset formed in memory and that missing values be recorded for them. The default is to exclude such observations, which can result in the loaded time-series dataset having gaps. Specifying fill has the same effect as issuing the tsfill command after loading; see [TS] tsll. clear clears the data in memory before loading the Haver dataset.
Remarks
Installation Determining the contents of a Haver dataset Loading a Haver dataset Combining variables Using subsets
Installation
Haver Analytics (http://www.haver.com) provides more than 100 economic and nancial databases in the form of .dat les to which you can purchase access. The haver command allows you to use those datasets with Stata. haver is provided only with Stata for Windows, and it uses the DLXAPI32.DLL le, which Haver provides. That le is probably already installed. If it is not, Stata will complain that DLXAPI32.DLL could not be found.
158
Stata accesses this le via the Windows operating system if it is installed where Windows expects. If Windows cannot nd the le, the location can be specied via Statas set command:
. set haverdll "C:\Windows\system32\DLXAPI32.DLL"
Determining the contents of a Haver dataset

haver describe displays the contents of a Haver dataset. If no varlist is specied, all variables are described:
. haver describe dailytst Dataset: dailytst (use) Variable FDB6 FFED FFED2 FFED3 FFED4 FFED5 FFED6 Description 6-Month Federal Federal Federal Federal Federal Federal Eurod.. Funds.. Funds.. Funds.. Funds.. Funds.. Funds.. Time span 02jan1996-01aug2003 01jan1996-01aug2003 01jan1988-29dec1995 01jan1980-31dec1987 03jan1972-31dec1979 01jan1964-31dec1971 02jan1962-31dec1963 Format Daily Daily Daily Daily Daily Daily Daily Group F01 F01 F70 F70 F70 F70 F70 Source FRB FRB FRB FRB FRB FRB FRB
Above, we describe the Haver dataset dailytst.dat, which we already have on our computer and in our current directory. By default, each line of the output corresponds to one Haver variable. Specifying detail would display more information about each variable, and specifying the optional varlist allows us to restrict the output to the variables that interest us:
. haver describe FDB6 using dailytst, detail FDB6 Frequency: Daily Group: F01 Source: FRB Agg Type: 1 Difference Type: 1 Data Type: % 6-Month Eurodollar Deposits (London Bid) (% p.a.) Time span: 02jan1996 - 01aug2003 Number of Observations: 1979 Date Modified: 04aug2003 Decimal Precision: 2 Magnitude: 0
Loading a Haver dataset

haver use loads Haver datasets. If no varlist is specied, all variables are loaded:
. haver use dailytst . list in 1/5 FDB6 1. 2. 3. 4. 5. . . . . . FFED . . . . . FFED2 . . . . . FFED3 . . . . . FFED4 . . . . . FFED5 . . . . . FFED6 2.75 2.5 2.75 2.5 2 time 02jan1962 03jan1962 04jan1962 05jan1962 08jan1962
159
The entire dataset contains 10,849 observations. The variables FDB6 through FFED5 contain missing because there is no information about them for these rst few dates; later in the dataset, nonmissing values for them will be found. haver use requires that the variables loaded all be of the same time frequency (for example, annual, quarterly) Havers weekly data are treated by Stata as daily data with a 7-day interval, so weekly data is loaded as daily data with gaps. Haver datasets include the following information about each variable:
Source the Haver code associated with the source for the data Group the Haver group to which the variable belongs DataType whether the variable is a ratio, index, etc. AggType the Haver code for the level of aggregation DifType the Haver code for the type of difference DecPrecision the number of decimals to which the variable is recorded Magnitude the Haver magnitude code DateTimeMod the date the series was last modied by Haver
When a variable is loaded, this information is stored in variable characteristics (see [P] char). Those characteristics can be viewed using char list:
. char list FDB6[] FDB6[Source]: FDB6[Group]: FDB6[DataType]: FDB6[AggType]: FDB6[DifType]: FDB6[DecPrecision]: FDB6[Magnitude]: FDB6[DateTimeMod]: FRB F01 % 1 1 2 0 04aug2003
Combining variables
In fact, the variables FFED through FFED6 all contain information about the same thing (the Federal Funds rate) but over different times:
. haver describe dailytst Dataset: dailytst (use) Variable FDB6 FFED FFED2 FFED3 FFED4 FFED5 FFED6 Description 6-Month Federal Federal Federal Federal Federal Federal Eurod.. Funds.. Funds.. Funds.. Funds.. Funds.. Funds.. Time span 02jan1996-01aug2003 01jan1996-01aug2003 01jan1988-29dec1995 01jan1980-31dec1987 03jan1972-31dec1979 01jan1964-31dec1971 02jan1962-31dec1963 Format Daily Daily Daily Daily Daily Daily Daily Group F01 F01 F70 F70 F70 F70 F70 Source FRB FRB FRB FRB FRB FRB FRB
In Haver datasets, variables may contain up to 2,100 data points, and if more are needed, Haver Analytics breaks the variables into multiple variables. You must put them back together. To determine when variables need to be combined, look at the description and time span. To combine variables, use egens rowfirst() function and then drop the originals:
160

. egen fedfunds = rowfirst(FFED FFED2 FFED3 FFED4 FFED5 FFED6) . drop FFED*
You can also combine variables by hand:

. gen fedfunds = FFED (8869 missing values generated) . replace fedfunds (2086 real changes . replace fedfunds (2088 real changes = FFED2 if fedfunds >= . made) = FFED3 if fedfunds >= . made)
. replace fedfunds = FFED4 if fedfunds >= . (2086 real changes made) . replace fedfunds = FFED5 if fedfunds >= . (2088 real changes made) . replace fedfunds = FFED6 if fedfunds >= . (521 real changes made) . drop FFED*
Using subsets
You can use a subset of the Haver dataset by specifying the variables to be loaded or the observations to be included:
. haver use FFED FFED2 using dailytst, tin(1jan1990,31dec1999) (data from dailytst.dat loaded)
Above, we load only the variables FFED and FFED2 and the observations from 1jan1990 to 31dec1999.
Also see
[TS] tsset Declare data to be time-series data [D] import Overview of importing data into Stata [D] insheet Read text data created by a spreadsheet [D] odbc Load, write, or view data from ODBC sources
Title
irf Create and analyze IRFs, dynamic-multiplier functions, and FEVDs
Syntax
irf subcommand . . . subcommand create set graph cgraph ograph table ctable describe add drop rename , ...
Description create IRF le containing IRFs, dynamic-multiplier functions, and FEVDs set the active IRF le graph results from active le combine graphs of IRFs, dynamic-multiplier functions, and FEVDs graph overlaid IRFs, dynamic-multiplier functions, and FEVDs create tables of IRFs, dynamic-multiplier functions, and FEVDs from active le combine tables of IRFs, dynamic-multiplier functions, and FEVDs describe contents of active le add results from an IRF le to the active IRF le drop IRF results from active le rename IRF results within a le
IRF stands for impulseresponse function; FEVD stands for forecast-error variance decomposition. irf can be used only after var, svar, or vec; see [TS] var, [TS] var svar, and [TS] vec. See [TS] irf create, [TS] irf set, [TS] irf graph, [TS] irf cgraph, [TS] irf ograph, [TS] irf table, [TS] irf ctable, [TS] irf describe, [TS] irf add, [TS] irf drop, and [TS] irf rename for details about subcommands.
Description
irf creates and manipulates IRF les that contain estimates of the IRFs, dynamic-multiplier functions, and forecast-error variance decompositions (FEVDs) created after estimation by var, svar, or vec; see [TS] var, [TS] var svar, or [TS] vec.
IRFs and FEVDs are described below, and the process of analyzing them is outlined. After reading this entry, please see [TS] irf create.
Remarks
An IRF measures the effect of a shock to an endogenous variable on itself or on another endogenous variable; see L tkepohl (2005, 5163) and Hamilton (1994, 318323) for formal denitions. Of the u many types of IRFs, irf create estimates the ve most important: simple IRFs, orthogonalized IRFs, cumulative IRFs, cumulative orthogonalized IRFs, and structural IRFs. A dynamic-multiplier function, or transfer function, measures the impact of a unit increase in an exogenous variable on the endogenous variables over time; see L tkepohl (2005, chap. 10) for formal u denitions. irf create estimates simple and cumulative dynamic-multiplier functions after var.
161
162
The forecast-error variance decomposition (FEVD) measures the fraction of the forecast-error variance of an endogenous variable that can be attributed to orthogonalized shocks to itself or to another endogenous variable; see L tkepohl (2005, 6366) and Hamilton (1994, 323324) for formal u denitions. Of the many types of FEVDs, irf create estimates the two most important: Cholesky and structural. To analyze IRFs and FEVDs in Stata, you rst t a model, then use irf create to estimate the IRFs and FEVDs and store them in a le, and nally use irf graph or any of the other irf analysis commands to examine results:
. use http://www.stata-press.com/data/r12/lutkepohl2 (Quarterly SA West German macro data, Bil DM, from Lutkepohl 1993 Table E.1) . var dln_inv dln_inc dln_consump if qtr<=tq(1978q4), lags(1/2) dfk (output omitted ) . irf create order1, step(10) set(myirf1) (file myirf1.irf created) (file myirf1.irf now active) (file myirf1.irf updated) . irf graph oirf, impulse(dln_inc) response(dln_consump)
order1, dln_inc, dln_consump

.006
.004
.002
.002 0 5 10
step 95% CI orthogonalized irf
Graphs by irfname, impulse variable, and response variable
Multiple sets of IRFs and FEVDs can be placed in the same le, with each set of results in a le bearing a distinct name. The irf create command above created le myirf1.irf and put one set of results in it, named order1. The order1 results include estimates of the simple IRFs, orthogonalized IRFs, cumulative IRFs, cumulative orthogonalized IRFs, and Cholesky FEVDs. Below we use the same estimated var but use a different Cholesky ordering to create a second set of IRF results, which we will store as order2 in the same le, and then we will graph both results:

. irf create order2, step(10) order(dln_inc dln_inv dln_consump) (file myirf1.irf updated) . irf graph oirf, irf(order1 order2) impulse(dln_inc) response(dln_consump)
163

.01
.005
.005 0 5 10 0 5 10
We have compared results for one model under two different identication schemes. We could just as well have compared results of two different models. We now use irf table to display the results tabularly:
. irf table oirf, irf(order1 order2) impulse(dln_inc) response(dln_consump) Results from order1 order2 (1) oirf .004934 .001309 .003573 -.000692 .000905 .000328 .000021 .000154 .000026 .000026 .000026 (1) Lower .003016 -.000931 .001285 -.002333 -.000541 -.0005 -.000675 -.000206 -.000248 -.000121 -.000061 (1) Upper .006852 .003549 .005862 .00095 .002351 .001156 .000717 .000515 .0003 .000174 .000113 (2) oirf .005244 .001235 .00391 -.000677 .00094 .000341 .000042 .000161 .000027 .00003 .000027 (2) Lower .003252 -.001011 .001542 -.002347 -.000576 -.000518 -.000693 -.000218 -.000261 -.000125 -.000065 (2) Upper .007237 .003482 .006278 .000993 .002456 .001201 .000777 .00054 .000315 .000184 .00012
step 0 1 2 3 4 5 6 7 8 9 10
95% lower and upper bounds reported (1) irfname = order1, impulse = dln_inc, and response = dln_consump (2) irfname = order2, impulse = dln_inc, and response = dln_consump
Both the table and the graph show that the two orthogonalized IRFs are essentially the same. In both functions, an increase in the orthogonalized shock to dln inc causes a short series of increases in dln consump that dies out after four or ve periods.
164
References
Hamilton, J. D. 1994. Time Series Analysis. Princeton: Princeton University Press. L tkepohl, H. 1993. Introduction to Multiple Time Series Analysis. 2nd ed. New York: Springer. u . 2005. New Introduction to Multiple Time Series Analysis. New York: Springer.
Also see
[TS] var Vector autoregressive models [TS] var svar Structural vector autoregressive models [TS] varbasic Fit a simple VAR and graph IRFs or FEVDs [TS] vec Vector error-correction models [TS] var intro Introduction to vector autoregressive models [TS] vec intro Introduction to vector error-correction models
Title
irf add Add results from an IRF le to the active IRF le
Syntax
irf add all | newname= oldname . . . , using(irf lename)
Menu
Statistics
>
>
Manage IRF results and les
>
Add IRF results
Description
irf add copies results from one IRF le to anotherfrom the specied using() le to the active IRF le, set by irf set; see [TS] irf set.
Option
using(irf lename) species the le from which results are to be obtained and is required. If irf lename is specied without an extension, .irf is assumed.
Remarks
If you have not read [TS] irf, please do so.
Example 1
After tting a VAR model, we create two separate IRF les:
. use http://www.stata-press.com/data/r12/lutkepohl2 (Quarterly SA West German macro data, Bil DM, from Lutkepohl 1993 Table E.1) . var dln_inv dln_inc dln_consump if qtr<=tq(1978q4), lags(1/2) dfk (output omitted ) . irf (file (file (file . irf (file (file (file create original, set(irf1, replace) irf1.irf created) irf1.irf now active) irf1.irf updated) create order2, order(dln_inc dln_inv dln_consump) set(irf2, replace) irf2.irf created) irf2.irf now active) irf2.irf updated)
We copy IRF results original to the active le giving them the name order1.
. irf add order1 = original, using(irf1) (file irf2.irf updated)
165
166
irf add Add results from an IRF le to the active IRF le
Here we create new IRF results and save them in the new le irf3.
. irf (file (file (file create order3, order(dln_inc dln_consump dln_inv) set(irf3, replace) irf3.irf created) irf3.irf now active) irf3.irf updated)
Now we copy all the IRF results in le irf2 into the active le.
. irf add _all, using(irf2) (file irf3.irf updated)

irf add is implemented as an ado-le.
Also see
[TS] irf Create and analyze IRFs, dynamic-multiplier functions, and FEVDs [TS] var intro Introduction to vector autoregressive models [TS] vec intro Introduction to vector error-correction models
Title
irf cgraph Combine graphs of IRFs, dynamic-multiplier functions, and FEVDs
Syntax
irf cgraph (spec1 ) where (speck ) is (irfname impulsevar responsevar stat , spec options ) (spec2 ) . . . (specN ) , options
irfname is the name of a set of IRF results in the active IRF le. impulsevar should be specied as an endogenous variable for all statistics except dm and cdm; for those, specify as an exogenous variable. responsevar is an endogenous variable name. stat is one or more statistics from the list below: stat
Main
Description impulseresponse function orthogonalized impulseresponse function dynamic-multiplier function cumulative impulseresponse function cumulative orthogonalized impulseresponse function cumulative dynamic-multiplier function Cholesky forecast-error variance decomposition structural impulseresponse function structural forecast-error variance decomposition
irf oirf dm cirf coirf cdm fevd sirf sfevd
Notes: 1. No statistic may appear more than once. 2. If condence intervals are included (the default), only two statistics may be included. 3. If condence intervals are suppressed (option noci), up to four statistics may be included.
options
Main
Description make lename active affect appearance of combined graph any options other than by() documented in [G-3] twoway options level, steps, and rendition of plots and their CIs graph each combination individually
set(lename)
Options
combine options
twoway options
spec options individual
spec options appear on multiple tabs in the dialog box. individual does not appear in the dialog box.
167
168
spec options
Main
Description suppress condence bands set condence level; default is level(95) use # for rst step use # for maximum step affect rendition of the line plotting the # stat affect rendition of the condence interval for the # stat
noci
Options
level(#) lstep(#) ustep(#)

Plots
plot#opts(line options)
CI plots
ci#opts(area options)
spec options may be specied within a graph specication, globally, or in both. When specied in a graph specication, the spec options affect only the specication in which they are used. When supplied globally, the spec options affect all graph specications. When supplied in both places, options in the graph specication take precedence.
Menu
Statistics
>
>
IRF and FEVD analysis
>
Combined graphs
Description
irf cgraph makes a graph or a combined graph of IRF results. Each block within a pair of matching parentheseseach (speck )species the information for a specic graph. irf cgraph combines these graphs into one image, unless the individual option is also specied, in which case separate graphs for each block are created. To become familiar with this command, we recommend that you type db irf cgraph.
Options

Main
noci suppresses graphing the condence interval for each statistic. noci is assumed when the model was t by vec because no condence intervals were estimated. set(lename) species the le to be made active; see [TS] irf set. If set() is not specied, the active le is used.
Options
level(#) species the default condence level, as a percentage, for condence intervals, when they are reported. The default is level(95) or as set by set level; see [U] 20.7 Specifying the width of condence intervals. The value set of an overall level() can be overridden by the level() inside a (speck ). lstep(#) species the rst step, or period, to be included in the graph. lstep(0) is the default. ustep(#), # 1, species the maximum step, or period, to be included in the graph. combine options affect the appearance of the combined graph; see [G-2] graph combine.
169
Plots
plot1opts(cline options), . . . , plot4opts(cline options) affect the rendition of the plotted statistics. plot1opts() affects the rendition of the rst statistic; plot2opts(), the second; and so on. cline options are as described in [G-3] cline options.
CI plots
ci1opts1(area options) and ci2opts2(area options) affect the rendition of the condence intervals for the rst (ci1opts()) and second (ci2opts()) statistics. See [TS] irf graph for a description of this option and [G-3] area options for the suboptions that change the look of the CI.
twoway options are any of the options documented in [G-3] twoway options, excluding by(). These include options for titling the graph (see [G-3] title options) and for saving the graph to disk (see [G-3] saving option). The following option is available with irf cgraph but is not shown in the dialog box: individual species that each graph be displayed individually. By default, irf cgraph combines the subgraphs into one image.
Remarks
If you have not read [TS] irf, please do so. The relationship between irf cgraph and irf graph is syntactically and conceptually the same as that between irf ctable and irf table; see [TS] irf ctable for a description of the syntax. irf cgraph is much the same as using irf graph to make individual graphs and then using graph combine to put them together. If you cannot use irf cgraph to do what you want, consider the other approach.
Example 1
You have previously issued the commands
. . . . . . . use http://www.stata-press.com/data/r12/lutkepohl2 mat a = (., 0, 0\0,.,0\.,.,.) mat b = I(3) svar dln_inv dln_inc dln_consump, aeq(a) beq(b) irf create modela, set(results3) step(8) svar dln_inc dln_inv dln_consump, aeq(a) beq(b) irf create modelb, step(8)
170
You now type

. irf cgraph (modela dln_inc dln_consump oirf sirf) > (modelb dln_inc dln_consump oirf sirf) > (modela dln_inc dln_consump fevd sfevd, lstep(1)) > (modelb dln_inc dln_consump fevd sfevd, lstep(1)), > title("Results from modela and modelb")
Results from modela and modelb

modela: dln_inc > dln_consump
.01 .006 .004 .002 0 0 .002 0 2 95% CI for oirf oirf 4 step 6 95% CI for sirf sirf 8 .005 0 2 95% CI for oirf oirf 4 step 6 95% CI for sirf sirf 8 .005
modelb: dln_inc > dln_consump
modela: dln_inc > dln_consump

.5 .4 .3 .2 .1 0 2 95% CI for fevd fevd 4 step 6 95% CI for sfevd sfevd 8 .5 .4 .3 .2 .1 0
modelb: dln_inc > dln_consump
2 95% CI for fevd fevd
4 step
6 95% CI for sfevd sfevd
Saved results
irf cgraph saves the following in r():
Scalars r(k) Macros r(individual) r(save) r(name) r(title) r(save#) r(name#) r(title#) r(ci#) r(response#) r(impulse#) r(irfname#) r(stats#) number of specic graph commands individual, if specied lename, replace from saving() option for combined graph name, replace from name() option for combined graph title of the combined graph lename, replace from saving() option for individual graphs name, replace from name() option for individual graphs title for the #th graph level applied to the #th condence interval or noci response specied in the #th command impulse specied in the #th command IRF name specied in the #th command statistics specied in the #th command

irf cgraph is implemented as an ado-le.
171
Also see
Title
irf create Obtain IRFs, dynamic-multiplier functions, and FEVDs
Syntax
After var irf create irfname After svar irf create irfname After vec irf create irfname , vec options , svar options , var options
irfname is any valid name that does not exceed 15 characters. var options
Main
Description make lename active replace irfname if it already exists set forecast horizon to #; default is step(8) specify Cholesky ordering of endogenous variables use previously saved results estname; default is to use active results do not calculate standard errors obtain standard errors from bootstrapped residuals obtain standard errors from parametric bootstrap do not display . for each bootstrap replication use # bootstrap replications; default is reps(200) save bootstrap results in lename
set(lename , replace ) replace step(#) order(varlist) estimates(estname)

Std. errors
nose bs bsp nodots reps(#) bsaving(lename , replace )
svar options
Main
Description make lename active replace irfname if it already exists set forecast horizon to #; default is step(8) use previously saved results estname; default is to use active results
set(lename , replace ) replace step(#) estimates(estname)
172

Std. errors
173
nose bs bsp nodots reps(#) bsaving(lename , replace ) vec options

Main
do not calculate standard errors obtain standard errors from bootstrapped residual obtain standard errors from parametric bootstrap do not display . for each bootstrap replication use # bootstrap replications; default is reps(200) save bootstrap results in lename Description make lename active replace irfname if it already exists set forecast horizon to #; default is step(8) use previously saved results estname; default is to use active results
set(lename , replace ) replace step(#) estimates(estname)
The default is to use asymptotic standard errors if no options are specied. irf create is for use after tting a model with the var, svar, or vec commands; see [TS] var, [TS] var svar, and [TS] vec. You must tsset your data before using var, svar, or vec and, hence, before using irf create; see [TS] tsset.
Menu
Statistics FEVDs
>
>
>
Obtain IRFs, dynamic-multiplier functions, and
Description
irf create estimates multiple sets of impulseresponse functions (IRFs), dynamic-multiplier functions, and forecast-error variance decompositions (FEVDs) after estimation by var, svar, or vec; see [TS] var, [TS] var svar, or [TS] vec. These estimates and their standard errors are known collectively as IRF results and are stored in an IRF le under the specied irfname. The following types of IRFs and dynamic-multiplier functions are stored: simple IRFs orthogonalized IRFs dynamic multipliers cumulative IRFs cumulative orthogonalized IRFs cumulative dynamic multipliers structural IRFs The following types of FEVDs are stored: Cholesky FEVDs structural FEVDs after svar, var, or vec after svar only after after after after after after after svar, var, svar, var, var svar, var, svar, var, var svar only or vec or vec or vec or vec
Once you have created a set of IRF results, use the other irf commands to analyze them.
174
Options

Main
set(lename[, replace]) species the IRF le to be used. If set() is not specied, the active IRF le is used; see [TS] irf set. If set() is specied, the specied le becomes the active le, just as if you had issued an irf set command. replace species that the results stored under irfname may be replaced, if they already exist. IRF results are saved in les, and one le may contain multiple IRF results. step(#) species the step (forecast) horizon; the default is eight periods. order(varlist) is allowed only after estimation by var; it species the Cholesky ordering of the endogenous variables to be used when estimating the orthogonalized IRFs. By default, the order in which the variables were originally specied on the var command is used. estimates(estname) species that estimation results previously estimated by svar, var, or vec, and stored by estimates, be used. This option is rarely specied; see [R] estimates.
Std. errors
nose, bs, and bsp are alternatives that specify how (whether) standard errors are to be calculated. If none of these options is specied, asymptotic standard errors are calculated, except in two cases: after estimation by vec and after estimation by svar in which long-run constraints were applied. In those two cases, the default is as if nose were specied, although in the second case, you could specify bs or bsp. After estimation by vec, standard errors are simply not available. nose species that no standard errors be calculated. bs species that standard errors be calculated by bootstrapping the residuals. bs may not be specied if there are gaps in the data. bsp species that standard errors be calculated via a multivariate-normal parametric bootstrap. bsp may not be specied if there are gaps in the data. nodots, reps(#), and bsaving(lename , replace ) are relevant only if bs or bsp is specied. nodots species that dots not be displayed each time irf create performs a bootstrap replication. reps(#), # > 50, species the number of bootstrap replications to be performed. reps(200) is the default. bsaving(lename , replace ) species that le lename be created and that the bootstrap replications be saved in it. New le lename is just a .dta dataset than can be loaded later using use; see [D] use. If lename is specied without an extension, .dta is assumed.
175
Remarks
If you have not read [TS] irf, please do so. An introductory example using IRFs is presented there. Remarks are presented under the following headings:
Introductory examples Technical aspects of IRF les IRFs and FEVDs IRF results for VARs An introduction to impulseresponse functions for VARs An introduction to dynamic-multiplier functions for VARs An introduction to forecast-error variance decompositions for VARs IRF results for VECMs An introduction to impulseresponse functions for VECMs An introduction to forecast-error variance decompositions for VECMs
Introductory examples Example 1: After var

Below we compare bootstrap and asymptotic standard errors for a specic FEVD. We begin by tting a VAR(2) model to the L tkepohl data (we use the var command). We next use the irf create u command twice, rst to create results with asymptotic standard errors (saved under the name asymp) and then to recreate the same results, this time with bootstrap standard errors (saved under the name bs). Because bootstrapping is a random process, we set the random-number seed (set seed 123456) before using irf create the second time; this makes our results reproducible. Finally, we compare results by using the IRF analysis command irf ctable.
. use http://www.stata-press.com/data/r12/lutkepohl2 (Quarterly SA West German macro data, Bil DM, from Lutkepohl 1993 Table E.1) . var dln_inv dln_inc dln_consump if qtr>=tq(1961q2) & qtr<=tq(1978q4), lags(1/2) (output omitted ) . irf create asymp, step(8) set(results1) (file results1.irf created) (file results1.irf now active) (file results1.irf updated) . set seed 123456 . irf create bs, step(8) bs reps(250) nodots (file results1.irf updated) . irf ctable (asymp dln_inc dln_consump fevd) > (bs dln_inc dln_consump fevd), noci stderror (1) fevd 0 .282135 .278777 .33855 .339942 .342813 .343119 .343079 .34315 (1) S.E. 0 .087373 .083782 .090006 .089207 .090494 .090517 .090499 .090569 (2) fevd 0 .282135 .278777 .33855 .339942 .342813 .343119 .343079 .34315 (2) S.E. 0 .104073 .096954 .100452 .099085 .099326 .09934 .099325 .099368
step 0 1 2 3 4 5 6 7 8
(1) irfname = asymp, impulse = dln_inc, and response = dln_consump (2) irfname = bs, impulse = dln_inc, and response = dln_consump
176
Point estimates are, of course, the same. The bootstrap estimates of the standard errors, however, are larger than the asymptotic estimates, which suggests that the sample size of 71 is not large enough for the distribution of the estimator of the FEVD to be well approximated by the asymptotic distribution. Here we would expect the bootstrap condence interval to be more reliable than the condence interval that is based on the asymptotic standard error.
Technical note
The details of the bootstrap algorithms are given in Methods and formulas. These algorithms are conditional on the rst p observations, where p is the order of the tted VAR. (In an SVAR model, p is the order of the VAR that underlies the SVAR.) The bootstrapped estimates are conditional on the rst p observations, just as the estimators of the coefcients in VAR models are conditional on the rst p observations. With bootstrap standard errors (option bs), the p initial observations are used with resampling the residuals to produce the bootstrap samples used for estimation. With the more parametric bootstrap (option bsp), the p initial observations are used with draws from a multivariate-normal distribution with variancecovariance matrix to generate the bootstrap samples.
Technical note
For var and svar e() results, irf uses , the estimated variance matrix of the disturbances, in computing the asymptotic standard errors of all the functions. The point estimates of the orthogonalized impulseresponse functions, the structural impulseresponse functions, and all the variance decompositions also depend on . As discussed in [TS] var, var and svar use the ML estimator of this matrix by default, but they have option dfk, which will instead use an estimator that includes a small-sample correction. Specifying dfk when the model is twhen the var or svar command is givenchanges the estimate of and will change the IRF results that depend on it.
Example 2: After var with exogenous variables

After tting a VAR, irf create computes estimates of the dynamic multipliers, which describe the impact of a unit change in an exogenous variable on each endogenous variable. For instance, below we estimate and report the cumulative dynamic multipliers from a model in which changes in investment are exogenous. The results indicate that both of the cumulative dynamic multipliers are signicant.
. var dln_inc dln_consump if qtr>=tq(1961q2) & qtr<=tq(1978q4), lags(1/2) > exog(L(0/2).dln_inv) (output omitted ) . irf create dm, step(8) (file results1.irf updated)

. irf table cdm, impulse(dln_inv) irf(dm) Results from dm (1) cdm .032164 .096568 .140107 .150527 .148979 .151247 .150267 .150336 .150525 (1) Lower -.027215 .003479 .022897 .032116 .031939 .033011 .033202 .032858 .033103 (1) Upper .091544 .189656 .257317 .268938 .26602 .269482 .267331 .267813 .267948
177
step 0 1 2 3 4 5 6 7 8
step 0 1 2 3 4 5 6 7 8
(2) cdm .058681 .062723 .126167 .136583 .146482 .146075 .145542 .146309 .145786
(2) Lower .012529 -.005058 .032497 .038691 .04442 .045201 .044988 .045315 .045206
(2) Upper .104832 .130504 .219837 .234476 .248543 .24695 .246096 .247304 .246365
95% lower and upper bounds reported (1) irfname = dm, impulse = dln_inv, and response = dln_inc (2) irfname = dm, impulse = dln_inv, and response = dln_consump
Example 3: After vec

Although all IRFs and orthogonalized IRFs (OIRFs) from models with stationary variables will taper off to zero, some of the IRFs and OIRFs from models with rst-difference stationary variables will not. This is the key difference between IRFs and OIRFs from systems of stationary variables t by var or svar and those obtained from systems of rst-difference stationary variables t by vec. When the effect of the innovations dies out over time, the shocks are said to be transitory. In contrast, when the effect does not taper off, shocks are said to be permanent. In this example, we look at the OIRF from one of the VECMs t to the unemployment-rate data analyzed in example 2 of [TS] vec. We see that an orthogonalized shock to Indiana has a permanent effect on the unemployment rate in Missouri:
. use http://www.stata-press.com/data/r12/urates . vec missouri indiana kentucky illinois, trend(rconstant) rank(2) lags(4) (output omitted ) . irf create vec1, set(vecirfs) step(50) (file vecirfs.irf created) (file vecirfs.irf now active) (file vecirfs.irf updated)
178
Now we can use irf graph to graph the OIRF of interest:

. irf graph oirf, impulse(indiana) response(missouri)
vec1, indiana, missouri

.3
.2
.1
0 0 50
step
The graph shows that the estimated OIRF converges to a positive asymptote, which indicates that an orthogonalized innovation to the unemployment rate in Indiana has a permanent effect on the unemployment rate in Missouri.
Technical aspects of IRF les

This section is included for programmers wishing to extend the irf system. irf create estimates a series of impulseresponse functions and their standard errors. Although these estimates are stored in an IRF le, most users will never need to look at the contents of this le. The IRF commands ll in, analyze, present, and manage IRF results.
IRF les are just Stata datasets that have names ending in .irf instead of .dta. The dataset in the le has a nested panel structure.
Variable irfname contains the irfname specied by the user. Variable impulse records the name of the endogenous variable whose innovations are the impulse. Variable response records the name of the endogenous variable that is responding to the innovations. In a model with K endogenous variables, there are K 2 combinations of impulse and response. Variable step records the periods for which these estimates were computed. Below is a catalog of the statistics that irf create estimates and the variable names under which they are stored in the IRF le.

Statistic Name
179
impulseresponse functions orthogonalized impulseresponse functions dynamic-multiplier functions cumulative impulseresponse functions cumulative orthogonalized impulseresponse functions cumulative dynamic-multiplier functions Cholesky forecast-error decomposition structural impulseresponse functions structural forecast-error decomposition standard error of the impulseresponse functions standard error of the orthogonalized impulseresponse functions standard error of the cumulative impulseresponse functions standard error of the cumulative orthogonalized impulseresponse functions standard error of the Cholesky forecast-error decomposition standard error of the structural impulseresponse functions standard error of the structural forecast-error decomposition
irf oirf dm cirf coirf cdm fevd sirf sfevd stdirf stdoirf stdcirf stdcoirf stdfevd stdsirf stdsfevd
In addition to the variables, information is stored in dta characteristics. Much of the following information is also available in r() after irf describe, where it is often more convenient to obtain the information. Characteristic dta[version] contains the version number of the IRF le, which is currently 1.1. Characteristic dta[irfnames] contains a list of all the irfnames in the IRF le. For each irfname, there are a series of additional characteristics:
Name Contents
dta[irfname dta[irfname dta[irfname dta[irfname dta[irfname dta[irfname dta[irfname dta[irfname dta[irfname dta[irfname dta[irfname dta[irfname
model] order] exog] exogvars] constant] lags] exlags] tmin] tmax] timevar] tsfmt] varcns]
dta[irfname svarcns] dta[irfname step] dta[irfname stderror] dta[irfname reps] dta[irfname version] dta[irfname dta[irfname dta[irfname dta[irfname rank] trend] veccns] sind]
var, sr var, lr var, or vec Cholesky order used in IRF estimates exogenous variables, and their lags, in VAR exogenous variables in VAR constant or noconstant, depending on whether noconstant was specied in var or svar lags in model lags of exogenous variables in model minimum value of timevar in the estimation sample maximum value of timevar in the estimation sample name of tsset timevar format of timevar constrained or colon-separated list of constraints placed on VAR coefcients constrained or colon-separated list of constraints placed on VAR coefcients maximum step in IRF estimates asymptotic, bs, bsp, or none, depending on the type of standard errors requested number of bootstrap replications performed version of the IRF le that originally held irfname IRF results number of cointegrating equations trend() specied in vec constraints placed on VECM parameters normalized seasonal indicators included in vec
180
IRFs and FEVDs

irf create can estimate several types of IRFs and FEVDs for VARs and VECMs. We rst discuss IRF results for VAR and SVAR models, and then we discuss them in the context of VECMs. Because the cointegrating VECM is an extension of the stationary VAR framework, the section that discusses the IRF results for VECMs draws on the earlier VAR material.
IRF results for VARs

An introduction to impulseresponse functions for VARs A pth-order vector autoregressive model (VAR) with exogenous variables is given by
yt = v + A1 yt1 + + Ap ytp + Bxt + ut

where
yt = (y1t , . . . , yKt ) is a K 1 random vector, the Ai are xed K K matrices of parameters, xt is an R0 1 vector of exogenous variables, B is a K R0 matrix of coefcients, v is a K 1 vector of xed parameters, and ut is assumed to be white noise; that is, E(ut ) = 0 E(ut ut ) = E(ut us ) = 0 for t = s
As discussed in [TS] varstable, a VAR can be rewritten in moving-average form only if it is stable. Any exogenous variables are assumed to be covariance stationary. Because the functions of interest in this section depend only on the exogenous variables through their effect on the estimated Ai , we can simplify the notation by dropping them from the analysis. All the formulas given below still apply, although the Ai are estimated jointly with B on the exogenous variables. Below we discuss conditions under which the IRFs and forecast-error variance decompositions have a causal interpretation. Although estimation requires only that the exogenous variables be predetermined, that is, that E[xjt uit ] = 0 for all i, j , and t, assigning a causal interpretation to IRFs and FEVDs requires that the exogenous variables be strictly exogenous, that is, that E[xjs uit ] = 0 for all i, j , s, and t.
IRFs describe how the innovations to one variable affect another variable after a given number of periods. For an example of how IRFs are interpreted, see Stock and Watson (2001). They use IRFs to investigate the effect of surprise shocks to the Federal Funds rate on ination and unemployment. In another example, Christiano, Eichenbaum, and Evans (1999) use IRFs to investigate how shocks to monetary policy affect other macroeconomic variables.
Consider a VAR without exogenous variables:
yt = v + A1 yt1 + + Ap ytp + ut
(1)
The VAR represents the variables in yt as functions of its own lags and serially uncorrelated innovations ut . All the information about contemporaneous correlations among the K variables in yt is contained in . In fact, as discussed in [TS] var svar, a VAR can be viewed as the reduced form of a dynamic simultaneous-equation model.
181
To see how the innovations affect the variables in yt after, say, i periods, rewrite the model in its moving-average form
yt = +
i=0
i uti
(2)
where is the K 1 time-invariant mean of yt , and
i =
i j=1
IK ij Aj
if i = 0 if i = 1, 2, . . .
We can rewrite a VAR in the moving-average form only if it is stable. Essentially, a VAR is stable if the variables are covariance stationary and none of the autocorrelations are too high (the issue of stability is discussed in greater detail in [TS] varstable). The i are the simple IRFs. The j, k element of i gives the effect of a 1time unit increase in the k th element of ut on the j th element of yt after i periods, holding everything else constant. Unfortunately, these effects have no causal interpretation, which would require us to be able to answer the question, How does an innovation to variable k , holding everything else constant, affect variable j after i periods? Because the ut are contemporaneously correlated, we cannot assume that everything else is held constant. Contemporaneous correlation among the ut implies that a shock to one variable is likely to be accompanied by shocks to some of the other variables, so it does not make sense to shock one variable and hold everything else constant. For this reason, (2) cannot provide a causal interpretation. This shortcoming may be overcome by rewriting (2) in terms of mutually uncorrelated innovations. Suppose that we had a matrix P, such that = PP . If we had such a P, then P1 P 1 = IK , and E{P1 ut (P1 ut ) } = P1 E{(ut ut )P 1 } = P1 P 1 = IK We can thus use P1 to orthogonalize the ut and rewrite (2) as
yt = +
i=0
i PP1 uti i P1 uti

i=0
=+ =+
i=0
i wti
where i = i P and wt = P1 ut . If we had such a P, the wk would be mutually orthogonal, and no information would be lost in the holding-everything-else-constant assumption, implying that the i would have the causal interpretation that we seek. Choosing a P is similar to placing identication restrictions on a system of dynamic simultaneous equations. The simple IRFs do not identify the causal relationships that we wish to analyze. Thus we seek at least as many identication restrictions as necessary to identify the causal IRFs. So, where do we get such a P? Sims (1980) popularized the method of choosing P to be the Cholesky decomposition of . The IRFs based on this choice of P are known as the orthogonalized IRFs. Choosing P to be the Cholesky decomposition of is equivalent to imposing a recursive structure for the corresponding dynamic structural equation model. The ordering of the recursive structure is the same as the ordering imposed in the Cholesky decomposition. Because this choice is arbitrary, some researchers will look at the OIRFs with different orderings assumed in the Cholesky decomposition. The order() option available with irf create facilitates this type of analysis.
182
The SVAR approach integrates the need to identify the causal IRFs into the model specication and estimation process. Sufcient identication restrictions can be obtained by placing either short-run or long-run restrictions on the model. The VAR in (1) can be rewritten as
yt v A1 yt1 Ap ytp = ut
Similarly, a short-run SVAR model can be written as
(3)
A(yt v A1 yt1 Ap ytp ) = Aut = Bet
(4)
where A and B are K K nonsingular matrices of parameters to be estimated, et is a K 1 vector of disturbances with et N (0, IK ), and E[et es ] = 0K for all s = t. Sufcient constraints must be placed on A and B so that P is identied. One way to see the connection is to draw out the implications of the latter equality in (4). From (4) it can be shown that
= A1 B(A1 B)
As discussed in [TS] var svar, the estimates A and B are obtained by maximizing the concentrated log-likelihood function on the basis of the obtained from the underlying VAR. The short-run SVAR approach chooses P = A1 B to identify the causal IRFs. The long-run SVAR approach works
1 1
similarly, with P = C = A B, where A effects of the reduced-form VAR shocks.
is the matrix of estimated long-run or accumulated
There is one important difference between long-run and short-run SVAR models. As discussed by Amisano and Giannini (1997, chap. 6), in the short-run model the constraints are applied directly to the parameters in A and B. Then A and B interact with the estimated parameters of the underlying VAR. In contrast, in a long-run model, the constraints are placed on functions of the estimated VAR parameters. Although estimation and inference of the parameters in C is straightforward, obtaining the asymptotic standard errors of the structural IRFs requires untenable assumptions. For this reason, irf create does not estimate the asymptotic standard errors of the structural IRFs generated by long-run SVAR models. However, bootstrap standard errors are still available. An introduction to dynamic-multiplier functions for VARs A dynamic-multiplier function measures the effect of a unit change in an exogenous variable on the endogenous variables over time. Per L tkepohl (2005, chap. 10), if the VAR with exogenous variables u is stable, it can be rewritten as

yt =
i=0
Di xti +
i=0
i uti
where the Di are the dynamic-multiplier functions. (See Methods and formulas for details.) Some authors refer to the dynamic-multiplier functions as transfer functions because they specify how a unit change in an exogenous variable is transferred to the endogenous variables.
183
Technical note
irf create computes dynamic-multiplier functions only after var. After short-run SVAR models, the dynamic multipliers from the VAR are the same as those from the SVAR. The dynamic multipliers for long-run SVARs have not yet been worked out.
An introduction to forecast-error variance decompositions for VARs Another measure of the effect of the innovations in variable k on variable j is the FEVD. This method, which is also known as innovation accounting, measures the fraction of the error in forecasting variable j after h periods that is attributable to the orthogonalized innovations in variable k . Because deriving the FEVD requires orthogonalizing the ut innovations, the FEVD is always predicated upon a choice of P. L tkepohl (2005, sec. 2.2.2) shows that the h-step forecast error can be written as u
h1
yt+h yt (h) =
i=0
i ut+hi
(6)
where yt+h is the value observed at time t + h and yt (h) is the h-step-ahead predicted value for yt+h that was made at time t. Because the ut are contemporaneously correlated, their distinct contributions to the forecast error cannot be ascertained. However, if we choose a P such that = PP , as above, we can orthogonalize the ut into wt = P1 ut . We can then ascertain the relative contribution of the distinct elements of wt . Thus we can rewrite (6) as
h1
yt+h yt (h) =
i=0 h1
i PP1 ut+hi
=
i=0
i wt+hi
(7)
Because the forecast errors can be written in terms of the orthogonalized errors, the forecasterror variance can be written in terms of the orthogonalized error variances. Forecast-error variance decompositions measure the fraction of the total forecast-error variance that is attributable to each orthogonalized shock.
Technical note
The details in this note are not critical to the discussion that follows. A forecast-error variance decomposition is derived for a given P. Per L tkepohl (2005, sec. 2.3.3), letting mn,i be the m, nth u element of i , we can express the h-step forecast error of the j th component of yt as
h1
yj,t+h yj (h) =
i=0 K
j1,1 w1,t+hi + + jK,i wK,t+hi jk,0 wk,t+h + + jk,h1 wk,t+1

k=1
184
The wt , which were constructed using P, are mutually orthogonal with unit variance. This allows us to compute easily the mean squared error (MSE) of the forecast of variable j at horizon h in terms of the contributions of the components of wt . Specically,
K 2 2 2 (jk,0 + + jk,h1 ) k=1
E[{yj,t+h yj,t (h)} ] =
The k th term in the sum above is interpreted as the contribution of the orthogonalized innovations in variable k to the h-step forecast error of variable j . Note that the k th element in the sum above can be rewritten as
h1 2 (jk,0
+ +
2 jk,h1 )
=
i=0
ej k ek
where ei is the ith column of IK . Normalizing by the forecast error for variable j at horizon h yields
jk,h =
where MSE[yj,t (h)] =
h1 i=0 K 2 k=1 jk,i .
ej k ek MSE[yj,t (h)]
h1 i=0
Because the FEVD depends on the choice of P, there are different forecast-error variance decompositions associated with each distinct P. irf create can estimate the FEVD for a VAR or an SVAR. For a VAR, P is the Cholesky decomposition of . For an SVAR, P is the estimated structural decomposition, P = A1 B for short-run models and P = C for long-run SVAR models. Due to the same complications that arose with the structural impulseresponse functions, the asymptotic standard errors of the structural FEVD are not available after long-run SVAR models, but bootstrap standard errors are still available.
IRF results for VECMs

An introduction to impulseresponse functions for VECMs As discussed in [TS] vec intro, the VECM is a reparameterization of the VAR that is especially useful for tting VARs with cointegrating variables. This implies that the estimated parameters for the corresponding VAR model can be backed out from the estimated parameters of the VECM model. This relationship means we can use the VAR form of the cointegrating VECM to discuss the IRFs for VECMs. Consider a cointegrating VAR with one lag with no constant or trend,
yt = Ayt1 + ut
(8)
where yt is a K 1 vector of endogenous, rst-difference stationary variables among which there are 1 r < K cointegration equations; A is K K matrix of parameters; and ut is a K 1 vector of i.i.d. disturbances.
185
We developed intuition for the IRFs from a stationary VAR by rewriting the VAR as an inniteorder vector moving-average (VMA) process. While the Granger representation theorem establishes the existence of a VMA formulation of this model, because the cointegrating VAR is not stable, the inversion is not nearly so intuitive. (See Johansen [1995, chapters 3 and 4] for more details.) For this reason, we use (8) to develop intuition for the IRFs from a cointegrating VAR. Suppose that K is 3, that u1 = (1, 0, 0), and that we want to analyze the time paths of the variables in y conditional on the initial values y0 = 0, A, and the condition that there are no more shocks to the system, that is, 0 = u2 = u3 = . These assumptions and (8) imply that
y1 = u1 y2 = Ay1 = Au1 y3 = Ay2 = A2 u1

and so on. The ith-row element of the rst column of As contains the effect of the unit shock to the rst variable after s periods. The rst column of As contains the IRF of a unit impulse to the rst variable after s periods. We could deduce the IRFs of a unit impulse to any of the other variables by administering the unit shock to one of them instead of to the rst variable. Thus we can see that the (i, j)th element of As contains the unit IRF from variable j to variable i after s periods. By starting with orthogonalized shocks of the form P1 ut , we can use the same logic to derive the OIRFs to be As P. For the stationary VAR, stability implies that all the eigenvalues of A have moduli strictly less than one, which in turn implies that all the elements of As 0 as s . This implies that all the IRFs from a stationary VAR taper off to zero as s . In contrast, in a cointegrating VAR, some of the eigenvalues of A are 1, while the remaining eigenvalues have moduli strictly less than 1. This implies that in cointegrating VARs some of the elements of As are not going to zero as s , which in turn implies that some of the IRFs and OIRFs are not going to zero as s . The fact that the IRFs and OIRFs taper off to zero for stationary VARs but not for cointegrating VARs is one of the key differences between the two models. When the IRF or OIRF from the innovation in one variable to another tapers off to zero as time goes on, the innovation to the rst variable is said to have a transitory effect on the second variable. When the IRF or OIRF does not go to zero, the effect is said to be permanent. Note that, because some of the IRFs and OIRFs do not taper off to zero, some of the cumulative
IRFs and OIRFs diverge over time.
An introduction to forecast-error variance decompositions for VECMs The results from An introduction to impulseresponse functions for VECMs can be used to show that the interpretation of FEVDs for a nite number of steps in cointegrating VARs is essentially the same as in the stationary case. Because the MSE of the forecast is diverging, this interpretation is valid only for a nite number of steps. (See [TS] vec intro and [TS] fcast compute for more information on this point.)

irf create is implemented as an ado-le.
186
Methods and formulas are presented under the following headings:

Impulseresponse function formulas for VARs Dynamic-multiplier function formulas for VARs Forecast-error variance decomposition formulas for VARs Impulseresponse function formulas for VECMs Algorithms for bootstrapping the VAR IRF and FEVD standard errors
Impulseresponse function formulas for VARs

The previous discussion implies that there are three different choices of P that can be used to obtain distinct i . P is the Cholesky decomposition of for the OIRFs. For the structural IRFs, P = A1 B for short-run models, and P = C for long-run models. We will distinguish between the three by dening o to be the OIRFs, sr to be the short-run structural IRFs, and lr to be the i i i long-run structural IRFs. We also dene Pc to be the Cholesky decomposition of , Psr = A1 B to be the short-run structural decomposition, and Plr = C to be the long-run structural decomposition. Given estimates of the Ai and from var or svar, the estimates of the simple IRFs and the
OIRFs are, respectively,
i
i =
j=1
ij Aj
and
o = i Pc i
where Aj = 0K for j > p. Given the estimates A and B, or C, from svar, the estimates of the structural IRFs are either
sr = i Psr i
or
lr = i Plr i
The estimated structural IRFs stored in an IRF le with the variable name sirf may be from either a short-run model or a long-run model, depending on the estimation results used to create the IRFs. As discussed in [TS] irf describe, you can easily determine whether the structural IRFs were generated from a short-run or a long-run SVAR model using irf describe. Following L tkepohl (2005, sec. 3.7), estimates of the cumulative IRFs and the cumulative u orthogonalized impulseresponse functions (COIRFs) at period n are, respectively,
n
n =
i=0
and
n =
i=0
187
The asymptotic standard errors of the different impulseresponse functions are obtained by applications of the delta method. See L tkepohl (2005, sec. 3.7) and Amisano and Giannini (1997, u chap. 4) for the derivations. See Sering (1980, sec. 3.3) for a discussion of the delta method. In presenting the variancecovariance matrix estimators, we make extensive use of the vec() operator, where vec(X) is the vector obtained by stacking the columns of X. L tkepohl (2005, sec. 3.7) derives the asymptotic VCEs of vec(i ), vec(o ), vec(n ), and u i vec(n ). Because vec(i ) is K 2 1, the asymptotic VCE of vec(i ) is K 2 K 2 , and it is given by Gi Gi where
Gi =
i1 m=0
J(M )(i1m) m
Gi is K 2 K 2 p J is KKp
J = (IK , 0K , . . . , 0K ) A1 A2 . . . Ap1 Ap IK 0K . . . 0K 0K 0K 0K M = 0K IK . . . .. . . . .
. . .
M is KpKp
0K 0K . . .
IK
0K
The Ai are the estimates of the coefcients on the lagged variables in the VAR, and is the VCE matrix of = vec(A1 , . . . , Ap ). is a K 2 p K 2 p matrix whose elements come from the VCE of the VAR coefcient estimator. As such, this VCE is the VCE of the constrained estimator if there are any constraints placed on the VAR coefcients. The K 2 K 2 asymptotic VCE matrix for vec(n ) after n periods is given by
Fn Fn
where
n
Fn =
i=1
Gi
The K 2 K 2 asymptotic VCE matrix of the vectorized, orthogonalized, IRFs at horizon i, vec(o ), i is
Ci Ci + Ci Ci
188
where
C0 = 0 Ci = (Pc IK )Gi , i = 1, 2, . . . Ci = (IK i )H, i = 0, 1, . . .

1
C0 is K 2 K 2 p Ci is K 2 K 2 p Ci is K 2 K 2 H is K 2 K LK is K
(K+1) 2
H = LK LK NK (Pc IK )LK LK solves

vech(F) =
LK vec(F)
for
(K+1) K 2 2
F K K and symmetric
KK is K 2 K 2 NK is K 2 K 2
KK solves
KK vec(G) = vec(G ) for any K K matrix G NK =

1 2
(IK 2 + KK )
= 2D+ ( )D+ K K D+ = (DK DK ) K DK solves DK vech(F) = vec(F) x11 x21 . . . xK1 vech(X) = x22 . . . xK2 . . . xKK
1
is K (K+1) K (K+1) 2 2
(K+1) K 2 2 (K+1) 2
DK
D+ is K K
for
F K K and symmetric
DK is K 2 K
for
X K K
vech(X) is K
(K+1) 1 2
Note that is the VCE of vech(). More details about LK , KK , DK and vech() are available in L tkepohl (2005, sec. A.12). Finally, as L tkepohl (2005, 113114) discusses, D+ is the Moore u u K Penrose inverse of DK . As discussed in Amisano and Giannini (1997, chap. 6), the asymptotic standard errors of the structural IRFs are available for short-run SVAR models but not for long-run SVAR models. Following Amisano and Giannini (1997, chap. 5), the asymptotic K 2 K 2 VCE of the short-run structural IRFs after i periods, when a maximum of h periods are estimated, is the i, i block of
(h)ij = Gi Gj + IK (JMi J ) (0) IK (JMj J )
189
where
G0 = 0K Gi =
i1 k=0
G0 is K 2 K 2 p
Psr J(M )i1k JMk J
Gi is K 2 K 2 p (0) is K 2 K 2 W is K 2 K 2 Q2 is K 2 K 2 Q1 is K 2 2K 2
(0) = Q2 W Q2 W = Q1 AB Q1 Q2 = Psr Psr Q1 =

1 (IK B1 ), (Psr B1 )
and AB is the 2K 2 2K 2 VCE of the estimator of vec(A, B).
Dynamic-multiplier function formulas for VARs

This section provides the details of how irf create estimates the dynamic-multiplier functions and their asymptotic standard errors. A pth order vector autoregressive model (VAR) with exogenous variables may be written as
yt = v + A1 yt1 + + Ap ytp + B0 xt + B1 xt1 + + Bs xts + ut

where all the notation is the same as above except that the s K R matrices B1 , B2 , . . . , Bs are explicitly included and s is the number of lags of the R exogenous variables in the model. L tkepohl (2005) shows that the dynamic-multipliers Di are consistently estimated by u
Di = Jx Ai Bx x
where
i {0, 1, . . .}
J is K(Kp+Rs) Ax is (Kp+Rs)(Kp+Rs)
Jx = (IK , 0K , . . . , 0K ) Ax = B= M B 0 I B1 B2 . . . Bs 0 0 ... 0 . . . .. . . . . . . . ... 0 0 0 0R 0R 0R 0R 0R 0R . . . . . . IR 0R
B is KpRs
0R 0R . . . IR 0R . . . I = 0 R IR . .. . . . 0R 0R . . . Bx = B I B = B0 0 0 = [ IR 0 R 0 R ] I
I is RsRs
Bx is R(Kp+Rs) B is RKp is RRs I
and 0 is a K R matrix of 0s and 0 is a Rs Kp matrix of 0s.
190
Consistent estimators of the cumulative dynamic-multiplier functions are given by

i
Di =
j=0
Dj
be the asymptotic variance Letting x = vec[A1 A2 Ap B1 B2 Bs B0 ] and letting x covariance estimator (VCE) of x , L tkepohl shows that an asymptotic VCE of Di is Gi Gi u x where
i1
Gi =
j=0
Bx Ai1j Jx Aj Jx , IR Jx Aj Jx x x x
i j=0
Similarly, an asymptotic VCE of Di is
Gj x
i j=0
Gj .
Forecast-error variance decomposition formulas for VARs

This section provides details of how irf create estimates the Cholesky FEVD, the structural FEVD, and their standard errors. Beginning with the Cholesky-based forecast-error decompositions, the fraction of the h-step-ahead forecast-error variance of variable j that is attributable to the Cholesky orthogonalized innovations in variable k can be estimated as
jk,h =
where MSEj (h) is the j th diagonal element of
h1
h1 2 i=0 (ej i ek )
MSEj (h)
i i
i=0
(See L tkepohl [2005, 109] for a discussion of this result.) jk,h and MSEj (h) are scalars. The u square of the standard error of jk,h is
djk,h djk,h + djk,h djk,h

where
djk,h =
2 MSEj (h)2
h1 i=0
MSEj (h)(ej i Pc ek )(ek Pc ej )Gi

h1 m=0 (ej m
(ej i Pc ek )2 djk,h =
h1 i=0
ej )Gm
djk,h is 1K 2 p
MSEj (h)(ej i Pc ek )(ek ej i )H

h1 m=0 (ej m
(ej i Pc ek )2 G0 = 0
ej m )DK
1 MSEj (h)2
djk,h is 1K
(K+1) 2
G0 is K 2 K 2 p
and DK is the K 2 K{(K + 1)/2} duplication matrix dened previously.
191
For the structural forecast-error decompositions, we follow Amisano and Giannini (1997, sec. 5.2). They dene the matrix of structural forecast-error decompositions at horizon s, when a maximum of h periods are estimated, as
Ws = F1 Ms s
s1
for s = 1, . . . , h + 1
Fs =
i=0 s1
sr sr i i sr i
i=0
IK
Ms =
where
sr i
is the Hadamard, or element-by-element, product.
The K 2 K 2 asymptotic VCE of vec(Ws ) is given by
Zs (h)Zs
where (h) is as derived previously, and
Zs = vec(Ws )
sr vec(j )
vec(Ws ) vec(Ws ) vec(Ws ) sr , sr , , sr vec(0 ) vec(1 ) vec(h )

sr sr
= 2 (IK F1 )D(j ) (Ws F1 )D(IK )NK (j IK ) s s
If X is an n n matrix, then D(X) is the n2 n2 matrix with vec(X) on the diagonal and zeros in all the off-diagonal elements, and NK is as dened previously.
Impulseresponse function formulas for VECMs

We begin by providing the formulas for backing out the estimates of the Ai from the i estimated by vec. As discussed in [TS] vec intro, the VAR in (1) can be rewritten as a VECM:
yt = v + yt1 + 1 yt1 + p1 yp2 +

vec estimates and the i . Johansen (1995, 25) notes that
p
=
i=1
Ai IK
(9)
where IK is the K -dimensional identity matrix, and

p
i =
j=i+1
Aj
(10)
192
Dening
p1
= IK
i=1
and using (9) and (10) allow us to solve for the Ai as
A1 = + 1 + IK Ai = i i1
and
for i = {2, . . . , p 1}
Ap = p1
Using these formulas, we can back out estimates of Ai from the estimates of the i and produced by vec. Then we simply use the formulas for the IRFs and OIRFs presented in Impulseresponse function formulas for VARs. The running sums of the IRFs and OIRFs over the steps within each impulseresponse pair are the cumulative IRFs and OIRFs.
Algorithms for bootstrapping the VAR IRF and FEVD standard errors
irf create offers two bootstrap algorithms for estimating the standard errors of the various IRFs and FEVDs. Both var and svar contain estimators for the coefcients in a VAR that are conditional on the rst p observations. The two bootstrap algorithms are also conditional on the rst p observations. Specifying the bs option calculates the standard errors by bootstrapping the residuals. For a bootstrap with R repetitions, this method uses the following algorithm: 1. Fit the model and save the estimated parameters. 2. Use the estimated coefcients to calculate the residuals. 3. Repeat steps 3a to 3c R times. 3a. Draw a simple random sample of size T with replacement from the residuals. The random samples are drawn over the K 1 vectors of residuals. When the tth vector is drawn, all K residuals are selected. This preserves the contemporaneous correlations among the residuals. 3b. Use the p initial observations, the sampled residuals, and the estimated coefcients to construct a new sample dataset. 3c. Fit the model and calculate the different IRFs and FEVDs. 3d. Save these estimates as observation r in the bootstrapped dataset. 4. For each IRF and FEVD, the estimated standard deviation from the R bootstrapped estimates is the estimated standard error of that impulseresponse function or forecast-error variance decomposition. Specifying the bsp option estimates the standard errors by a multivariate normal parametric bootstrap. The algorithm for the multivariate normal parametric bootstrap is identical to the one above, with the exception that 3a is replaced by 3a(bsp): 3a(bsp). Draw T pseudovariates from a multivariate normal distribution with covariance matrix .
193
References
Amisano, G., and C. Giannini. 1997. Topics in Structural VAR Econometrics. 2nd ed. Heidelberg: Springer. Christiano, L. J., M. Eichenbaum, and C. L. Evans. 1999. Monetary policy shocks: What have we learned and to what end? In Handbook of Macroeconomics: Volume 1A, ed. J. B. Taylor and M. Woodford. New York: Elsevier. Hamilton, J. D. 1994. Time Series Analysis. Princeton: Princeton University Press. Johansen, S. 1995. Likelihood-Based Inference in Cointegrated Vector Autoregressive Models. Oxford: Oxford University Press. L tkepohl, H. 1993. Introduction to Multiple Time Series Analysis. 2nd ed. New York: Springer. u . 2005. New Introduction to Multiple Time Series Analysis. New York: Springer. Sering, R. J. 1980. Approximation Theorems of Mathematical Statistics. New York: Wiley. Sims, C. A. 1980. Macroeconomics and reality. Econometrica 48: 148. Stock, J. H., and M. W. Watson. 2001. Vector autoregressions. Journal of Economic Perspectives 15: 101115.
Also see
Title
irf ctable Combine tables of IRFs, dynamic-multiplier functions, and FEVDs
Syntax
irf ctable (spec1 ) where (speck ) is (irfname impulsevar responsevar stat , spec options ) (spec2 ) . . . (specN ) , options
irfname is the name of a set of IRF results in the active IRF le. impulsevar should be specied as an endogenous variable for all statistics except dm and cdm; for those, specify as an exogenous variable. responsevar is an endogenous variable name. stat is one or more statistics from the list below: stat irf oirf dm cirf coirf cdm fevd sirf sfevd options set(lename) noci stderror individual title("text") step(#) level(#) spec options noci stderror level(#) ititle("text") Description impulseresponse function orthogonalized impulseresponse function dynamic-multiplier function cumulative impulseresponse function cumulative orthogonalized impulseresponse function cumulative dynamic-multiplier function Cholesky forecast-error variance decomposition structural impulseresponse function structural forecast-error variance decomposition Description make lename active do not report condence intervals include standard errors for each statistic make an individual table for each combination use text as overall table title set common maximum step set condence level; default is level(95) Description do not report condence intervals include standard errors for each statistic set condence level; default is level(95) use text as individual subtitle for specic table
spec options may be specied within a table specication, globally, or both. When specied in a table specication, the spec options affect only the specication in which they are used. When supplied globally, the spec options affect all table specications. When specied in both places, options for the table specication take precedence. ititle() does not appear in the dialog box.
194
195
Menu
Statistics
>
>
>
Combined tables
Description
irf ctable makes a table or a combined table of IRF results. Each block within a pair of matching parentheseseach (speck )species the information for a specic table. irf ctable combines these tables into one table, unless the individual option is specied, in which case separate tables for each block are created. irf ctable operates on the active IRF le; see [TS] irf set.
Options
set(lename) species the le to be made active; see [TS] irf set. If set() is not specied, the active le is used. noci suppresses reporting of the condence intervals for each statistic. noci is assumed when the model was t by vec because no condence intervals were estimated. stderror species that standard errors for each statistic also be included in the table. individual places each block, or (speck ), in its own table. By default, irf ctable combines all the blocks into one table. title("text") species a title for the table or the set of tables. step(#) species the maximum number of steps to use for all tables. By default, each table is constructed using all steps available. level(#) species the default condence level, as a percentage, for condence intervals, when they are reported. The default is level(95) or as set by set level; see [U] 20.7 Specifying the width of condence intervals. The following option is available with irf ctable but is not shown in the dialog box: ititle("text") species an individual subtitle for a specic table. ititle() may be specied only when the individual option is also specied.
Remarks
If you have not read [TS] irf, please do so. Also see [TS] irf table for a slightly easier to use, but less powerful, table command. irf ctable creates a series of tables from IRF results. The information enclosed within each set of parentheses, (irfname impulsevar responsevar stat forms a request for a specic table. The rst partirfname impulsevar responsevaridenties a set of IRF estimates or a set of variance decomposition estimates. The next partstatspecies which statistics are to be included in the table. The last partspec optionsincludes the noci, level(), and stderror options, and places (or suppresses) additional columns in the table. , spec options )
196
Each specic table displays the requested statistics corresponding to the specied combination of irfname, impulsevar, and responsevar over the step horizon. By default, all the individual tables are combined into one table. Also by default, all the steps, or periods, available are included in the table. You can use the step() option to impose a common maximum for all tables.
Example 1
In example 1 of [TS] irf table, we t a model using var and we saved the IRFs for two different orderings. The commands we used were
. . . . . use var irf irf irf http://www.stata-press.com/data/r12/lutkepohl2 dln_inv dln_inc dln_consump set results4 create ordera, step(8) create orderb, order(dln_inc dln_inv dln_consump) step(8)
We then formed the desired table by typing

. irf table oirf fevd, impulse(dln_inc) response(dln_consump) noci std > title("Ordera versus orderb")
Using irf ctable, we can form the equivalent table by typing

. irf ctable (ordera dln_inc dln_consump oirf fevd) > (orderb dln_inc dln_consump oirf fevd), > noci std title("Ordera versus orderb") Ordera versus orderb (1) oirf .005123 .001635 .002948 -.000221 .000811 .000462 .000044 .000151 .000091 (1) S.E. .000878 .000984 .000993 .000662 .000586 .000333 .000275 .000162 .000114 (1) fevd 0 .288494 .294288 .322454 .319227 .322579 .323552 .323383 .323499 (1) S.E. 0 .077483 .073722 .075562 .074063 .075019 .075371 .075314 .075386
step 0 1 2 3 4 5 6 7 8
step 0 1 2 3 4 5 6 7 8
(2) oirf .005461 .001578 .003307 -.00019 .000846 .000491 .000069 .000158 .000096
(2) S.E. .000925 .000988 .001042 .000676 .000617 .000349 .000292 .000172 .000122
(2) fevd 0 .327807 .328795 .370775 .366896 .370399 .371487 .371315 .371438
(2) S.E. 0 .08159 .077519 .080604 .079019 .079941 .080323 .080287 .080366
(1) irfname = ordera, impulse = dln_inc, and response = dln_consump (2) irfname = orderb, impulse = dln_inc, and response = dln_consump
The output is displayed in one table. Because the table did not t horizontally, it automatically wrapped. At the bottom of the table is a list of keys that appear at the top of each column. The
197
results in the table above indicate that the orthogonalized IRFs do not change by much. Because the estimated forecast-error variances do change, we might want to produce two tables that contain the estimated forecast-error variance decompositions and their 95% condence intervals:
. irf ctable (ordera dln_inc dln_consump fevd) > (orderb dln_inc dln_consump fevd), individual Table 1 (1) fevd 0 .288494 .294288 .322454 .319227 .322579 .323552 .323383 .323499 (1) Lower 0 .13663 .149797 .174356 .174066 .175544 .175826 .17577 .175744 (1) Upper 0 .440357 .43878 .470552 .464389 .469613 .471277 .470995 .471253
step 0 1 2 3 4 5 6 7 8
95% lower and upper bounds reported (1) irfname = ordera, impulse = dln_inc, and response = dln_consump Table 2 (2) fevd 0 .327807 .328795 .370775 .366896 .370399 .371487 .371315 .371438 (2) Lower 0 .167893 .17686 .212794 .212022 .213718 .214058 .213956 .213923 (2) Upper 0 .487721 .48073 .528757 .52177 .52708 .528917 .528674 .528953
step 0 1 2 3 4 5 6 7 8
95% lower and upper bounds reported (2) irfname = orderb, impulse = dln_inc, and response = dln_consump
Because we specied the individual option, the output contains two tables, one for each specic table command. At the bottom of each table is a list of the keys used in that table and a note indicating the level of the condence intervals that we requested. The results from table 1 and table 2 indicate that each estimated function is well within the condence interval of the other, so we conclude that the functions are not signicantly different.
198
Saved results
irf ctable saves the following in r():
Scalars r(ncols) r(k umax) r(k) Macros r(key#) r(tnotes) number of columns in all tables number of distinct keys number of specic table commands #th key list of keys applied to each column

irf ctable is implemented as an ado-le.
Also see
Title
irf describe Describe an IRF le
Syntax
irf describe options set(lename) using(irf lename) detail variables irf resultslist , options
Description make lename active describe irf lename without making active show additional details of IRF results show underlying structure of the IRF dataset
Menu
Statistics
>
>
>
Describe IRF le
Description
irf describe describes the IRF results stored in an IRF le. If set() or using() is not specied, the IRF results of the active IRF le are described.
Options
set(lename) species the IRF le to be described and set; see [TS] irf set. If lename is specied without an extension, .irf is assumed. using(irf lename) species the IRF le to be described. The active IRF le, if any, remains unchanged. If irf lename is specied without an extension, .irf is assumed. detail species that irf describe display detailed information about each set of IRF results. detail is implied when irf resultslist is specied. variables is a programmers option; additionally displays the output produced by the describe command.
Remarks
Example 1
. use http://www.stata-press.com/data/r12/lutkepohl2 (Quarterly SA West German macro data, Bil DM, from Lutkepohl 1993 Table E.1) . var dln_inv dln_inc dln_consump if qtr<=tq(1978q4), lags(1/2) dfk (output omitted )
199
200
We create three sets of IRF results:

. irf create order1, set(myirfs, replace) (file myirfs.irf created) (file myirfs.irf now active) (file myirfs.irf updated) . irf create order2, order(dln_inc dln_inv dln_consump) (file myirfs.irf updated) . irf create order3, order(dln_inc dln_consump dln_inv) (file myirfs.irf updated) . irf describe Contains irf results from myirfs.irf (dated 4 Apr 2011 12:36) irfname order1 order2 order3 model var var var endogenous variables and order (*) dln_inv dln_inc dln_consump dln_inc dln_inv dln_consump dln_inc dln_consump dln_inv
(*) order is relevant only when model is var
The output reveals the order in which we specied the variables.

. irf describe order1
irf results for order1

Estimation specification model: var endog: dln_inv dln_inc dln_consump sample: quarterly data from 1960q4 to 1978q4 lags: 1 2 constant: constant exog: none exogvars: none exlags: none varcns: unconstrained IRF specification step: 8 order: dln_inv dln_inc dln_consump std error: asymptotic reps: none
Here we see a summary of the model we t as well as the specication of the IRFs.
201
Saved results
irf describe saves the following in r():
Scalars r(N) r(k) r(width) r(N max) r(k max) r(widthmax) r(changed) Macros r( version) r(irfnames) r(irfname model) r(irfname order) r(irfname exog) r(irfname exogvar) r(irfname constant) r(irfname lags) r(irfname exlags) r(irfname tmin) r(irfname tmax) r(irfname timevar) r(irfname tsfmt) r(irfname varcns) r(irfname svarcns) r(irfname step) r(irfname stderror) r(irfname r(irfname r(irfname r(irfname r(irfname r(irfname reps) version) rank) trend) veccns) sind) number of observations in the IRF le number of variables in the IRF le width of dataset in the IRF le maximum number of observations maximum number of variables maximum width of the dataset ag indicating that data have changed since last saved version of IRF results le names of IRF results in the IRF le var, sr var, lr var, or vec Cholesky order assumed in IRF estimates exogenous variables, and their lags, in VAR or underlying VAR exogenous variables in VAR or underlying VAR constant or noconstant lags in model lags of exogenous variables in model minimum value of timevar in the estimation sample maximum value of timevar in the estimation sample name of tsset timevar format of timevar in the estimation sample unconstrained or colon-separated list of constraints placed on VAR coefcients "." or colon-separated list of constraints placed on SVAR coefcients maximum step in IRF estimates asymptotic, bs, bsp, or none, depending on type of standard errors specied to irf create "." or number of bootstrap replications performed version of IRF le that originally held irfname IRF results "." or number of cointegrating equations "." or trend() specied in vec "." or constraints placed on VECM parameters "." or normalized seasonal indicators included in vec

irf describe is implemented as an ado-le.
Also see
Title
irf drop Drop IRF results from the active IRF le
Syntax
irf drop irf resultslist , set(lename)
Menu
Statistics
>
>
>
Drop IRF results
Description
irf drop removes IRF results from the active IRF le.
Option
set(lename) species the le to be made active; see [TS] irf set. If set() is not specied, the active le is used.
Remarks
Example 1

. irf create order1, set(myirfs, replace) (file myirfs.irf created) (file myirfs.irf now active) (file myirfs.irf updated) . irf create order2, order(dln_inc dln_inv dln_consump) (file myirfs.irf updated) . irf create order3, order(dln_inc dln_consump dln_inv) (file myirfs.irf updated) . irf describe Contains irf results from myirfs.irf (dated 4 Apr 2011 12:59) irfname model endogenous variables and order (*) order1 order2 order3 var var var dln_inv dln_inc dln_consump dln_inc dln_inv dln_consump dln_inc dln_consump dln_inv
202
irf drop Drop IRF results from the active IRF le
203
Now lets remove order1 and order2 from myirfs.irf.

. irf drop order1 order2 (order1 dropped) (order2 dropped) file myirfs.irf updated . irf describe Contains irf results from myirfs.irf (dated 4 Apr 2011 12:59) model endogenous variables and order (*) irfname order3 var dln_inc dln_consump dln_inv
order1 and order2 have been dropped.

irf drop is implemented as an ado-le.
Also see
Title
irf graph Graph IRFs, dynamic-multiplier functions, and FEVDs
Syntax
irf graph stat stat irf oirf dm cirf coirf cdm fevd sirf sfevd , options Description impulseresponse function orthogonalized impulseresponse function dynamic-multiplier function cumulative impulseresponse function cumulative orthogonalized impulseresponse function cumulative dynamic-multiplier function Cholesky forecast-error variance decomposition structural impulseresponse function structural forecast-error variance decomposition
Notes: 1. No statistic may appear more than once. 2. If condence intervals are included (the default), only two statistics may be included. 3. If condence intervals are suppressed (option noci), up to four statistics may be included.
options
Main
Description make lename active use irfnames IRF result sets use impulsevar as impulse variables use endogenous variables as response variables suppress condence bands set condence level; default is level(95) use # for rst step use # for maximum step graph each combination individually stub for naming the individual graphs stub for saving the individual graphs to les affect rendition of the line plotting the # stat affect rendition of the condence interval for the # stat
set(lename) irf(irfnames) impulse(impulsevar) response(endogvars) noci level(#) lstep(#) ustep(#)

Advanced
individual iname(namestub , replace ) isaving(lenamestub , replace )

Plots
plot#opts(cline options)
CI plots
ci#opts(area options)
204

205
twoway options byopts(by option)
any options other than by() documented in [G-3] twoway options how subgraphs are combined, labeled, etc.
Menu
Statistics
>
>
>
Graphs by impulse or response
Description
irf graph graphs impulseresponse functions (IRFs), dynamic-multiplier functions, and forecasterror variance decompositions (FEVDs) over time.
Options

Main
set(lename) species the le to be made active; see [TS] irf set. If set() is not specied, the active le is used. irf(irfnames) species the IRF result sets to be used. If irf() is not specied, each of the results in the active IRF le is used. (Files often contain just one set of IRF results stored under one irfname; in that case, those results are used.) impulse(impulsevar) and response(endogvars) specify the impulse and response variables. Usually one of each is specied, and one graph is drawn. If multiple variables are specied, a separate subgraph is drawn for each impulseresponse combination. If impulse() and response() are not specied, subgraphs are drawn for all combination of impulse and response variables. impulsevar should be specied as an endogenous variable for all statistics except dm or cdm; for those, specify as an exogenous variable. noci suppresses graphing the condence interval for each statistic. noci is assumed when the model was t by vec because no condence intervals were estimated. level(#) species the default condence level, as a percentage, for condence intervals, when they are reported. The default is level(95) or as set by set level; see [U] 20.7 Specifying the width of condence intervals. Also see [TS] irf cgraph for a graph command that allows the condence level to vary over the graphs. lstep(#) species the rst step, or period, to be included in the graphs. lstep(0) is the default. ustep(#), # 1, species the maximum step, or period, to be included in the graphs.
Advanced
individual species that each graph be displayed individually. By default, irf graph combines the subgraphs into one image. When individual is specied, byopts() may not be specied, but the isaving() and iname() options may be specied. iname(namestub , replace ) species that the ith individual graph be saved in memory under the name namestubi, which must be a valid Stata name of 24 characters or fewer. iname() may be specied only with the individual option. isaving(lenamestub , replace ) species that the ith individual graph should be saved to disk in the current working directory under the name lenamestubi.gph. isaving() may be specied only when the individual option is also specied.
206
Plots
plot1opts(cline options), . . . , plot4opts(cline options) affect the rendition of the plotted statistics (the stat). plot1opts() affects the rendition of the rst statistic; plot2opts(), the second; and so on. cline options are as described in [G-3] cline options.
CI plots
ci1opts(area options) and ci2opts(area options) affect the rendition of the condence intervals for the rst (ci1opts()) and second (ci2opts()) statistics in stat. area options are as described in [G-3] area options.
twoway options are any of the options documented in [G-3] twoway options, excluding by(). These include options for titling the graph (see [G-3] title options) and for saving the graph to disk (see [G-3] saving option). Note that the saving() and name() options may not be combined with the individual option. byopts(by option) is as documented in [G-3] by option and may not be specied when individual is specied. byopts() affects how the subgraphs are combined, labeled, etc.
Remarks
If you have not read [TS] irf, please do so. Also see [TS] irf cgraph, which produces combined graphs; [TS] irf ograph, which produces overlaid graphs; and [TS] irf table, which displays results in tabular form. irf graph produces one or more graphs and displays them arrayed into one image unless the individual option is specied, in which case the individual graphs are displayed separately. Each individual graph consists of all the specied stat and represents one impulseresponse combination. Because all the specied stat appear on the same graph, putting together statistics with very different scales is not recommended. For instance, sometimes sirf and oirf are on similar scales while irf is on a different scale. In such cases, combining sirf and oirf on the same graph looks ne, but combining either with irf produces an uninformative graph.
Example 1
Suppose that we have results generated from two different SVAR models. We want to know whether the shapes of the structural IRFs and the structural FEVDs are similar in the two models. We are also interested in knowing whether the structural IRFs and the structural FEVDs differ signicantly from their Cholesky counterparts. Filling in the background, we have previously issued the commands
. . . . . . . use http://www.stata-press.com/data/r12/lutkepohl2 mat a = (., 0, 0\0,.,0\.,.,.) mat b = I(3) svar dln_inv dln_inc dln_consump, aeq(a) beq(b) irf create modela, set(results3) step(8) svar dln_inc dln_inv dln_consump, aeq(a) beq(b) irf create modelb, step(8)
To see whether the shapes of the structural IRFs and the structural FEVDs are similar in the two models, we type

. irf graph oirf sirf, impulse(dln_inc) response(dln_consump)
207
modela, dln_inc, dln_consump

.01
modelb, dln_inc, dln_consump
.005
.005 0 2 4 6 8 0 2 4 6 8
step 95% CI for oirf orthogonalized irf

95% CI for sirf structural irf
The graph reveals that the oirf and the sirf estimates are essentially the same for both models and that the shapes of the functions are very similar for the two models. To see whether the structural IRFs and the structural FEVDs differ signicantly from their Cholesky counterparts, we type
. irf graph fevd sfevd, impulse(dln_inc) response(dln_consump) lstep(1) > legend(cols(1))

.5
modelb, dln_inc, dln_consump
.4
.3
.2
.1 0 2 4 6 8 0 2 4 6 8
step 95% CI for fevd 95% CI for sfevd fraction of mse due to impulse (structural) fraction of mse due to impulse
This combined graph reveals that the shapes of these functions are also similar for the two models. However, the graph illuminates one minor difference between them: In modela, the estimated structural
208
FEVD is slightly larger than the Cholesky-based estimates, whereas in modelb the Cholesky-based
estimates are slightly larger than the structural estimates. For both models, however, the structural estimates are close to the center of the wide condence intervals for the two estimates.
Example 2
Lets focus on the results from modela. Suppose that we were interested in examining how dln consump responded to impulses in its own structural innovations, structural innovations to dln inc, and structural innovations to dln inv. We type
. irf graph sirf, irf(modela) response(dln_consump)
modela, dln_consump, dln_consump

.01
.005
.005 0 2 4 6 8
modela, dln_inv, dln_consump

.01
.005
.005 0 2 4 6 8
step 95% CI
structural irf
The upper-left graph shows the structural IRF of an innovation in dln consump on dln consump. It indicates that the identication restrictions used in modela imply that a positive shock to dln consump causes an increase in dln consump, followed by a decrease, followed by an increase, and so on, until the effect dies out after roughly 5 periods. The upper-right graph shows the structural IRF of an innovation in dln inc on dln consump, indicating that a positive shock to dln inc causes an increase in dln consump, which dies out after 4 or 5 periods.
Technical note
[TS] irf table contains a technical note warning you to be careful in naming variables when you t models. What is said there applies equally here.
209
Saved results
irf graph saves the following in r():
Scalars r(k) Macros r(stats) r(irfname) r(impulse) r(response) r(plot#) r(ci) r(ciopts#) number of graphs statlist resultslist impulselist responselist contents of plot#opts() level applied to condence intervals or noci contents of ci#opts() r(byopts) r(saving) r(name) r(individual) r(isaving) r(iname) r(subtitle#) contents of byopts() supplied saving() option supplied name() option individual or blank contents of saving() contents of name() subtitle for individual graph #

irf graph is implemented as an ado-le.
Also see
Title
irf ograph Graph overlaid IRFs, dynamic-multiplier functions, and FEVDs
Syntax
irf ograph (spec1 ) where (speck ) is (irfname impulsevar responsevar stat , spec options ) (spec2 ) . . . (spec15 ) , options
irfname is the name of a set of IRF results in the active IRF le or ., which means the rst named result in the active IRF le. impulsevar should be specied as an endogenous variable for all statistics except dm and cdm; for those, specify as an exogenous variable. responsevar is an endogenous variable name. stat is one or more statistics from the list below: stat irf oirf dm cirf coirf cdm fevd sirf sfevd options
Plots
Description impulseresponse function orthogonalized impulseresponse function dynamic-multiplier function cumulative impulseresponse function cumulative orthogonalized impulseresponse function cumulative dynamic-multiplier function Cholesky forecast-error variance decomposition structural impulseresponse function structural forecast-error variance decomposition Description dene the IRF plots make lename active level and steps any options other than by() documented in [G-3] twoway options Description make lename active use irfnames IRF result sets use impulsevar as impulse variables use endogenous variables as response variables add condence bands to the graph
210
plot options set(lename)

Options
common options twoway options plot options

Main
set(lename) irf(irfnames) impulse(impulsevar) response(endogvars) ci
211
spec options
Options
Description level and steps affect rendition of the plotted lines affect rendition of the condence intervals
common options
Plot
cline options
CI plot
common options
Options
Description set condence level; default is level(95) use # for rst step use # for maximum step
level(#) lstep(#) ustep(#)
common options may be specied within a plot specication, globally, or in both. When specied in a plot specication, the common options affect only the specication in which they are used. When supplied globally, the common options affect all plot specications. When supplied in both places, options in the plot specication take precedence.
Menu
Statistics
>
>
>
Overlaid graph
Description
irf ograph displays plots of irf results on one graph (one pair of axes). To become familiar with this command, type db irf ograph.
Options

Plots
plot options denes the IRF plots and are found under the Main, Plot, and CI plot tabs. set(lename) species the le to be made active; see [TS] irf set. If set() is not specied, the active le is used.
Main
set(lename) species the le to be made active; see [TS] irf set. If set() is not specied, the active le is used. irf(irfnames) species the IRF result sets to be used. If irf() is not specied, each of the results in the active IRF le is used. (Files often contain just one set of IRF results stored under one irfname; in that case, those results are used.)
212
impulse(varlist) and response(endogvars) specify the impulse and response variables. Usually one of each is specied, and one graph is drawn. If multiple variables are specied, a separate subgraph is drawn for each impulseresponse combination. If impulse() and response() are not specied, subgraphs are drawn for all combination of impulse and response variables. ci adds condence bands to the graph. The noci option may be used within a plot specication to suppress its condence bands when the ci option is supplied globally.
Plot
cline options affect the rendition of the plotted lines; see [G-3] cline options.
CI plot
ciopts(area options) affects the rendition of the condence bands for the plotted statistic; see [G-3] area options. ciopts() implies ci.
Options
level(#) species the condence level, as a percentage, for condence bands; see [U] 20.7 Specifying the width of condence intervals. lstep(#) species the rst step, or period, to be included in the graph. lstep(0) is the default. ustep(#), # 1, species the maximum step, or period, to be included.
Remarks
If you have not read [TS] irf, please do so. irf ograph overlays plots of IRFs and FEVDs on one graph.
Example 1
We have previously issued the commands
. . . . use var irf irf http://www.stata-press.com/data/r12/lutkepohl2 dln_inv dln_inc dln_consump if qtr<=tq(1978q4), lags(1/2) dfk create order1, step(10) set(myirf1, new) create order2, step(10) order(dln_inc dln_inv dln_consump)
We now wish to compare the oirf for impulse dln inc and response dln consump for two different Cholesky orderings:

. irf ograph (order1 dln_inc dln_consump oirf) > (order2 dln_inc dln_consump oirf)
.006
213
.002
0 0
.002
.004
4 step
10
order1: oirf of dln_inc > dln_consump order2: oirf of dln_inc > dln_consump
Technical note
Graph options allow you to change the appearance of each plot. The following graph contains the plots of the FEVDs (FEVDs) for impulse dln inc and each response using the results from the rst collection of results in the active IRF le (using the . shortcut). In the second plot, we supply the clpat(dash) option (an abbreviation for clpattern(dash)) to give the line a dashed pattern. In the third plot, we supply the m(o) clpat(dashdot) recast(connected) options to get small circles connected by a line with a dashdot pattern; the cilines option plots the condence bands by using lines instead of areas. We use the title() option to add a descriptive title to the graph and supply the ci option globally to add condence bands to all the plots.
214

. irf ograph (. dln_inc dln_inc fevd) > (. dln_inc dln_consump fevd, clpat(dash)) > (. dln_inc dln_inv fevd, cilines m(o) clpat(dashdot) > recast(connected)) > , ci title("Comparison of forecast-error variance decomposition")
Comparison of forecasterror variance decomposition

.8 0 0 .2 .4 .6 1
4 step
10
95% CI of fevd of dln_inc > dln_inc 95% CI of fevd of dln_inc > dln_consump 95% CI of fevd of dln_inc > dln_inv fevd of dln_inc > dln_inc fevd of dln_inc > dln_consump fevd of dln_inc > dln_inv
The clpattern() option is described in [G-3] connect options, msymbol() is described in [G-3] marker options, title() is described in [G-3] title options, and recast() is described in [G-3] advanced options.
Saved results
irf ograph saves the following in r():
Scalars r(plots) r(ciplots) Macros r(irfname#) r(impulse#) r(response#) r(stat#) r(ci#) number of plot specications number of plotted condence bands irfname from (spec#) impulse from (spec#) response from (spec#) statistics from (spec#) level from (spec#) or noci

irf ograph is implemented as an ado-le.
215
Also see
Title
irf rename Rename an IRF result in an IRF le
Syntax
irf rename oldname newname , set(lename)
Menu
Statistics
>
>
>
Rename IRF results
Description
irf rename changes the name of a set of IRF results stored in the active IRF le.
Option
set(lename) species the le to be made active; see [TS] irf set. If set() is not specied, the active le is used.
Remarks
Example 1

. irf create original, set(myirfs, replace) (file myirfs.irf created) (file myirfs.irf now active) (file myirfs.irf updated) . irf create order2, order(dln_inc dln_inv dln_consump) (file myirfs.irf updated) . irf create order3, order(dln_inc dln_consump dln_inv) (file myirfs.irf updated) . irf describe Contains irf results from myirfs.irf (dated 4 Apr 2011 13:06) irfname model endogenous variables and order (*) original order2 order3 var var var dln_inv dln_inc dln_consump dln_inc dln_inv dln_consump dln_inc dln_consump dln_inv
216
irf rename Rename an IRF result in an IRF le
217
Now lets rename IRF result original to order1.

. irf rename original order1 (81 real changes made) original renamed to order1 . irf describe Contains irf results from myirfs.irf (dated 4 Apr 2011 13:06) model endogenous variables and order (*) irfname order1 order2 order3 var var var dln_inv dln_inc dln_consump dln_inc dln_inv dln_consump dln_inc dln_consump dln_inv
original has been renamed to order1.
Saved results
irf rename saves the following in r():
Macros r(irfnames) r(oldnew) irfnames after rename oldname newname

irf rename is implemented as an ado-le.
Also see
Title
irf set Set the active IRF le
Syntax
Report identity of active le irf set Set, and if necessary create, active le irf set irf lename Create, and if necessary replace, active le irf set irf lename, replace Clear any active IRF le irf set, clear
Menu
Statistics
>
>
>
Set active IRF le
Description
In the rst syntax, irf set reports the identity of the active le, if there is one. Also see [TS] irf describe for obtaining reports on the contents of an IRF le. In the second syntax, irf set irf lename species that the le be set as the active le and, if the le does not exist, that it be created as well. In the third syntax, irf set irf lename, replace species that even if le irf lename exists, a new, empty le is to be created and set. In the rarely used fourth syntax, irf set, clear species that, if any IRF le is set, it be unset and that there be no active IRF le.
IRF les are just les: they can be erased by erase, listed by dir, and copied by copy; see [D] erase, [D] dir, and [D] copy.
If irf lename is specied without an extension, .irf is assumed.
Options
replace species that if irf lename already exists, the le is to be erased and a new, empty IRF le is to be created in its place. If it does not already exist, a new, empty le is created. clear unsets the active IRF le.
218
irf set Set the active IRF le
219
Remarks
If you have not read [TS] irf, please do so. irf set reports the identity of the active IRF le:
. irf set no irf file active
irf set irf lename creates and sets an IRF le:

. irf set results1 (file results1.irf now active)
We specied the name results1, and results1.irf became the active le. The sufx .irf was added for us. irf set irf lename can also be used to create a new le:
. use http://www.stata-press.com/data/r12/lutkepohl2 (Quarterly SA West German macro data, Bil DM, from Lutkepohl 1993 Table E.1) . var dln_inc dln_consump, exog(l.dln_inv) (output omitted ) . irf set results2 (file results2.irf created) (file results2.irf now active) . irf create order1 (file results2.irf updated)
Saved results
irf set saves the following in r():
Macros r(Orville) name of active IRF le, if there is an active IRF

irf set is implemented as an ado-le.
Also see
Title
irf table Create tables of IRFs, dynamic-multiplier functions, and FEVDs
Syntax
irf table stat
Main
stat
, options Description impulseresponse function orthogonalized impulseresponse function dynamic-multiplier function cumulative impulseresponse function cumulative orthogonalized impulseresponse function cumulative dynamic-multiplier function Cholesky forecast-error variance decomposition structural impulseresponse function structural forecast-error variance decomposition
irf oirf dm cirf coirf cdm fevd sirf sfevd
If stat is not specied, all statistics are included, unless option nostructural is also specied, in which case sirf and sfevd are excluded. You may specify more than one stat.
options
Main
Description make lename active use irfnames IRF result sets use impulsevar as impulse variables use endogenous variables as response variables make an individual table for each result set use text for overall table title set condence level; default is level(95) suppress condence intervals include standard errors in the tables suppress sirf and sfevd from the default list of statistics use common maximum step horizon # for all tables
set(lename) irf(irfnames) impulse(impulsevar) response(endogvars) individual title("text")

Options
level(#) noci stderror nostructural step(#)
Menu
Statistics
>
>
>
Tables by impulse or response
220
221
Description
irf table makes a table from the specied IRF results. The rows of the tables are the time since impulse. Each column represents a combination of impulse() variable and response() variable for a stat from the irf() results.
Options

Main
set(lename) species the le to be made active; see [TS] irf set. If set() is not specied, the active le is used. All results are obtained from one IRF le. If you have results in different les that you want in one table, use irf add to copy results into one le; see [TS] irf add. irf(irfnames) species the IRF result sets to be used. If irf() is not specied, all the results in the active IRF le are used. (Files often contain just one set of IRF results, stored under one irfname; in that case, those results are used. When there are multiple IRF results, you may also wish to specify the individual option.) impulse(impulsevar) species the impulse variables for which the statistics are to be reported. If impulse() is not specied, each model variable, in turn, is used. impulsevar should be specied as an endogenous variable for all statistics except dm or cdm; for those, specify as an exogenous variable. response(endogvars) species the response variables for which the statistics are to be reported. If response() is not specied, each endogenous variable, in turn, is used. individual species that each set of IRF results be placed in its own table, with its own title and footer. By default, irf table places all the IRF results in one table with one title and one footer. individual may not be combined with title(). title("text") species a title for the overall table.
Options
level(#) species the default condence level, as a percentage, for condence intervals, when they are reported. The default is level(95) or as set by set level; see [U] 20.7 Specifying the width of condence intervals. noci suppresses reporting of the condence intervals for each statistic. noci is assumed when the model was t by vec because no condence intervals were estimated. stderror species that standard errors for each statistic also be included in the table. nostructural species that stat, when not specied, exclude sirf and sfevd. step(#) species the maximum step horizon for all tables. If step() is not specied, each table is constructed using all steps available.
Remarks
If you have not read [TS] irf, please do so. Also see [TS] irf graph, which produces output in graphical form, and see [TS] irf ctable, which also produces tabular output. irf ctable is more difcult to use but provides more control over how tables are formed.
222
Example 1
We have t a model with var, and we saved the IRFs from two different orderings. The commands we previously used were
. . . . . use var irf irf irf http://www.stata-press.com/data/r12/lutkepohl2 dln_inv dln_inc dln_consump set results4 create ordera, step(8) create orderb, order(dln_inc dln_inv dln_consump) step(8)
We now wish to compare the two orderings:

. irf table oirf fevd, impulse(dln_inc) response(dln_consump) noci std > title("Ordera versus orderb") Ordera versus orderb (1) oirf .005123 .001635 .002948 -.000221 .000811 .000462 .000044 .000151 .000091 (1) S.E. .000878 .000984 .000993 .000662 .000586 .000333 .000275 .000162 .000114 (1) fevd 0 .288494 .294288 .322454 .319227 .322579 .323552 .323383 .323499 (1) S.E. 0 .077483 .073722 .075562 .074063 .075019 .075371 .075314 .075386
step 0 1 2 3 4 5 6 7 8
step 0 1 2 3 4 5 6 7 8
(2) oirf .005461 .001578 .003307 -.00019 .000846 .000491 .000069 .000158 .000096
(2) S.E. .000925 .000988 .001042 .000676 .000617 .000349 .000292 .000172 .000122
(2) fevd 0 .327807 .328795 .370775 .366896 .370399 .371487 .371315 .371438
(2) S.E. 0 .08159 .077519 .080604 .079019 .079941 .080323 .080287 .080366
(1) irfname = ordera, impulse = dln_inc, and response = dln_consump (2) irfname = orderb, impulse = dln_inc, and response = dln_consump
The output is displayed as a single table; because the table did not t horizontally, it wrapped automatically. At the bottom of the table is a denition of the keys that appear at the top of each column. The results in the table above indicate that the orthogonalized IRFs do not change by much.
Example 2
Because the estimated FEVDs do change signicantly, we might want to produce two tables that contain the estimated FEVDs and their 95% condence intervals:

. irf table fevd, impulse(dln_inc) response(dln_consump) individual Results from ordera (1) fevd 0 .288494 .294288 .322454 .319227 .322579 .323552 .323383 .323499 (1) Lower 0 .13663 .149797 .174356 .174066 .175544 .175826 .17577 .175744 (1) Upper 0 .440357 .43878 .470552 .464389 .469613 .471277 .470995 .471253
223
step 0 1 2 3 4 5 6 7 8
95% lower and upper bounds reported (1) irfname = ordera, impulse = dln_inc, and response = dln_consump Results from orderb (1) fevd 0 .327807 .328795 .370775 .366896 .370399 .371487 .371315 .371438 (1) Lower 0 .167893 .17686 .212794 .212022 .213718 .214058 .213956 .213923 (1) Upper 0 .487721 .48073 .528757 .52177 .52708 .528917 .528674 .528953
step 0 1 2 3 4 5 6 7 8
95% lower and upper bounds reported (1) irfname = orderb, impulse = dln_inc, and response = dln_consump
Because we specied the individual option, the output contains two tables, one for each set of
IRF results. Examining the results in the tables indicates that each of the estimated functions is well
within the condence interval of the other, so we conclude that the functions are not signicantly different.
Technical note
Be careful in how you name variables when you t models. Say that you t one model with var and used time-series operators to form one of the endogenous variables
. var d.ln_inv
...
and in another model, you created a new variable:

. gen dln_inv = d.ln_inv . var dln_inv . . .
Say that you saved IRF results from both (perhaps they differ in the number of lags). Now you wish to use irf table to compare them. You would not be able to specify response(d.ln inv) or response(dln inv) because neither variable is in both models. Similarly, you could not specify impulse(d.ln inv) or impulse(dln inv) for the same reason.
224
All is not lost; if impulse() is not specied, all endogenous variables are used, and similarly if response() is not specied, so you could obtain the result you desired by simply not specifying the options, but you will also obtain a lot more, besides. If you want to specify the impulse() or response() options, be sure to name variables consistently. Also, you may forget how the endogenous variables were named. If so, irf describe, detail can provide the answer. In irf describes output, the endogenous variables are listed next to endog.
Saved results
If the individual option is not specied, irf table saves the following in r():
Scalars r(ncols) r(k umax) r(k) Macros r(key#) r(tnotes) number of columns in table number of distinct keys number of specic table commands #th key list of keys applied to each column
If the individual option is specied, then for each irfname, irf table saves the following in r():
Scalars r(irfname r(irfname r(irfname Macros r(irfname r(irfname ncols) k umax) k) key#) tnotes) number of columns in table for irfname number of distinct keys in table for irfname number of specic table commands used to create table for irfname #th key for irfname table list of keys applied to each column in table for irfname

irf table is implemented as an ado-le.
Also see
Title
mgarch Multivariate GARCH models
Syntax
mgarch model eq eq . . . eq if in , ...
Family Vech Diagonal vech Conditional correlation constant conditional correlation dynamic conditional correlation varying conditional correlation
model dvech ccc dcc vcc
See [TS] mgarch dvech, [TS] mgarch ccc, [TS] mgarch dcc, and [TS] mgarch vcc for details.
Description
mgarch estimates the parameters of multivariate generalized autoregressive conditionalheteroskedasticity (MGARCH) models. MGARCH models allow both the conditional mean and the conditional covariance to be dynamic. The general MGARCH model is so exible that not all the parameters can be estimated. For this reason, there are many MGARCH models that parameterize the problem more parsimoniously. mgarch implements four commonly used parameterizations: the diagonal vech model, the constant conditional correlation model, the dynamic conditional correlation model, and the time-varying conditional correlation model.
Remarks
An introduction to MGARCH models Diagonal vech MGARCH models Conditional correlation MGARCH models Constant conditional correlation MGARCH model Dynamic conditional correlation MGARCH model Varying conditional correlation MGARCH model Error distributions and quasimaximum likelihood Treatment of missing data
225
226
An introduction to MGARCH models

Multivariate GARCH models allow the conditional covariance matrix of the dependent variables to follow a exible dynamic structure and allow the conditional mean to follow a vector-autoregressive (VAR) structure. The general MGARCH model is too exible for most problems. There are many restricted MGARCH models in the literature because there is no parameterization that always provides an optimal trade-off between exibility and parsimony. mgarch implements four commonly used parameterizations: the diagonal vech (DVECH) model, the constant conditional correlation (CCC) model, the dynamic conditional correlation (DCC) model, and the time-varying conditional correlation (VCC) model. Bollerslev, Engle, and Wooldridge (1988); Bollerslev, Engle, and Nelson (1994); Bauwens, Laurent, and Rombouts (2006); Silvennoinen and Ter svirta (2009); and Engle (2009) provide general a introductions to MGARCH models. We provide a quick introduction organized around the models implemented in mgarch. We give a formal denition of the general MGARCH model to establish notation that facilitates comparisons of the models. The general MGARCH model is given by
yt = Cxt +
t
1/2 Ht t
where
yt is an m 1 vector of dependent variables; C is an m k matrix of parameters; xt is a k 1 vector of independent variables, which may contain lags of yt ; Ht
1/2
is the Cholesky factor of the time-varying conditional covariance matrix Ht ; and
t is an m 1 vector of zero-mean, unit-variance, and independent and identically distributed innovations. In the general MGARCH model, Ht is a matrix generalization of univariate GARCH models. For example, in a general MGARCH model with one autoregressive conditional heteroskedastic (ARCH) term and one GARCH term,
vech (Ht ) = s + Avech
t1 t1
+ Bvech (Ht1 )
(1)
where the vech() function stacks the unique elements that lie on or below the main diagonal in a symmetric matrix into a vector, s is a vector of parameters, and A and B are conformable matrices of parameters. Because this model uses the vech() function to extract and model the unique elements of Ht , it is also known as the VECH model. Because it is a conditional covariance matrix, Ht must be positive denite. Equation (1) can be used to show that the parameters in s, A, and B are not uniquely identied and that further restrictions must be placed on s, A, and B to ensure that Ht is positive denite for all t.
227
The various MGARCH models proposed in the literature differ in how they trade off exibility and parsimony in their specications for Ht . Increased exibility allows a model to capture more complex Ht processes. Increased parsimony makes parameter estimation feasible for more datasets. An important measure of the exibilityparsimony trade-off is how fast the number of model parameters increases with the number of time series m, because many applied models use multiple time series.
Diagonal vech MGARCH models

Bollerslev, Engle, and Wooldridge (1988) derived the diagonal vech (DVECH) model by restricting A and B to be diagonal. Although the DVECH model is much more parsimonious than the general model, it can only handle a few series because the number of parameters grows quadratically with the number of series. For example, there are 3m(m + 1)/2 parameters in a DVECH(1,1) model for Ht . Despite the large number of parameters, the diagonal structure implies that each conditional variance and each conditional covariance depends on its own past but not on the past of the other conditional variances and covariances. Formally, in the DVECH(1,1) model each element of Ht is modeled by
hij,t = sij + aij
i,(t1) j,(t1)
+ bij hij,(t1)
Parameter estimation can be difcult because it requires that Ht be positive denite for each t. The requirement that Ht be positive denite for each t imposes complicated restrictions on the off-diagonal elements. See [TS] mgarch dvech for more details about this model.
Conditional correlation MGARCH models

Conditional correlation (CC) models use nonlinear combinations of univariate GARCH models to represent the conditional covariances. In each of the conditional correlation models, the conditional covariance matrix is positive denite by construction and has a simple structure, which facilitates parameter estimation. CC models have a slower parameter growth rate than DVECH models as the number of time series increases. In CC models, Ht is decomposed into a matrix of conditional correlations Rt and a diagonal matrix of conditional variances Dt :
Ht = Dt Rt Dt
1/2
1/2
(2)
where each conditional variance follows a univariate GARCH process and the parameterizations of Rt vary across models. Equation (2) implies that
hij,t = ij,t i,t j,t
(3)
2 where i,t is modeled by a univariate GARCH process. Equation (3) highlights that CC models use nonlinear combinations of univariate GARCH models to represent the conditional covariances and that the parameters in the model for ij,t describe the extent to which the errors from equations i and j move together.
228
Comparing (1) and (2) shows that the number of parameters increases more slowly with the number of time series in a CC model than in a DVECH model. The three CC models implemented in mgarch differ in how they parameterize Rt .
Constant conditional correlation MGARCH model
Bollerslev (1990) proposed a CC MGARCH model in which the correlation matrix is time invariant. It is for this reason that the model is known as a constant conditional correlation (CCC) MGARCH model. Restricting Rt to a constant matrix reduces the number of parameters and simplies the estimation but may be too strict in many empirical applications. See [TS] mgarch ccc for more details about this model.
Dynamic conditional correlation MGARCH model
Engle (2002) introduced a dynamic conditional correlation (DCC) MGARCH model in which the conditional quasicorrelations Rt follow a GARCH(1,1)-like process. (As described by Engle [2009] and Aielli [2009], the parameters in Rt are not standardized to be correlations and are thus known as quasicorrelations.) To preserve parsimony, all the conditional quasicorrelations are restricted to follow the same dynamics. The DCC model is signicantly more exible than the CCC model without introducing an unestimable number of parameters for a reasonable number of series. See [TS] mgarch dcc for more details about this model.
Varying conditional correlation MGARCH model
Tse and Tsui (2002) derived the varying conditional correlation (VCC) MGARCH model in which the conditional correlations at each period are a weighted sum of a time-invariant component, a measure of recent correlations among the residuals, and last periods conditional correlations. For parsimony, all the conditional correlations are restricted to follow the same dynamics. See [TS] mgarch vcc for more details about this model.
Error distributions and quasimaximum likelihood

By default, mgarch dvech, mgarch ccc, mgarch dcc, and mgarch vcc estimate the parameters of MGARCH models by maximum likelihood (ML), assuming that the errors come from a multivariate normal distribution. Both the ML estimator and the quasimaximum likelihood (QML) estimator, which drops the normality assumption, are assumed to be consistent and normally distributed in large samples; see Jeantheau (1998), Berkes and Horv th (2003), Comte and Lieberman (2003), Ling and a McAleer (2003), and Fiorentini and Sentana (2007). Specify vce(robust) to estimate the parameters by QML. The QML parameter estimates are the same as the ML estimates, but the VCEs are different. Based on low-level assumptions, Jeantheau (1998), Comte and Lieberman (2003), and Ling and McAleer (2003) prove that some of the ML and QML estimators implemented in mgarch are consistent and asymptotically normal. Based on higher-level assumptions, Fiorentini and Sentana (2007) prove that all the ML and QML estimators implemented in mgarch are consistent and asymptotically normal. The low-level assumption proofs specify the technical restrictions on the data-generating processes more precisely than the high-level proofs, but they do not cover as many models or cases as the high-level proofs.
229
It is generally accepted that there could be more low-level theoretical work done to substantiate the claims that the ML and QML estimators are consistent and asymptotically normally distributed. These widely applied estimators have been subjected to many Monte Carlo studies that show that the large-sample theory performs well in nite samples. The distribution(t) option causes the mgarch commands to estimate the parameters of the corresponding model by ML assuming that the errors come from a multivariate Student t distribution. The choice between the multivariate normal and the multivariate t distributions is one between robustness and efciency. If the disturbances come from a multivariate Student t, then the ML estimates based on the multivariate Student t assumption will be consistent and efcient, while the QML estimates based on the multivariate normal assumption will be consistent but not efcient. In contrast, if the disturbances come from a well-behaved distribution that is neither multivariate Student t nor multivariate normal, then the ML estimates based on the multivariate Student t assumption will not be consistent, while the QML estimates based on the multivariate normal assumption will be consistent but not efcient. Fiorentini and Sentana (2007) compare the ML and QML estimators implemented in mgarch and provide many useful technical results pertaining to the estimators.
Treatment of missing data

mgarch allows for gaps due to missing data. The unconditional expectations are substituted for the dynamic components that cannot be computed because of gaps. This method of handling gaps can only handle the case in which g/T goes to zero as T goes to innity, where g is the number of observations lost to gaps in the data and T is the number of nonmissing observations.
References
Aielli, G. P. 2009. Dynamic Conditional Correlations: On Properties and Estimation. Working paper, Dipartimento di Statistica, University of Florence, Florence, Italy. Bauwens, L., S. Laurent, and J. V. K. Rombouts. 2006. Multivariate GARCH models: A survey. Journal of Applied Econometrics 21: 79109. Berkes, I., and L. Horv th. 2003. The rate of consistency of the quasi-maximum likelihood estimator. Statistics and a Probability Letters 61: 133143. Bollerslev, T. 1990. Modelling the coherence in short-run nominal exchange rates: A multivariate generalized ARCH model. Review of Economics and Statistics 72: 498505. Bollerslev, T., R. F. Engle, and D. B. Nelson. 1994. ARCH models. In Handbook of Econometrics, Volume IV, ed. R. F. Engle and D. L. McFadden. New York: Elsevier. Bollerslev, T., R. F. Engle, and J. M. Wooldridge. 1988. A capital asset pricing model with time-varying covariances. Journal of Political Economy 96: 116131. Comte, F., and O. Lieberman. 2003. Asymptotic theory for multivariate GARCH processes. Journal of Multivariate Analysis 84: 6184. Engle, R. F. 2002. Dynamic conditional correlation: A simple class of multivariate generalized autoregressive conditional heteroskedasticity models. Journal of Business & Economic Statistics 20: 339350. . 2009. Anticipating Correlations: A New Paradigm for Risk Management. Princeton, NJ: Princeton University Press. Fiorentini, G., and E. Sentana. 2007. On the efciency and consistency of likelihood estimation in multivariate conditionally heteroskedastic dynamic regression models. Working paper 0713, CEMFI, Madrid, Spain. ftp://ftp.cem.es/wp/07/0713.pdf. Jeantheau, T. 1998. Strong consistency of estimators for multivariate ARCH models. Economic Theory 14: 7086.
230
Ling, S., and M. McAleer. 2003. Asymptotic theory for a vector ARMGARCH model. Economic Theory 19: 280310. Silvennoinen, A., and T. Ter svirta. 2009. Multivariate GARCH models. In Handbook of Financial Time Series, ed. a T. G. Andersen, R. A. Davis, J.-P. Krei, and T. Mikosch, 201229. New York: Springer. Tse, Y. K., and A. K. C. Tsui. 2002. A multivariate generalized autoregressive conditional heteroscedasticity model with time-varying correlations. Journal of Business & Economic Statistics 20: 351362.
Also see
[TS] arch Autoregressive conditional heteroskedasticity (ARCH) family of estimators [TS] var Vector autoregressive models [U] 20 Estimation and postestimation commands
Title
mgarch ccc Constant conditional correlation multivariate GARCH models
Syntax
mgarch ccc eq eq . . . eq if in , options where each eq has the form (depvars = indepvars options
Model
, eqoptions ) Description
ARCH terms for all equations GARCH terms for all equations
arch(numlist) garch(numlist) het(varlist) distribution(dist # ) unconcentrated constraints(numlist)

SE/Robust
include varlist in the specication of the conditional variance for all equations use dist distribution for errors [may be gaussian (synonym normal) or t; default is gaussian] perform optimization on unconcentrated log likelihood apply linear constraints vcetype may be oim or robust set condence level; default is level(95) do not display constraints control column formats, row spacing, line width, and display of omitted variables and base and empty cells control the maximization process; seldom used initial values for the coefcients; seldom used display legend instead of statistics
vce(vcetype)
Reporting

Maximization
maximize options from(matname) coeflegend
eqoptions noconstant arch(numlist) garch(numlist) het(varlist)
Description suppress constant term in the mean equation

include varlist in the specication of the conditional variance
You must tsset your data before using mgarch ccc; see [TS] tsset. indepvars and varlist may contain factor variables; see [U] 11.4.3 Factor variables. depvars, indepvars, and varlist may contain time-series operators; see [U] 11.4.4 Time-series varlists. by, statsby, and rolling are allowed; see [U] 11.1.10 Prex commands. coeflegend does not appear in the dialog box. See [U] 20 Estimation and postestimation commands for more capabilities of estimation commands. 231
232
Menu
Statistics
>
>
Multivariate GARCH
Description
mgarch ccc estimates the parameters of constant conditional correlation (CCC) multivariate generalized autoregressive conditionally heteroskedastic (MGARCH) models in which the conditional variances are modeled as univariate generalized autoregressive conditionally heteroskedastic (GARCH) models and the conditional covariances are modeled as nonlinear functions of the conditional variances. The conditional correlation parameters that weight the nonlinear combinations of the conditional variance are constant in the CCC MGARCH model. The CCC MGARCH model is less exible than the dynamic conditional correlation MGARCH model (see [TS] mgarch dcc) and varying conditional correlation MGARCH model (see [TS] mgarch vcc), which specify GARCH-like processes for the conditional correlations. The conditional correlation MGARCH models are more parsimonious than the diagonal vech MGARCH model (see [TS] mgarch dvech).
Options

Model
arch(numlist) species the ARCH terms for all equations in the model. By default, no ARCH terms are specied. garch(numlist) species the GARCH terms for all equations in the model. By default, no GARCH terms are specied. het(varlist) species that varlist be included in the model in the specication of the conditional variance for all equations. This varlist enters the variance specication collectively as multiplicative heteroskedasticity. distribution(dist # ) species the assumed distribution for the errors. dist may be gaussian, normal, or t. gaussian and normal are synonyms; each causes mgarch ccc to assume that the errors come from a multivariate normal distribution. # cannot be specied with either of them. t causes mgarch ccc to assume that the errors follow a multivariate Student t distribution, and the degree-of-freedom parameter is estimated along with the other parameters of the model. If distribution(t #) is specied, then mgarch ccc uses a multivariate Student t distribution with # degrees of freedom. # must be greater than 2. unconcentrated species that optimization be performed on the unconcentrated log likelihood. The default is to start with the concentrated log likelihood. constraints(numlist) species linear constraints to apply to the parameter estimates.
SE/Robust
vce(vcetype) species the estimator for the variancecovariance matrix of the estimator. vce(oim), the default, species to use the observed information matrix (OIM) estimator. vce(robust) species to use the Huber/White/sandwich estimator.
233
Reporting
level(#); see [R] estimation options. nocnsreport; see [R] estimation options. display options: noomitted, vsquish, noemptycells, baselevels, allbaselevels, cformat(% fmt), pformat(% fmt), sformat(% fmt), and nolstretch; see [R] estimation options.
Maximization
maximize options: difficult, technique(algorithm spec), iterate(#), no log, trace, gradient, showstep, hessian, showtolerance, tolerance(#), ltolerance(#), nrtolerance(#), and nonrtolerance, and from(matname); see [R] maximize for all options except from(), and see below for information on from(). These options are seldom used. from(matname) species initial values for the coefcients. from(b0) causes mgarch ccc to begin the optimization algorithm with the values in b0. b0 must be a row vector, and the number of columns must equal the number of parameters in the model. The following option is available with mgarch ccc but is not shown in the dialog box: coeflegend; see [R] estimation options.
Eqoptions
noconstant suppresses the constant term in the mean equation. arch(numlist) species the ARCH terms in the equation. By default, no ARCH terms are specied. This option may not be specied with model-level arch(). garch(numlist) species the GARCH terms in the equation. By default, no GARCH terms are specied. This option may not be specied with model-level garch(). het(varlist) species that varlist be included in the specication of the conditional variance. This varlist enters the variance specication collectively as multiplicative heteroskedasticity. This option may not be specied with model-level het().
Remarks
We assume that you have already read [TS] mgarch, which provides an introduction to MGARCH models and the methods implemented in mgarch ccc.
MGARCH models are dynamic multivariate regression models in which the conditional variances and covariances of the errors follow an autoregressive-moving-average structure. The CCC MGARCH model uses a nonlinear combination of univariate GARCH models in which the cross-equation weights are time invariant to model the conditional covariance matrix of the disturbances.
As discussed in [TS] mgarch, MGARCH models differ in the parsimony and exibility of their specications for a time-varying conditional covariance matrix of the disturbances, denoted by Ht . In the conditional correlation family of MGARCH models, the diagonal elements of Ht are modeled as univariate GARCH models, whereas the off-diagonal elements are modeled as nonlinear functions of the diagonal terms. In the CCC MGARCH model,
hij,t = ij
hii,t hjj,t
234
where the diagonal elements hii,t and hjj,t follow univariate GARCH processes and ij is a timeinvariate weight interpreted as a conditional correlation. In the dynamic conditional correlation (DCC) and varying conditional correlation (VCC) MGARCH models discussed in [TS] mgarch dcc and [TS] mgarch vcc, the ij are allowed to vary over time. Although the conditional-correlation structure provides a useful trade-off between parsimony and exibility in the DCC MGARCH and VCC MGARCH models, the time-invariant parameterization used in the CCC MGARCH model is generally viewed as too restrictive for many applications; see Silvennoinen and Ter svirta (2009). The baseline CCC MGARCH estimates are frequently compared a with DCC MGARCH and VCC MGARCH estimates.
Technical note
Formally, the CCC MGARCH model derived by Bollerslev (1990) can be written as
yt = Cxt +
t
= Ht t
1/2 1/2
1/2
Ht = Dt RDt
where
1/2
is the Cholesky factor of the time-varying conditional covariance matrix Ht ;
t is an m 1 vector of normal, independent, and identically distributed innovations;
Dt is a diagonal matrix of conditional variances,

2 1,t 0 Dt = . . . 0
0 2 2,t . . . 0
0 0 . .. . . . 2 m,t
2 in which each i,t evolves according to a univariate GARCH model of the form 2 i,t = si + pi j=1
2 i,tj
+ j
qi j=1
2 j i,tj qi j=1
by default, or
2 i,t = exp(i zi,t ) + pi j=1 2 i,tj
2 j i,tj
when the het() option is specied, where t is a 1 p vector of parameters, zi is a p 1 vector of independent variables including a constant term, the j s are ARCH parameters, and the j s are GARCH parameters; and
R is a matrix of time-invariant unconditional correlations of the standardized residuals 1/2 Dt t, 1 12 1m 1 2m 12 R= . . . .. . . . . . . . 1m 2m 1
235
This model is known as the constant conditional correlation MGARCH model because R is time invariant.

We have daily data on the stock returns of three car manufacturersToyota, Nissan, and Honda, from January 2, 2003, to December 31, 2010in the variables toyota, nissan, and honda. We model the conditional means of the returns as a rst-order vector autoregressive process and the conditional covariances as a CCC MGARCH process in which the variance of each disturbance term follows a GARCH(1,1) process. We specify the noconstant option, because the returns have mean zero. The estimated constants in the variance equations are near zero in this example because of how the data are scaled.
. use http://www.stata-press.com/data/r12/stocks (Data from Yahoo! Finance) . mgarch ccc (toyota nissan honda = L.toyota L.nissan L.honda, noconstant), > arch(1) garch(1) Calculating starting values.... Optimizing concentrated log likelihood (setting technique to bhhh) Iteration 0: log likelihood = 16898.994 Iteration 1: log likelihood = 17008.914 Iteration 2: log likelihood = 17156.946 Iteration 3: log likelihood = 17249.527 Iteration 4: log likelihood = 17287.251 Iteration 5: log likelihood = 17313.5 Iteration 6: log likelihood = 17335.087 Iteration 7: log likelihood = 17356.534 Iteration 8: log likelihood = 17376.051 Iteration 9: log likelihood = 17400.035 (switching technique to nr) Iteration 10: log likelihood = 17423.634 Iteration 11: log likelihood = 17440.774 Iteration 12: log likelihood = 17446.838 Iteration 13: log likelihood = 17447.635 Iteration 14: log likelihood = 17447.645 Iteration 15: log likelihood = 17447.645 Optimizing unconcentrated log likelihood Iteration 0: Iteration 1: Iteration 2: log likelihood = log likelihood = log likelihood = 17447.645 17447.651 17447.651 Number of obs Wald chi2(9) Prob > chi2 = = = 2014 17.46 0.0420
Constant conditional correlation MGARCH model Sample: 1 - 2015 Distribution: Gaussian Log likelihood = 17447.65
236
Coef. toyota toyota L1. nissan L1. honda L1. ARCH_toyota arch L1. garch L1. _cons nissan toyota L1. nissan L1. honda L1. ARCH_nissan arch L1. garch L1. _cons honda toyota L1. nissan L1. honda L1. ARCH_honda arch L1. garch L1. _cons -.0378616
Std. Err.
P>|z|
-.0537817
.0353211
-1.52
0.128
-.1230098
.0154463
.026686
.024841
1.07
0.283
-.0220015
.0753734
-.0043073
.0302761
-0.14
0.887
-.0636473
.0550327
.0615321
.0087313
7.05
0.000
.0444191
.0786452
.9213798 4.42e-06
.0110412 1.12e-06
83.45 3.93
0.000 0.000
.8997395 2.21e-06
.9430201 6.62e-06
-.0232321
.0400563
-0.58
0.562
-.1017411
.0552769
-.0299552
.0309362
-0.97
0.333
-.0905891
.0306787
.0369229
.0360532
1.02
0.306
-.0337402
.1075859
.0740294
.0119353
6.20
0.000
.0506366
.0974222
.9102548 6.36e-06
.0142328 1.76e-06
63.95 3.61
0.000 0.000
.8823589 2.91e-06
.9381506 9.81e-06
.036792
-1.03
0.303
-.1099727
.0342495
.0551649
.0272559
2.02
0.043
.0017444
.1085855
-.0431919
.0331268
-1.30
0.192
-.1081193
.0217354
.0433036
.0070224
6.17
0.000
.0295399
.0570673
.939117 5.02e-06
.010131 1.31e-06
92.70 3.83
0.000 0.000
.9192605 2.45e-06
.9589735 7.59e-06
237
Correlation toyota nissan honda nissan honda
.6532264 .7185412
.0128035 .0108132
51.02 66.45
0.000 0.000
.628132 .6973477
.6783208 .7397346
.6298972
.0135336
46.54
0.000
.6033717
.6564226
The iteration log has three parts: the dots from the search for initial values, the iteration log from optimizing the concentrated log likelihood, and the iteration log from maximizing the unconcentrated log likelihood. A detailed discussion of the optimization methods can be found in Methods and formulas. The header describes the estimation sample and reports a Wald test against the null hypothesis that all the coefcients on the independent variables in the mean equations are zero. Here the null hypothesis is rejected at the 5% level. The output table rst presents results for the mean or variance parameters used to model each dependent variable. Subsequently, the output table presents results for the conditional correlation parameters. For example, the conditional correlation between the standardized residuals for Toyota and Nissan is estimated to be 0.65. The output above indicates that we may not need all the vector autoregressive parameters, but that each of the univariate ARCH, univariate GARCH, and conditional correlation parameters are statistically signicant. That the estimated conditional correlation parameters are positive and signicant indicates that the returns on these stocks rise or fall together. That the conditional correlations are time invariant is a restrictive assumption. The DCC MGARCH model and the VCC MGARCH model nest the CCC MGARCH model. When we test the time-invariance assumption with Wald tests on the parameters of these more general models in [TS] mgarch dcc and [TS] mgarch vcc, we reject the null hypothesis that these conditional correlations are time invariant.
Example 2
We improve the previous example by removing the insignicant parameters from the model. To remove these parameters, we specify the honda equation separately from the toyota and nissan equations:
. mgarch ccc (toyota nissan = , noconstant) (honda = L.nissan, noconstant), > arch(1) garch(1) Calculating starting values.... Optimizing concentrated log likelihood (setting technique to bhhh) Iteration 0: log likelihood Iteration 1: log likelihood Iteration 2: log likelihood Iteration 3: log likelihood Iteration 4: log likelihood Iteration 5: log likelihood Iteration 6: log likelihood Iteration 7: log likelihood Iteration 8: log likelihood Iteration 9: log likelihood (switching technique to nr) = = = = = = = = = = 16886.88 16974.779 17147.893 17247.473 17285.549 17311.153 17333.588 17353.717 17374.895 17400.669
238

Iteration 10: log likelihood = 17425.661 Iteration 11: log likelihood = 17436.768 Iteration 12: log likelihood = 17439.738 Iteration 13: log likelihood = 17439.865 Iteration 14: log likelihood = 17439.866 Optimizing unconcentrated log likelihood Iteration 0: log likelihood = 17439.866 Iteration 1: log likelihood = 17439.872 Iteration 2: log likelihood = 17439.872 Constant conditional correlation MGARCH model Sample: 1 - 2015 Distribution: Gaussian Log likelihood = 17439.87 Coef. ARCH_toyota arch L1. garch L1. _cons ARCH_nissan arch L1. garch L1. _cons honda nissan L1. ARCH_honda arch L1. garch L1. _cons Correlation toyota nissan honda nissan honda .0186628 .0138575 1.35 0.178 -.0084975 .0458231 Std. Err. z
Number of obs Wald chi2(1) Prob > chi2 P>|z|
= = =
2014 1.81 0.1781
.0619604
.0087942
7.05
0.000
.044724
.0791968
.9208961 4.43e-06
.0110996 1.13e-06
82.97 3.94
0.000 0.000
.8991414 2.23e-06
.9426508 6.64e-06
.0773095
.012328
6.27
0.000
.0531471
.1014719
.906088 6.77e-06
.0147303 1.85e-06
61.51 3.66
0.000 0.000
.8772171 3.14e-06
.9349589 .0000104
.0433741
.006996
6.20
0.000
.0296622
.0570861
.9391094 5.02e-06
.0100707 1.31e-06
93.25 3.83
0.000 0.000
.9193712 2.45e-06
.9588477 7.60e-06
.652299 .7189531
.0128271 .0108005
50.85 66.57
0.000 0.000
.6271583 .6977845
.6774396 .7401218
.628435
.0135653
46.33
0.000
.6018475
.6550225
It turns out that the coefcient on L1.nissan in the honda equation is now statistically insignicant. We could further improve the model by removing L1.nissan from the model. As expected, removing the insignicant parameters from conditional mean equations had almost no effect on the estimated conditional variance parameters.
239
There is no mean equation for Toyota or Nissan. In [TS] mgarch ccc postestimation, we discuss prediction from models without covariates.
Example 3
Here we t a bivariate CCC MGARCH model for the Toyota and Nissan shares. We believe that the shares of these car manufacturers follow the same process, so we impose the constraints that the ARCH and the GARCH coefcients are the same for the two companies.
. constraint 1 _b[ARCH_toyota:L.arch] = _b[ARCH_nissan:L.arch] . constraint 2 _b[ARCH_toyota:L.garch] = _b[ARCH_nissan:L.garch] . mgarch ccc (toyota nissan = , noconstant), arch(1) garch(1) constraints(1 2) Calculating starting values.... Optimizing concentrated log likelihood (setting technique to bhhh) Iteration 0: log likelihood = 10317.225 Iteration 1: log likelihood = 10630.464 Iteration 2: log likelihood = 10865.964 Iteration 3: log likelihood = 11063.329 (output omitted ) Iteration 8: log likelihood = 11273.962 Iteration 9: log likelihood = 11274.409 (switching technique to nr) Iteration 10: log likelihood = 11274.494 Iteration 11: log likelihood = 11274.499 Iteration 12: log likelihood = 11274.499 Optimizing unconcentrated log likelihood Iteration 0: log likelihood = 11274.499 Iteration 1: log likelihood = 11274.501 Iteration 2: log likelihood = 11274.501 Constant conditional correlation MGARCH model Sample: 1 - 2015 Number of obs = 2015 Distribution: Gaussian Wald chi2(.) = . Log likelihood = 11274.5 Prob > chi2 = . ( 1) [ARCH_toyota]L.arch - [ARCH_nissan]L.arch = 0 ( 2) [ARCH_toyota]L.garch - [ARCH_nissan]L.garch = 0 Coef. ARCH_toyota arch L1. garch L1. _cons ARCH_nissan arch L1. garch L1. _cons Std. Err. z P>|z| [95% Conf. Interval]
.0742678
.0095464
7.78
0.000
.0555572
.0929785
.9131674 3.77e-06
.0111558 1.02e-06
81.86 3.71
0.000 0.000
.8913024 1.78e-06
.9350323 5.77e-06
.0742678
.0095464
7.78
0.000
.0555572
.0929785
.9131674 5.30e-06
.0111558 1.36e-06
81.86 3.89
0.000 0.000
.8913024 2.63e-06
.9350323 7.97e-06
240
Correlation toyota nissan
.651389
.0128482
50.70
0.000
.6262071
.6765709
We could test our constraints by tting the unconstrained model and performing a likelihood-ratio test. The results indicate that the restricted model is preferable.
Example 4
In this example, we have data on ctional stock returns for the Acme and Anvil corporations and we believe that the movement of the two stocks is governed by different processes. We specify one ARCH and one GARCH term for the conditional variance equation for Acme and two ARCH terms for the conditional variance equation for Anvil. In addition, we include the lagged value of the stock return for Apex, the main subsidiary of Anvil corporation, in the variance equation of Anvil. For Acme, we have data on the changes in an index of futures prices of products related to those produced by Acme in afrelated. For Anvil, we have data on the changes in an index of futures prices of inputs used by Anvil in afinputs.
. use http://www.stata-press.com/data/r12/acmeh . mgarch ccc (acme = afrelated, noconstant arch(1) > (anvil = afinputs, arch(1/2) het(L.apex)) Calculating starting values.... Optimizing concentrated log likelihood (setting technique to bhhh) Iteration 0: log likelihood = -12996.245 Iteration 1: log likelihood = -12609.982 Iteration 2: log likelihood = -12563.103 Iteration 3: log likelihood = -12554.73 Iteration 4: log likelihood = -12554.542 Iteration 5: log likelihood = -12554.534 Iteration 6: log likelihood = -12554.534 Iteration 7: log likelihood = -12554.534 Optimizing unconcentrated log likelihood Iteration 0: log likelihood = -12554.534 Iteration 1: log likelihood = -12554.533 Constant conditional correlation MGARCH model Sample: 1 - 2500 Distribution: Gaussian Log likelihood = -12554.53 Coef. acme afrelated ARCH_acme arch L1. garch L1. _cons Std. Err. z garch(1))
= = =
2499 2212.30 0.0000
P>|z|
.9175148
.0651088
14.09
0.000
.7899039
1.045126
.0798719
.0169526
4.71
0.000
.0466455
.1130983
.7336823 2.880836
.0601569 .760206
12.20 3.79
0.000 0.000
.6157768 1.390859
.8515877 4.370812
241
anvil afinputs _cons ARCH_anvil arch L1. L2. apex L1. _cons Correlation acme anvil
-1.015561 .0703606
.0226437 .0211689
-44.85 3.32
0.000 0.001
-1.059942 .0288703
-.97118 .1118508
.4893288 .2782296
.0286012 .0208172
17.11 13.37
0.000 0.000
.4332714 .2374287
.5453862 .3190305
1.894972 .1034111
.0616293 .0735512
30.75 1.41
0.000 0.160
1.774181 -.0407466
2.015763 .2475688
-.5354047
.0143275
-37.37
0.000
-.563486
-.5073234
The results indicate that increases in the futures prices for related products lead to higher returns on the Acme stock, and increased input prices lead to lower returns on the Anvil stock. In the conditional variance equation for Anvil, the coefcient on L1.apex is positive and signicant, which indicates that an increase in the return on the Apex stock leads to more variability in the return on the Anvil stock. That the estimated conditional correlation between the two returns is 0.54 indicates that these returns tend to move in opposite directions; in other words, an increase in the return for the Acme stock tends to be associated with a decrease in the return for the Anvil stock, and vice versa.
Saved results
mgarch ccc saves the following in e():
Scalars e(N) e(k) e(k aux) e(k extra) e(k eq) e(k dv) e(df m) e(ll) e(chi2) e(p) e(estdf) e(usr) e(tmin) e(tmax) e(N gaps) e(rank) e(ic) e(rc) e(converged) number of observations number of parameters number of auxiliary parameters number of extra estimates added to number of equations in e(b) number of dependent variables model degrees of freedom log likelihood
2
signicance 1 if distribution parameter was estimated, 0 otherwise user-provided distribution parameter minimum time in sample maximum time in sample number of gaps rank of e(V) number of iterations return code 1 if converged, 0 otherwise
242
Macros e(cmd) e(model) e(cmdline) e(depvar) e(covariates) e(dv eqs) e(indeps) e(tvar) e(title) e(chi2type) e(vce) e(vcetype) e(tmins) e(tmaxs) e(dist) e(arch) e(garch) e(technique) e(properties) e(estat cmd) e(predict) e(marginsok) e(marginsnotok) Matrices e(b) e(Cns) e(ilog) e(gradient) e(hessian) e(V) e(pinfo) Functions e(sample)
mgarch ccc command as typed names of dependent variables list of covariates dependent variables with mean equations independent variables in each equation time variable title in estimation output Wald; type of model 2 test vcetype specied in vce() title used to label Std. Err. formatted minimum time formatted maximum time distribution for error term: gaussian or t specied ARCH terms specied GARCH terms maximization technique b V program used to implement estat program used to implement predict predictions allowed by margins predictions disallowed by margins
coefcient vector constraints matrix iteration log (up to 20 iterations) gradient vector Hessian matrix variancecovariance matrix of the estimators parameter information, used by predict marks estimation sample

mgarch ccc is implemented as an ado-le. mgarch ccc estimates the parameters of the CCC MGARCH model by maximum likelihood. The unconcentrated log-likelihood function based on the multivariate normal distribution for observation t is
lt = 0.5m log(2) 0.5log {det (R)} log det Dt

where t = Dt t is an m 1 vector of standardized residuals, T function is t=1 lt .
1/2 t
1/2
0.5 t R1
(1)
= yt Cxt . The log-likelihood
If we assume that t follow a multivariate t distribution with degrees of freedom (df) greater than 2, then the unconcentrated log-likelihood function for observation t is
lt = log
df + m 2
log
df 2
1/2
m log {(df 2)} 2

1 df + m tR t log 1 + 2 df 2
0.5log {det (R)} log det Dt
(2)
243
The correlation matrix R can be concentrated out of (1) and (2) by dening the (i, j)th element of R as
T T it jt t=1 t=1 2 it 1 2 T 2 jt t=1
1 2
ij =
mgarch ccc starts the optimization process with the concentrated log-likelihood function. The starting values for the parameters in the mean equations and the initial residuals t are obtained by least-squares regression. The starting values for the parameters in the variance equations are obtained by a procedure proposed by Gourieroux and Monfort (1997, sec. 6.2.2). If the optimization is started with the unconcentrated log likelihood, then the initial values for the parameters in R are calculated from the standardized residuals t .
GARCH estimators require initial values that can be plugged in for ti ti and Htj when t i < 1 and t j < 1. mgarch ccc substitutes an estimator of the unconditional covariance of the disturbances
T
= T 1
t=1
t t
(3)
for ti ti when t i < 1 and for Htj when t j < 1, where calculated using the estimated parameters.
is the vector of residuals
mgarch ccc requires a sample size that at the minimum is equal to the number of parameters in the model plus twice the number of equations. mgarch ccc uses numerical derivatives in maximizing the log-likelihood function.
References
Bollerslev, T. 1990. Modelling the coherence in short-run nominal exchange rates: A multivariate generalized ARCH model. Review of Economics and Statistics 72: 498505. Gourieroux, C., and A. Monfort. 1997. Time Series and Dynamic Models. Trans. ed. G. M. Gallo. Cambridge: Cambridge University Press. Silvennoinen, A., and T. Ter svirta. 2009. Multivariate GARCH models. In Handbook of Financial Time Series, ed. a T. G. Andersen, R. A. Davis, J.-P. Krei, and T. Mikosch, 201229. Berlin: Springer. s
Also see
[TS] mgarch ccc postestimation Postestimation tools for mgarch ccc [TS] mgarch Multivariate GARCH models [TS] tsset Declare data to be time-series data [TS] arch Autoregressive conditional heteroskedasticity (ARCH) family of estimators [TS] var Vector autoregressive models [U] 20 Estimation and postestimation commands
Title
mgarch ccc postestimation Postestimation tools for mgarch ccc
Description
The following standard postestimation commands are available after mgarch ccc: Command contrast estat estimates lincom Description contrasts and ANOVA-style joint tests of estimates AIC, BIC, VCE, and estimation sample summary
cataloging estimation results point estimates, standard errors, testing, and inference for linear combinations of coefcients lrtest likelihood-ratio test margins marginal means, predictive margins, marginal effects, and average marginal effects marginsplot graph the results from margins (prole plots, interaction plots, etc.) nlcom point estimates, standard errors, testing, and inference for nonlinear combinations of coefcients predict predictions, residuals, inuence statistics, and other diagnostic measures predictnl point estimates, standard errors, testing, and inference for generalized predictions pwcompare pairwise comparisons of estimates test Wald tests of simple and composite linear hypotheses testnl Wald tests of nonlinear hypotheses See the corresponding entries in the Base Reference Manual for details.
Syntax for predict

predict statistic
Main
type
Description
if
in
, statistic options
xb residuals variance
linear prediction; the default residuals conditional variances and covariances
options
Options
Description
names of equations for which predictions are made equation(eqnames) dynamic(time constant) begin dynamic forecast at specied time
244
245
Menu
Statistics
>
Postestimation
>
Options for predict

Main
xb, the default, calculates the linear predictions of the dependent variables. residuals calculates the residuals. variance predicts the conditional variances and conditional covariances.
Options
equation(eqnames) species the equation for which the predictions are calculated. Use this option to predict a statistic for a particular equation. Equation names, such as equation(income), are used to identify equations. One equation name may be specied when predicting the dependent variable, the residuals, or the conditional variance. For example, specifying equation(income) causes predict to predict income, and specifying variance equation(income) causes predict to predict the conditional variance of income. Two equations may be specied when predicting a conditional variance or covariance. For example, specifying equation(income, consumption) variance causes predict to predict the conditional covariance of income and consumption. dynamic(time constant) species when predict starts producing dynamic forecasts. The specied time constant must be in the scale of the time variable specied in tsset, and the time constant must be inside a sample for which observations on the dependent variables are available. For example, dynamic(tq(2008q4)) causes dynamic predictions to begin in the fourth quarter of 2008, assuming that your time variable is quarterly; see [D] datetime. If the model contains exogenous variables, they must be present for the whole predicted sample. dynamic() may not be specied with residuals.
Remarks
We assume that you have already read [TS] mgarch ccc. In this entry, we use predict after mgarch ccc to make in-sample and out-of-sample forecasts.
Example 1
In this example, we obtain dynamic forecasts for the Toyota, Nissan, and Honda stock returns modeled in example 2 of [TS] mgarch ccc. In the output below, we reestimate the parameters of the model, use tsappend (see [TS] tsappend) to extend the data, and use predict to obtain in-sample one-step-ahead forecasts and dynamic forecasts of the conditional variances of the returns. We graph the forecasts below.
. use http://www.stata-press.com/data/r12/stocks (Data from Yahoo! Finance) . quietly mgarch ccc (toyota nissan = , noconstant) (honda = L.nissan, noconstant), > arch(1) garch(1) . tsappend, add(50) . predict H*, variance dynamic(2016)
246

.003 01jan2009 0 .001 .002
01jul2009
01jan2010 Date
01jul2010
01jan2011
Variance prediction (toyota,toyota), dynamic(2016) Variance prediction (nissan,nissan), dynamic(2016) Variance prediction (honda,honda), dynamic(2016)
Recent in-sample one-step-ahead forecasts are plotted to the left of the vertical line in the above graph, and the dynamic out-of-sample forecasts appear to the right of the vertical line. The graph shows the tail end of the huge increase in return volatility that took place in 2008 and 2009. It also shows that the dynamic forecasts quickly converge.

All postestimation commands listed above are implemented as ado-les. All one-step predictions are obtained by substituting the parameter estimates into the model. The estimated unconditional variance matrix of the disturbances, , is the initial value for the ARCH and GARCH terms. The postestimation routines recompute using the prediction sample, the parameter estimates stored in e(b), and (3) in Methods and formulas of [TS] mgarch ccc. For observations in which the residuals are missing, the estimated unconditional variance matrix of the disturbances is used in place of the outer product of the residuals. Dynamic predictions of the dependent variables use previously predicted values beginning in the period specied by dynamic(). Dynamic variance predictions are implemented by substituting for the outer product of the residuals beginning in the period specied in dynamic().
Also see
[TS] mgarch ccc Constant conditional correlation multivariate GARCH models [U] 20 Estimation and postestimation commands
Title
mgarch dcc Dynamic conditional correlation multivariate GARCH models
Syntax
mgarch dcc eq eq . . . eq if in , options where each eq has the form (depvars = indepvars options
Model
arch(numlist) garch(numlist) het(varlist) distribution(dist # ) constraints(numlist)

SE/Robust
include varlist in the specication of the conditional variance for all equations use dist distribution for errors [may be gaussian (synonym normal) or t; default is gaussian] apply linear constraints vcetype may be oim or robust set condence level; default is level(95) do not display constraints control column formats, row spacing, line width, and display of omitted variables and base and empty cells control the maximization process; seldom used initial values for the coefcients; seldom used display legend instead of statistics
vce(vcetype)
Reporting

Maximization

ARCH terms GARCH terms include varlist in the specication of the conditional variance
You must tsset your data before using mgarch dcc; see [TS] tsset. indepvars and varlist may contain factor variables; see [U] 11.4.3 Factor variables. depvars, indepvars, and varlist may contain time-series operators; see [U] 11.4.4 Time-series varlists. by, statsby, and rolling are allowed; see [U] 11.1.10 Prex commands. coeflegend does not appear in the dialog box. See [U] 20 Estimation and postestimation commands for more capabilities of estimation commands.
247
248
Menu
Statistics
>
>
Multivariate GARCH
Description
mgarch dcc estimates the parameters of dynamic conditional correlation (DCC) multivariate generalized autoregressive conditionally heteroskedastic (MGARCH) models in which the conditional variances are modeled as univariate generalized autoregressive conditionally heteroskedastic (GARCH) models and the conditional covariances are modeled as nonlinear functions of the conditional variances. The conditional quasicorrelation parameters that weight the nonlinear combinations of the conditional variances follow the GARCH-like process specied in Engle (2002). The DCC MGARCH model is about as exible as the closely related varying conditional correlation MGARCH model (see [TS] mgarch vcc), more exible than the conditional correlation MGARCH model (see [TS] mgarch ccc), and more parsimonious than the diagonal vech MGARCH model (see [TS] mgarch dvech).
Options

Model
arch(numlist) species the ARCH terms for all equations in the model. By default, no ARCH terms are specied. garch(numlist) species the GARCH terms for all equations in the model. By default, no GARCH terms are specied. het(varlist) species that varlist be included in the model in the specication of the conditional variance for all equations. This varlist enters the variance specication collectively as multiplicative heteroskedasticity. distribution(dist # ) species the assumed distribution for the errors. dist may be gaussian, normal, or t. gaussian and normal are synonyms; each causes mgarch dcc to assume that the errors come from a multivariate normal distribution. # may not be specied with either of them. t causes mgarch dcc to assume that the errors follow a multivariate Student t distribution, and the degree-of-freedom parameter is estimated along with the other parameters of the model. If distribution(t #) is specied, then mgarch dcc uses a multivariate Student t distribution with # degrees of freedom. # must be greater than 2. constraints(numlist) species linear constraints to apply to the parameter estimates.
SE/Robust
Reporting
level(#); see [R] estimation options. nocnsreport; see [R] estimation options.
249
display options: noomitted, vsquish, noemptycells, baselevels, allbaselevels, cformat(% fmt), pformat(% fmt), sformat(% fmt), and nolstretch; see [R] estimation options.
Maximization
maximize options: difficult, technique(algorithm spec), iterate(#), no log, trace, gradient, showstep, hessian, showtolerance, tolerance(#), ltolerance(#), nrtolerance(#), and nonrtolerance, and from(matname); see [R] maximize for all options except from(), and see below for information on from(). These options are seldom used. from(matname) species initial values for the coefcients. from(b0) causes mgarch dcc to begin the optimization algorithm with the values in b0. b0 must be a row vector, and the number of columns must equal the number of parameters in the model. The following option is available with mgarch dcc but is not shown in the dialog box: coeflegend; see [R] estimation options.
Eqoptions
Remarks
We assume that you have already read [TS] mgarch, which provides an introduction to MGARCH models and the methods implemented in mgarch dcc.
MGARCH models are dynamic multivariate regression models in which the conditional variances and covariances of the errors follow an autoregressive-moving-average structure. The DCC MGARCH model uses a nonlinear combination of univariate GARCH models with time-varying cross-equation weights to model the conditional covariance matrix of the errors.
As discussed in [TS] mgarch, MGARCH models differ in the parsimony and exibility of their specications for a time-varying conditional covariance matrix of the disturbances, denoted by Ht . In the conditional correlation family of MGARCH models, the diagonal elements of Ht are modeled as univariate GARCH models, whereas the off-diagonal elements are modeled as nonlinear functions of the diagonal terms. In the DCC MGARCH model,
hij,t = ij,t
hii,t hjj,t
where the diagonal elements hii,t and hjj,t follow univariate GARCH processes and ij,t follows the dynamic process specied in Engle (2002) and discussed below. Because the ij,t varies with time, this model is known as the DCC GARCH model.
250
Technical note
The DCC GARCH model proposed by Engle (2002) can be written as
yt = Cxt +
t
Ht = Rt =
t 1/2 Ht t 1/2 1/2 Dt Rt Dt 1/2 1/2 diag(Qt ) Qt diag(Qt ) t1 t1
Qt = (1 1 2 )R + 1
where
+ 2 Qt1
(1)
1/2

2 1,t 0 Dt = . . . 0
0 2 2,t . . . 0
0 0 . .. . . . 2 m,t
2 i,tj
+ j
qi j=1
2 j i,tj
by default, or
qi j=1
2 j i,tj
when the het() option is specied, where t is a 1 p vector of parameters, zi is a p 1 vector of independent variables including a constant term, the j s are ARCH parameters, and the j s are GARCH parameters;
Rt is a matrix of conditional quasicorrelations, 1 12,t 1 . . . 2m,t 1m,t 2m,t . .. . . . 1

1/2 t;
12,t Rt = . . . 1m,t
t
is an m 1 vector of standardized residuals, Dt
and
1 and 2 are parameters that govern the dynamics of conditional quasicorrelations. 1 and 2 are nonnegative and satisfy 0 1 + 2 < 1.
251
When Qt is stationary, the R matrix in (1) is a weighted average of the unconditional covariance matrix of the standardized residuals t , denoted by R, and the unconditional mean of Qt , denoted by Q. Because R = Q, as shown by Aielli (2009), R is neither the unconditional correlation matrix nor the unconditional mean of Qt . For this reason, the parameters in R are known as quasicorrelations; see Aielli (2009) and Engle (2009) for discussions.

We have daily data on the stock returns of three car manufacturersToyota, Nissan, and Honda, from January 2, 2003, to December 31, 2010in the variables toyota, nissan and honda. We model the conditional means of the returns as a rst-order vector autoregressive process and the conditional covariances as a DCC MGARCH process in which the variance of each disturbance term follows a GARCH(1,1) process.
. use http://www.stata-press.com/data/r12/stocks (Data from Yahoo! Finance) . mgarch dcc (toyota nissan honda = L.toyota L.nissan L.honda, noconstant), > arch(1) garch(1) Calculating starting values.... Optimizing log likelihood (setting technique to bhhh) Iteration 0: log likelihood = 16902.435 Iteration 1: log likelihood = 17005.448 Iteration 2: log likelihood = 17157.958 Iteration 3: log likelihood = 17267.363 Iteration 4: log likelihood = 17318.29 Iteration 5: log likelihood = 17353.029 Iteration 6: log likelihood = 17369.115 Iteration 7: log likelihood = 17388.035 Iteration 8: log likelihood = 17401.254 Iteration 9: log likelihood = 17435.556 (switching technique to nr) Iteration 10: log likelihood = 17451.739 Iteration 11: log likelihood = 17475.511 Iteration 12: log likelihood = 17483.259 Iteration 13: log likelihood = 17484.906 Iteration 14: log likelihood = 17484.95 Iteration 15: log likelihood = 17484.95 Refining estimates Iteration 0: log likelihood = 17484.95 Iteration 1: log likelihood = 17484.95
252

Dynamic conditional correlation MGARCH model Sample: 1 - 2015 Distribution: Gaussian Log likelihood = 17484.95 Coef. toyota toyota L1. nissan L1. honda L1. ARCH_toyota arch L1. garch L1. _cons nissan toyota L1. nissan L1. honda L1. ARCH_nissan arch L1. garch L1. _cons honda toyota L1. nissan L1. honda L1. -.0272421 .0361819 -0.75 0.451 -.0981573 .0436731 Std. Err. z
= = =
2014 19.54 0.0210
-.0510871
.0339825
-1.50
0.133
-.1176915
.0155174
.0297832
.0247455
1.20
0.229
-.0187171
.0782835
-.0162822
.0300323
-0.54
0.588
-.0751445
.0425801
.0608225
.0086687
7.02
0.000
.0438321
.0778128
.9222202 4.47e-06
.0111055 1.15e-06
83.04 3.90
0.000 0.000
.9004538 2.22e-06
.9439865 6.72e-06
-.0056724
.0389348
-0.15
0.884
-.0819832
.0706384
-.0287095
.0309379
-0.93
0.353
-.0893467
.0319276
.0154978
.0358802
0.43
0.666
-.0548261
.0858217
.0844239
.0128192
6.59
0.000
.0592988
.109549
.8994209 7.21e-06
.0151124 1.93e-06
59.52 3.74
0.000 0.000
.8698011 3.43e-06
.9290407 .000011
.0617495
.0271378
2.28
0.023
.0085603
.1149387
-.063507
.0332919
-1.91
0.056
-.1287578
.0017438
253
ARCH_honda arch L1. garch L1. _cons Correlation toyota nissan honda nissan honda Adjustment lambda1 lambda2
.0490142
.0073697
6.65
0.000
.0345698
.0634586
.9331114 5.35e-06
.0103689 1.35e-06
89.99 3.95
0.000 0.000
.9127888 2.69e-06
.9534341 8.00e-06
.6689551 .7259628
.016802 .0140155
39.81 51.80
0.000 0.000
.6360239 .6984929
.7018864 .7534327
.6335662
.018041
35.12
0.000
.5982064
.6689259
.0315267 .8704173
.008838 .0613324
3.57 14.19
0.000 0.000
.0142045 .750208
.0488489 .9906266
The iteration log has three parts: the dots from the search for initial values, the iteration log from optimizing the log likelihood, and the iteration log from the rening step. A detailed discussion of the optimization methods is in Methods and formulas. The header describes the estimation sample and reports a Wald test against the null hypothesis that all the coefcients on the independent variables in the mean equations are zero. Here the null hypothesis is rejected at the 5% level. The output table rst presents results for the mean or variance parameters used to model each dependent variable. Subsequently, the output table presents results for the conditional quasicorrelations. For example, the conditional quasicorrelation between the standardized residuals for Toyota and Nissan is estimated to be 0.67. Finally, the output table presents results for the adjustment parameters 1 and 2 . In the example at hand, the estimates for both 1 and 2 are statistically signicant. The DCC MGARCH model reduces to the CCC MGARCH model when 1 = 2 = 0. The output below shows that a Wald test rejects the null hypothesis that 1 = 2 = 0 at all conventional levels.
. test _b[Adjustment:lambda1] = _b[Adjustment:lambda2] = 0 ( 1) [Adjustment]lambda1 - [Adjustment]lambda2 = 0 ( 2) [Adjustment]lambda1 = 0 chi2( 2) = 1102.33 Prob > chi2 = 0.0000
These results indicate that the assumption of time-invariant conditional correlations maintained in the CCC MGARCH model is too restrictive for these data.
Example 2
We improve the previous example by removing the insignicant parameters from the model. To remove these parameters, we specify the honda equation separately from the toyota and nissan equations:
254

. mgarch dcc (toyota nissan = , noconstant) (honda = L.nissan, noconstant), > arch(1) garch(1) Calculating starting values.... Optimizing log likelihood (setting technique to bhhh) Iteration 0: log likelihood = 16884.502 Iteration 1: log likelihood = 16970.755 Iteration 2: log likelihood = 17140.318 Iteration 3: log likelihood = 17237.807 Iteration 4: log likelihood = 17306.12 Iteration 5: log likelihood = 17342.533 Iteration 6: log likelihood = 17363.511 Iteration 7: log likelihood = 17392.501 Iteration 8: log likelihood = 17407.242 Iteration 9: log likelihood = 17448.702 (switching technique to nr) Iteration 10: log likelihood = 17472.199 Iteration 11: log likelihood = 17475.839 Iteration 12: log likelihood = 17476.345 Iteration 13: log likelihood = 17476.35 Iteration 14: log likelihood = 17476.35 Refining estimates Iteration 0: log likelihood = 17476.35 Iteration 1: log likelihood = 17476.35 Dynamic conditional correlation MGARCH model Sample: 1 - 2015 Distribution: Gaussian Log likelihood = 17476.35 Coef. ARCH_toyota arch L1. garch L1. _cons ARCH_nissan arch L1. garch L1. _cons honda nissan L1. .019978 .0134488 1.49 0.137 -.0063811 .0463371 Std. Err. z Number of obs Wald chi2(1) Prob > chi2 P>|z| = = = 2014 2.21 0.1374
.0608188
.0086675
7.02
0.000
.0438309
.0778067
.9219957 4.49e-06
.0111066 1.14e-06
83.01 3.95
0.000 0.000
.9002271 2.27e-06
.9437642 6.72e-06
.0876162
.01302
6.73
0.000
.0620974
.1131349
.8950964 7.69e-06
.0152908 1.99e-06
58.54 3.86
0.000 0.000
.865127 3.79e-06
.9250657 .0000116
255
ARCH_honda arch L1. garch L1. _cons Correlation toyota nissan honda nissan honda Adjustment lambda1 lambda2
.0488799
.0073767
6.63
0.000
.0344218
.063338
.9330047 5.42e-06
.0103944 1.36e-06
89.76 3.98
0.000 0.000
.912632 2.75e-06
.9533774 8.08e-06
.6668433 .7258101
.0163209 .0137072
40.86 52.95
0.000 0.000
.6348548 .6989446
.6988317 .7526757
.6313515
.0175454
35.98
0.000
.5969631
.6657399
.0324493 .8574681
.0074013 .0476273
4.38 18.00
0.000 0.000
.0179429 .7641202
.0469556 .950816
It turns out that the coefcient on L1.nissan in the honda equation is now statistically insignicant. We could further improve the model by removing L1.nissan from the model. There is no mean equation for Toyota or Nissan. In [TS] mgarch dcc postestimation, we discuss prediction from models without covariates.
Example 3
Here we t a bivariate DCC MGARCH model for the Toyota and Nissan shares. We believe that the shares of these car manufacturers follow the same process, so we impose the constraints that the ARCH coefcients are the same for the two companies and that the GARCH coefcients are also the same.
. constraint 1 _b[ARCH_toyota:L.arch] = _b[ARCH_nissan:L.arch] . constraint 2 _b[ARCH_toyota:L.garch] = _b[ARCH_nissan:L.garch] . mgarch dcc (toyota nissan = , noconstant), arch(1) garch(1) constraints(1 2) Calculating starting values.... Optimizing log likelihood (setting technique to bhhh) Iteration 0: log likelihood Iteration 1: log likelihood Iteration 2: log likelihood Iteration 3: log likelihood Iteration 4: log likelihood Iteration 5: log likelihood Iteration 6: log likelihood Iteration 7: log likelihood Iteration 8: log likelihood Iteration 9: log likelihood (switching technique to nr) Iteration 10: log likelihood Iteration 11: log likelihood Iteration 12: log likelihood Iteration 13: log likelihood = = = = = = = = = = = = = = 10307.609 10656.153 10862.137 10987.457 11062.347 11135.207 11245.619 11253.56 11294 11296.364 11296.76 11297.087 11297.091 11297.091
256

Refining estimates Iteration 0: log likelihood = 11297.091 Iteration 1: log likelihood = 11297.091 Dynamic conditional correlation MGARCH model Sample: 1 - 2015 Distribution: Gaussian Log likelihood = 11297.09 ( 1) ( 2) Number of obs Wald chi2(.) Prob > chi2 = = = 2015 . .
[ARCH_toyota]L.arch - [ARCH_nissan]L.arch = 0 [ARCH_toyota]L.garch - [ARCH_nissan]L.garch = 0 Coef. Std. Err. z P>|z| [95% Conf. Interval]
ARCH_toyota arch L1. garch L1. _cons ARCH_nissan arch L1. garch L1. _cons Correlation toyota nissan Adjustment lambda1 lambda2
.080889
.0103227
7.84
0.000
.060657
.1011211
.9060711 4.21e-06
.0119107 1.10e-06
76.07 3.83
0.000 0.000
.8827267 2.05e-06
.9294156 6.36e-06
.080889
.0103227
7.84
0.000
.060657
.1011211
.9060711 5.92e-06
.0119107 1.47e-06
76.07 4.03
0.000 0.000
.8827267 3.04e-06
.9294156 8.80e-06
.6646283
.0187793
35.39
0.000
.6278215
.7014351
.0446559 .8686054
.0123017 .0510884
3.63 17.00
0.000 0.000
.020545 .7684739
.0687668 .968737
Example 4
In this example, we have data on ctional stock returns for the Acme and Anvil corporations, and we believe that the movement of the two stocks is governed by different processes. We specify one ARCH and one GARCH term for the conditional variance equation for Acme and two ARCH terms for the conditional variance equation for Anvil. In addition, we include the lagged value of the stock return for Apex, the main subsidiary of Anvil corporation, in the variance equation of Anvil. For Acme, we have data on the changes in an index of futures prices of products related to those produced by Acme in afrelated. For Anvil, we have data on the changes in an index of futures prices of inputs used by Anvil in afinputs.

. use http://www.stata-press.com/data/r12/acmeh . mgarch dcc (acme = afrelated, noconstant arch(1) garch(1)) > (anvil = afinputs, arch(1/2) het(L.apex)) Calculating starting values.... Optimizing log likelihood (setting technique to bhhh) Iteration 0: log likelihood = -13260.522 (output omitted ) Iteration 9: log likelihood = -12362.876 (switching technique to nr) Iteration 10: log likelihood = -12362.876 Refining estimates Iteration 0: Iteration 1: log likelihood = -12362.876 log likelihood = -12362.876 Number of obs Wald chi2(2) Prob > chi2 Std. Err. z P>|z| = = = 2499 2596.18 0.0000
257
Dynamic conditional correlation MGARCH model Sample: 1 - 2500 Distribution: Gaussian Log likelihood = -12362.88 Coef. acme afrelated ARCH_acme arch L1. garch L1. _cons anvil afinputs _cons ARCH_anvil arch L1. L2. apex L1. _cons Correlation acme anvil Adjustment lambda1 lambda2
.950805
.0557082
17.07
0.000
.841619
1.059991
.1063295
.015716
6.77
0.000
.0755266
.1371324
.7556294 2.197566
.0391568 .458343
19.30 4.79
0.000 0.000
.6788836 1.29923
.8323752 3.095901
-1.015657 .0808653
.0209959 .019445
-48.37 4.16
0.000 0.000
-1.056808 .0427538
-.9745054 .1189767
.5261675 .2866454
.0281586 .0196504
18.69 14.59
0.000 0.000
.4709777 .2481314
.5813572 .3251595
1.953173 -.0062964
.0594862 .0710842
32.83 -0.09
0.000 0.929
1.836582 -.1456188
2.069763 .1330261
-.5600358
.0326358
-17.16
0.000
-.6240008
-.4960708
.1904321 .7147267
.0154449 .0226204
12.33 31.60
0.000 0.000
.1601607 .6703916
.2207035 .7590618
The results indicate that increases in the futures prices for related products lead to higher returns on the Acme stock, and increased input prices lead to lower returns on the Anvil stock. In the conditional variance equation for Anvil, the coefcient on L1.apex is positive and signicant, which indicates
258
that an increase in the return on the Apex stock leads to more variability in the return on the Anvil stock.
Saved results
mgarch dcc saves the following in e():
Scalars e(N) e(k) e(k aux) e(k extra) e(k eq) e(k dv) e(df m) e(ll) e(chi2) e(p) e(estdf) e(usr) e(tmin) e(tmax) e(N gaps) e(rank) e(ic) e(rc) e(converged) Macros e(cmd) e(model) e(cmdline) e(depvar) e(covariates) e(dv eqs) e(indeps) e(tvar) e(title) e(chi2type) e(vce) e(vcetype) e(tmins) e(tmaxs) e(dist) e(arch) e(garch) e(technique) e(properties) e(estat cmd) e(predict) e(marginsok) e(marginsnotok) Matrices e(b) e(Cns) e(ilog) e(gradient) e(hessian) e(V) e(pinfo) Functions e(sample) number of observations number of parameters number of auxiliary parameters number of extra estimates added to number of equations in e(b) number of dependent variables model degrees of freedom log likelihood
2
signicance 1 if distribution parameter was estimated, 0 otherwise user-provided distribution parameter minimum time in sample maximum time in sample number of gaps rank of e(V) number of iterations return code 1 if converged, 0 otherwise mgarch dcc command as typed names of dependent variables list of covariates dependent variables with mean equations independent variables in each equation time variable title in estimation output Wald; type of model 2 test vcetype specied in vce() title used to label Std. Err. formatted minimum time formatted maximum time distribution for error term: gaussian or t specied ARCH terms specied GARCH terms maximization technique b V program used to implement estat program used to implement predict predictions allowed by margins predictions disallowed by margins coefcient vector constraints matrix iteration log (up to 20 iterations) gradient vector Hessian matrix variancecovariance matrix of the estimators parameter information, used by predict marks estimation sample
259

mgarch dcc is implemented as an ado-le. mgarch dcc estimates the parameters of the DCC MGARCH model by maximum likelihood. The log-likelihood function based on the multivariate normal distribution for observation t is
lt = 0.5m log(2) 0.5log {det (Rt )} log det Dt

1/2 t
1/2
0.5 t R1 t
If we assume that t follow a multivariate t distribution with degrees of freedom (df) greater than 2, then the log-likelihood function for observation t is
lt = log
df + m 2
log
df 2
1/2
m log {(df 2)} 2

1 df + m t Rt t log 1 + 2 df 2
0.5log {det (Rt )} log det Dt
The starting values for the parameters in the mean equations and the initial residuals t are obtained by least-squares regression. The starting values for the parameters in the variance equations are obtained by a procedure proposed by Gourieroux and Monfort (1997, sec. 6.2.2). The starting values for the quasicorrelation parameters are calculated from the standardized residuals t . Given the starting values for the mean and variance equations, the starting values for the parameters 1 and 2 are obtained from a grid search performed on the log likelihood. The initial optimization step is performed in the unconstrained space. Once the maximum is found, we impose the constraints 1 0, 2 0, and 0 1 + 2 < 1, and maximize the log likelihood in the constrained space. This step is reported in the iteration log as the rening step.
GARCH estimators require initial values that can be plugged in for ti ti and Htj when t i < 1 and t j < 1. mgarch dcc substitutes an estimator of the unconditional covariance of the disturbances
T
= T 1
t=1
t t
(2)
mgarch dcc uses numerical derivatives in maximizing the log-likelihood function.
References
Aielli, G. P. 2009. Dynamic Conditional Correlations: On Properties and Estimation. Working paper, Dipartimento di Statistica, University of Florence, Florence, Italy. Engle, R. F. 2002. Dynamic conditional correlation: A simple class of multivariate generalized autoregressive conditional heteroskedasticity models. Journal of Business & Economic Statistics 20: 339350. . 2009. Anticipating Correlations: A New Paradigm for Risk Management. Princeton, NJ: Princeton University Press.
260
Gourieroux, C., and A. Monfort. 1997. Time Series and Dynamic Models. Trans. ed. G. M. Gallo. Cambridge: Cambridge University Press.
Also see
[TS] mgarch dcc postestimation Postestimation tools for mgarch dcc [TS] mgarch Multivariate GARCH models [TS] tsset Declare data to be time-series data [TS] arch Autoregressive conditional heteroskedasticity (ARCH) family of estimators [TS] var Vector autoregressive models [U] 20 Estimation and postestimation commands
Title
mgarch dcc postestimation Postestimation tools for mgarch dcc
Description
The following standard postestimation commands are available after mgarch dcc: Command contrast estat estimates lincom Description contrasts and ANOVA-style joint tests of estimates AIC, BIC, VCE, and estimation sample summary
Syntax for predict

predict statistic
Main
type
Description
if
in
, statistic options
options
Options
Description
equation(eqnames) names of equations for which predictions are made dynamic(time constant) begin dynamic forecast at specied time
261
262
Menu
Statistics
>
Postestimation
>
Options for predict

Main
Options
Remarks
We assume that you have already read [TS] mgarch dcc. In this entry, we use predict after mgarch dcc to make in-sample and out-of-sample forecasts.
Example 1
In this example, we obtain dynamic forecasts for the Toyota, Nissan, and Honda stock returns modeled in example 2 of [TS] mgarch dcc. In the output below, we reestimate the parameters of the model, use tsappend (see [TS] tsappend) to extend the data, and use predict to obtain in-sample one-step-ahead forecasts and dynamic forecasts of the conditional variances of the returns. We graph the forecasts below.
. use http://www.stata-press.com/data/r12/stocks (Data from Yahoo! Finance) . quietly mgarch dcc (toyota nissan = , noconstant) (honda = L.nissan, noconstant), > arch(1) garch(1) . tsappend, add(50) . predict H*, variance dynamic(2016)

.003
263
01jan2009
.001
.002
01jul2009
01jan2010 Date
01jul2010
01jan2011

All postestimation commands listed above are implemented as ado-les. All one-step predictions are obtained by substituting the parameter estimates into the model. The estimated unconditional variance matrix of the disturbances, , is the initial value for the ARCH and GARCH terms. The postestimation routines recompute using the prediction sample, the parameter estimates stored in e(b), and (2) in Methods and formulas of [TS] mgarch dcc. For observations in which the residuals are missing, the estimated unconditional variance matrix of the disturbances is used in place of the outer product of the residuals. Dynamic predictions of the dependent variables use previously predicted values beginning in the period specied by dynamic(). Dynamic variance predictions are implemented by substituting for the outer product of the residuals beginning in the period specied in dynamic().
Also see
[TS] mgarch dcc Dynamic conditional correlation multivariate GARCH models [U] 20 Estimation and postestimation commands
Title
mgarch dvech Diagonal vech multivariate GARCH models
Syntax
mgarch dvech eq eq . . . eq if in , options
where each eq has the form (depvars = indepvars options

Model
, noconstant ) Description
arch(numlist) garch(numlist) distribution(dist # ) constraints(numlist)

SE/Robust
use dist distribution for errors (may be gaussian, normal, or t; default is gaussian) apply linear constraints vcetype may be oim or robust set condence level; default is level(95) do not display constraints control column formats, row spacing, line width, and display of omitted variables and base and empty cells control the maximization process; seldom used initial values for the coefcients; seldom used starting-value maximization algorithm number of starting-value iterations; default is sviterate(25) display legend instead of statistics
vce(vcetype)
Reporting

Maximization
maximize options from(matname) svtechnique(algorithm spec) sviterate(#) coeflegend
You must tsset your data before using mgarch dvech; see [TS] tsset. indepvars may contain factor variables; see [U] 11.4.3 Factor variables. depvars and indepvars may contain time-series operators; see [U] 11.4.4 Time-series varlists. by, statsby, and rolling are allowed; see [U] 11.1.10 Prex commands. coeflegend does not appear in the dialog box. See [U] 20 Estimation and postestimation commands for more capabilities of estimation commands.
Menu
Statistics
>
>
Multivariate GARCH
264
265
Description
mgarch dvech estimates the parameters of diagonal vech (DVECH) multivariate generalized autoregressive conditionally heteroskedastic (MGARCH) models in which each element of the conditional correlation matrix is parameterized as a linear function of its own past and past shocks.
DVECH MGARCH models are less parsimonious than the conditional correlation models discussed in [TS] mgarch ccc, [TS] mgarch dcc, and [TS] mgarch vcc because the number of parameters in DVECH MGARCH models increases more rapidly with the number of series modeled.
Options

Model
noconstant suppresses the constant term(s). arch(numlist) species the ARCH terms in the model. By default, no ARCH terms are specied. garch(numlist) species the GARCH terms in the model. By default, no GARCH terms are specied. distribution(dist # ) species the assumed distribution for the errors. dist may be gaussian, normal, or t. gaussian and normal are synonyms; each causes mgarch dvech to assume that the errors come from a multivariate normal distribution. # cannot be specied with either of them. t causes mgarch dvech to assume that the errors follow a multivariate Student t distribution, and the degree-of-freedom parameter is estimated along with the other parameters of the model. If distribution(t #) is specied, then mgarch dvech uses a multivariate Student t distribution with # degrees of freedom. # must be greater than 2. constraints(numlist) species linear constraints to apply to the parameter estimates.
SE/Robust
Reporting
level(#); see [R] estimation options. nocnsreport; see [R] estimation options. display options: noomitted, vsquish, noemptycells, baselevels, allbaselevels, cformat(% fmt), pformat(% fmt), sformat(% fmt), and nolstretch; see [R] estimation options.
Maximization
maximize options: difficult, technique(algorithm spec), iterate(#), no log, trace, gradient, showstep, hessian, showtolerance, tolerance(#), ltolerance(#), nrtolerance(#), and nonrtolerance, and from(matname); see [R] maximize for all options except from(), and see below for information on from(). These options are seldom used. from(matname) species initial values for the coefcients. from(b0) causes mgarch dvech to begin the optimization algorithm with the values in b0. b0 must be a row vector, and the number of columns must equal the number of parameters in the model.
266
svtechnique(algorithm spec) and sviterate(#) specify options for the starting-value search process. svtechnique(algorithm spec) species the algorithm used to search for initial values. The syntax for algorithm spec is the same as for the technique() option; see [R] maximize. svtechnique(bhhh 5 nr 16000) is the default. This option may not be specied with from(). sviterate(#) species the maximum number of iterations that the search algorithm may perform. The default is sviterate(25). This option may not be specied with from(). The following option is available with mgarch dvech but is not shown in the dialog box: coeflegend; see [R] estimation options.
Remarks
We assume that you have already read [TS] mgarch, which provides an introduction to MGARCH models and the methods implemented in mgarch dvech.
MGARCH models are dynamic multivariate regression models in which the conditional variances and covariances of the errors follow an autoregressive-moving-average structure. The DVECH MGARCH model parameterizes each element of the current conditional covariance matrix as a linear function of its own past and past shocks.
As discussed in [TS] mgarch, MGARCH models differ in the parsimony and exibility of their specications for a time-varying conditional covariance matrix of the disturbances, denoted by Ht . In a DVECH MGARCH model with one ARCH term and one GARCH term, the (i, j)th element of conditional covariance matrix is modeled by
hij,t = sij + aij
i,t1 j,t1
+ bij hij,t1
where sij , aij , and bij are parameters and t1 is the vector of errors from the previous period. This expression shows the linear form in which each element of the current conditional covariance matrix is a function of its own past and past shocks.
Technical note
The general vech MGARCH model developed by Bollerslev, Engle, and Wooldridge (1988) can be written as
yt = Cxt +
t
(1) (2)
q
1/2 Ht t p
ht = s +
i=1
Ai vech(
ti ti )
+
j=1
Bj htj
(3)
where
yt is an m 1 vector of dependent variables; C is an m k matrix of parameters; xt is a k 1 vector of independent variables, which may contain lags of yt ;
267
Ht
1/2
t is an m 1 vector of independent and identically distributed innovations;
ht = vech(Ht );
the vech() function stacks the lower diagonal elements of a symmetric matrix into a column vector, for example, vech
1 2 2 3
= (1, 2, 3)
s is an m(m + 1)/2 1 vector of parameters;

each Ai is an {m(m + 1)/2} {m(m + 1)/2} matrix of parameters; and each Bj is an {m(m + 1)/2} {m(m + 1)/2} matrix of parameters. Bollerslev, Engle, and Wooldridge (1988) argued that the general-vech MGARCH model in (1)(3) was too exible to t to data, so they proposed restricting the matrices Ai and Bj to be diagonal matrices. It is for this restriction that the model is known as a diagonal vech MGARCH model. The diagonal vech MGARCH model can also be expressed by replacing (3) with
p q
Ht = S +
i=1
Ai
ti
ti + j=1
Bj
Htj
(3 )
where S is an m m symmetric parameter matrix; each Ai is an m m symmetric parameter matrix; is the elementwise or Hadamard product; and each Bj is an m m symmetric parameter matrix. In (3 ), A and B are symmetric but not diagonal matrices because we used the Hadamard product. The matrices are diagonal in the vech representation of (3) but not in the Hadamard-product representation of (3 ). The Hadamard-product representation in (3 ) claries that each element in Ht depends on its past values and the past values of the corresponding ARCH terms. Although this representation does not allow cross-covariance effects, it is still quite exible. The rapid rate at which the number of parameters grows with m, p, or q is one aspect of the models exibility.

We have data on a secondary market rate of a six-month U.S. Treasury bill, tbill, and on Moodys seasoned AAA corporate bond yield, bond. We model the rst-differences of tbill and the rst-differences of bond as a VAR(1) with an ARCH(1) term.
268

. use http://www.stata-press.com/data/r12/irates4 (St. Louis Fed (FRED) financial data) . mgarch dvech (D.bond D.tbill = LD.bond LD.tbill), arch(1) Getting starting values (setting technique to bhhh) Iteration 0: log likelihood = 3569.2723 Iteration 1: log likelihood = 3708.4561 (output omitted ) Iteration 6: log likelihood = 4183.8853 Iteration 7: log likelihood = 4184.2424 (switching technique to nr) Iteration 8: log likelihood = 4184.4141 Iteration 9: log likelihood = 4184.5973 Iteration 10: log likelihood = 4184.5975 Estimating parameters (setting technique to bhhh) Iteration 0: log likelihood = 4184.5975 Iteration 1: log likelihood = 4200.6303 Iteration 2: log likelihood = 4208.5342 Iteration 3: log likelihood = 4212.426 Iteration 4: log likelihood = 4215.2373 (switching technique to nr) Iteration 5: log likelihood = 4217.0676 Iteration 6: log likelihood = 4221.5706 Iteration 7: log likelihood = 4221.6576 Iteration 8: log likelihood = 4221.6577 Diagonal vech MGARCH model Sample: 3 - 2456 Number of obs Distribution: Gaussian Wald chi2(4) Log likelihood = 4221.658 Prob > chi2 Coef. D.bond bond LD. tbill LD. _cons D.tbill bond LD. tbill LD. _cons Sigma0 1_1 2_1 2_2 L.ARCH 1_1 2_1 2_2 .4514942 .2518879 .843368 .0456835 .036736 .0608055 9.88 6.86 13.87 0.000 0.000 0.000 .3619562 .1798866 .7241914 .5410323 .3238893 .9625446 .004894 .0040986 .0115149 .0002006 .0002396 .0005227 24.40 17.10 22.03 0.000 0.000 0.000 .0045008 .0036289 .0104904 .0052871 .0045683 .0125395 .0108373 .0301501 0.36 0.719 -.0482558 .0699304 .2967674 .0247149 12.01 0.000 .2483271 .3452077 Std. Err. z P>|z|
= = =
2454 1183.52 0.0000
.0947949 .0003991
.0098683 .00143
9.61 0.28
0.000 0.780
.0754533 -.0024036
.1141364 .0032019
.4344747 .0011611
.0176497 .0021033
24.62 0.55
0.000 0.581
.3998819 -.0029612
.4690675 .0052835
269
The output has three parts: an iteration log, a header, and an output table. The iteration log has two parts: the rst part reports the iterations from the process of searching for starting values, and the second part reports the iterations from maximizing the log-likelihood function. The header describes the estimation sample and reports a Wald test against the null hypothesis that all the coefcients on the independent variables in each equation are zero. Here the null hypothesis is rejected at all conventional levels. The output table reports point estimates, standard errors, tests against zero, and condence intervals for the estimated coefcients, the estimated elements of S, and any estimated elements of A or B. Here the output indicates that in the equation for D.tbill, neither the coefcient on LD.bond nor the constant are statistically signicant. The elements of S are reported in the Sigma0 equation. The estimate of S[1, 1] is 0.005, and the estimate of S[2, 1] is 0.004. The ARCH term results are reported in the L.ARCH equation. In the L.ARCH equation, 1 1 is the coefcient on the ARCH term for the conditional variance of the rst dependent variable, 2 1 is the coefcient on the ARCH term for the conditional covariance between the rst and second dependent variables, and 2 2 is the coefcient on the ARCH term for the conditional variance of the second dependent variable.
Example 2
We improve the previous example by removing the insignicant parameters from the model:
. mgarch dvech (D.bond = LD.bond LD.tbill, noconstant) > (D.tbill = LD.tbill, noconstant), arch(1) Getting starting values (setting technique to bhhh) Iteration 0: log likelihood = 3566.8824 Iteration 1: log likelihood = 3701.6181 Iteration 2: log likelihood = 3952.8048 Iteration 3: log likelihood = 4076.5164 Iteration 4: log likelihood = 4166.6842 Iteration 5: log likelihood = 4180.2998 Iteration 6: log likelihood = 4182.4545 Iteration 7: log likelihood = 4182.9563 (switching technique to nr) Iteration 8: log likelihood = 4183.0293 Iteration 9: log likelihood = 4183.1112 Iteration 10: log likelihood = 4183.1113 Estimating parameters (setting technique to bhhh) Iteration 0: log likelihood = 4183.1113 Iteration 1: log likelihood = 4202.0304 Iteration 2: log likelihood = 4210.2929 Iteration 3: log likelihood = 4215.7798 Iteration 4: log likelihood = 4217.7755 (switching technique to nr) Iteration 5: log likelihood = 4219.0078 Iteration 6: log likelihood = 4221.4197 Iteration 7: log likelihood = 4221.433 Iteration 8: log likelihood = 4221.433
270

Diagonal vech MGARCH model Sample: 3 - 2456 Distribution: Gaussian Log likelihood = 4221.433 Coef. D.bond bond LD. tbill LD. D.tbill tbill LD. Sigma0 1_1 2_1 2_2 L.ARCH 1_1 2_1 2_2 .4519233 .2515474 .8437212 .045671 .0366701 .0600839 9.90 6.86 14.04 0.000 0.000 0.000 .3624099 .1796752 .7259589 .5414368 .3234195 .9614836 .0048922 .0040949 .0115043 .0002005 .0002394 .0005184 24.40 17.10 22.19 0.000 0.000 0.000 .0044993 .0036256 .0104883 .0052851 .0045641 .0125203 .2941649 .0234734 12.53 0.000 .2481579 .3401718 Std. Err. z
= = =
2454 1197.76 0.0000
.0953158
.0098077
9.72
0.000
.076093
.1145386
.4385945
.0136672
32.09
0.000
.4118072
.4653817
We specied each equation separately to remove the insignicant parameters. All the parameter estimates are statistically signicant.
Example 3
Here we analyze some ctional weekly data on the percentages of bad widgets found in the factories of Acme Inc. and Anvil Inc. We model the levels as a rst-order autoregressive process. We believe that the adaptive management style in these companies causes the variances to follow a diagonal vech MGARCH process with one ARCH term and one GARCH term. Furthermore, these close competitors follow essentially the same process, so we impose the constraints that the ARCH coefcients are the same for the two companies and that the GARCH coefcients are also the same.
271
Imposing these constraints yields

. use http://www.stata-press.com/data/r12/acme, clear . constraint 1 [L.ARCH]1_1 = [L.ARCH]2_2 . constraint 2 [L.GARCH]1_1 = [L.GARCH]2_2 . mgarch dvech (acme = L.acme) (anvil = L.anvil), arch(1) garch(1) constraints(1 2) Getting starting values (setting technique to bhhh) Iteration 0: log likelihood Iteration 1: log likelihood Iteration 2: log likelihood Iteration 3: log likelihood Iteration 4: log likelihood Iteration 5: log likelihood Iteration 6: log likelihood Iteration 7: log likelihood Estimating parameters (setting technique to bhhh) Iteration 0: log likelihood Iteration 1: log likelihood Iteration 2: log likelihood Iteration 3: log likelihood Iteration 4: log likelihood Diagonal vech MGARCH model Sample: 1969w35 - 1998w25 Distribution: Gaussian Log likelihood = -5973.8 ( 1) ( 2) [L.ARCH]1_1 - [L.ARCH]2_2 = 0 [L.GARCH]1_1 - [L.GARCH]2_2 = 0 Coef. acme acme L1. _cons anvil anvil L1. _cons Sigma0 1_1 2_1 2_2 L.ARCH 1_1 2_1 2_2 L.GARCH 1_1 2_1 2_2 .1487581 .085404 .1487581 .0697531 .1446524 .0697531 2.13 0.59 2.13 0.033 0.555 0.033 .0120445 -.1981094 .0120445 .2854716 .3689175 .2854716 .2813443 .181877 .2813443 .0299124 .0335393 .0299124 9.41 5.42 9.41 0.000 0.000 0.000 .222717 .1161412 .222717 .3399716 .2476128 .3399716 1.889237 .4599576 2.063113 .2168733 .1139843 .2454633 8.71 4.04 8.40 0.000 0.000 0.000 1.464173 .2365525 1.582014 2.314301 .6833626 2.544213 .3151955 1.215786 .0263287 .0642052 11.97 18.94 0.000 0.000 .2635922 1.089947 .3667988 1.341626 .3365278 1.124611 .0255134 .060085 13.19 18.72 0.000 0.000 .2865225 1.006847 .3865331 1.242376 Std. Err. z P>|z| [95% Conf. Interval] Number of obs Wald chi2(2) Prob > chi2 = = = 1499 272.47 0.0000
= = = = = = = =
-6087.0665 -6022.2046 -5986.6152 -5976.5739 -5974.4342 -5974.4046 -5974.4036 -5974.4035
(not concave)
= = = = =
-5974.4035 -5973.812 -5973.8004 -5973.7999 -5973.7999
272
We could test our constraints by tting the unconstrained model and performing either a Wald or a likelihood-ratio test. The results indicate that we might further restrict the time-invariant components of the conditional variances to be the same across companies.
Example 4
Some models of nancial data include no covariates or constant terms. For example, in modeling ctional data on the stock returns of Acme Inc. and Anvil Inc., we found it best not to include any covariates or constant terms. We include two ARCH terms and one GARCH term to model the conditional variances.
. use http://www.stata-press.com/data/r12/aacmer . mgarch dvech (acme anvil = , noconstant), arch(1/2) garch(1) Getting starting values (setting technique to bhhh) Iteration 0: log likelihood = -18417.243 (not concave) Iteration 1: log likelihood = -18215.005 Iteration 2: log likelihood = -18199.691 Iteration 3: log likelihood = -18136.699 Iteration 4: log likelihood = -18084.256 Iteration 5: log likelihood = -17993.662 Iteration 6: log likelihood = -17731.1 Iteration 7: log likelihood = -17629.505 (switching technique to nr) Iteration 8: log likelihood = -17548.172 Iteration 9: log likelihood = -17544.987 Iteration 10: log likelihood = -17544.937 Iteration 11: log likelihood = -17544.937 Estimating parameters (setting technique to bhhh) Iteration 0: log likelihood = -17544.937 Iteration 1: log likelihood = -17544.937 Diagonal vech MGARCH model Sample: 1 - 5000 Number of obs Distribution: Gaussian Wald chi2(.) Log likelihood = -17544.94 Prob > chi2 Coef. Sigma0 1_1 2_1 2_2 L.ARCH 1_1 2_1 2_2 L2.ARCH 1_1 2_1 2_2 L.GARCH 1_1 2_1 2_2 .0782467 .2888104 .201618 .053944 .0818303 .0470584 1.45 3.53 4.28 0.147 0.000 0.000 -.0274816 .1284261 .1093853 .183975 .4491948 .2938508 .1837825 .0884425 .2025718 .0274555 .02208 .0272639 6.69 4.01 7.43 0.000 0.000 0.000 .1299706 .0451665 .1491355 .2375943 .1317185 .256008 .2878739 .1036685 .2034196 .02157 .0161446 .019855 13.35 6.42 10.25 0.000 0.000 0.000 .2455975 .0720256 .1645044 .3301504 .1353114 .2423347 1.026283 .4300997 1.019753 .0823348 .0590294 .0837146 12.46 7.29 12.18 0.000 0.000 0.000 .8649096 .3144042 .8556751 1.187656 .5457952 1.18383 Std. Err. z P>|z|
= = =
5000 . .
273
The model test is omitted from the output, because there are no covariates in the model. The univariate tests indicate that the included parameters t the data well. In [TS] mgarch dvech postestimation, we discuss prediction from models without covariates.
Saved results
mgarch dvech saves the following in e():
Scalars e(N) e(k) e(k extra) e(k eq) e(k dv) e(df m) e(ll) e(chi2) e(p) e(estdf) e(usr) e(tmin) e(tmax) e(N gaps) e(rank) e(ic) e(rc) e(converged) Macros e(cmd) e(model) e(cmdline) e(depvar) e(covariates) e(dv eqs) e(indeps) e(tvar) e(title) e(chi2type) e(vce) e(vcetype) e(tmins) e(tmaxs) e(dist) e(arch) e(garch) e(svtechnique) e(technique) e(properties) e(estat cmd) e(predict) e(marginsok) e(marginsnotok) number of observations number of parameters number of extra estimates added to number of equations in e(b) number of dependent variables model degrees of freedom log likelihood
2
signicance 1 if distribution parameter was estimated, 0 otherwise user-provided distribution parameter minimum time in sample maximum time in sample number of gaps rank of e(V) number of iterations return code 1 if converged, 0 otherwise mgarch dvech command as typed names of dependent variables list of covariates dependent variables with mean equations independent variables in each equation time variable title in estimation output Wald; type of model 2 test vcetype specied in vce() title used to label Std. Err. formatted minimum time formatted maximum time distribution for error term: gaussian or t specied ARCH terms specied GARCH terms maximization technique(s) for starting values maximization technique b V program used to implement estat program used to implement predict predictions allowed by margins predictions disallowed by margins
274
Matrices e(b) e(Cns) e(ilog) e(gradient) e(hessian) e(A) e(B) e(S) e(Sigma) e(V) e(pinfo) Functions e(sample)
coefcient vector constraints matrix iteration log (up to 20 iterations) gradient vector Hessian matrix estimates of A matrices estimates of B matrices estimates of Sigma0 matrix Sigma hat variancecovariance matrix of the estimators parameter information, used by predict marks estimation sample

mgarch dvech is implemented as an ado-le. Recall that the diagonal vech MGARCH model can be written as
yt = Cxt +
t
= Ht t
p q
1/2
Ht = S +
i=1
Ai
ti ti
+
j=1
Bj
Htj
where
1/2
S is an m m symmetric matrix of parameters;

each Ai is an m m symmetric matrix of parameters; is the elementwise or Hadamard product; and each Bj is an m m symmetric matrix of parameters. mgarch dvech estimates the parameters by maximum likelihood. The log-likelihood function based on the multivariate normal distribution for observation t is
lt = 0.5m log(2) 0.5log {det (Ht )} 0.5 t H1 t

where
t
= yt Cxt . The log-likelihood function is
T t=1 lt .
275
lt = log
df + m 2
log
df 2
m log {(df 2)} 2
0.5log {det (Ht )}
1 df + m t Ht t log 1 + 2 df 2
mgarch dvech ensures that Ht is positive denite for each t. By default, mgarch dvech performs an iterative search for starting values. mgarch dvech estimates starting values for C by seemingly unrelated regression, uses these estimates to compute residuals t , plugs t into the above log-likelihood function, and optimizes this log-likelihood function over the parameters in Ht . This starting-value method plugs in consistent estimates of the parameters for the conditional means of the dependent variables and then iteratively searches for the variance parameters that maximize the log-likelihood function. L tkepohl (2005, chap. 16) discusses this method as an u estimator for the variance parameters.
GARCH estimators require initial values that can be plugged in for ti ti and Htj when t i < 1 and t j < 1. mgarch dvech substitutes an estimator of the unconditional covariance of the disturbances,
T
= T 1
t=1
t t
(4)
mgarch dvech uses analytic rst and second derivatives in maximizing the log-likelihood function based on the multivariate normal distribution. mgarch dvech uses numerical derivatives in maximizing the log-likelihood function based on the multivariate t distribution.
References
Bollerslev, T., R. F. Engle, and J. M. Wooldridge. 1988. A capital asset pricing model with time-varying covariances. Journal of Political Economy 96: 116131. L tkepohl, H. 2005. New Introduction to Multiple Time Series Analysis. New York: Springer. u
Also see
[TS] mgarch dvech postestimation Postestimation tools for mgarch dvech [TS] mgarch Multivariate GARCH models [TS] tsset Declare data to be time-series data [TS] arch Autoregressive conditional heteroskedasticity (ARCH) family of estimators [TS] var Vector autoregressive models [U] 20 Estimation and postestimation commands
Title
mgarch dvech postestimation Postestimation tools for mgarch dvech
Description
The following standard postestimation commands are available after mgarch dvech: Command contrast estat estimates lincom Description contrasts and ANOVA-style joint tests of estimates AIC, BIC, VCE, and estimation sample summary
Syntax for predict

predict statistic
Main
type
Description
if
in
, statistic options
options
Options
Description
276
277
Menu
Statistics
>
Postestimation
>
Options for predict

Main
Options
Remarks
We assume that you have already read [TS] mgarch dvech. In this entry, we illustrate some of the features of predict after using mgarch dvech to estimate the parameters of diagonal vech MGARCH models.
Example 1
In this example, we obtain dynamic predictions for the Acme Inc. and Anvil Inc. ctional widget data modeled in example 3 of [TS] mgarch dvech. We begin by reestimating the parameters of the model.
278

. use http://www.stata-press.com/data/r12/acme . constraint 1 [L.ARCH]1_1 = [L.ARCH]2_2 . constraint 2 [L.GARCH]1_1 = [L.GARCH]2_2 . mgarch dvech (acme = L.acme) (anvil = L.anvil), arch(1) garch(1) Getting starting values (setting technique to bhhh) Iteration 0: log likelihood = -6087.0665 (not concave) Iteration 1: log likelihood = -6022.2046 Iteration 2: log likelihood = -5986.6152 Iteration 3: log likelihood = -5976.5739 Iteration 4: log likelihood = -5974.4342 Iteration 5: log likelihood = -5974.4046 Iteration 6: log likelihood = -5974.4036 Iteration 7: log likelihood = -5974.4035 Estimating parameters (setting technique to bhhh) Iteration 0: log likelihood = -5974.4035 Iteration 1: log likelihood = -5973.812 Iteration 2: log likelihood = -5973.8004 Iteration 3: log likelihood = -5973.7999 Iteration 4: log likelihood = -5973.7999 Diagonal vech MGARCH model Sample: 1969w35 - 1998w25 Number of obs Distribution: Gaussian Wald chi2(2) Log likelihood = -5973.8 Prob > chi2 ( 1) [L.ARCH]1_1 - [L.ARCH]2_2 = 0 ( 2) [L.GARCH]1_1 - [L.GARCH]2_2 = 0 Coef. acme acme L1. _cons anvil anvil L1. _cons Sigma0 1_1 2_1 2_2 L.ARCH 1_1 2_1 2_2 L.GARCH 1_1 2_1 2_2 .1487581 .085404 .1487581 .0697531 .1446524 .0697531 2.13 0.59 2.13 0.033 0.555 0.033 .0120445 -.1981094 .0120445 .2854716 .3689175 .2854716 .2813443 .181877 .2813443 .0299124 .0335393 .0299124 9.41 5.42 9.41 0.000 0.000 0.000 .222717 .1161412 .222717 .3399716 .2476128 .3399716 1.889237 .4599576 2.063113 .2168733 .1139843 .2454633 8.71 4.04 8.40 0.000 0.000 0.000 1.464173 .2365525 1.582014 2.314301 .6833626 2.544213 .3151955 1.215786 .0263287 .0642052 11.97 18.94 0.000 0.000 .2635922 1.089947 .3667988 1.341626 .3365278 1.124611 .0255134 .060085 13.19 18.72 0.000 0.000 .2865225 1.006847 .3865331 1.242376 Std. Err. z P>|z|
constraints(1 2)
= = =
1499 272.47 0.0000
Now we use tsappend (see [TS] tsappend) to extend the data, use predict to obtain the dynamic predictions, and graph the predictions.

. tsappend, add(12) . predict H*, variance dynamic(tw(1998w26)) . tsline H_acme_acme H_anvil_anvil if t>=tw(1995w25), legend(rows(2))
279
2 1995w26
10
1996w27 t
1997w26
1998w26
variance prediction (acme, acme) dynamic(tw(1998w26)) variance prediction (anvil, anvil) dynamic(tw(1998w26))
The graph shows that the in-sample predictions are similar for the conditional variances of Acme Inc. and Anvil Inc. and that the dynamic forecasts converge to similar levels. It also shows that the ARCH and GARCH parameters cause substantial time-varying volatility. The predicted conditional variance of acme ranges from lows of just over 2 to highs above 10.
Example 2
In this example, we obtain the in-sample predicted conditional variances of the returns for the ctional Acme Inc., which we modeled in example 4 of [TS] mgarch dvech. First, we reestimate the parameters of the model.
. use http://www.stata-press.com/data/r12/aacmer, clear . mgarch dvech (acme anvil = , noconstant), arch(1/2) garch(1) Getting starting values (setting technique to bhhh) Iteration 0: log likelihood = -18417.243 (not concave) Iteration 1: log likelihood = -18215.005 Iteration 2: log likelihood = -18199.691 Iteration 3: log likelihood = -18136.699 Iteration 4: log likelihood = -18084.256 Iteration 5: log likelihood = -17993.662 Iteration 6: log likelihood = -17731.1 Iteration 7: log likelihood = -17629.505 (switching technique to nr) Iteration 8: log likelihood = -17548.172 Iteration 9: log likelihood = -17544.987 Iteration 10: log likelihood = -17544.937 Iteration 11: log likelihood = -17544.937 Estimating parameters (setting technique to bhhh) Iteration 0: log likelihood = -17544.937 Iteration 1: log likelihood = -17544.937
280

Diagonal vech MGARCH model Sample: 1 - 5000 Distribution: Gaussian Log likelihood = -17544.94 Coef. Sigma0 1_1 2_1 2_2 L.ARCH 1_1 2_1 2_2 L2.ARCH 1_1 2_1 2_2 L.GARCH 1_1 2_1 2_2 .0782467 .2888104 .201618 .053944 .0818303 .0470584 1.45 3.53 4.28 0.147 0.000 0.000 -.0274816 .1284261 .1093853 .183975 .4491948 .2938508 .1837825 .0884425 .2025718 .0274555 .02208 .0272639 6.69 4.01 7.43 0.000 0.000 0.000 .1299706 .0451665 .1491355 .2375943 .1317185 .256008 .2878739 .1036685 .2034196 .02157 .0161446 .019855 13.35 6.42 10.25 0.000 0.000 0.000 .2455975 .0720256 .1645044 .3301504 .1353114 .2423347 1.026283 .4300997 1.019753 .0823348 .0590294 .0837146 12.46 7.29 12.18 0.000 0.000 0.000 .8649096 .3144042 .8556751 1.187656 .5457952 1.18383 Std. Err. z Number of obs Wald chi2(.) Prob > chi2 P>|z| = = = 5000 . .
Now we use predict to obtain the in-sample conditional variances of acme and use tsline (see [TS] tsline) to graph the results.
. predict h_acme, variance eq(acme, acme) . tsline h_acme
50 0 0 variance prediction (acme, acme) 10 20 30 40
1000
2000 t
3000
4000
5000
The graph shows that the predicted conditional variances vary substantially over time, as the parameter estimates indicated. Because there are no covariates in the model for acme, specifying xb puts a prediction of 0 in each observation, and specifying residuals puts the value of the dependent variable into the prediction.
281

All postestimation commands listed above are implemented as ado-les. All one-step predictions are obtained by substituting the parameter estimates into the model. The estimated unconditional variance matrix of the disturbances, , is the initial value for the ARCH and GARCH terms. The postestimation routines recompute using the prediction sample, the parameter estimates stored in e(b), and (4) in Methods and formulas of [TS] mgarch dvech. For observations in which the residuals are missing, the estimated unconditional variance matrix of the disturbances is used in place of the outer product of the residuals. Dynamic predictions of the dependent variables use previously predicted values beginning in the period specied by dynamic(). Dynamic variance predictions are implemented by substituting for the outer product of the residuals beginning in the period specied by dynamic().
Also see
[TS] mgarch dvech Diagonal vech multivariate GARCH models [U] 20 Estimation and postestimation commands
Title
mgarch vcc Varying conditional correlation multivariate GARCH models
Syntax
mgarch vcc eq eq . . . eq if in , options where each eq has the form (depvars = indepvars options
Model
arch(numlist) garch(numlist) het(varlist) distribution(dist # ) constraints(numlist)

SE/Robust
include varlist in the specication of the conditional variance for all equations use dist distribution for errors [may be gaussian (synonym normal) or t; default is gaussian] apply linear constraints vcetype may be oim or robust set condence level; default is level(95) do not display constraints control column formats, row spacing, line width, and display of omitted variables and base and empty cells control the maximization process; seldom used initial values for the coefcients; seldom used display legend instead of statistics
vce(vcetype)
Reporting

Maximization

ARCH terms GARCH terms include varlist in the specication of the conditional variance
You must tsset your data before using mgarch vcc; see [TS] tsset. indepvars and varlist may contain factor variables; see [U] 11.4.3 Factor variables. depvars, indepvars, and varlist may contain time-series operators; see [U] 11.4.4 Time-series varlists. by, statsby, and rolling are allowed; see [U] 11.1.10 Prex commands. coeflegend does not appear in the dialog box. See [U] 20 Estimation and postestimation commands for more capabilities of estimation commands.
282
283
Menu
Statistics
>
>
Multivariate GARCH
Description
mgarch vcc estimates the parameters of varying conditional correlation (VCC) multivariate generalized autoregressive conditionally heteroskedastic (MGARCH) models in which the conditional variances are modeled as univariate generalized autoregressive conditionally heteroskedastic (GARCH) models and the conditional covariances are modeled as nonlinear functions of the conditional variances. The conditional correlation parameters that weight the nonlinear combinations of the conditional variance follow the GARCH-like process specied in Tse and Tsui (2002). The VCC MGARCH model is about as exible as the closely related dynamic conditional correlation MGARCH model (see [TS] mgarch dcc), more exible than the conditional correlation MGARCH model (see [TS] mgarch ccc), and more parsimonious than the diagonal vech model (see [TS] mgarch
dvech).
Options

Model
arch(numlist) species the ARCH terms for all equations in the model. By default, no ARCH terms are specied. garch(numlist) species the GARCH terms for all equations in the model. By default, no GARCH terms are specied. het(varlist) species that varlist be included in the model in the specication of the conditional variance for all equations. This varlist enters the variance specication collectively as multiplicative heteroskedasticity. distribution(dist # ) species the assumed distribution for the errors. dist may be gaussian, normal, or t. gaussian and normal are synonyms; each causes mgarch vcc to assume that the errors come from a multivariate normal distribution. # may not be specied with either of them. t causes mgarch vcc to assume that the errors follow a multivariate Student t distribution, and the degree-of-freedom parameter is estimated along with the other parameters of the model. If distribution(t #) is specied, then mgarch vcc uses a multivariate Student t distribution with # degrees of freedom. # must be greater than 2. constraints(numlist) species linear constraints to apply to the parameter estimates.
SE/Robust
Reporting
level(#); see [R] estimation options. nocnsreport; see [R] estimation options.
284
display options: noomitted, vsquish, noemptycells, baselevels, allbaselevels, cformat(% fmt), pformat(% fmt), sformat(% fmt), and nolstretch; see [R] estimation options.
Maximization
maximize options: difficult, technique(algorithm spec), iterate(#), no log, trace, gradient, showstep, hessian, showtolerance, tolerance(#), ltolerance(#), nrtolerance(#), and nonrtolerance, and from(matname); see [R] maximize for all options except from(), and see below for information on from(). These options are seldom used. from(matname) species initial values for the coefcients. from(b0) causes mgarch vcc to begin the optimization algorithm with the values in b0. b0 must be a row vector, and the number of columns must equal the number of parameters in the model. The following option is available with mgarch vcc but is not shown in the dialog box: coeflegend; see [R] estimation options.
Eqoptions
Remarks
We assume that you have already read [TS] mgarch, which provides an introduction to MGARCH models and the methods implemented in mgarch vcc.
MGARCH models are dynamic multivariate regression models in which the conditional variances and covariances of the errors follow an autoregressive-moving-average structure. The VCC MGARCH model uses a nonlinear combination of univariate GARCH models with time-varying cross-equation weights to model the conditional covariance matrix of the errors.
As discussed in [TS] mgarch, MGARCH models differ in the parsimony and exibility of their specications for a time-varying conditional covariance matrix of the disturbances, denoted by Ht . In the conditional correlation family of MGARCH models, the diagonal elements of Ht are modeled as univariate GARCH models, whereas the off-diagonal elements are modeled as nonlinear functions of the diagonal terms. In the VCC MGARCH model,
hij,t = ij,t
hii,t hjj,t
where the diagonal elements hii,t and hjj,t follow univariate GARCH processes and ij,t follows the dynamic process specied in Tse and Tsui (2002) and discussed below. Because the ij,t varies with time, this model is known as the VCC GARCH model.
285
Technical note
The VCC GARCH model proposed by Tse and Tsui (2002) can be written as
yt = Cxt +
t
Ht = Rt = (1 1 2 )R + 1 t1 + 2 Rt1
where
t 1/2 Ht t 1/2 1/2 Dt Rt Dt
(1)
1/2
t is an m 1 vector of independent and identically distributed innovations;

2 1,t 0 Dt = . . . 0
0 2 2,t . . . 0
0 0 . .. . . . 2 m,t
2 i,tj
+ j
qi j=1
2 j i,tj qi j=1
by default, or
2 j i,tj
when the het() option is specied, where t is a 1 p vector of parameters, zi is a p 1 vector of independent variables including a constant term, the j s are ARCH parameters, and the j s are GARCH parameters;
Rt is a matrix of conditional correlations, 1 12,t 1 . . . 2m,t 1m,t 2m,t . .. . . . 1
12,t Rt = . . . 1m,t
R is the matrix of means to which the dynamic process in (1) reverts; t is the rolling estimator of the correlation matrix of observations; and
t,
which uses the previous m + 1
1 and 2 are parameters that govern the dynamics of conditional correlations. 1 and 2 are nonnegative and satisfy 0 1 + 2 < 1.
286
To differentiate this model from Engle (2002), Tse and Tsui (2002) call their model a VCC MGARCH model.

We have daily data on the stock returns of three car manufacturersToyota, Nissan, and Honda, from January 2, 2003, to December 31, 2010in the variables toyota, nissan, and honda. We model the conditional means of the returns as a rst-order vector autoregressive process and the conditional covariances as a VCC MGARCH process in which the variance of each disturbance term follows a GARCH(1,1) process.
. use http://www.stata-press.com/data/r12/stocks (Data from Yahoo! Finance) . mgarch vcc (toyota nissan honda = L.toyota L.nissan L.honda, noconstant), > arch(1) garch(1) Calculating starting values.... Optimizing log likelihood (setting technique to bhhh) Iteration 0: log likelihood = 16901.2 Iteration 1: log likelihood = 17028.644 Iteration 2: log likelihood = 17145.905 Iteration 3: log likelihood = 17251.485 Iteration 4: log likelihood = 17306.115 Iteration 5: log likelihood = 17332.59 Iteration 6: log likelihood = 17353.617 Iteration 7: log likelihood = 17374.86 Iteration 8: log likelihood = 17398.526 Iteration 9: log likelihood = 17418.748 (switching technique to nr) Iteration 10: log likelihood = 17442.552 Iteration 11: log likelihood = 17455.66 Iteration 12: log likelihood = 17463.597 Iteration 13: log likelihood = 17463.922 Iteration 14: log likelihood = 17463.925 Iteration 15: log likelihood = 17463.925 Refining estimates Iteration 0: log likelihood = 17463.925 Iteration 1: log likelihood = 17463.925 Varying conditional correlation MGARCH model Sample: 1 - 2015 Number of obs = 2014 Distribution: Gaussian Wald chi2(9) = 17.67 Log likelihood = 17463.92 Prob > chi2 = 0.0392 Coef. toyota toyota L1. nissan L1. honda L1. Std. Err. z P>|z| [95% Conf. Interval]
-.0565645
.0335696
-1.68
0.092
-.1223597
.0092307
.0248101
.0252701
0.98
0.326
-.0247184
.0743385
.0035836
.0298895
0.12
0.905
-.0549986
.0621659
287
ARCH_toyota arch L1. garch L1. _cons nissan toyota L1. nissan L1. honda L1. ARCH_nissan arch L1. garch L1. _cons honda toyota L1. nissan L1. honda L1. ARCH_honda arch L1. garch L1. _cons Correlation toyota nissan honda nissan honda Adjustment lambda1 lambda2
.0602807
.0086799
6.94
0.000
.0432684
.077293
.9224689 4.38e-06
.0110316 1.12e-06
83.62 3.91
0.000 0.000
.9008473 2.18e-06
.9440906 6.58e-06
-.0196399
.0387112
-0.51
0.612
-.0955124
.0562326
-.0306663
.031051
-0.99
0.323
-.091525
.0301925
.038315
.0354691
1.08
0.280
-.0312031
.1078331
.0774228
.0119642
6.47
0.000
.0539733
.1008723
.9076856 6.20e-06
.0139339 1.70e-06
65.14 3.65
0.000 0.000
.8803756 2.87e-06
.9349956 9.53e-06
-.0358292
.0340492
-1.05
0.293
-.1025645
.030906
.0544071
.0276156
1.97
0.049
.0002814
.1085327
-.0424383
.0326249
-1.30
0.193
-.106382
.0215054
.0458673
.0072714
6.31
0.000
.0316157
.0601189
.9369253 4.99e-06
.0101755 1.29e-06
92.08 3.85
0.000 0.000
.9169816 2.45e-06
.9568689 7.52e-06
.6643028 .7302093
.0151086 .0126361
43.97 57.79
0.000 0.000
.6346905 .705443
.6939151 .7549755
.6347321
.0159738
39.74
0.000
.603424
.6660401
.0277374 .8255525
.0086942 .0755882
3.19 10.92
0.001 0.000
.010697 .6774024
.0447778 .9737026
The output has three parts: an iteration log, a header, and an output table.
288
The iteration log has three parts: the dots from the search for initial values, the iteration log from optimizing the log likelihood, and the iteration log from the rening step. A detailed discussion of the optimization methods is in Methods and formulas. The header describes the estimation sample and reports a Wald test against the null hypothesis that all the coefcients on the independent variables in the mean equations are zero. Here the null hypothesis is rejected at the 5% level. The output table rst presents results for the mean or variance parameters used to model each dependent variable. Subsequently, the output table presents results for the parameters in R. For example, the estimate of the mean of the process that associates Toyota and Nissan is 0.66. Finally, the output table presents results for the adjustment parameters 1 and 2 . In the example at hand, the estimates for both 1 and 2 are statistically signicant. The VCC MGARCH model reduces to the CCC MGARCH model when 1 = 2 = 0. The output below shows that a Wald test rejects the null hypothesis that 1 = 2 = 0 at all conventional levels.
. test _b[Adjustment:lambda1] = _b[Adjustment:lambda2] = 0 ( 1) ( 2) [Adjustment]lambda1 - [Adjustment]lambda2 = 0 [Adjustment]lambda1 = 0 chi2( 2) = Prob > chi2 = 482.80 0.0000
These results indicate that the assumption of time-invariant conditional correlations maintained in the CCC MGARCH model is too restrictive for these data.
Example 2
We improve the previous example by removing the insignicant parameters from the model. To accomplish that, we specify the honda equation separately from the toyota and nissan equations:
. mgarch vcc (toyota nissan = , noconstant) (honda = L.nissan, noconstant), > arch(1) garch(1) Calculating starting values.... Optimizing log likelihood (setting technique to bhhh) Iteration 0: log likelihood Iteration 1: log likelihood Iteration 2: log likelihood Iteration 3: log likelihood Iteration 4: log likelihood Iteration 5: log likelihood Iteration 6: log likelihood Iteration 7: log likelihood Iteration 8: log likelihood Iteration 9: log likelihood (switching technique to nr) Iteration 10: log likelihood Iteration 11: log likelihood Iteration 12: log likelihood Iteration 13: log likelihood Iteration 14: log likelihood Iteration 15: log likelihood Refining estimates Iteration 0: Iteration 1: log likelihood = log likelihood = 17455.98 17455.98 (backed up) = = = = = = = = = = = = = = = = 16889.43 17002.567 17134.525 17233.192 17295.342 17326.347 17348.063 17363.988 17387.216 17404.734 17438.432 17450.001 17455.443 17455.971 17455.98 17455.98 (not concave)

Varying conditional correlation MGARCH model Sample: 1 - 2015 Distribution: Gaussian Log likelihood = 17455.98 Coef. ARCH_toyota arch L1. garch L1. _cons ARCH_nissan arch L1. garch L1. _cons honda nissan L1. ARCH_honda arch L1. garch L1. _cons Correlation toyota nissan honda nissan honda Adjustment lambda1 lambda2 .0175566 .0137982 1.27 0.203 -.0094874 .0446005 Std. Err. z
289
= = =
2014 1.62 0.2032
.0609064
.0087784
6.94
0.000
.0437011
.0781117
.921703 4.42e-06
.0111493 1.13e-06
82.67 3.91
0.000 0.000
.8998509 2.20e-06
.9435552 6.64e-06
.0806598
.0123529
6.53
0.000
.0564486
.104871
.9035239 6.61e-06
.014421 1.79e-06
62.65 3.70
0.000 0.000
.8752592 3.11e-06
.9317886 .0000101
.0461398
.0073048
6.32
0.000
.0318226
.060457
.9366096 5.03e-06
.0102021 1.31e-06
91.81 3.85
0.000 0.000
.9166139 2.47e-06
.9566052 7.59e-06
.6635251 .7299703
.0150293 .0124828
44.15 58.48
0.000 0.000
.6340682 .7055045
.692982 .754436
.6338207
.0158681
39.94
0.000
.6027198
.6649217
.0285319 .8113924
.0092448 .0854955
3.09 9.49
0.002 0.000
.0104124 .6438243
.0466514 .9789604
It turns out that the coefcient on L1.nissan in the honda equation is now statistically insignicant. We could further improve the model by removing L1.nissan from the model. There is no mean equation for Toyota or Nissan. In [TS] mgarch vcc postestimation, we discuss prediction from models without covariates.
290
Example 3
Here we t a bivariate VCC MGARCH model for the Toyota and Nissan shares. We believe that the shares of these car manufacturers follow the same process, so we impose the constraints that the ARCH coefcients are the same for the two companies and that the GARCH coefcients are also the same.
. constraint 1 _b[ARCH_toyota:L.arch] = _b[ARCH_nissan:L.arch] . constraint 2 _b[ARCH_toyota:L.garch] = _b[ARCH_nissan:L.garch] . mgarch vcc (toyota nissan = , noconstant), arch(1) garch(1) constraints(1 2) Calculating starting values.... Optimizing log likelihood (setting technique to bhhh) Iteration 0: log likelihood = 10326.298 Iteration 1: log likelihood = 10680.73 Iteration 2: log likelihood = 10881.388 Iteration 3: log likelihood = 11043.345 Iteration 4: log likelihood = 11122.459 Iteration 5: log likelihood = 11202.411 Iteration 6: log likelihood = 11253.657 Iteration 7: log likelihood = 11276.325 Iteration 8: log likelihood = 11279.823 Iteration 9: log likelihood = 11281.704 (switching technique to nr) Iteration 10: log likelihood = 11282.313 Iteration 11: log likelihood = 11282.46 Iteration 12: log likelihood = 11282.461 Refining estimates Iteration 0: log likelihood = 11282.461 Iteration 1: log likelihood = 11282.461 (backed up) Varying conditional correlation MGARCH model Sample: 1 - 2015 Distribution: Gaussian Log likelihood = 11282.46 ( 1) ( 2) Number of obs Wald chi2(.) Prob > chi2 = = = 2015 . .
[ARCH_toyota]L.arch - [ARCH_nissan]L.arch = 0 [ARCH_toyota]L.garch - [ARCH_nissan]L.garch = 0 Coef. Std. Err. z P>|z| [95% Conf. Interval]
ARCH_toyota arch L1. garch L1. _cons ARCH_nissan arch L1. garch L1. _cons
.0797459
.0101634
7.85
0.000
.059826
.0996659
.9063808 4.24e-06
.0118211 1.10e-06
76.67 3.85
0.000 0.000
.883212 2.08e-06
.9295497 6.40e-06
.0797459
.0101634
7.85
0.000
.059826
.0996659
.9063808 5.91e-06
.0118211 1.47e-06
76.67 4.03
0.000 0.000
.883212 3.03e-06
.9295497 8.79e-06
291
Correlation toyota nissan Adjustment lambda1 lambda2
.6720056
.0162585
41.33
0.000
.6401394
.7038718
.0343012 .7945548
.0128097 .101067
2.68 7.86
0.007 0.000
.0091945 .5964671
.0594078 .9926425
Example 4
In this example, we have data on ctional stock returns for the Acme and Anvil corporations, and we believe that the movement of the two stocks is governed by different processes. We specify one ARCH and one GARCH term for the conditional variance equation for Acme and two ARCH terms for the conditional variance equation for Anvil. In addition, we include the lagged value of the stock return for Apex, the main subsidiary of Anvil corporation, in the variance equation of Anvil. For Acme, we have data on the changes in an index of futures prices of products related to those produced by Acme in afrelated. For Anvil, we have data on the changes in an index of futures prices of inputs used by Anvil in afinputs.
. use http://www.stata-press.com/data/r12/acmeh . mgarch vcc (acme = afrelated, noconstant arch(1) garch(1)) > (anvil = afinputs, arch(1/2) het(L.apex)) Calculating starting values.... Optimizing log likelihood (setting technique to bhhh) Iteration 0: log likelihood = -13252.793 Iteration 1: log likelihood = -12859.124 Iteration 2: log likelihood = -12522.14 Iteration 3: log likelihood = -12406.487 Iteration 4: log likelihood = -12304.275 Iteration 5: log likelihood = -12273.103 Iteration 6: log likelihood = -12256.104 Iteration 7: log likelihood = -12254.55 Iteration 8: log likelihood = -12254.482 Iteration 9: log likelihood = -12254.478 (switching technique to nr) Iteration 10: log likelihood = -12254.478 Iteration 11: log likelihood = -12254.478 Refining estimates Iteration 0: Iteration 1: log likelihood = -12254.478 log likelihood = -12254.478
292

Varying conditional correlation MGARCH model Sample: 1 - 2500 Distribution: Gaussian Log likelihood = -12254.48 Coef. acme afrelated ARCH_acme arch L1. garch L1. _cons anvil afinputs _cons ARCH_anvil arch L1. L2. apex L1. _cons Correlation acme anvil Adjustment lambda1 lambda2 Std. Err. z
= = =
2499 5226.19 0.0000
.9672465
.0510066
18.96
0.000
.8672753
1.067218
.0949142
.0147302
6.44
0.000
.0660435
.1237849
.7689442 2.129468
.038885 .464916
19.77 4.58
0.000 0.000
.6927309 1.218249
.8451574 3.040687
-1.018629 .1015986
.0145027 .0177952
-70.24 5.71
0.000 0.000
-1.047053 .0667205
-.9902037 .1364766
.4990272 .2839812
.0243531 .0181966
20.49 15.61
0.000 0.000
.4512959 .2483165
.5467584 .3196459
1.897144 .0682724
.0558791 .0662257
33.95 1.03
0.000 0.303
1.787623 -.0615276
2.006665 .1980724
-.6574256
.0294259
-22.34
0.000
-.7150994
-.5997518
.2375029 .6492072
.0179114 .0254493
13.26 25.51
0.000 0.000
.2023971 .5993274
.2726086 .6990869
The results indicate that increases in the futures prices for related products lead to higher returns on the Acme stock, and increased input prices lead to lower returns on the Anvil stock. In the conditional variance equation for Anvil, the coefcient on L1.apex is positive and signicant, which indicates that an increase in the return on the Apex stock leads to more variability in the return on the Anvil stock.
293
Saved results
mgarch vcc saves the following in e():
Scalars e(N) e(k) e(k aux) e(k extra) e(k eq) e(k dv) e(df m) e(ll) e(chi2) e(p) e(estdf) e(usr) e(tmin) e(tmax) e(N gaps) e(rank) e(ic) e(rc) e(converged) Macros e(cmd) e(model) e(cmdline) e(depvar) e(covariates) e(dv eqs) e(indeps) e(tvar) e(title) e(chi2type) e(vce) e(vcetype) e(tmins) e(tmaxs) e(dist) e(arch) e(garch) e(technique) e(properties) e(estat cmd) e(predict) e(marginsok) e(marginsnotok) Matrices e(b) e(Cns) e(ilog) e(gradient) e(hessian) e(V) e(pinfo) Functions e(sample) number of observations number of parameters number of auxiliary parameters number of extra estimates added to number of equations in e(b) number of dependent variables model degrees of freedom log likelihood
2
signicance 1 if distribution parameter was estimated, 0 otherwise user-provided distribution parameter minimum time in sample maximum time in sample number of gaps rank of e(V) number of iterations return code 1 if converged, 0 otherwise mgarch vcc command as typed names of dependent variables list of covariates dependent variables with mean equations independent variables in each equation time variable title in estimation output Wald; type of model 2 test vcetype specied in vce() title used to label Std. Err. formatted minimum time formatted maximum time distribution for error term: gaussian or t specied ARCH terms specied GARCH terms maximization technique b V program used to implement estat program used to implement predict predictions allowed by margins predictions disallowed by margins coefcient vector constraints matrix iteration log (up to 20 iterations) gradient vector Hessian matrix variancecovariance matrix of the estimators parameter information, used by predict marks estimation sample
294

mgarch vcc is implemented as an ado-le. mgarch vcc estimates the parameters of the varying conditional correlation MGARCH model by maximum likelihood. The log-likelihood function based on the multivariate normal distribution for observation t is
lt = 0.5m log(2) 0.5log {det (Rt )} log det Dt

1/2 t
1/2
0.5 t R1 t
lt = log
df + m 2
log
df 2
1/2
m log {(df 2)} 2

1 df + m t Rt t log 1 + 2 df 2
0.5log {det (Rt )} log det Dt
The starting values for the parameters in the mean equations and the initial residuals t are obtained by least-squares regression. The starting values for the parameters in the variance equations are obtained by a procedure proposed by Gourieroux and Monfort (1997, sec. 6.2.2). The starting values for the parameters in R are calculated from the standardized residuals t . Given the starting values for the mean and variance equations, the starting values for the parameters 1 and 2 are obtained from a grid search performed on the log likelihood. The initial optimization step is performed in the unconstrained space. Once the maximum is found, we impose the constraints 1 0, 2 0, and 0 1 + 2 < 1, and maximize the log likelihood in the constrained space. This step is reported in the iteration log as the rening step.
GARCH estimators require initial values that can be plugged in for ti ti and Htj when t i < 1 and t j < 1. mgarch vcc substitutes an estimator of the unconditional covariance of the disturbances
T
= T 1
t=1
t t
(2)
mgarch vcc uses numerical derivatives in maximizing the log-likelihood function.
References
Engle, R. F. 2002. Dynamic conditional correlation: A simple class of multivariate generalized autoregressive conditional heteroskedasticity models. Journal of Business & Economic Statistics 20: 339350. Gourieroux, C., and A. Monfort. 1997. Time Series and Dynamic Models. Trans. ed. G. M. Gallo. Cambridge: Cambridge University Press.
295
Tse, Y. K., and A. K. C. Tsui. 2002. A multivariate generalized autoregressive conditional heteroscedasticity model with time-varying correlations. Journal of Business & Economic Statistics 20: 351362.
Also see
[TS] mgarch vcc postestimation Postestimation tools for mgarch vcc [TS] mgarch Multivariate GARCH models [TS] tsset Declare data to be time-series data [TS] arch Autoregressive conditional heteroskedasticity (ARCH) family of estimators [TS] var Vector autoregressive models [U] 20 Estimation and postestimation commands
Title
mgarch vcc postestimation Postestimation tools for mgarch vcc
Description
The following standard postestimation commands are available after mgarch vcc: Command contrast estat estimates lincom Description contrasts and ANOVA-style joint tests of estimates AIC, BIC, VCE, and estimation sample summary
cataloging estimation results point estimates, standard errors, testing, and inference for linear combinations of coefcients lrtest likelihood-ratio test margins marginal means, predictive margins, marginal effects, and average marginal effects marginsplot graph the results from margins (prole plots, interaction plots, etc.) nlcom point estimates, standard errors, testing, and inference for nonlinear combinations of coefcients predict predictions and residuals predictnl point estimates, standard errors, testing, and inference for generalized predictions pwcompare pairwise comparisons of estimates test Wald tests of simple and composite linear hypotheses testnl Wald tests of nonlinear hypotheses See the corresponding entries in the Base Reference Manual for details.
Syntax for predict

predict statistic
Main
type
Description
if
in
, statistic options
options
Options
Description
296
297
Menu
Statistics
>
Postestimation
>
Options for predict

Main
Options
Remarks
We assume that you have already read [TS] mgarch vcc. In this entry, we use predict after mgarch vcc to make in-sample and out-of-sample forecasts.
Example 1
In this example, we obtain dynamic forecasts for the Toyota, Nissan, and Honda stock returns modeled in example 2 of [TS] mgarch vcc. In the output below, we reestimate the parameters of the model, use tsappend (see [TS] tsappend) to extend the data, and use predict to obtain in-sample one-step-ahead forecasts and dynamic forecasts of the conditional variances of the returns. We graph the forecasts below.
. use http://www.stata-press.com/data/r12/stocks (Data from Yahoo! Finance) . quietly mgarch vcc (toyota nissan = , noconstant) (honda = L.nissan, noconstant), > arch(1) garch(1) . tsappend, add(50) . predict H*, variance dynamic(2016)
298

.003 01jan2009 0 .001 .002
01jul2009
01jan2010 Date
01jul2010
01jan2011

All postestimation commands listed above are implemented as ado-les. All one-step predictions are obtained by substituting the parameter estimates into the model. The estimated unconditional variance matrix of the disturbances, , is the initial value for the ARCH and GARCH terms. The postestimation routines recompute using the prediction sample, the parameter estimates stored in e(b), and (2) in Methods and formulas of [TS] mgarch vcc. For observations in which the residuals are missing, the estimated unconditional variance matrix of the disturbances is used in place of the outer product of the residuals. Dynamic predictions of the dependent variables use previously predicted values beginning in the period specied by dynamic(). Dynamic variance predictions are implemented by substituting for the outer product of the residuals beginning in the period specied in dynamic().
Also see
[TS] mgarch vcc Varying conditional correlation multivariate GARCH models [U] 20 Estimation and postestimation commands
Title
newey Regression with NeweyWest standard errors
Syntax
newey depvar options
Model
indepvars
if
in
weight , lag(#)
options
Description set maximum lag order of autocorrelation suppress constant term set condence level; default is level(95) control column formats, row spacing, line width, and display of omitted variables and base and empty cells display legend instead of statistics
lag(#) noconstant level(#) display options
Reporting
coeflegend
lag(#) is required. You must tsset your data before using newey; see [TS] tsset. indepvars may contain factor variables; see [U] 11.4.3 Factor variables. depvar and indepvars may contain time-series operators; see [U] 11.4.4 Time-series varlists. by, rolling, and statsby are allowed; see [U] 11.1.10 Prex commands. aweights are allowed; see [U] 11.1.6 weight. coeflegend does not appear in the dialog box. See [U] 20 Estimation and postestimation commands for more capabilities of estimation commands.
Menu
Statistics
>
Time series
>
Regression with Newey-West std. errors
Description
newey produces Newey West standard errors for coefcients estimated by OLS regression. The error structure is assumed to be heteroskedastic and possibly autocorrelated up to some lag.
Options

Model
lag(#) species the maximum lag to be considered in the autocorrelation structure. If you specify lag(0), the output is the same as regress, vce(robust). lag() is required. noconstant; see [R] estimation options.
299
300
Reporting
level(#); see [R] estimation options. display options: noomitted, vsquish, noemptycells, baselevels, allbaselevels, cformat(% fmt), pformat(% fmt), sformat(% fmt), and nolstretch; see [R] estimation options. The following option is available with newey but is not shown in the dialog box: coeflegend; see [R] estimation options.
Remarks
The Huber/White/sandwich robust variance estimator (see White [1980]) produces consistent standard errors for OLS regression coefcient estimates in the presence of heteroskedasticity. The Newey West (1987) variance estimator is an extension that produces consistent estimates when there is autocorrelation in addition to possible heteroskedasticity. The Newey West variance estimator handles autocorrelation up to and including a lag of m, where m is specied by stipulating the lag() option. Thus, it assumes that any autocorrelation at lags greater than m can be ignored. If lag(0) is specied, the variance estimates produced by newey are simply the Huber/White/sandwich robust variances estimates calculated by regress, vce(robust); see [R] regress.
Example 1
newey, lag(0) is equivalent to regress, vce(robust):
. use http://www.stata-press.com/data/r12/auto (1978 Automobile Data) . regress price weight displ, vce(robust) Linear regression Number of obs = F( 2, 71) = Prob > F = R-squared = Root MSE = Robust Std. Err. .7808755 7.436967 1129.602 74 14.44 0.0000 0.2909 2518.4
price weight displacement _cons
Coef. 1.823366 2.087054 247.907
t 2.34 0.28 0.22
P>|t| 0.022 0.780 0.827
[95% Conf. Interval] .2663445 -12.74184 -2004.455 3.380387 16.91595 2500.269
. generate t = _n . tsset t time variable: delta:
t, 1 to 74 1 unit

. newey price weight displ, lag(0) Regression with Newey-West standard errors maximum lag: 0
301
Number of obs F( 2, 71) Prob > F
= = =
74 14.44 0.0000
price weight displacement _cons
Coef. 1.823366 2.087054 247.907
Newey-West Std. Err. .7808755 7.436967 1129.602
t 2.34 0.28 0.22
P>|t| 0.022 0.780 0.827
[95% Conf. Interval] .2663445 -12.74184 -2004.455 3.380387 16.91595 2500.269
Because newey requires the dataset to be tsset, we generated a dummy time variable t, which in this example played no role in the estimation.
Example 2
Say that we have time-series measurements on variables usr and idle and now wish to t an
OLS model but obtain Newey West standard errors allowing for a lag of up to 3: . use http://www.stata-press.com/data/r12/idle2, clear . tsset time time variable: time, 1 to 30 delta: 1 unit . newey usr idle, lag(3) Regression with Newey-West standard errors Number of obs maximum lag: 3 F( 1, 28) Prob > F Newey-West Std. Err. .0690927 6.327031
= = =
30 10.90 0.0026
usr idle _cons
Coef. -.2281501 23.13483
t -3.30 3.66
P>|t| 0.003 0.001
[95% Conf. Interval] -.3696801 10.17449 -.08662 36.09516
302
Saved results
newey saves the following in e():
Scalars e(N) e(df m) e(df r) e(F) e(lag) e(rank) Macros e(cmd) e(cmdline) e(depvar) e(wtype) e(wexp) e(title) e(vcetype) e(properties) e(estat cmd) e(predict) e(asbalanced) e(asobserved) Matrices e(b) e(Cns) e(V) Functions e(sample) number of observations model degrees of freedom residual degrees of freedom F statistic maximum lag rank of e(V) newey command as typed name of dependent variable weight type weight expression title in estimation output title used to label Std. Err. b V program used to implement estat program used to implement predict factor variables fvset as asbalanced factor variables fvset as asobserved coefcient vector constraints matrix variancecovariance matrix of the estimators marks estimation sample

newey is implemented as an ado-le. newey calculates the estimates OLS = (X X)1 X y
Var(OLS ) = (X X)1 X X(X X)1

That is, the coefcient estimates are simply those of OLS linear regression. For lag(0) (no autocorrelation), the variance estimates are calculated using the White formulation:
X X = X 0 X =
n nk
e2 xi xi i
i
Here ei = yi xi OLS , where xi is the ith row of the X matrix, n is the number of observations, and k is the number of predictors in the model, including the constant if there is one. The above formula is the same as that used by regress, vce(robust) with the regression-like formula (the default) for the multiplier qc ; see Methods and formulas of [R] regress.
303
For lag(m), m > 0, the variance estimates are calculated using the Newey West (1987) formulation
X X = X 0 X +
n nk
1
l=1
l m+1
et etl (xt xtl + xtl xt )

t=l+1
where xt is the row of the X matrix observed at time t.
Whitney K. Newey (1954 ) earned degrees in economics at Brigham Young University and MIT. After a period at Princeton, he returned to MIT as a professor in 1990. His interests in theoretical and applied econometrics include bootstrapping, nonparametric estimation of models, semiparametric models, and choosing the number of instrumental variables. Kenneth D. West (1953 ) earned a bachelors degree in economics and mathematics at Wesleyan University and then a PhD in economics at MIT. After a period at Princeton, he joined the University of Wisconsin in 1988. His interests include empirical macroeconomics and timeseries econometrics.
References
Hardin, J. W. 1997. sg72: NeweyWest standard errors for probit, logit, and poisson models. Stata Technical Bulletin 39: 3235. Reprinted in Stata Technical Bulletin Reprints, vol. 7, pp. 182186. College Station, TX: Stata Press. Newey, W. K., and K. D. West. 1987. A simple, positive semi-denite, heteroskedasticity and autocorrelation consistent covariance matrix. Econometrica 55: 703708. White, H. 1980. A heteroskedasticity-consistent covariance matrix estimator and a direct test for heteroskedasticity. Econometrica 48: 817838.
Also see
[TS] newey postestimation Postestimation tools for newey [TS] tsset Declare data to be time-series data [TS] arima ARIMA, ARMAX, and other dynamic regression models [R] regress Linear regression [U] 20 Estimation and postestimation commands
Title
newey postestimation Postestimation tools for newey
Description
The following postestimation commands are available after newey: Command contrast estat estimates lincom linktest margins marginsplot nlcom predict predictnl pwcompare test testnl Description contrasts and ANOVA-style joint tests of estimates VCE and estimation sample summary cataloging estimation results point estimates, standard errors, testing, and inference for linear combinations of coefcients link test for model specication marginal means, predictive margins, marginal effects, and average marginal effects graph the results from margins (prole plots, interaction plots, etc.) point estimates, standard errors, testing, and inference for nonlinear combinations of coefcients predictions, residuals, inuence statistics, and other diagnostic measures point estimates, standard errors, testing, and inference for generalized predictions pairwise comparisons of estimates Wald tests of simple and composite linear hypotheses Wald tests of nonlinear hypotheses
Syntax for predict

predict statistic
Main
type
newvar
if
in
, statistic
Description linear prediction; the default standard error of the linear prediction residuals
xb stdp residuals
Menu
Statistics
>
Postestimation
>
304
newey postestimation Postestimation tools for newey
305
Options for predict

Main
xb, the default, calculates the linear prediction. stdp calculates the standard error of the linear prediction. residuals calculates the residuals.

Also see
[TS] newey Regression with NeweyWest standard errors [U] 20 Estimation and postestimation commands
Title
pergram Periodogram
Syntax
pergram varname options
Main
if
in
, options
Description generate newvar to contain the raw periodogram values affect rendition of the plotted points connected by lines change look of markers (color, size, etc.) add marker labels; change look or position add other plots to the generated graph any options other than by() documented in [G-3] twoway options suppress the graph
generate(newvar)
Plot
cline options marker options marker label options

Add plots
addplot(plot) twoway options nograph
You must tsset your data before using pergram; see [TS] tsset. Also, the time series must be dense (nonmissing with no gaps in the time variable) in the specied sample. varname may contain time-series operators; see [U] 11.4.4 Time-series varlists. nograph does not appear in the dialog box.
Menu
Statistics
>
Time series
>
Graphs
>
Periodogram
Description
pergram plots the log-standardized periodogram for a dense time series.
Options

Main
generate(newvar) species a new variable to contain the raw periodogram values. The generated graph log-transforms and scales the values by the sample variance and then truncates them to the [ 6, 6 ] interval before graphing them.
306
pergram Periodogram
307
Plot
cline options affect the rendition of the plotted points connected by lines; see [G-3] cline options. marker options specify the look of markers. This look includes the marker symbol, the marker size, and its color and outline; see [G-3] marker options. marker label options specify if and how the markers are to be labeled; see [G-3] marker label options.
Add plots
twoway options are any of the options documented in [G-3] twoway options, excluding by(). These include options for titling the graph (see [G-3] title options) and for saving the graph to disk (see [G-3] saving option). The following option is available with pergram but is not shown in the dialog box: nograph prevents pergram from constructing a graph.
Remarks
A good discussion of the periodogram is provided in Chateld (2004), Hamilton (1994), and Newton (1988). Chateld is also a good introductory reference for time-series analysis. Another classic reference is Box, Jenkins, and Reinsel (2008). pergram produces a scatterplot in which the points of the scatterplot are connected. The points themselves represent the log-standardized periodogram, and the connections between points represent the (continuous) log-standardized sample spectral density. In the following examples, we present the periodograms with an interpretation of the main features of the plots.
Example 1
We have time-series data consisting of 144 observations on the monthly number of international airline passengers (in thousands) between 1949 and 1960 (Box, Jenkins, and Reinsel 2008, Series G). We can graph the raw series and the log periodogram for these data by typing
308
pergram Periodogram
. use http://www.stata-press.com/data/r12/air2 (TIMESLAB: Airline passengers) . scatter air time, m(o) c(l)
100
Airline Passengers (19491960) 200 300 400 500
600
1950
1955 Time (in months)
1960
. pergram air
6.00
Airline Passengers (19491960) Log Periodogram 6.00 4.00 2.00 0.00 2.00 4.00
0.00
0.10
0.20 0.30 Frequency
0.40
0.50
Evaluated at the natural frequencies
The periodogram highlights the annual cycle together with the harmonics. Notice the peak at a frequency of about 0.08 cycles per month (cpm). The period is the reciprocal of frequency, and the reciprocal of 0.08 cpm is approximately 12 months per cycle. The similarity in shape of each group of 12 observations reveals the annual cycle. The magnitude of the cycle is increasing, resulting in the peaks in the periodogram at the harmonics of the principal annual cycle.
Example 2
This example uses 215 observations on the annual number of sunspots from 1749 to 1963 (Box and Jenkins 1976, Series E). The graph of the raw series and the log periodogram for these data are given as
6.00 4.00 2.00
0.00
2.00
4.00
6.00
Sample spectral density function
pergram Periodogram
. use http://www.stata-press.com/data/r12/sunspot (TIMESLAB: Wolfer sunspot data) . scatter spot time, m(o) c(l)
200
309
0 1750
50
Number of sunspots 100 150
1800
1850 Year
1900
1950
. pergram spot
6.00
4.00
Number of sunspots Log Periodogram 6.00 4.00 2.00 0.00 2.00
0.00
0.10
0.20 0.30 Frequency
0.40
0.50
The periodogram peaks at a frequency of slightly less than 0.10 cycles per year, indicating a 10to 12-year cycle in sunspot activity.
Example 3
Here we examine the number of trapped Canadian lynx from 1821 through 1934 (Newton 1988, 587). The raw series and the log periodogram are given as
6.00 4.00 2.00
0.00
2.00
4.00
6.00
310
pergram Periodogram
. use http://www.stata-press.com/data/r12/lynx2 (TIMESLAB: Canadian lynx) . scatter lynx time, m(o) c(l)
8000 0 0 Number of lynx trapped 2000 4000 6000
50 Time
100
150
. pergram lynx
6.00
Number of lynx trapped Log Periodogram 6.00 4.00 2.00 0.00 2.00 4.00
0.00
0.10
0.20 0.30 Frequency
0.40
0.50
The periodogram indicates that there is a cycle with a duration of about 10 years for these data but that it is otherwise random.
Example 4
To more clearly highlight what the periodogram depicts, we present the result of analyzing a time series of the sum of four sinusoids (of different periods). The periodogram should be able to decompose the time series into four different sinusoids whose periods may be determined from the plot.
6.00 4.00 2.00
0.00
2.00
4.00
6.00
pergram Periodogram
. use http://www.stata-press.com/data/r12/cos4 (TIMESLAB: Sum of 4 Cosines) . scatter sumfc time, m(o) c(l)
20
311
20 0
10
Sum of 4 cosines 0
10
50 Time
100
150
. pergram sumfc, gen(ordinate)
6.00
4.00
Sum of 4 cosines Log Periodogram 6.00 4.00 2.00 0.00 2.00
0.00
0.10
0.20 0.30 Frequency
0.40
0.50
The periodogram clearly shows the four contributions to the original time series. From the plot, we can see that the periods of the summands were 3, 6, 12, and 36, although you can conrm this by using
6.00 4.00 2.00
0.00
2.00
4.00
6.00
312
pergram Periodogram
. generate double omega = (_n-1)/144 . generate double period = 1/omega (1 missing value generated) . list period omega if ordinate> 1e-5 & omega <=.5 period 5. 13. 25. 49. 36 12 6 3 omega .02777778 .08333333 .16666667 .33333333

pergram is implemented as an ado-le. We use the notation of Newton (1988) in the following discussion. A time series of interest is decomposed into a unique set of sinusoids of various frequencies and amplitudes. A plot of the sinusoidal amplitudes (ordinates) versus the frequencies for the sinusoidal decomposition of a time series gives us the spectral density of the time series. If we calculate the sinusoidal amplitudes for a discrete set of natural frequencies (1/n, 2/n, . . . , q/n), we obtain the periodogram. Let x(1), . . . , x(n) be a time series, and let k = (k 1)/n denote the natural frequencies for k = 1, . . . , ( n/2 ) + 1. Dene
2 Ck
1 = 2 n
x(t)e
t=1
2i(t1)k
2 A plot of nCk versus k is then called the periodogram.
The sample spectral density is dened for a continuous frequency as
1 n x(t)e2i(t1) f () = n t=1 f (1 )
if [ 0, .5 ] if [ .5, 1 ]
The periodogram (and sample spectral density) is symmetric about = 0.5. Further standardize the periodogram such that
1 n
k=2
2 nCk =1 2
where 2 is the sample variance of the time series so that the average value of the ordinate is one. Once the amplitudes are standardized, we may then take the natural log of the values and produce the log periodogram. In doing so, we truncate the graph at 6. We drop the word log and simply refer to the log periodogram as the periodogram in text.
pergram Periodogram
313
References
Box, G. E. P., and G. M. Jenkins. 1976. Time Series Analysis: Forecasting and Control. Oakland, CA: HoldenDay. Box, G. E. P., G. M. Jenkins, and G. C. Reinsel. 2008. Time Series Analysis: Forecasting and Control. 4th ed. Hoboken, NJ: Wiley. Chateld, C. 2004. The Analysis of Time Series: An Introduction. 6th ed. Boca Raton, FL: Chapman & Hall/CRC. Hamilton, J. D. 1994. Time Series Analysis. Princeton: Princeton University Press. Newton, H. J. 1988. TIMESLAB: A Time Series Analysis Laboratory. Belmont, CA: Wadsworth.
Also see
[TS] tsset Declare data to be time-series data [TS] corrgram Tabulate and graph autocorrelations [TS] cumsp Cumulative spectral distribution [TS] wntestb Bartletts periodogram-based test for white noise
Title
pperron PhillipsPerron unit-root test
Syntax
pperron varname options
Main
if
in
, options
Description suppress constant term include trend term in regression display regression table use # NeweyWest lags
noconstant trend regress lags(#)
You must tsset your data before using pperron; see [TS] tsset. varname may contain time-series operators; see [U] 11.4.4 Time-series varlists.
Menu
Statistics
>
Time series
>
Tests
>
Phillips-Perron unit-root test
Description
pperron performs the PhillipsPerron (1988) test that a variable has a unit root. The null hypothesis is that the variable contains a unit root, and the alternative is that the variable was generated by a stationary process. pperron uses NeweyWest (1987) standard errors to account for serial correlation, whereas the augmented DickeyFuller test implemented in dfuller (see [TS] dfuller) uses additional lags of the rst-differenced variable.
Options

Main
noconstant suppresses the constant term (intercept) in the model. trend species that a trend term be included in the associated regression. This option may not be specied if noconstant is specied. regress species that the associated regression table appear in the output. By default, the regression table is not produced. lags(#) species the number of NeweyWest lags to use in calculating the standard error. The default is to use int 4(T /100)2/9 lags.
314
315
Remarks
As noted in [TS] dfuller, the DickeyFuller test involves tting the regression model
yt = + yt1 + t + ut
(1)
by ordinary least squares (OLS), but serial correlation will present a problem. To account for this, the augmented DickeyFuller tests regression includes lags of the rst differences of yt . The PhillipsPerron test involves tting (1), and the results are used to calculate the test statistics. Phillips and Perron (1988) proposed two alternative statistics, which pperron presents. Phillips and Perrons test statistics can be viewed as DickeyFuller statistics that have been made robust to serial correlation by using the NeweyWest (1987) heteroskedasticity- and autocorrelation-consistent covariance matrix estimator. Hamilton (1994, chap. 17) and [TS] dfuller discuss four different cases into which unit-root tests can be classied. The PhillipsPerron test applies to cases one, two, and four but not to case three. Cases one and two assume that the variable has a unit root without drift under the null hypothesis, the only difference being whether the constant term is included in regression (1). Case four assumes that the variable has a random walk, with or without drift, under the null hypothesis. Case three, which assumes that the variable has a random walk with drift under the null hypothesis, is just a special case of case four, so the fact that the PhillipsPerron test does not apply is not restrictive. The table below summarizes the relevant cases: Process under null hypothesis Random walk without drift Random walk without drift Random walk with or without drift Regression restrictions dfuller option noconstant (default) trend
Case
1 2 4
= 0, = 0 =0 (none)
The critical values for the PhillipsPerron test are the same as those for the augmented DickeyFuller test. See Hamilton (1994, chap. 17) for more information.
Example 1
Here we use the international airline passengers dataset (Box, Jenkins, and Reinsel 2008, Series G). This dataset has 144 observations on the monthly number of international airline passengers from 1949 through 1960. Because the data exhibit a clear upward trend over time, we will use the trend option.
316

. use http://www.stata-press.com/data/r12/air2 (TIMESLAB: Airline passengers) . pperron air, lags(4) trend regress Phillips-Perron test for unit root
Test Statistic Z(rho) Z(t) -46.405 -5.049
Number of obs = 143 Newey-West lags = 4 Interpolated Dickey-Fuller 1% Critical 5% Critical 10% Critical Value Value Value -27.687 -4.026 -20.872 -3.444 -17.643 -3.144
MacKinnon approximate p-value for Z(t) = 0.0002 air air L1. _trend _cons Coef. Std. Err. t P>|t| [95% Conf. Interval]
.7318116 .7107559 25.95168
.0578092 .1670563 7.325951
12.66 4.25 3.54
0.000 0.000 0.001
.6175196 .3804767 11.46788
.8461035 1.041035 40.43547
Just as in the example in [TS] dfuller, we reject the null hypothesis of a unit root at all common signicance levels. The interpolated critical values for Zt differ slightly from those shown in the example in [TS] dfuller because the sample sizes are different: with the augmented DickeyFuller regression we lose observations because of the inclusion of lagged difference terms as regressors.
Saved results
pperron saves the following in r():
Scalars r(N) r(lags) r(pval) number of observations number of lagged differences used MacKinnon approximate p-value r(Zt) r(Zrho) PhillipsPerron test statistic Phillips Perron test statistic (not included if noconstant specied)

pperron is implemented as an ado-le. In the OLS estimation of an AR(1) process with Gaussian errors,
yi = yi1 +
where i are independently and identically distributed as N (0, 2 ) and y0 = 0, the OLS estimate (based on an n-observation time series) of the autocorrelation parameter is given by
n
yi1 yi n =
i=1 n 2 yi i=1
317
If || < 1, then n(n ) N (0, 1 2 ). If this result were valid for when = 1, then the resulting distribution would have a variance of zero. When = 1, the OLS estimate still converges to one, though we need to nd a nondegenerate distribution so that we can test H0 : = 1. See Hamilton (1994, chap. 17).
The PhillipsPerron test involves tting the regression
yi = + yi1 +
where we may exclude the constant or include a trend term. There are two statistics, Z and Z , calculated as 1 n2 2 2 Z = n(n 1) n 0,n 2 s2 n
Z = j,n =
0,n n 1 1 1 n 2 n 0,n 2 2 n n sn 1 ui uij n i=j+1

q n
2 n = 0,n + 2
1 ui2
s2 = n
1 nk
j=1 n
j q+1
j,n
i=1
where ui is the OLS residual, k is the number of covariates in the regression, q is the number of 2 Newey West lags to use in calculating n , and is the OLS standard error of . The critical values, which have the same distribution as the Dickey Fuller statistic (see Dickey and Fuller 1979) included in the output, are linearly interpolated from the table of values that appear in Fuller (1996), and the MacKinnon approximate p-values use the regression surface published in MacKinnon (1994).
Peter Charles Bonest Phillips (1948 ) was born in Weymouth, England, and earned degrees in economics at the University of Auckland in New Zealand, and the London School of Economics. After periods at the Universities of Essex and Birmingham, Phillips moved to Yale in 1979. He also holds appointments at the University of Auckland and the University of York. His main research interests are in econometric theory, nancial econometrics, time-series and panel-data econometrics, and applied macroeconomics. Pierre Perron (1959 ) was born in Qu bec, Canada, and earned degrees at McGill, Queens, and e Yale in economics. After posts at Princeton and the Universit de Montr al, he joined Boston e e University in 1997. His research interests include time-series analysis, econometrics, and applied macroeconomics.
References
Box, G. E. P., G. M. Jenkins, and G. C. Reinsel. 2008. Time Series Analysis: Forecasting and Control. 4th ed. Hoboken, NJ: Wiley.
318
Dickey, D. A., and W. A. Fuller. 1979. Distribution of the estimators for autoregressive time series with a unit root. Journal of the American Statistical Association 74: 427431. Fuller, W. A. 1996. Introduction to Statistical Time Series. 2nd ed. New York: Wiley. Hamilton, J. D. 1994. Time Series Analysis. Princeton: Princeton University Press. MacKinnon, J. G. 1994. Approximate asymptotic distribution functions for unit-root and cointegration tests. Journal of Business and Economic Statistics 12: 167176. Newey, W. K., and K. D. West. 1987. A simple, positive semi-denite, heteroskedasticity and autocorrelation consistent covariance matrix. Econometrica 55: 703708. Phillips, P. C. B., and P. Perron. 1988. Testing for a unit root in time series regression. Biometrika 75: 335346.
Also see
[TS] tsset Declare data to be time-series data [TS] dfgls DF-GLS unit-root test [TS] dfuller Augmented DickeyFuller unit-root test [XT] xtunitroot Panel-data unit-root tests
Title
prais Prais Winsten and Cochrane Orcutt regression
Syntax
prais depvar options
Model
indepvars
if
in
, options
Description base on single-lag OLS of residuals; the default base on single-lead OLS of residuals base on autocorrelation of residuals base on autocorrelation based on DurbinWatson base on adjusted autocorrelation base on adjusted DurbinWatson use CochraneOrcutt transformation search for that minimizes SSE stop after the rst iteration suppress constant term has user-dened constant conserve memory during estimation vcetype may be ols, robust, cluster clustvar, hc2, or hc3 set condence level; default is level(95) do not report the DurbinWatson statistic control column formats, row spacing, line width, and display of omitted variables and base and empty cells control the optimization process; seldom used display legend instead of statistics
rhotype(regress) rhotype(freg) rhotype(tscorr) rhotype(dw) rhotype(theil) rhotype(nagar) corc ssesearch twostep noconstant hascons savespace
SE/Robust
vce(vcetype)
Reporting
level(#) nodw display options

Optimization
optimize options coeflegend
You must tsset your data before using prais; see [TS] tsset. indepvars may contain factor variables; see [U] 11.4.3 Factor variables. depvar and indepvars may contain time-series operators; see [U] 11.4.4 Time-series varlists. by, rolling, and statsby are allowed; see [U] 11.1.10 Prex commands. coeflegend does not appear in the dialog box. See [U] 20 Estimation and postestimation commands for more capabilities of estimation commands.
Menu
Statistics
>
Time series
>
Prais-Winsten regression
319
320
Description
prais uses the generalized least-squares method to estimate the parameters in a linear regression model in which the errors are serially correlated. Specically, the errors are assumed to follow a rst-order autoregressive process.
Options

Model
rhotype(rhomethod) selects a specic computation for the autocorrelation parameter , where rhomethod can be regress freg tscorr dw theil nagar
reg = from the residual regression t = t1 freg = from the residual regression t = t+1 tscorr = t1 / , where is the vector of residuals dw = 1 dw/2, where dw is the Durbin Watson d statistic theil = tscorr (N k)/N nagar = (dw N 2 + k 2 )/(N 2 k 2 )
The prais estimator can use any consistent estimate of to transform the equation, and each of these estimates meets that requirement. The default is regress, which produces the minimum sum-of-squares solution (ssesearch option) for the Cochrane Orcutt transformation none of these computations will produce the minimum sum-of-squares solution for the full Prais Winsten transformation. See Judge et al. (1985) for a discussion of each estimate of . corc species that the Cochrane Orcutt transformation be used to estimate the equation. With this option, the Prais Winsten transformation of the rst observation is not performed, and the rst observation is dropped when estimating the transformed equation; see Methods and formulas below. ssesearch species that a search be performed for the value of that minimizes the sum-of-squared errors of the transformed equation (Cochrane Orcutt or Prais Winsten transformation). The search method is a combination of quadratic and modied bisection searches using golden sections. twostep species that prais stop on the rst iteration after the equation is transformed by the two-step efcient estimator. Although iterating these estimators to convergence is customary, they are efcient at each step. noconstant; see [R] estimation options. hascons indicates that a user-dened constant, or a set of variables that in linear combination forms a constant, has been included in the regression. For some computational concerns, see the discussion in [R] regress. savespace species that prais attempt to save as much space as possible by retaining only those variables required for estimation. The original data are restored after estimation. This option is rarely used and should be used only if there is insufcient space to t a model without the option.
SE/Robust
vce(vcetype) species the type of standard error reported, which includes types that are derived from asymptotic theory, that are robust to some kinds of misspecication, and that allow for intragroup correlation; see [R] vce option. vce(ols), the default, uses the standard variance estimator for ordinary least squares regression.
321
prais also allows the following: vce(hc2) and vce(hc3) specify an alternative bias correction for the vce(robust) variance calculation; for more information, see [R] regress. You may specify only one of vce(hc2), vce(hc3), or vce(robust). All estimates from prais are conditional on the estimated value of . Robust variance estimates here are robust only to heteroskedasticity and are not generally robust to misspecication of the functional form or omitted variables. The estimation of the functional form is intertwined with the estimation of , and all estimates are conditional on . Thus estimates cannot be robust to misspecication of functional form. For these reasons, it is probably best to interpret vce(robust) in the spirit of Whites (1980) original paper on estimation of heteroskedastic-consistent covariance matrices.
Reporting
level(#); see [R] estimation options. nodw suppresses reporting of the Durbin Watson statistic. display options: noomitted, vsquish, noemptycells, baselevels, allbaselevels, cformat(% fmt), pformat(% fmt), sformat(% fmt), and nolstretch; see [R] estimation options.
Optimization
optimize options: iterate(#), no log, tolerance(#). iterate() species the maximum number of iterations. log/nolog species whether to show the iteration log. tolerance() species the tolerance for the coefcient vector; tolerance(1e-6) is the default. These options are seldom used. The following option is available with prais but is not shown in the dialog box: coeflegend; see [R] estimation options.
Remarks
prais ts a linear regression of depvar on indepvars that is corrected for rst-order serially correlated residuals by using the Prais Winsten (1954) transformed regression estimator, the Cochrane Orcutt (1949) transformed regression estimator, or a version of the search method suggested by Hildreth and Lu (1960). Davidson and MacKinnon (1993) provide theoretical details on the three methods (see pages 333335 for the latter two and pages 343351 for PraisWinsten). The most common autocorrelated error process is the rst-order autoregressive process. Under this assumption, the linear regression model can be written as
yt = xt + ut
where the errors satisfy
ut = ut1 + et
and the et are independently and identically distributed as N (0, 2 ). The covariance matrix of the error term e can then be written as 1 2 T 1 T 2 1 1 2 1 T 3 = 2 1 . . . . .. . . . . . . . . . T 1 T 2 T 3 1
322
The Prais Winsten estimator is a generalized least squares (GLS) estimator. The Prais Winsten method (as described in Judge et al. 1985) is derived from the AR(1) model for the error term described above. Whereas the Cochrane Orcutt method uses a lag denition and loses the rst observation in the iterative method, the Prais Winsten method preserves that rst observation. In small samples, this can be a signicant advantage.
Technical note
To t a model with autocorrelated errors, you must specify your data as time series and have (or create) a variable denoting the time at which an observation was collected. The data for the regression should be equally spaced in time.
Example 1
Say that we wish to t a time-series model of usr on idle but are concerned that the residuals may be serially correlated. We will declare the variable t to represent time by typing
. use http://www.stata-press.com/data/r12/idle . tsset t time variable: t, 1 to 30 delta: 1 unit
We can obtain CochraneOrcutt estimates by specifying the corc option:

. prais usr idle, corc Iteration 0: rho = 0.0000 Iteration 1: rho = 0.3518 (output omitted ) Iteration 13: rho = 0.5708 Cochrane-Orcutt AR(1) regression -- iterated estimates Source SS df MS Number of obs F( 1, 27) Model 40.1309584 1 40.1309584 Prob > F Residual 166.898474 27 6.18142498 R-squared Adj R-squared Total 207.029433 28 7.39390831 Root MSE usr idle _cons rho Coef. -.1254511 14.54641 .5707918 Std. Err. .0492356 4.272299 t -2.55 3.40 P>|t| 0.017 0.002 = = = = = = 29 6.49 0.0168 0.1938 0.1640 2.4862
[95% Conf. Interval] -.2264742 5.78038 -.024428 23.31245
Durbin-Watson statistic (original) 1.295766 Durbin-Watson statistic (transformed) 1.466222
The tted model is
usrt = 0.1255 idlet + 14.55 + ut
and
ut = 0.5708 ut1 + et
We can also t the model with the Prais Winsten method,

. prais usr idle Iteration 0: rho = 0.0000 Iteration 1: rho = 0.3518 (output omitted ) Iteration 14: rho = 0.5535 Prais-Winsten AR(1) regression -- iterated estimates SS df MS Source Model Residual Total usr idle _cons rho 43.0076941 169.165739 212.173433 Coef. -.1356522 15.20415 .5535476 1 28 29 43.0076941 6.04163354 7.31632528 t -2.87 3.65 P>|t| 0.008 0.001
323
Number of obs F( 1, 28) Prob > F R-squared Adj R-squared Root MSE
= = = = = =
30 7.12 0.0125 0.2027 0.1742 2.458
Std. Err. .0472195 4.160391
[95% Conf. Interval] -.2323769 6.681978 -.0389275 23.72633
where the Prais Winsten tted model is
usrt = .1357 idlet + 15.20 + ut
and
ut = .5535 ut1 + et
As the results indicate, for these data there is little difference between the Cochrane Orcutt and Prais Winsten estimators, whereas the OLS estimate of the slope parameter is substantially different.
Example 2
We have data on quarterly sales, in millions of dollars, for 5 years, and we would like to use this information to model sales for company X. First, we t a linear model by OLS and obtain the DurbinWatson statistic by using estat dwatson; see [R] regress postestimation time series.
. use http://www.stata-press.com/data/r12/qsales . regress csales isales Source SS Model Residual Total csales isales _cons 110.256901 .133302302 110.390204 Coef. .1762828 -1.454753 df 1 18 19 MS 110.256901 .007405683 5.81001072 t 122.02 -6.79 P>|t| 0.000 0.000 Number of obs F( 1, 18) Prob > F R-squared Adj R-squared Root MSE = 20 =14888.15 = 0.0000 = 0.9988 = 0.9987 = .08606
Std. Err. .0014447 .2141461
[95% Conf. Interval] .1732475 -1.904657 .1793181 -1.004849
. estat dwatson Durbin-Watson d-statistic(
2,
20) =
.7347276
324
Because the DurbinWatson statistic is far from 2 (the expected value under the null hypothesis of no serial correlation) and well below the 5% lower limit of 1.2, we conclude that the disturbances are serially correlated. (Upper and lower bounds for the d statistic can be found in most econometrics texts; for example, Harvey [1990]. The bounds have been derived for only a limited combination of regressors and observations.) To reinforce this conclusion, we use two other tests to test for serial correlation in the error distribution.
. estat bgodfrey, lags(1) Breusch-Godfrey LM test for autocorrelation lags(p) 1 chi2 7.998 df 1 Prob > chi2 0.0047
H0: no serial correlation . estat durbinalt Durbins alternative test for autocorrelation lags(p) 1 chi2 11.329 df 1 Prob > chi2 0.0008
H0: no serial correlation
estat bgodfrey reports the BreuschGodfrey Lagrange-multiplier test statistic, and estat durbinalt reports the Durbins alternative test statistic. Both tests give a small p-value and thus reject the null hypothesis of no serial correlation. These two tests are asymptotically equivalent when testing for AR(1) process. See [R] regress postestimation time series if you are not familiar with these two tests. We correct for autocorrelation with the ssesearch option of prais to search for the value of that minimizes the sum-of-squared residuals of the Cochrane Orcutt transformed equation. Normally, the default Prais Winsten transformations is used with such a small dataset, but the less-efcient Cochrane Orcutt transformation allows us to demonstrate an aspect of the estimators convergence.
. prais csales isales, corc ssesearch Iteration 1: rho = 0.8944, criterion = -.07298558 Iteration 2: rho = 0.8944, criterion = -.07298558 (output omitted ) Iteration 15: rho = 0.9588, criterion = -.07167037 Cochrane-Orcutt AR(1) regression -- SSE search estimates Source Model Residual Total csales isales _cons rho SS 2.33199178 .071670369 2.40366215 Coef. .1605233 1.738946 .9588209 df 1 17 18 MS 2.33199178 .004215904 .133536786 t 23.52 1.21 P>|t| 0.000 0.241 Number of obs F( 1, 17) Prob > F R-squared Adj R-squared Root MSE = = = = = = 19 553.14 0.0000 0.9702 0.9684 .06493
Std. Err. .0068253 1.432674
[95% Conf. Interval] .1461233 -1.283732 .1749234 4.761624
325
We noted in Options that, with the default computation of , the Cochrane Orcutt method produces an estimate of that minimizes the sum-of-squared residuals the same criterion as the ssesearch option. Given that the two methods produce the same results, why would the search method ever be preferred? It turns out that the back-and-forth iterations used by Cochrane Orcutt may have difculty converging if the value of is large. Using the same data, the Cochrane Orcutt iterative procedure requires more than 350 iterations to converge, and a higher tolerance must be specied to prevent premature convergence:
. prais csales isales, corc tol(1e-9) iterate(500) Iteration 0: rho = 0.0000 Iteration 1: rho = 0.6312 Iteration 2: rho = 0.6866 (output omitted ) Iteration 377: rho = 0.9588 Iteration 378: rho = 0.9588 Iteration 379: rho = 0.9588 Cochrane-Orcutt AR(1) regression -- iterated estimates Source Model Residual Total csales isales _cons rho SS 2.33199171 .071670369 2.40366208 Coef. .1605233 1.738946 .9588209 df 1 17 18 MS 2.33199171 .004215904 .133536782 t 23.52 1.21 P>|t| 0.000 0.241 Number of obs F( 1, 17) Prob > F R-squared Adj R-squared Root MSE = = = = = = 19 553.14 0.0000 0.9702 0.9684 .06493
Std. Err. .0068253 1.432674
[95% Conf. Interval] .1461233 -1.283732 .1749234 4.761625
Once convergence is achieved, the two methods produce identical results.
326
Saved results
prais saves the following in e():
Scalars e(N) e(N gaps) e(mss) e(df m) e(rss) e(df r) e(r2) e(r2 a) e(F) e(rmse) e(ll) e(N clust) e(rho) e(dw) e(dw 0) e(tol) e(max ic) e(ic) Macros e(cmd) e(cmdline) e(depvar) e(title) e(clustvar) e(cons) e(method) e(tranmeth) e(rhotype) e(vce) e(vcetype) e(properties) e(predict) e(marginsok) e(asbalanced) e(asobserved) Matrices e(b) e(V) e(V modelbased) Functions e(sample) number of observations number of gaps model sum of squares model degrees of freedom residual sum of squares residual degrees of freedom
R2
adjusted R2 F statistic root mean squared error log likelihood number of clusters autocorrelation parameter DurbinWatson d statistic for transformed regression DurbinWatson d statistic of untransformed regression target tolerance maximum number of iterations number of iterations prais command as typed name of dependent variable title in estimation output name of cluster variable noconstant or not reported twostep, iterated, or SSE search corc or prais method specied in rhotype() option vcetype specied in vce() title used to label Std. Err. b V program used to implement predict predictions allowed by margins factor variables fvset as asbalanced factor variables fvset as asobserved coefcient vector variancecovariance matrix of the estimators model-based variance estimation sample

prais is implemented as an ado-le. Consider the command prais y x z . The 0th iteration is obtained by estimating a, b, and c from the standard linear regression:
yt = axt + bzt + c + ut
An estimate of the correlation in the residuals is then obtained. By default, prais uses the auxiliary regression: ut = ut1 + et This can be changed to any computation noted in the rhotype() option.
327
Next we apply a Cochrane Orcutt transformation (1) for observations t = 2, . . . , n
yt yt1 = a(xt xt1 ) + b(zt zt1 ) + c(1 ) + vt

and the transformation (1 ) for t = 1
(1)
1 2 y1 = a( 1 2 x1 ) + b( 1 2 z1 ) + c 1 2 +
1 2 v1
(1 )
Thus the differences between the Cochrane Orcutt and the Prais Winsten methods are that the latter uses (1 ) in addition to (1), whereas the former uses only (1), necessarily decreasing the sample size by one. Equations (1) and (1 ) are used to transform the data and obtain new estimates of a, b, and c. When the twostep option is specied, the estimation process stops at this point and reports these estimates. Under the default behavior of iterating to convergence, this process is repeated until the change in the estimate of is within a specied tolerance. The new estimates are used to produce tted values
yt = axt + bzt + c
and then is reestimated using, by default, the regression dened by
yt yt = (yt1 yt1 ) + ut
(2)
We then reestimate (1) by using the new estimate of and continue to iterate between (1) and (2) until the estimate of converges. Convergence is declared after iterate() iterations or when the absolute difference in the estimated correlation between two iterations is less than tol(); see [R] maximize. Sargan (1964) has shown that this process will always converge. Under the ssesearch option, a combined quadratic and bisection search using golden sections searches for the value of that minimizes the sum-of-squared residuals from the transformed equation. The transformation may be either the Cochrane Orcutt (1 only) or the Prais Winsten (1 and 1 ). All reported statistics are based on the -transformed variables, and is assumed to be estimated without error. See Judge et al. (1985) for details.
328
The Durbin Watson d statistic reported by prais and estat dwatson is

n1
(uj+1 uj )2
n j=1
d=
j=1
u2 j
where uj represents the residual of the j th observation. This command supports the Huber/White/sandwich estimator of the variance and its clustered version using vce(robust) and vce(cluster clustvar), respectively. See [P] robust, particularly Introduction and Methods and formulas. All estimates from prais are conditional on the estimated value of . Robust variance estimates here are robust only to heteroskedasticity and are not generally robust to misspecication of the functional form or omitted variables. The estimation of the functional form is intertwined with the estimation of , and all estimates are conditional on . Thus estimates cannot be robust to misspecication of functional form. For these reasons, it is probably best to interpret vce(robust) in the spirit of Whites original paper on estimation of heteroskedastic-consistent covariance matrices.
Acknowledgment
We thank Richard Dickens of the Centre for Economic Performance at the London School of Economics and Political Science for testing and assistance with an early version of this command.
Sigbert Jon Prais (1928 ) was born in Frankfurt and moved to Britain in 1934 as a refugee. After earning degrees at the universities of Birmingham and Cambridge and serving in various posts in research and industry, he settled at the National Institute of Economic and Social Research. Praiss interests extend widely across economics, including studies of the inuence of education on economic progress. Christopher Blake Winsten (19232005) was born in Welwyn Garden City, England; the son of the writer Stephen Winsten and the painter and sculptress Clare Blake. He was educated at the University of Cambridge and worked with the Cowles Commission at the University of Chicago and at the universities of Oxford, London (Imperial College) and Essex, making many contributions to economics and statistics, including the PraisWinsten transformation and joint authorship of a celebrated monograph on transportation economics. Donald Cochrane (19171983) was an Australian economist and econometrician. He was born in Melbourne and earned degrees at Melbourne and Cambridge. After wartime service in the Royal Australian Air Force, he held chairs at Melbourne and Monash, being active also in work for various international organizations and national committees. Guy Henderson Orcutt (1917 ) was born in Michigan and earned degrees in physics and economics at the University of Michigan. He worked at Harvard, the University of Wisconsin, and Yale. He has contributed to econometrics and economics in several elds, most distinctively in developing microanalytical models of economic behavior.
329
References
Chatterjee, S., and A. S. Hadi. 2006. Regression Analysis by Example. 4th ed. New York: Wiley. Cochrane, D., and G. H. Orcutt. 1949. Application of least squares regression to relationships containing auto-correlated error terms. Journal of the American Statistical Association 44: 3261. Davidson, R., and J. G. MacKinnon. 1993. Estimation and Inference in Econometrics. New York: Oxford University Press. Durbin, J., and G. S. Watson. 1950. Testing for serial correlation in least squares regression. I. Biometrika 37: 409428. . 1951. Testing for serial correlation in least squares regression. II. Biometrika 38: 159177. Hardin, J. W. 1995. sts10: PraisWinsten regression. Stata Technical Bulletin 25: 2629. Reprinted in Stata Technical Bulletin Reprints, vol. 5, pp. 234237. College Station, TX: Stata Press. Harvey, A. C. 1990. The Econometric Analysis of Time Series. 2nd ed. Cambridge, MA: MIT Press. Hildreth, C., and J. Y. Lu. 1960. Demand relations with autocorrelated disturbances. Reprinted in Agricultural Experiment Station Technical Bulletin, No. 276. East Lansing, MI: Michigan State University Press. Judge, G. G., W. E. Grifths, R. C. Hill, H. L tkepohl, and T.-C. Lee. 1985. The Theory and Practice of Econometrics. u 2nd ed. New York: Wiley. King, M. L., and D. E. A. Giles, ed. 1987. Specication Analysis in the Linear Model: Essays in Honor of Donald Cochrane. London: Routledge & Kegan Paul. Kmenta, J. 1997. Elements of Econometrics. 2nd ed. Ann Arbor: University of Michigan Press. Prais, S. J., and C. B. Winsten. 1954. Trend estimators and serial correlation. Working paper 383, Cowles Commission. http://cowles.econ.yale.edu/P/ccdp/st/s-0383.pdf. Sargan, J. D. 1964. Wages and prices in the United Kingdom: A study in econometric methodology. In Econometric Analysis for National Economic Planning, ed. P. E. Hart, G. Mills, and J. K. Whitaker, 2564. London: Butterworths. Theil, H. 1971. Principles of Econometrics. New York: Wiley. White, H. 1980. A heteroskedasticity-consistent covariance matrix estimator and a direct test for heteroskedasticity. Econometrica 48: 817838. Wooldridge, J. M. 2009. Introductory Econometrics: A Modern Approach. 4th ed. Cincinnati, OH: South-Western. Zellner, A. 1990. Guy H. Orcutt: Contributions to economic statistics. Journal of Economic Behavior and Organization 14: 4351.
Also see
[TS] prais postestimation Postestimation tools for prais [TS] tsset Declare data to be time-series data [TS] arima ARIMA, ARMAX, and other dynamic regression models [R] regress Linear regression [R] regress postestimation time series Postestimation tools for regress with time series [U] 20 Estimation and postestimation commands
Title
prais postestimation Postestimation tools for prais
Description
The following standard postestimation commands are available after prais: Command contrast estat estimates lincom linktest margins marginsplot nlcom predict predictnl pwcompare test testnl Description contrasts and ANOVA-style joint tests of estimates AIC, BIC, VCE, and estimation sample summary cataloging estimation results point estimates, standard errors, testing, and inference for linear combinations of coefcients link test for model specication marginal means, predictive margins, marginal effects, and average marginal effects graph the results from margins (prole plots, interaction plots, etc.) point estimates, standard errors, testing, and inference for nonlinear combinations of coefcients predictions, residuals, inuence statistics, and other diagnostic measures point estimates, standard errors, testing, and inference for generalized predictions pairwise comparisons of estimates Wald tests of simple and composite linear hypotheses Wald tests of nonlinear hypotheses
Syntax for predict

predict statistic
Main
type
newvar
if
in
, statistic
xb stdp residuals
Menu
Statistics
>
Postestimation
>
330
prais postestimation Postestimation tools for prais
331
Options for predict

Main
xb, the default, calculates the tted values the prediction of xj b for the specied equation. This is the linear predictor from the tted regression model; it does not apply the estimate of to prior residuals. stdp calculates the standard error of the prediction for the specied equation, that is, the standard error of the predicted expected value or mean for the observations covariate pattern. The standard error of the prediction is also referred to as the standard error of the tted value. As computed for prais, this is strictly the standard error from the variance in the estimates of the parameters of the linear model and assumes that is estimated without error. residuals calculates the residuals from the linear prediction.

Also see
[TS] prais Prais Winsten and Cochrane Orcutt regression [U] 20 Estimation and postestimation commands
Title
psdensity Parametric spectral density estimation after arima, arma, and ucm
Syntax
psdensity type newvarsd newvarf if in , options
where newvarsd is the name of the new variable that will contain the estimated spectral density and newvarf is the name of the new variable that will contain the frequencies at which the spectral density estimate is computed. options
Model
Description
pspectrum estimate the power spectrum rather than the spectral density range(a b) limit the frequency range to [a, b) estimate the spectral density from the specied stochastic cycle; only allowed cycle(#) after ucm smemory estimate the spectral density of the short-memory component of the ARFIMA process; only allowed after arfima
Menu
Statistics
>
Time series
>
Postestimation
>
Parametric spectral density
Description
psdensity estimates the spectral density of a stationary process using the parameters of a previously estimated parametric model. psdensity works after arfima and ucm, and after arima when there are no multiplicative autoregressive moving-average (ARMA) terms; see [TS] arma, [TS] ucm, and [TS] arima.
Options

Model
pspectrum causes psdensity to estimate the power spectrum rather than the spectral density. The power spectrum is equal to the spectral density times the variance of the process. range(a b) limits the frequency range. By default, the spectral density is computed over [0, ). Specifying range(a b) causes the spectral density to be computed over [a, b). We require that 0 a < b < . cycle(#) causes psdensity to estimate the spectral density from the specied stochastic cycle after ucm. By default, the spectral density from the rst stochastic cycle is estimated. cycle(#) must specify an integer that corresponds to a cycle in the model t by ucm.
332
333
smemory causes psdensity to ignore the ARFIMA fractional integration parameter. The spectral density computed is for the short-memory ARMA component of the model.
Remarks
The frequency-domain approach to time series Some ARMA examples
The frequency-domain approach to time series

A stationary process can be decomposed into random components that occur at the frequencies [0, ]. The spectral density of a stationary process describes the relative importance of these random components. psdensity uses the estimated parameters of a parametric model to estimate the spectral density of a stationary process. We need some concepts from the frequency-domain approach to time-series analysis to interpret estimated spectral densities. Here we provide a simple, intuitive explanation. More technical presentations can be found in Priestley (1981), Harvey (1989, 1993), Hamilton (1994), Fuller (1996), and Wei (2006). In the time domain, the dependent variable evolves over time because of random shocks. The autocovariances j , j {0, 1, . . . , }, of a covariance-stationary process yt specify its variance and dependence structure, and the autocorrelations j , j {1, 2, . . . , }, provide a scale-free measure of its dependence structure. The autocorrelation at lag j species whether realizations at time t and realizations at time t j are positively related, unrelated, or negatively related. In the frequency domain, the dependent variable is generated by an innite number of random components that occur at the frequencies [0, ]. The spectral density species the relative importance of these random components. The area under the spectral density in the interval (, +d) is the fraction of the variance of the process than can be attributed to the random components that occur at the frequencies in the interval (, + d). The spectral density and the autocorrelations provide the same information about the dependence structure, albeit in different domains. The spectral density can be written as a weighted average of the autocorrelations of yt , and it can be inverted to retrieve the autocorrelations as a function of the spectral density. Like autocorrelations, the spectral density is normalized by 0 , the variance of yt . Multiplying the spectral density by 0 yields the power spectrum of yt , which changes with the units of yt . A peak in the spectral density around frequency implies that the random components around make an important contribution to the variance of yt . A random variable primarily generated by low-frequency components will tend to have more runs above or below its mean than an independent and identically distributed (i.i.d.) random variable, and its plot may look smoother than the plot of the i.i.d. variable. A random variable primarily generated by high-frequency components will tend to have fewer runs above or below its mean than an i.i.d. random variable, and its plot may look more jagged than the plot of the i.i.d. variable.
Technical note
A more formal specication of the spectral density allows us to be more specic about how the spectral density species the relative importance of the random components.
334
If yt is a covariance-stationary process with absolutely summable autocovariances, its spectrum is given by
gy () =
1 1 0 + 2
k cos(k)
k=1
(1)
where gy () is the spectrum of yt at frequency and k is the k th autocovariance of yt . Taking the inverse Fourier transform of each side of (1) yields
k =
gy ()eik d
(2)
where i is the imaginary number i = Evaluating (2) at k = 0 yields
1.
0 =
gy ()d
which means that the variance of yt can be decomposed in terms of the spectrum gy (). In particular, gy ()d is the contribution to the variance of yt attributable to the random components in the interval (, + d). The spectrum depends on the units in which yt is measured, because it depends on the 0 . Dividing both sides of (1) by 0 gives us the scale-free spectral density of yt :
1 1 + fy () = 2
By construction,
k cos(k)
k=1
fy ()d = 1
so fy ()d is the fraction of the variance of yt attributable to the random components in the interval (, + d).
Some ARMA examples

In this section, we estimate and interpret the spectral densities implied by the estimated ARMA parameters. The examples illustrate some of the essential relationships between covariance-stationary processes, the parameters of ARMA models, and the spectral densities implied by the ARMA-model parameters. See [TS] ucm for a discussion of unobserved-components models and the stochastic-cycle model derived by Harvey (1989) for stationary processes. The stochastic-cycle model has a different parameterization of the spectral density, and it tends to produce spectral densities that look more like probability densities than ARMA models. See Remarks in [TS] ucm for an introduction to these models, some examples, and some comparisons between the stochastic-cycle model and ARMA models.
335
Example 1
Lets consider the changes in the number of manufacturing employees in the United States, which we plot below.
. use http://www.stata-press.com/data/r12/manemp2 (FRED data: Number of manufacturing employees in U.S.) . tsline D.manemp, yline(-0.206)
5 1950m1
Change in number of mfg. employees, D 0 5
1960m1
1970m1
1980m1 Month
1990m1
2000m1
2010m1
We added a horizontal line at the sample mean of 0.0206 to highlight that there appear to be more runs above or below the mean than we would expect in data generated by an i.i.d. process. As a rst pass at modeling this dependence, we use arima to estimate the parameters of a rst-order autoregressive (AR(1)) model. Formally, the AR(1) model is given by
yt = yt1 +
where yt is the dependent variable, is the autoregressive coefcient, and t is an i.i.d. error term. See [TS] arima for an introduction to ARMA modeling and the arima command.
336
. arima D.manemp, ar(1) noconstant (setting optimization to BHHH) Iteration 0: log likelihood = -870.64844 Iteration 1: log likelihood = -870.64794 Iteration 2: log likelihood = -870.64789 Iteration 3: log likelihood = -870.64787 Iteration 4: log likelihood = -870.64786 (switching optimization to BFGS) Iteration 5: log likelihood = -870.64786 Iteration 6: log likelihood = -870.64786 ARIMA regression Sample: 1950m2 - 2011m2 Log likelihood = -870.6479 OPG Std. Err.
= = =
733 730.51 0.0000
D.manemp ARMA ar L1. /sigma
Coef.
P>|z|
.5179561 .7934554
.0191638 .0080636
27.03 98.40
0.000 0.000
.4803959 .777651
.5555164 .8092598
The statistically signicant estimate of 0.518 for the autoregressive coefcient indicates that there is an important amount of positive autocorrelation in this series. The spectral density of a covariance-stationary process is symmetric around 0. Following convention, psdensity estimates the spectral density over the interval [0, ) at the points given in Methods and formulas. Now we use psdensity to estimate the spectral density of the process implied by the estimated
ARMA parameters. We specify the names of two new variables in the call to psdensity. The rst
new variable will contain the estimated spectral density. The second new variable will contain the frequencies at which the spectral density is estimated.
. psdensity psden1 omega . line psden1 omega
337
0 0
.1
ARMA spectral density .2 .3
.4
.5
1 Frequency
The above graph is typical of a spectral density of an AR(1) process with a positive coefcient. The curve is highest at frequency 0, and it tapers off toward zero or a positive asymptote. The estimated spectral density is telling us that the low-frequency random components are the most important random components of an AR(1) process with a positive autoregressive coefcient. The closer the is to 1, the more important are the low-frequency components relative to the high-frequency components. To illustrate this point, we plot the spectral densities implied by AR(1) models with = 0.1 and = 0.9.
100 1.2 Spectral Density (=.1) 1 1.1 .9 .8 0 1 2 Frequency 3 0 0 20 Spectral Density (=.9) 40 60 80
2 Frequency
As gets closer to 1, the plot of the spectral density gets closer to being a spike at frequency 0, implying that only the lowest-frequency components are important.
338
Example 2
Now lets consider a dataset for which the estimated coefcient from an AR(1) model is negative. Below we plot the changes in initial claims for unemployment insurance in the United States.
. use http://www.stata-press.com/data/r12/icsa1, clear . tsline D.icsa, yline(0.08)
200 200 01jan1970 Change in initial claims, D 100 0 100
01jan1980
01jan1990 Date
01jan2000
01jan2010
The plot looks a little more jagged than we would expect from an i.i.d. process, but it is hard to tell. Below we estimate the AR(1) coefcient.
. arima D.icsa, ar(1) noconstant (setting optimization to BHHH) Iteration 0: log likelihood = -9934.0659 Iteration 1: log likelihood = -9934.0657 Iteration 2: log likelihood = -9934.0657 ARIMA regression Sample: 14jan1967 - 19feb2011 Log likelihood = -9934.066 OPG Std. Err.
= = =
2302 666.06 0.0000
D.icsa ARMA ar L1. /sigma
Coef.
P>|z|
-.2756024 18.10988
.0106789 .1176556
-25.81 153.92
0.000 0.000
-.2965326 17.87928
-.2546722 18.34048
The estimated coefcient is negative and statistically signicant.
339
The spectral density implied by the estimated parameters is

. psdensity psden2 omega2 . line psden2 omega2
.1 0
ARMA spectral density .15 .2 .25
.3
1 Frequency
The above graph is typical of a spectral density of an AR(1) process with a negative coefcient. The curve is lowest at frequency 0, and it monotonically increases to its highest point, which occurs when the frequency is . When the coefcient of an AR(1) model is negative, the high-frequency random components are the most important random components of the process. The closer the is to 1, the more important are the high-frequency components relative to the low-frequency components. To illustrate this point, we plot the spectral densities implied by AR(1) models with = 0.1, and = 0.9.
100 1.2 Spectral Density (=.1) 1 1.1 .9 .8 0 1 2 Frequency 3 0 0 20 Spectral Density (=.9) 40 60 80
2 Frequency
As gets closer to 1, the plot of the spectral density shifts toward becoming a spike at frequency , implying that only the highest-frequency components are important.
For examples of psdensity after arfima and ucm, see [TS] arma and [TS] ucm.
340

psdensity is implemented as an ado-le. Methods and formulas are presented under the following headings:
Introduction Spectral density after arima or arma Spectral density after ucm
Introduction
The spectral density f () is estimated at the values {1 , 2 , . . . , N } using one of the formulas given below. Given a sample of size N , after accounting for any if or in restrictions, the N values of are given by i = (i 1)/(N 1) for i {1, 2, . . . , N }. In the rare case in which the dataset in memory has insufcient observations for the desired resolution of the estimated spectral density, you may use tsappend or set obs (see [TS] tsappend or [D] obs) to increase the number of observations in the current dataset. You may use an if restriction or an in restriction to restrict the observations to handle panel data or to compute the estimates for a subset of the observations.
Spectral density after arima or arma

Let k and k denote the p autoregressive and q moving-average parameters of an ARMA model, respectively. Box, Jenkins, and Reinsel (2008) show that the spectral density implied by the ARMA parameters is
fARMA (; , , 2 , 0 ) =
2 |1 + 1 ei + 2 ei2 + + q eiq | 20 |1 1 ei 2 ei2 p eip |2
where [0, ] and 2 is the variance of the idiosyncratic error and 0 is the variance of the dependent variable. We estimate 0 using the arima parameter estimates. The spectral density for the ARFIMA model is
fARFIMA (; , , d, 2 , 0 ) = |1 ei |2d fARMA (; , , 2 ) where d, 1/2 < d < 1/2, is the fractional integration parameter. The spectral density goes to innity as the frequency approaches 0 for 0 < d < 1/2, and it is zero at frequency 0 for 1/2 < d < 0.
The smemory option causes psdensity to perform the estimation with d = 0, which is equivalent to estimating the spectral density of the fractionally differenced series. The power spectrum omits scaling by 0 .
Spectral density after ucm

The spectral density of an order-k stochastic cycle with frequency and damping is (Trimbur 2006)
2 f (; , , ) =
(1 2 )2k1
2 k1 k1 2 2i i=0 i k j=0
k i
k j+i k i=0 (1) j
j+i cos (j i) cos (j i)

k
2 {1 + 42 cos2 + 4 4(1 + 2 ) cos cos + 22 cos 2}

2 where is the variance of the cycle error term.
341
The variance of the cycle is

2
k1 k1 2 2i i=0 i (1 2 )2k1
2 and the power spectrum omits scaling by .
References
Box, G. E. P., G. M. Jenkins, and G. C. Reinsel. 2008. Time Series Analysis: Forecasting and Control. 4th ed. Hoboken, NJ: Wiley. Fuller, W. A. 1996. Introduction to Statistical Time Series. 2nd ed. New York: Wiley. Hamilton, J. D. 1994. Time Series Analysis. Princeton: Princeton University Press. Harvey, A. C. 1989. Forecasting, Structural Time Series Models and the Kalman Filter. Cambridge: Cambridge University Press. . 1993. Time Series Models. 2nd ed. Cambridge, MA: MIT Press. Priestley, M. B. 1981. Spectral Analysis and Time Series. London: Academic Press. Trimbur, T. M. 2006. Properties of higher order stochastic cycles. Journal of Time Series Analysis 27: 117. Wei, W. W. S. 2006. Time Series Analysis: Univariate and Multivariate Methods. 2nd ed. Boston: Pearson.
Also see
[TS] arma Autoregressive fractionally integrated moving-average models [TS] arima ARIMA, ARMAX, and other dynamic regression models [TS] ucm Unobserved-components model
Title
rolling Rolling-window and recursive estimation
Syntax
rolling options
Main
exp list
if
in
, options Description
: command
window(#) recursive rrecursive clear saving(lename, . . .) stepsize(#) start(time constant) end(time constant) keep(varname , start )
number of consecutive data points in each sample use recursive samples use reverse recursive samples replace data in memory with results save results to lename; save statistics in double precision; save results to lename every # replications number of periods to advance window period at which rolling is to start period at which rolling is to end save varname along with results; optionally, use value at left edge of window suppress replication dots display any output from command trace commands execution identify invalid results
Options
Reporting
nodots noisily trace

Advanced
reject(exp)
window(#) is required. You must tsset your data before using rolling; see [TS] tsset. aweights are allowed in command if command accepts aweights; see [U] 11.1.6 weight.
exp list contains
elist contains eexp is specname is
(name: elist) elist eexp newvar = (exp) (exp) specname [eqno]specname b b[] se se[]
342
343
eqno is
## name , which indicate optional arguments.
exp is a standard Stata expression; see [U] 13 Functions and expressions. Distinguish between [ ], which are to be typed, and
Menu
Statistics
>
Time series
>
Rolling-window and recursive estimation
Description
rolling is a moving sampler that collects statistics from command after executing command on subsets of the data in memory. Typing . rolling exp list, window(50) clear: command executes command on sample windows of span 50. That is, rolling will rst execute command by using periods 150 of the dataset, and then using periods 251, 352, and so on. rolling can also perform recursive and reverse recursive analyses, in which the starting or ending period is held xed and the window size grows. command denes the statistical command to be executed. Most Stata commands and user-written programs can be used with rolling, as long as they follow standard Stata syntax and allow the if qualier; see [U] 11 Language syntax. The by prex cannot be part of command. exp list species the statistics to be collected from the execution of command. If no expressions are given, exp list assumes a default of b if command stores results in e() and of all the scalars if command stores results in r() and not in e(). Otherwise, not specifying an expression in exp list is an error.
Options

Main
window(#) denes the window size used each time command is executed. The window size refers to calendar periods, not the number of observations. If there are missing data (for example, because of weekends), the actual number of observations used by command may be less than window(#). window(#) is required. recursive species that a recursive analysis be done. The starting period is held xed, the ending period advances, and the window size grows. rrecursive species that a reverse recursive analysis be done. Here the ending period is held xed, the starting period advances, and the window size shrinks.
Options
clear species that Stata replace the data in memory with the collected statistics even though the current data in memory have not been saved to disk. saving( lename , suboptions ) creates a Stata data le (.dta le) consisting of (for each statistic in exp list) a variable containing the window replicates.
344
double species that the results for each replication be stored as doubles, meaning 8-byte reals. By default, they are stored as floats, meaning 4-byte reals. every(#) species that results be written to disk every #th replication. every() should be specied in conjunction only with saving() when command takes a long time for each replication. This will allow recovery of partial results should your computer crash. See [P] postle. stepsize(#) species the number of periods the window is to be advanced each time command is executed. start(time constant) species the date on which rolling is to start. start() may be specied as an integer or as a date literal. end(time constant) species the date on which rolling is to end. end() may be specied as an integer or as a date literal. keep(varname , start ) species a variable to be posted along with the results. The value posted is the value that corresponds to the right edge of the window. Specifying the start() option requests that the value corresponding to the left edge of the window be posted instead. This option is often used to record calendar dates.
Reporting
nodots suppresses display of the replication dot for each window on which command is executed. By default, one dot character is printed for each window. A red x is printed if command returns with an error or if any value in exp list is missing. noisily causes the output of command to be displayed for each window on which command is executed. This option implies the nodots option. trace causes a trace of the execution of command to be displayed. This option implies the noisily and nodots options.
Advanced
reject(exp) identies an expression that indicates when results should be rejected. When exp is true, the saved statistics are set to missing values.
Remarks
rolling executes a command on each of a series of windows of observations and stores the results. rolling can perform what are commonly called rolling regressions, recursive regressions, and reverse recursive regressions. However, rolling is not limited to just linear regression analysis: any command that saves results in e() or r() can be used with rolling. Suppose that you have data collected at 100 consecutive points in time, numbered 1100, and you wish to perform a rolling regression with a window size of 20 periods. Typing
. rolling _b, window(20) clear: regress depvar indepvar
causes Stata to regress depvar on indepvar using periods 120, store the regression coefcients ( b), run the regression using periods 221, and so on, nishing with a regression using periods 81100 (the last 20 periods). The stepsize() option species how far ahead the window is moved each time. For example, if you specify step(2), then command is executed on periods 120, and then 322, 524, etc. By default, rolling replaces the dataset in memory with the computed statistics unless the saving() option is specied, in which case the computed statistics are stored in the lename specied. If the
345
dataset in memory has been changed since it was last saved and you do not specify saving(), you must use clear. rolling can also perform recursive and reverse recursive analyses. In a recursive analysis, the starting date is held xed, and the window size grows as the ending date is advanced. In a reverse recursive analysis, the ending date is held xed, and the window size shrinks as the starting date is advanced.
Example 1
We have data on the daily returns to IBM stock (ibm), the S&P 500 (spx), and short-term interest rates (irx), and we want to create a series containing the beta of IBM by using the previous 200 trading days at each date. We will also record the standard errors, so that we can obtain 95% condence intervals for the betas. See, for example, Stock and Watson (2011, 118) for more information on estimating betas. We type
. use http://www.stata-press.com/data/r12/ibm (Source: Yahoo! Finance) . tsset t time variable: t, 1 to 494 delta: 1 unit . generate ibmadj = ibm - irx (1 missing value generated) . generate spxadj = spx - irx (1 missing value generated) . rolling _b _se, window(200) saving(betas, replace) > spxadj (running regress on estimation sample) (note: file betas.dta not found) Rolling replications (295) 1 2 3 4 5 .................................................. .................................................. .................................................. .................................................. .................................................. ............................................. file betas.dta saved
keep(date): regress ibmadj
50 100 150 200 250
Our dataset has both a time variable t that runs consecutively and a date variable date that measures the calendar date and therefore has gaps at weekends and holidays. Had we used the date variable as our time variable, rolling would have used windows consisting of 200 calendar days instead of 200 trading days, and each window would not have exactly 200 observations. We used the keep(date) option so that we could refer to the date variable when working with the results dataset. We can list a portion of the dataset created by rolling to see what it contains:
. use betas, clear (rolling: regress) . sort date . list in 1/3, abbrev(10) start 1. 2. 3. 1 2 3 end 200 201 202 date 16oct2003 17oct2003 20oct2003 _b_spxadj 1.043422 1.039024 1.038371 _b_cons -.0181504 -.0126876 -.0235616 _se_spxadj .0658531 .0656893 .0654591 _se_cons .0748295 .074609 .0743851
346
The variables start and end indicate the rst and last observations used each time that rolling called regress, and the date variable contains the calendar date corresponding the period represented by end. The remaining variables are the estimated coefcients and standard errors from the regression. In our example , b spxadj contains the estimated betas, and b cons contains the estimated alphas. The variables se spxadj and se cons have the corresponding standard errors. Finally, we compute the condence intervals for the betas and examine how they have changed over time:
. . . > generate lower = _b_spxadj - 1.96*_se_spxadj generate upper = _b_spxadj + 1.96*_se_spxadj twoway (line _b_spxadj date) (rline lower upper date) if date>=td(1oct2003), ytitle("Beta")
Beta .6 01oct2003 .8
1.2
01jan2004
01apr2004 01jul2004 date _b[spxadj]
01oct2004
01jan2005
lower/upper
As 2004 progressed, IBMs stock returns were less inuenced by returns in the broader market. Beginning in June of 2004, IBMs beta became signicantly different from unity at the 95% condence level, as indicated by the fact that the condence interval does not contain one from then onward.
In addition to rolling-window analyses, rolling can also perform recursive ones. Suppose again that you have data collected at 100 consecutive points in time, and now you type
. rolling _b, window(20) recursive clear: regress depvar indepvar
Stata will rst regress depvar on indepvar by using observations 120, store the coefcients, run the regression using observations 121, observations 122, and so on, nishing with a regression using all 100 observations. Unlike a rolling regression, in which case the number of observations is held constant and the starting and ending points are shifted, a recursive regression holds the starting point xed and increases the number of observations. Recursive analyses are often used in forecasting situations. As time goes by, more information becomes available that can be used in making forecasts. See Kmenta (1997, 423424).
347
Example 2
Using the same dataset, we type
. use http://www.stata-press.com/data/r12/ibm, clear (Source: Yahoo! Finance) . tsset t time variable: t, 1 to 494 delta: 1 unit . generate ibmadj = ibm - irx (1 missing value generated) . generate spxadj = spx - irx (1 missing value generated) . rolling _b _se, recursive window(200) clear: regress ibmadj spxadj (output omitted ) . list in 1/3, abbrev(10) start 1. 2. 3. 1 1 1 end 200 201 202 _b_spxadj 1.043422 1.039024 1.037687 _b_cons -.0181504 -.0126876 -.016475 _se_spxadj .0658531 .0656893 .0655896 _se_cons .0748295 .074609 .0743481
Here the starting period remains xed and the window grows larger.
In a reverse recursive analysis, the ending date is held xed, and the window size becomes smaller as the starting date is advanced. For example, with a dataset that has observations numbered 1100, typing
. rolling _b, window(20) reverse recursive clear: regress depvar indepvar
creates a dataset in which the rst observation has the results based on periods 1100, the second observation has the results based on 2100, the third having 3100, and so on, up to the last observation having results based on periods 81100 (the last 20 observations).
Example 3
Using the data on stock returns, we want to build a model in which we predict todays IBM stock return on the basis of yesterdays returns on IBM and the S&P 500. That is, letting it and st denote the returns to IBM and the S&P 500 on date t, we want to t the regression model
it = 0 + 1 it1 + 2 st1 +
where
t
is a regression error term, and then compute
it+1 = 0 + 1 it + 2 st
We will use recursive regression because we suspect that the more data we have to t the regression model, the better the model will predict returns. We will use at least 20 periods in tting the regression.
. use http://www.stata-press.com/data/r12/ibm, clear (Source: Yahoo! Finance) . tsset t time variable: t, 1 to 494 delta: 1 unit
348
One alternative would be to use rolling with the recursive option to t the regressions, collect the coefcients, and then compute the predicted values afterward. However, we will instead write a short program that computes the forecasts automatically and then use rolling, recursive on that program. The program must accept an if expression so that rolling can indicate to the program which observations are to be used. Our program is
program myforecast, rclass syntax [if] regress ibm L.ibm L.spx if // Find last time period of estimation sample and // make forecast for period just after that summ t if e(sample) local last = r(max) local fcast = _b[_cons] + _b[L.ibm]*ibm[last] + /// _b[L.spx]*spx[last] return scalar forecast = fcast // Next periods actual return // Will return missing value for final period return scalar actual = ibm[last+1] end
Now we call rolling:

. rolling actual=r(actual) forecast=r(forecast), recursive window(20): myforecast (output omitted ) . corr actual forecast (obs=474) actual forecast actual forecast 1.0000 -0.0957
1.0000
Our model does not work too wellthe correlation between actual returns and our forecasts is negative.
Saved results
rolling sets no r- or e-class macros. The results from the command used with rolling, depending on the last window of data used, are available after rolling has nished.

rolling is implemented as an ado-le.
Acknowledgment
We thank Christopher F. Baum for an earlier rolling regression command.
References
Kmenta, J. 1997. Elements of Econometrics. 2nd ed. Ann Arbor: University of Michigan Press. Stock, J. H., and M. W. Watson. 2011. Introduction to Econometrics. 3rd ed. Boston: AddisonWesley.
349
Also see
[D] statsby Collect statistics for a command across a by list [R] saved results Saved results
Title
sspace State-space models
Syntax
Covariance-form syntax sspace state ceq if in state ceq . . . state ceq obs ceq obs ceq . . . obs ceq
, options
where each state ceq is of the form (statevar lagged statevars and each obs ceq is of the form (depvar statevars Error-form syntax sspace state efeq if in state efeq . . . state efeq obs efeq obs efeq . . . obs efeq indepvars , noerror noconstant ) indepvars , state noerror noconstant )
, options
where each state efeq is of the form (statevar lagged statevars and each obs efeq is of the form (depvar statevars indepvars obs errors , noconstant ) indepvars state errors , state noconstant )
statevar is the name of an unobserved state, not a variable. If there happens to be a variable of the same name, the variable is ignored and plays no role in the estimation. lagged statevars is a list of lagged statevars. Only rst lags are allowed. state errors is a list of state-equation errors that enter a state equation. Each state error has the form e.statevar, where statevar is the name of a state in the model. obs errors is a list of observation-equation errors that enter an equation for an observed variable. Each error has the form e.depvar, where depvar is an observed dependent variable in the model. equation-level options
Model
Description species that the equation is a state equation species that there is no error term in the equation suppresses the constant term from the equation
state noerror noconstant
350
351
options
Model
Description species the covariance structure for the errors in the state variables species the covariance structure for the errors in the observed dependent variables apply specied linear constraints vcetype may be oim or robust set condence level; default is level(95) do not display constraints control column formats, row spacing, and display of omitted variables and base and empty cells control the maximization process; seldom used specify the method for calculating the log likelihood; seldom used display legend instead of statistics
covstate(covform) covobserved(covform) constraints(constraints)

SE/Robust
vce(vcetype)
Reporting

Maximization
maximize options
Advanced
method(method) coeflegend
covform identity dscalar diagonal unstructured
Description identity matrix; the default for error-form syntax diagonal scalar matrix diagonal matrix; the default for covariance-form syntax symmetric, positive-denite matrix; not allowed with error-form syntax
method hybrid dejong kdiffuse
Description use the stationary Kalman lter and the De Jong diffuse Kalman lter; the default use the stationary De Jong Kalman lter and the De Jong diffuse Kalman lter use the stationary Kalman lter and the nonstationary large- diffuse Kalman lter; seldom used
You must tsset your data before using sspace; see [TS] tsset. indepvars may contain factor variables; see [U] 11.4.3 Factor variables. indepvars and depvar may contain time-series operators; see [U] 11.4.4 Time-series varlists. by, statsby, and rolling are allowed; see [U] 11.1.10 Prex commands. coeflegend does not appear in the dialog box. See [U] 20 Estimation and postestimation commands for more capabilities of estimation commands.
352
Menu
Statistics
>
State-space models
Description
sspace estimates the parameters of linear state-space models by maximum likelihood. Linear state-space models are very exible and many linear time-series models can be written as linear state-space models. sspace uses two forms of the Kalman lter to recursively obtain conditional means and variances of both the unobserved states and the measured dependent variables that are used to compute the likelihood. The covariance-form syntax and the error-form syntax of sspace reect the two different forms in which researchers specify state-space models. Choose the syntax that is easier for you; the two forms are isomorphic.
Options
Equation-level options
Model
state species that the equation is a state equation. noerror species that there is no error term in the equation. noerror may not be specied in the error-form syntax. noconstant suppresses the constant term from the equation. Options
Model
covstate(covform) species the covariance structure for the state errors. covstate(identity) species a covariance matrix equal to an identity matrix, and it is the default for the error-form syntax.
2 covstate(dscalar) species a covariance matrix equal to state times an identity matrix.
covstate(diagonal) species a diagonal covariance matrix, and it is the default for the covarianceform syntax. covstate(unstructured) species a symmetric, positive-denite covariance matrix with parameters for all variances and covariances. covstate(unstructured) may not be specied with the error-form syntax. covobserved(covform) species the covariance structure for the observation errors. covobserved(identity) species a covariance matrix equal to an identity matrix, and it is the default for the error-form syntax.
2 covobserved(dscalar) species a covariance matrix equal to observed times an identity matrix.
covobserved(diagonal) species a diagonal covariance matrix, and it is the default for the covariance-form syntax.
353
covobserved(unstructured) species a symmetric, positive-denite covariance matrix with parameters for all variances and covariances. covobserved(unstructured) may not be specied with the error-form syntax. constraints(constraints); see [R] estimation options.
SE/Robust
vce(vcetype) species the estimator for the variancecovariance matrix of the estimator. vce(oim), the default, causes sspace to use the observed information matrix estimator. vce(robust) causes sspace to use the Huber/White/sandwich estimator.
Reporting
level(#), nocnsreport; see [R] estimation options. display options: noomitted, vsquish, noemptycells, baselevels, allbaselevels, cformat(% fmt), pformat(% fmt), and sformat(% fmt); see [R] estimation options.
Maximization
maximize options: difficult, technique(algorithm spec), iterate(#), no log, trace, gradient, showstep, hessian, showtolerance, tolerance(#), ltolerance(#), nrtolerance(#), and from(matname); see [R] maximize for all options except from(), and see below for information on from(). These options are seldom used. from(matname) species initial values for the maximization process. from(b0) causes sspace to begin the maximization algorithm with the values in b0. b0 must be a row vector; the number of columns must equal the number of parameters in the model; and the values in b0 must be in the same order as the parameters in e(b).
Advanced
method(method) species how to compute the log likelihood. This option is seldom used. method(hybrid), the default, uses the Kalman lter with model-based initial values for the states when the model is stationary and uses the De Jong (1988, 1991) diffuse Kalman lter when the model is nonstationary. method(dejong) uses the Kalman lter with the De Jong (1988) method for estimating the initial values for the states when the model is stationary and uses the De Jong (1988, 1991) diffuse Kalman lter when the model is nonstationary. method(kdiffuse) is a seldom used method that uses the Kalman lter with model-based initial values for the states when the model is stationary and uses the large- diffuse Kalman lter when the model is nonstationary. The following option is available with sspace but is not shown in the dialog box: coeflegend; see [R] estimation options.
Remarks
An introduction to state-space models Some stationary state-space models Some nonstationary state-space models
354
An introduction to state-space models

Many linear time-series models can be written as linear state-space models, including vector autoregressive moving-average (VARMA) models, dynamic-factor (DF) models, and structural timeseries (STS) models. The solutions to some stochastic dynamic-programming problems can also be written in the form of linear state-space models. We can estimate the parameters of a linear statespace model by maximum likelihood (ML). The Kalman lter or a diffuse Kalman lter is used to write the likelihood function in prediction-error form, assuming normally distributed errors. The quasimaximum likelihood (QML) estimator, which drops the normality assumption, is consistent and asymptotically normal when the model is stationary. Chang, Miller, and Park (2009) establish consistency and asymptotic normality of the QML estimator for a class of nonstationary state-space models. The QML estimator differs from the ML estimator only in the VCE; specify the vce(robust) option to obtain the QML estimator. Hamilton (1994a, 1994b), Harvey (1989), and Brockwell and Davis (1991) provide good introductions to state-space models. Anderson and Moores (1979) text is a classic reference; they produced many results used subsequently. Caines (1988) and Hannan and Deistler (1988) provide excellent, more advanced, treatments. sspace estimates linear state-space models with time-invariant coefcient matrices, which cover the models listed above and many others. sspace can estimate parameters from state-space models of the form zt = Azt1 + Bxt + C t
yt = Dzt + Fwt + Gt
where
zt is an m 1 vector of unobserved state variables; xt is a kx 1 vector of exogenous variables;

t
is a q 1 vector of state-error terms, (q m);
yt is an n 1 vector of observed endogenous variables; wt is a kw 1 vector of exogenous variables;

t is an r 1 vector of observation-error terms, (r n); and
A, B, C, D, F, and G are parameter matrices.

The equations for zt are known as the state equations, and the equations for yt are known as the observation equations. The error terms are assumed to be zero mean, normally distributed, serially uncorrelated, and uncorrelated with each other;
t
N (0, Q) = 0 for all s = t
t N (0, R)
E[
t s]
E[ t s ] = 0 for all s and t

The state-space form is used to derive the log likelihood of the observed endogenous variables conditional on their own past and any exogenous variables. When the model is stationary, a method for recursively predicting the current values of the states and the endogenous variables, known as the Kalman lter, is used to obtain the prediction error form of the log-likelihood function. When the model is nonstationary, a diffuse Kalman lter is used. How the Kalman lter and the diffuse
355
Kalman lter initialize their recursive computations depends on the method() option; see Methods and formulas. The linear state-space models with time-invariant coefcient matrices dened above can be specied in the covariance-form syntax and the error-form syntax. The covariance-form syntax requires that C and G be selection matrices, but places no restrictions on Q or R. In contrast, the error-form syntax places no restrictions C or G, but requires that Q and R be either diagonal, diagonal-scalar, or identity matrices. Some models are more easily specied in the covariance-form syntax, while others are more easily specied in the error-form syntax. Choose the syntax that is easiest for your application.
Some stationary state-space models Example 1

Following Hamilton (1994b, 373374), we can write the rst-order autoregressive (AR(1)) model
yt = (yt1 ) +
as a state-space model with the observation equation
yt = + ut
and the state equation
ut = ut1 +
where the unobserved state is ut = yt .
Here we t this model to data on the capacity utilization rate. The variable lncaputil contains data on the natural log of the capacity utilization rate for the manufacturing sector of the U.S. economy. We treat the series as rst-difference stationary and t its rst-difference to an AR(1) process. Here we estimate the parameters of the above state-space form of the AR(1) model:
356

. use http://www.stata-press.com/data/r12/manufac (St. Louis Fed (FRED) manufacturing data) . constraint 1 [D.lncaputil]u = 1 . sspace (u L.u, state noconstant) (D.lncaputil u, noerror), constraints(1) searching for initial values .......... (setting technique to bhhh) Iteration 0: log likelihood = 1505.36 Iteration 1: log likelihood = 1512.0581 (output omitted ) Refining estimates: Iteration 0: log likelihood = 1516.44 Iteration 1: log likelihood = 1516.44 State-space model Sample: 1972m2 - 2008m12 Number of obs = 443 Wald chi2(1) = 61.73 Log likelihood = 1516.44 Prob > chi2 = 0.0000 ( 1) [D.lncaputil]u = 1 OIM Std. Err.
lncaputil u u L1. D.lncaputil u _cons Variance u
Coef.
P>|z|
.3523983
.0448539
7.86
0.000
.2644862
.4403104
1 -.0003558
(constrained) .0005781 -0.62
0.538
-.001489
.0007773
.0000622
4.18e-06
14.88
0.000
.000054
.0000704
The iteration log has three parts: the dots from the search for initial values, the log from nding the maximum, and the log from a rening step. Here is a description of the logic behind each part: 1. The quality of the initial values affect the speed and robustness of the optimization algorithm. sspace takes a few iterations in a nonlinear least-squares (NLS) algorithm to nd good initial values and reports a dot for each (NLS) iteration. 2. This iteration log is the standard method by which Stata reports the search for the maximum likelihood estimates of the parameters in a nonlinear model. 3. Some of the parameters are transformed in the maximization process that sspace reports in part 2. After a maximum candidate is found in part 2, sspace looks for a maximum in the unconstrained space, checks that the Hessian of the log-likelihood function is of full rank, and reports these iterations as the rening step. The header in the output describes the estimation sample, reports the log-likelihood function at the maximum, and gives the results of a Wald test against the null hypothesis that the coefcients on all the independent variables, state variables, and lagged state variables are zero. In this example, the null hypothesis that the coefcient on L1.u is zero is rejected at all conventional levels. The estimation table reports results for the state equations, the observation equations, and the variancecovariance parameters. The estimated autoregressive coefcient of 0.3524 indicates that there is persistence in the rst-differences of the log of the manufacturing rate. The estimated mean of the differenced series is 0.0004, which is smaller in magnitude than its standard error, indicating that there is no deterministic linear trend in the series.
357
Typing
. arima D.lncaputil, ar(1) technique(nr) (output omitted )
produces nearly identical parameter estimates and standard errors for the mean and the autoregressive parameter. Because sspace estimates the variance of the state error while arima estimates the standard deviation, calculations are required to obtain the same results. The different parameterization of the variance parameter can cause small numerical differences.
Technical note
In some situations, the second part of the iteration log terminates but the rening step never converges. Only when the rening step converges does the maximization algorithm nd interpretable estimates. If the rening step iterates without convergence, the parameters of the specied model are not identied by the data. (See Rothenberg [1971], Drukker and Wiggins [2004], and Davidson and MacKinnon [1993, sec. 5.2] for discussions of identication.)
Example 2
Following Harvey (1993, 9596), we can write a zero-mean, rst-order, autoregressive movingaverage (ARMA(1,1)) model
yt = yt1 +
as a state-space model with state equations
t1
(1)
yt t
and observation equation
1 0 0
yt1 t1
(2)
yt = ( 1 0 )
yt t
(3)
The unobserved states in this model are u1t = yt and u2t = t . We set the process mean to zero because economic theory and the previous example suggest that we should do so. Below we estimate the parameters in the state-space model by using the error-form syntax:
358

. constraint 2 [u1]L.u2 = 1 . constraint 3 [u1]e.u1 = 1 . constraint 4 [D.lncaputil]u1 = 1 . sspace (u1 L.u1 L.u2 e.u1, state noconstant) (u2 e.u1, state noconstant) > (D.lncaputil u1, noconstant), constraints(2/4) covstate(diagonal) searching for initial values ........... (setting technique to bhhh) Iteration 0: log likelihood = 1506.0947 Iteration 1: log likelihood = 1514.014 (output omitted ) Refining estimates: Iteration 0: log likelihood = 1531.255 Iteration 1: log likelihood = 1531.255 State-space model Sample: 1972m2 - 2008m12 Number of obs = 443 Wald chi2(2) = 333.84 Log likelihood = 1531.255 Prob > chi2 = 0.0000 ( 1) [u1]L.u2 = 1 ( 2) [u1]e.u1 = 1 ( 3) [D.lncaputil]u1 = 1 OIM Std. Err.
lncaputil u1 u1 L1. u2 L1. e.u1 u2 e.u1 D.lncaputil u1 Variance u1
Coef.
P>|z|
.8056815
.0522661
15.41
0.000
.7032418
.9081212
1 1
(constrained) (constrained)
-.5188453
.0701985
-7.39
0.000
-.6564317
-.3812588
(constrained)
.0000582
3.91e-06
14.88
0.000
.0000505
.0000659
The command in the above output species two state equations, one observation equation, and two options. The rst state equation denes u1t and the second denes u2t according to (2) above. The observation equation denes the process for D.lncaputil according to the one specied in (3) above. Several coefcients in (2) and (3) are set to 1, and constraints 24 place these restrictions on the model. The estimated coefcient on L.u1 in equation u1, 0.806, is the estimate of in (2), which is the autoregressive coefcient in the ARMA model in (1). The estimated coefcient on e.u1 in equation u2, 0.519, is the estimate of , which is the moving-average term in the ARMA model in (1). This example highlights a difference between the error-form syntax and the covariance-form syntax. The error-form syntax used in this example includes only explicitly included errors. In contrast, the covariance-form syntax includes an error term in each equation, unless the noerror option is specied. The default for covstate() also differs between the error-form syntax and the covarianceform syntax. Because the coefcients on the errors in the error-form syntax are frequently used to
359
estimate the standard deviation of the errors, covstate(identity) is the default for the errorform syntax. In contrast, unit variances are less common in the covariance-form syntax, for which covstate(diagonal) is the default. In this example, we specied covstate(diagonal) to estimate a nonunitary variance for the state. Typing
. arima D.lncaputil, noconstant ar(1) ma(1) technique(nr) (output omitted )
produces nearly identical results. As in the AR(1) example above, arima estimates the standard deviation of the error term, while sspace estimates the variance. Although they are theoretically equivalent, the different parameterizations give rise to small numerical differences in the other parameters.
Example 3
The variable lnhours contains data on the log of manufacturing hours, which we treat as rstdifference stationary. We have a theory in which the process driving the changes in the log utilization rate affects the changes in the log of hours, but changes in the log hours do not affect changes in the log utilization rate. In line with this theory, we estimate the parameters of a lower triangular, rst-order vector autoregressive (VAR(1)) process
lncaputilt lnhourst
1 2
0 3
lncaputilt1 lnhourst1
1t 2t
(4)
where yt = yt yt1 , t = ( 1t , 2t ) and Var( ) = . We can write this VAR(1) process as a state-space model with state equations
u1t u2t
1 2
0 3
u1(t1) u2(t1)
1t 2t
(5)
with Var( ) = and observation equations
lncaputil lnhours
u1t u2t
Below we estimate the parameters of the state-space model:
360

. constraint 5 [D.lncaputil]u1 = 1 . constraint 6 [D.lnhours]u2 = 1 . sspace (u1 L.u1, state noconstant) > (u2 L.u1 L.u2, state noconstant) > (D.lncaputil u1, noconstant noerror) > (D.lnhours u2, noconstant noerror), > constraints(5/6) covstate(unstructured) searching for initial values ........... (setting technique to bhhh) Iteration 0: log likelihood = 2993.6647 Iteration 1: log likelihood = 3088.7416 (output omitted ) Refining estimates: Iteration 0: log likelihood = 3211.7532 Iteration 1: log likelihood = 3211.7532 State-space model Sample: 1972m2 - 2008m12 Log likelihood = 3211.7532 ( 1) [D.lncaputil]u1 = 1 ( 2) [D.lnhours]u2 = 1 OIM Std. Err.
= = =
443 166.87 0.0000
Coef. u1 u1 L1. u2 u1 L1. u2 L1. D.lncaputil u1 D.lnhours u2 Variance u1 Covariance u1 u2 Variance u2 1 .1286218 .353257
P>|z|
.0448456
7.88
0.000
.2653612
.4411528
.0394742
3.26
0.001
.0512537
.2059899
-.3707083
.0434255
-8.54
0.000
-.4558208
-.2855959
(constrained)
(constrained)
.0000623
4.19e-06
14.88
0.000
.0000541
.0000705
.000026
2.67e-06
9.75
0.000
.0000208
.0000312
.0000386
2.61e-06
14.76
0.000
.0000335
.0000437
Specifying covstate(unstructured) caused sspace to estimate the off-diagonal element of . The output indicates that this parameter, cov(u2,u1): cons, is small but statistically signicant. The estimated coefcient on L.u1 in equation u1, 0.353, is the estimate of 1 in (5). The estimated coefcient on L.u1 in equation u2, 0.129, is the estimate of 2 in (5). The estimated coefcient on L.u1 in equation u2, 0.371, is the estimate of 3 in (5).
361
For the VAR(1) model in (4), the estimated autoregressive coefcient for D.lncaputil is similar to the corresponding estimate in the univariate results in example 1. The estimated effect of LD.lncaputil on D.lnhours is 0.129, the estimated autoregressive coefcient of D.lnhours is 0.371, and both are statistically signicant. These estimates can be compared with those produced by typing
. constraint 101 [D_lncaputil]LD.lnhours = 0 . var D.lncaputil D.lnhours, lags(1) noconstant constraints(101) (output omitted ) . matrix list e(Sigma) (output omitted )
The var estimates are not the same as the sspace estimates because the generalized least-squares estimator implemented in var is only asymptotically equivalent to the ML estimator implemented in sspace, but the point estimates are similar. The comparison is useful for pedagogical purposes because the var estimator is relatively simple. Some problems require constraining a covariance term to zero. If we wanted to constrain cov(u2,u1): cons to zero, we could type
. constraint 7 [cov(u2,u1)]_cons = 0 . sspace (u1 L.u1, state noconstant) > (u2 L.u1 L.u2, state noconstant) > (D.lncaputil u1, noconstant noerror) > (D.lnhours u2, noconstant noerror), > constraints(5/7) covstate(unstructured) (output omitted )
Example 4
We now extend the previous example by modeling D.lncaputil and D.lnhours as a rst-order vector autoregressive moving-average (VARMA(1,1)) process. Building on the previous examples, we allow the lag of D.lncaputil to affect D.lnhours but we do not allow the lag of D.lnhours to affect the lag of D.lncaputil. Previous univariate analysis revealed that D.lnhours is better modeled as an autoregressive process than as an ARMA(1,1) process. As a result, we estimate the parameters of
lncaputilt lnhourst
1 2
0 3
lncaputilt1 lnhourst1
1 0
0 0
1(t1) 2(t1)
1t 2t
We can write this VARMA(1,1) process as s1t 1 1 s2t = 0 0 s3t 2 0 where the states are
a state-space model with state equations 0 s1(t1) 1 0 0 s2(t1) + 1 0 1t 2t 3 s3(t1) 0 1
s1t lncaputilt s2t = 1 1t s3t lnhourst

2 1 0
and we simplify the problem by assuming that
Var
1t 2t
0 2 2
Below we estimate the parameters of this model by using sspace:
362

. constraint 7 [u1]L.u2 = 1 . constraint 8 [u1]e.u1 = 1 . constraint 9 [u3]e.u3 = 1 . constraint 10 [D.lncaputil]u1 = 1 . constraint 11 [D.lnhours]u3 = 1 . sspace (u1 L.u1 L.u2 e.u1, state noconstant) > (u2 e.u1, state noconstant) > (u3 L.u1 L.u3 e.u3, state noconstant) > (D.lncaputil u1, noconstant) > (D.lnhours u3, noconstant), > constraints(7/11) technique(nr) covstate(diagonal) searching for initial values .......... (output omitted ) Refining estimates: Iteration 0: log likelihood = 3156.0564 Iteration 1: log likelihood = 3156.0564 State-space model Sample: 1972m2 - 2008m12 Number of obs Wald chi2(4) Log likelihood = 3156.0564 Prob > chi2 ( 1) [u1]L.u2 = 1 ( 2) [u1]e.u1 = 1 ( 3) [u3]e.u3 = 1 ( 4) [D.lncaputil]u1 = 1 ( 5) [D.lnhours]u3 = 1 OIM Std. Err.
= = =
443 427.55 0.0000
Coef. u1 u1 L1. u2 L1. e.u1 u2 e.u1 u3 u1 L1. u3 L1. e.u3 D.lncaputil u1 D.lnhours u3 Variance u1 u3 1 .1734868 -.518907 .8058031
P>|z|
.0522493
15.42
0.000
.7033964
.9082098
1 1
(constrained) (constrained)
.0701848
-7.39
0.000
-.6564667
-.3813474
.0405156
4.28
0.000
.0940776
.252896
-.4809376 1
.0498574 -9.65 (constrained)
0.000
-.5786563
-.3832188
(constrained)
(constrained)
.0000582 .0000382
3.91e-06 2.56e-06
14.88 14.88
0.000 0.000
.0000505 .0000331
.0000659 .0000432
363
The estimates of the parameters in the model for D.lncaputil are similar to those in the univariate model t in example 2. The estimates of the parameters in the model for D.lnhours indicate that the lag of D.lncaputil has a positive effect on D.lnhours.
Technical note
The technique(nr) option facilitates convergence in example 4. Fitting state-space models is notoriously difcult. Convergence problems are common. Four methods for overcoming convergence problems are 1) selecting an alternate optimization algorithm by using the technique() option, 2) using alternative starting values by specifying the from() option, 3) using starting values obtained by estimating the parameters of a restricted version of the model of interest, or 4) putting the variables on the same scale.
Example 5
Stock and Watson (1989, 1991) wrote a simple macroeconomic model as a dynamic-factor model, estimated the parameters by ML, and extracted an economic indicator. In this example, we estimate the parameters of a dynamic-factor model. In [TS] sspace postestimation, we extend this example and extract an economic indicator for the differenced series. We have data on an industrial-production index, ipman; an aggregate weekly hours index, hours; and aggregate unemployment, unemp. income is real disposable income divided by 100. We rescaled real disposable income to avoid convergence problems. We postulate a latent factor that follows an AR(2) process. Each measured variable is then related to the current value of that latent variable by a parameter. The state-space form of our model is
ft ft1
1 1
ipmant 1 incomet 2 = ft + hourst 3 unempt 4

where
2 1 0 Var 2t = 0 3t 0 4t
2 0
ft1 + ft2
1t 2t 3t 4t
t 0
1t
0 2 2 0 0
0 0 2 3 0
0 0 0 2 4
The parameter estimates are
364

. use http://www.stata-press.com/data/r12/dfex (St. Louis Fed (FRED) macro data) . constraint 12 [lf]L.f = 1 . sspace (f L.f L.lf, state noconstant) > (lf L.f, state noconstant noerror) > (D.ipman f, noconstant) > (D.income f, noconstant) > (D.hours f, noconstant) > (D.unemp f, noconstant), > covstate(identity) constraints(12) searching for initial values ................ (setting technique to bhhh) Iteration 0: log likelihood = -676.30903 Iteration 1: log likelihood = -665.61104 (output omitted ) Refining estimates: Iteration 0: log likelihood = -662.09507 Iteration 1: log likelihood = -662.09507 State-space model Sample: 1972m2 - 2008m11 Log likelihood = -662.09507 ( 1) [lf]L.f = 1 OIM Std. Err. Number of obs Wald chi2(6) Prob > chi2 = = = 442 751.95 0.0000
Coef. f f L1. lf L1. lf f L1. D.ipman f D.income f D.hours f D.unemp f Variance D.ipman D.income D.hours D.unemp -.0676016 .2177469 .0746338 .3502249 1 .2651932
P>|z|
.0568663
4.66
0.000
.1537372
.3766491
.4820398
.0624635
7.72
0.000
.3596136
.604466
(constrained)
.0287389
12.19
0.000
.2938976
.4065522
.0217319
3.43
0.001
.0320401
.1172276
.0186769
11.66
0.000
.1811407
.254353
.0071022
-9.52
0.000
-.0815217
-.0536816
.1383158 .2773808 .0911446 .0237232
.0167086 .0188302 .0080847 .0017932
8.28 14.73 11.27 13.23
0.000 0.000 0.000 0.000
.1055675 .2404743 .0752988 .0202086
.1710641 .3142873 .1069903 .0272378
The output indicates that the unobserved factor is quite persistent and that it is a signicant predictor for each of the observed variables.
365
These models are frequently used to forecast the dependent variables and to estimate the unobserved factors. We present some illustrative examples in [TS] sspace postestimation. The dfactor command estimates the parameters of dynamic-factor models; see [TS] dfactor.
Some nonstationary state-space models Example 6

Harvey (1989) advocates the use of STS models. These models parameterize the trends and seasonal components of a set of time series. The simplest STS model is the local-level model, which is given by
yt = t +
where
t = t1 + t
The model is called a local-level model because the level of the series is modeled as a random walk plus an idiosyncratic noise term. (The model is also known as the random-walk-plus-noise model.) The local-level model is nonstationary because of the random-walk component. When the variance of the idiosyncratic-disturbance t is zero and the variance of the level-disturbance t is not zero, the local-level model reduces to a random walk. When the variance of the level-disturbance t is zero and the variance of the idiosyncratic-disturbance t is not zero,
t = t1 =
and the local-level model reduces to
yt = +
which is a simple regression with a time-invariant mean. The parameter is not estimated in the state-space formulation below. In this example, we t weekly levels of the Standard and Poors 500 Index to a local-level model. Because this model is already in state-space form, we t close by typing
366

. use http://www.stata-press.com/data/r12/sp500w, clear . constraint 13 [z]L.z = 1 . constraint 14 [close]z = 1 . sspace (z L.z, state noconstant) (close z, noconstant), constraints(13 14) searching for initial values .......... (setting technique to bhhh) Iteration 0: log likelihood = -12581.763 Iteration 1: log likelihood = -12577.727 (output omitted ) Refining estimates: Iteration 0: log likelihood = -12576.99 Iteration 1: log likelihood = -12576.99 State-space model Sample: 1 - 3093 Log likelihood = -12576.99 ( 1) [z]L.z = 1 ( 2) [close]z = 1 OIM Std. Err. Number of obs = 3093
close z z L1. close z Variance z close
Coef.
P>|z|
(constrained)
(constrained)
170.3456 15.24858
7.584909 3.392457
22.46 4.49
0.000 0.000
155.4794 8.599486
185.2117 21.89767
Note: Model is not stationary. Note: Tests of variances against zero are one sided, and the two-sided confidence intervals are truncated at zero.
The results indicate that both components have nonzero variances. The output footer informs us that the model is nonstationary at the estimated parameter values.
Technical note
In the previous example, we estimated the parameters of a nonstationary state-space model. The model is nonstationary because one of the eigenvalues of the A matrix has unit modulus. That all the coefcients in the A matrix are xed is also important. See L tkepohl (2005, 636637) for why u the ML estimator for the parameters of a nonstationary state-model that is nonstationary because of eigenvalues with unit moduli from a xed A matrix is still consistent and asymptotically normal.
Example 7
In another basic STS model, known as the local linear-trend model, both the level and the slope of a linear time trend are random walks. Here are the state equations and the observation equation for a local linear-trend model for the level of industrial production contained in variable ipman:
367
t t
1 1 0 1
t
t1 t1
1t 2t
ipmant = t +
The estimated parameters are
. use http://www.stata-press.com/data/r12/dfex, clear (St. Louis Fed (FRED) macro data) . constraint 15 [f1]L.f1 = 1 . constraint 16 [f1]L.f2 = 1 . constraint 17 [f2]L.f2 = 1 . constraint 18 [ipman]f1 = 1 . sspace (f1 L.f1 L.f2, state noconstant) > (f2 L.f2, state noconstant) > (ipman f1, noconstant), constraints(15/18) searching for initial values .......... (setting technique to bhhh) Iteration 0: log likelihood = -362.93861 Iteration 1: log likelihood = -362.12048 (output omitted ) Refining estimates: Iteration 0: log likelihood = -359.1266 Iteration 1: log likelihood = -359.1266 State-space model Sample: 1972m1 - 2008m11 Log likelihood = -359.1266 ( 1) [f1]L.f1 = 1 ( 2) [f1]L.f2 = 1 ( 3) [f2]L.f2 = 1 ( 4) [ipman]f1 = 1 OIM Std. Err. Number of obs = 443
ipman f1 f1 L1. f2 L1. f2 f2 L1. ipman f1 Variance f1 f2 ipman
Coef.
P>|z|
(constrained)
(constrained)
(constrained)
(constrained)
.1473071 .0178752 .0354429
.0407156 .0065743 .0148186
3.62 2.72 2.39
0.000 0.003 0.008
.067506 .0049898 .0063989
.2271082 .0307606 .0644868
368
There is little evidence that either of the variance parameters are zero. The t obtained indicates that we could now proceed with specication testing and checks to see how well this model forecasts these data.
Saved results
sspace saves the following in e():
Scalars e(N) e(k) e(k aux) e(k eq) e(k dv) e(k obser) e(k state) e(k obser err) e(k state err) e(df m) e(ll) e(chi2) e(p) e(tmin) e(tmax) e(stationary) e(rank) e(ic) e(rc) e(converged) Macros e(cmd) e(cmdline) e(depvar) e(obser deps) e(state deps) e(covariates) e(indeps) e(tvar) e(eqnames) e(title) e(tmins) e(tmaxs) e(R structure) e(Q structure) e(chi2type) e(vce) e(vcetype) e(opt) e(method) e(initial values) e(technique) e(tech steps) e(datasignature) e(datasignaturevars) e(properties) e(estat cmd) e(predict) e(marginsok) e(marginsnotok) number of observations number of parameters number of auxiliary parameters number of equations in e(b) number of dependent variables number of observation equations number of state equations number of observation-error terms number of state-error terms model degrees of freedom log likelihood
2
signicance minimum time in sample maximum time in sample 1 if the estimated parameters indicate a stationary model, 0 otherwise rank of VCE number of iterations return code 1 if converged, 0 otherwise sspace command as typed unoperated names of dependent variables in observation equations names of dependent variables in observation equations names of dependent variables in state equations list of covariates independent variables variable denoting time within groups names of equations title in estimation output formatted minimum time formatted maximum time structure of observed-variable-error covariance matrix structure of state-error covariance matrix Wald; type of model 2 test vcetype specied in vce() title used to label Std. Err. type of optimization likelihood method type of initial values maximization technique iterations taken in maximization technique the checksum variables used in calculation of checksum b V program used to implement estat program used to implement predict predictions allowed by margins predictions disallowed by margins

Matrices e(b) e(Cns) e(ilog) e(gradient) e(gamma) e(A) e(B) e(C) e(D) e(F) e(G) e(chol R) e(chol Q) e(chol Sz0) e(z0) e(d) e(T) e(M) e(V) e(V modelbased) Functions e(sample)
369
parameter vector constraints matrix iteration log (up to 20 iterations) gradient vector mapping from parameter vector to state-space matrices estimated A matrix estimated B matrix estimated C matrix estimated D matrix estimated F matrix estimated G matrix Cholesky factor of estimated R matrix Cholesky factor of estimated Q matrix Cholesky factor of initial state covariance matrix initial state vector augmented with a matrix identifying nonstationary components additional term in diffuse initial state vector, if nonstationary model inner part of quadratic form for initial state covariance in a partially nonstationary model outer part of quadratic form for initial state covariance in a partially nonstationary model variancecovariance matrix of the estimators model-based variance marks estimation sample

sspace is implemented as an ado-le. Recall that our notation for linear state-space models with time-invariant coefcient matrices is
zt = Azt1 + Bxt + C yt = Dzt + Fwt + Gt

where
zt is an m 1 vector of unobserved state variables; xt is a kx 1 vector of exogenous variables;

t
is a q 1 vector of state-error terms, (q m);
yt is an n 1 vector of observed endogenous variables; wt is a kw 1 vector of exogenous variables;

t is an r 1 vector of observation-error terms, (r n); and
A, B, C, D, F, and G are parameter matrices.

The equations for zt are known as the state equations, and the equations for yt are known as the observation equations. The error terms are assumed to be zero mean, normally distributed, serially uncorrelated, and uncorrelated with each other;
370

t
N (0, Q) = 0 for all s = t
t N (0, R)
E[
t s]
E[ t s ] = 0 for all s and t

sspace estimates the parameters of linear state-space models by maximum likelihood. The Kalman lter is a method for recursively obtaining linear, least-squares forecasts of yt conditional on past information. These forecasts are used to construct the log likelihood, assuming normality and stationarity. When the model is nonstationary, a diffuse Kalman lter is used. Hamilton (1994a; 1994b, 389) shows that the QML estimator, obtained when the normality assumption is dropped, is consistent and asymptotically normal, although the variancecovariance matrix of the estimator (VCE) must be estimated by the Huber/White/sandwich estimator. Hamiltons discussion applies to stationary models, and specifying vce(robust) produces a consistent estimator of the VCE when the errors are not normal. Methods for computing the log likelihood differ in how they calculate initial values for the Kalman lter when the model is stationary, how they compute a diffuse Kalman lter when the model is nonstationary, and whether terms for initial states are included. sspace offers the method(hybrid), method(dejong), and method(kdiffuse) options for computing the log likelihood. All three methods handle both stationary and nonstationary models. method(hybrid), the default, uses the initial values for the states implied by stationarity to initialize the Kalman lter when the model is stationary. Hamilton (1994b, 378) discusses this method of computing initial values for the states and derives a log-likelihood function that does not include terms for the initial states. When the model is nonstationary, method(hybrid) uses the De Jong (1988, 1991) diffuse Kalman lter and log-likelihood function, which includes terms for the initial states. method(dejong) uses the stationary De Jong (1988) method when the model is stationary and the De Jong (1988, 1991) diffuse Kalman lter when the model is nonstationary. The stationary De Jong (1988) method estimates initial values for the Kalman lter as part of the log-likelihood computation, as in De Jong (1988). method(kdiffuse) implements the seldom-used large- diffuse approximation to the diffuse Kalman lter when the model is nonstationary and uses initial values for the states implied by stationarity when the model is stationary. The log likelihood does not include terms for the initial states in either case. We recommend that you do not use method(kdiffuse) except to replicate older results computed using this method. De Jong (1988, 1991) and De Jong and Chu-Chun-Lin (1994) derive the log likelihood and a diffuse Kalman lter for handling nonstationary data. De Jong (1988) replaces the stationarity assumption with a time-immemorial assumption, which he uses to derive the log-likelihood function, an initial state vector, and a covariance of the initial state vector when the model is nonstationary. By default, and when method(hybrid) or method(dejong) is specied, sspace uses the diffuse Kalman lter given in denition 5 of De Jong and Chu-Chun-Lin (1994). This method uses theorem 3 of De Jong and Chu-Chun-Lin (1994) to compute the covariance of the initial states. When using this method, sspace saves the matrices from their theorem 3 in e(), although the names are changed. e(Z) is their U1 , e(T) is their U2 , e(A) is their T, and e(M) is their M. See De Jong (1988, 1991) and De Jong and Chu-Chun-Lin (1994) for the details of the De Jong diffuse Kalman lter. Practical estimation and inference require that the maximum likelihood estimator be consistent and normally distributed in large samples. These statistical properties of the maximum likelihood estimator
371
are well established when the model is stationary; see Caines (1988, chap. 5 and 7), Hamilton (1994b, 388389), and Hannan and Deistler (1988, chap. 4). When the model is nonstationary, additional assumptions must hold for the maximum likelihood estimator to be consistent and asymptotically normal; see Harvey (1989, sec. 3.4), L tkepohl (2005, 636637), and Schneider (1988). Chang, u Miller, and Park (2009) show that the ML and the QML estimators are consistent and asymptotically normal for a class of nonstationary state-space models. We now give an intuitive version of the Kalman lter. sspace uses theoretically equivalent, but numerically more stable, methods. For each time t, the Kalman lter produces the conditional expected state vector zt|t and the conditional covariance matrix t|t ; both are conditional on information up to and including time t. Using the model and previous period results, for each t we begin with
zt|t1 = Azt1|t1 + Bxt

t|t1 = At1|t1 A + CQC
(6)
yt|t1 = Dzt|t1 + Fwt

The residuals and the mean squared error (MSE) matrix of the forecast error are t|t = yt yt|t1
t|t = Dt|t1 D + GRG
(7)
In the last steps, we update the conditional expected state vector and the conditional covariance with the time t information:
zt|t = zt|t1 + t|t1 D1 t|t t|t

t|t = t|t1 t|t1 D1 D t|t1 t|t
(8)
Equations (6)(8) are the Kalman lter. The equations denoted by (6) are the one-step predictions. The one-step predictions do not use contemporaneous values of yt ; only past values of yt , past values of the exogenous xt , and contemporaneous values of xt are used. Equations (7) and (8) form the update step of the Kalman lter; they incorporate the contemporaneous dependent variable information into the predicted states. The Kalman lter requires initial values for the states and a covariance matrix for the initial states to start off the recursive process. Hamilton (1994b) discusses how to compute initial values for the Kalman lter assuming stationarity. This method is used by default when the model is stationary. De Jong (1988) discusses how to estimate initial values by maximum likelihood; this method is used when method(dejong) is specied. Letting be the vector of parameters in the model, L tkepohl (2005) and Harvey (1989) show u that the log-likelihood function for the parameters of a stationary model is given by
372
lnL() = 0.5
nT ln(2) +
t=1
ln(|t|t1 |) +
t=1
et 1 et t|t1
where et = (yt yt|t1 ) depends on and also depends on . The variancecovariance matrix of the estimator (VCE) is estimated by the observed information matrix (OIM) estimator by default. Specifying vce(robust) causes sspace to use the Huber/White/sandwich estimator. Both estimators of the VCE are standard and documented in Hamilton (1994b). Hamilton (1994b), Hannan and Deistler (1988), and Caines (1988) show that the ML estimator is consistent and asymptotically normal when the model is stationary. Schneider (1988) establishes consistency and asymptotic normality when the model is nonstationary because A has some eigenvalues with modulus 1 and there are no unknown parameters in A. Not all state-space models are identied, as discussed in Hamilton (1994b) and L tkepohl (2005). u sspace checks for local identication at the optimum. sspace will not declare convergence unless the Hessian is full rank. This check for local identiability is due to Rothenberg (1971). Specifying method(dejong) causes sspace to maximize the log-likelihood function given in section 2 (vii) of De Jong (1988). This log-likelihood function includes the initial states as parameters to be estimated. We use some of the methods in Casals, Sotoca, and Jerez (1999) for computing the De Jong (1988) log-likelihood function.
References
Anderson, B. D. O., and J. B. Moore. 1979. Optimal Filtering. Englewood Cliffs, NJ: Prentice Hall. Brockwell, P. J., and R. A. Davis. 1991. Time Series: Theory and Methods. 2nd ed. New York: Springer. Caines, P. E. 1988. Linear Stochastic Systems. New York: Wiley. Casals, J., S. Sotoca, and M. Jerez. 1999. A fast and stable method to compute the likelihood of time invariant state-space models. Economics Letters 65: 329337. Chang, Y., J. I. Miller, and J. Y. Park. 2009. Extracting a common stochastic trend: Theory with some applications. Journal of Econometrics 150: 231247. Davidson, R., and J. G. MacKinnon. 1993. Estimation and Inference in Econometrics. New York: Oxford University Press. De Jong, P. 1988. The likelihood for a state space model. Biometrika 75: 165169. . 1991. The diffuse Kalman lter. Annals of Statistics 19: 10731083. De Jong, P., and S. Chu-Chun-Lin. 1994. Stationary and non-stationary state space models. Journal of Time Series Analysis 15: 151166. Drukker, D. M., and V. L. Wiggins. 2004. Verifying the solution from a nonlinear solver: A case study: Comment. American Economic Review 94: 397399. Hamilton, J. D. 1994a. State-space models. In Vol. 4 of Handbook of Econometrics, ed. R. F. Engle and D. L. McFadden, 30393080. New York: Elsevier. . 1994b. Time Series Analysis. Princeton: Princeton University Press. Hannan, E. J., and M. Deistler. 1988. The Statistical Theory of Linear Systems. New York: Wiley. Harvey, A. C. 1989. Forecasting, Structural Time Series Models and the Kalman Filter. Cambridge: Cambridge University Press. . 1993. Time Series Models. 2nd ed. Cambridge, MA: MIT Press. L tkepohl, H. 2005. New Introduction to Multiple Time Series Analysis. New York: Springer. u

Rothenberg, T. J. 1971. Identication in parametric models. Econometrica 39: 577591. Schneider, W. 1988. Analytical uses of Kalman ltering in econometrics: A survey. Statistical Papers 29: 333.
373
Stock, J. H., and M. W. Watson. 1989. New indexes of coincident and leading economic indicators. In NBER Macroeconomics Annual 1989, ed. O. J. Blanchard and S. Fischer, vol. 4, 351394. Cambridge, MA: MIT Press. . 1991. A probability model of the coincident economic indicators. In Leading Economic Indicators: New Approaches and Forecasting Records, ed. K. Lahiri and G. H. Moore, 6389. Cambridge: Cambridge University Press.
Also see
[TS] sspace postestimation Postestimation tools for sspace [TS] tsset Declare data to be time-series data [TS] dfactor Dynamic-factor models [TS] ucm Unobserved-components model [TS] var Vector autoregressive models [TS] arima ARIMA, ARMAX, and other dynamic regression models [U] 20 Estimation and postestimation commands
Title
sspace postestimation Postestimation tools for sspace
Description
The following standard postestimation commands are available after sspace: Command estat estimates lincom lrtest nlcom predict predictnl test testnl Description
cataloging estimation results point estimates, standard errors, testing and inference for linear combinations of coefcients likelihood-ratio test point estimates, standard errors, testing and inference for nonlinear combinations of coefcients predictions, residuals, inuence statistics, and other diagnostic measures point estimates, standard errors, testing, and inference for generalized predictions Wald tests of simple and composite linear hypotheses Wald tests of nonlinear hypotheses
Syntax for predict

predict statistic
Main
type
Description
if
in
, statistic options
xb states residuals rstandard
observable variables latent state variables residuals standardized residuals
374
375
options
Options
Description
name(s) of equation(s) for which predictions are to be made put estimated root mean squared errors of predicted statistics in new variables dynamic(time constant) begin dynamic forecast at specied time equation(eqnames) rmse(stub* | newvarlist)
Advanced
method for predicting unobserved states Description predict using past information predict using all sample information predict using past and contemporaneous information
Menu
Statistics
>
Postestimation
>
Options for predict

Main
xb, states, residuals, and rstandard specify the statistic to be predicted. xb, the default, calculates the linear predictions of the observed variables. states calculates the linear predictions of the latent state variables. residuals calculates the residuals in the equations for observable variables. residuals may not be specied with dynamic(). rstandard calculates the standardized residuals, which are the residuals normalized to be uncorrelated and to have unit variances. rstandard may not be specied with smethod(filter), smethod(smooth), or dynamic().
Options
equation(eqnames) species the equation(s) for which the predictions are to be calculated. If you do not specify equation() or stub*, the results are the same as if you had specied the name of the rst equation for the predicted statistic. You specify a list of equation names, such as equation(income consumption) or equation(factor1 factor2), to identify the equations. Specify names of state equations when predicting states and names of observable equations in all other cases. equation() may not be specied with stub*. rmse(stub* | newvarlist) puts the root mean squared errors of the predicted statistics into the specied new variables. The root mean squared errors measure the variances due to the disturbances but do not account for estimation error.
376
dynamic(time constant) species when predict starts producing dynamic forecasts. The specied time constant must be in the scale of the time variable specied in tsset, and the time constant must be inside a sample for which observations on the dependent variables are available. For example, dynamic(tq(2008q4)) causes dynamic predictions to begin in the fourth quarter of 2008, assuming that your time variable is quarterly; see [D] datetime. If the model contains exogenous variables, they must be present for the whole predicted sample. dynamic() may not be specied with rstandard, residuals, or smethod(smooth).
Advanced
smethod(method) species the method for predicting the unobserved states; smethod(onestep), smethod(filter), and smethod(smooth) cause different amounts of information on the dependent variables to be used in predicting the states at each time period. smethod(onestep), the default, causes predict to estimate the states at each time period using previous information on the dependent variables. The Kalman lter is performed on previous periods, but only the one-step predictions are made for the current period. smethod(smooth) causes predict to estimate the states at each time period using all the sample data by the Kalman smoother. smethod(smooth) may not be specied with rstandard. smethod(filter) causes predict to estimate the states at each time period using previous and contemporaneous data by the Kalman lter. The Kalman lter is performed on previous periods and the current period. smethod(filter) may be specied only with states.
Remarks
We assume that you have already read [TS] sspace. In this entry, we illustrate some of the features of predict after using sspace to estimate the parameters of a state-space model. All the predictions after sspace depend on the unobserved states, which are estimated recursively. Changing the sample can alter the state estimates, which can change all other predictions.
Example 1
In example 5 of [TS] sspace, we estimated the parameters of the dynamic-factor model
ft ft1
1 1
ipmant 1 incomet 2 = ft + hourst 3 unempt 4

where
2 1 0 Var 2t = 0 3t 0 4t
2 0
ft1 + ft2
1t 2t 3t 4t
t 0
1t
0 2 2 0 0
0 0 2 3 0
0 0 0 2 4
377
by typing
. use http://www.stata-press.com/data/r12/dfex (St. Louis Fed (FRED) macro data) . constraint 1 [lf]L.f = 1 . sspace (f L.f L.lf, state noconstant) > (lf L.f, state noconstant noerror) > (D.ipman f, noconstant) > (D.income f, noconstant) > (D.hours f, noconstant) > (D.unemp f, noconstant), > covstate(identity) constraints(1) (output omitted )
Below we obtain the one-step predictions for each of the four dependent variables in the model, and then we graph the actual and predicted ipman:
. predict dep* (option xb assumed; fitted values) . tsline D.ipman dep1, lcolor(gs10) xtitle("") legend(rows(2))
4 1970m1
1980m1
1990m1
2000m1
2010m1
Industrial production; manufacturing (NAICS), D xb prediction, D.ipman, onestep
The graph shows that the one-step predictions account for only a small part of the swings in the realized ipman.
Example 2
We use the estimates from the previous example to make out-of-sample predictions. After using tsappend to extend the dataset by six periods, we use predict with the dynamic() option and graph the result.
. tsappend, add(6) . predict Dipman_f, dynamic(tm(2008m12)) equation(D.ipman) . tsline D.ipman Dipman_f if month>=tm(2008m1), xtitle("") legend(rows(2))
378
6 2008m1
2008m4
2008m7
2008m10
2009m1
2009m4
Industrial production; manufacturing (NAICS), D xb prediction, D.ipman, dynamic(tm(2008m12))
The model predicts that the changes in industrial production will remain negative for the forecast horizon, although they increase toward zero.
Example 3
In this example, we want to estimate the unobserved factor instead of predicting a dependent variable. Specifying smethod(smooth) causes predict to use all sample information in estimating the states by the Kalman smoother. Below we estimate the unobserved factor by using the estimation sample, and we graph ipman and the estimated factor:
. predict fac if e(sample), states smethod(smooth) equation(f) . tsline D.ipman fac, xtitle("") legend(rows(2))
6 1970m1
1980m1
1990m1
2000m1
2010m1
Industrial production; manufacturing (NAICS), D states, f, smooth
379
Example 4
The residuals and the standardized residuals are frequently used to review the specication of the model. Below we calculate the standardized residuals for each of the series and display them in a combined graph:
. . . . . . predict sres1-sres4 if e(sample), rstandard tsline sres1, xtitle("") name(sres1) tsline sres2, xtitle("") name(sres2) tsline sres3, xtitle("") name(sres3) tsline sres4, xtitle("") name(sres4) graph combine sres1 sres2 sres3 sres4, name(combined)
rstandard, D.ipman, onestep 10 5 0 5
1970m1 1980m1 1990m1 2000m1 2010m1
rstandard, D.hours, onestep 4 2 0 2 4 6
1970m1 1980m1 1990m1 2000m1 2010m1

All postestimation commands listed above are implemented as ado-les. Estimating the unobserved states is key to predicting the dependent variables. By default and with the smethod(onestep) option, predict estimates the states in each period by applying the Kalman lter to all previous periods and only making the one-step predictions to the current period. (See Methods and formulas of [TS] sspace for the Kalman lter equations.) With the smethod(filter) option, predict estimates the states in each period by applying the Kalman lter on all previous periods and the current period. The computational difference between smethod(onestep) and smethod(filter) is that smethod(filter) performs the update
rstandard, D.unemp, onestep 4 2 0 2 4 1970m1 1980m1 1990m1 2000m1 2010m1
rstandard, D.income, onestep 5 0 5 10 1970m1 1980m1 1990m1 2000m1 2010m1
380
step on the current period while smethod(onestep) does not. The statistical difference between smethod(onestep) and smethod(filter) is that smethod(filter) uses contemporaneous information on the dependent variables while smethod(onestep) does not. As noted in [TS] sspace, sspace has both a stationary and a diffuse Kalman lter. predict uses the same Kalman lter used for estimation. With the smethod(smooth) option, predict estimates the states in each period using all the sample information by applying the Kalman smoother. predict uses the Harvey (1989, sec. 3.6.2) xed-interval smoother with model-based initial values to estimate the states when the estimated parameters imply a stationary model. De Jong (1989) provides a computationally efcient method. Hamilton (1994) discusses the model-based initial values for stationary state-space models. When the model is nonstationary, the De Jong (1989) diffuse Kalman smoother is used to predict the states. The smoothed estimates of the states are subsequently used to predict the dependent variables. The dependent variables are predicted by plugging in the estimated states. The residuals are calculated as the differences between the predicted and the realized dependent variables. The root mean squared errors are the square roots of the diagonal elements of the mean squared error matrices that are computed by the Kalman lter. The standardized residuals are the residuals normalized by the Cholesky factor of their mean squared error produced by the Kalman lter. predict uses the Harvey (1989, sec. 3.5) methods to compute the dynamic forecasts and the root mean squared errors. Let be the period at which the dynamic forecasts begin; must either be in the specied sample or be in the period immediately following the specied sample. The dynamic forecasts depend on the predicted states in the period 1, which predict obtains by running the Kalman lter or the diffuse Kalman lter on the previous sample observations. The states in the periods prior to starting the dynamic predictions may be estimated using smethod(onestep) or smethod(smooth). Using an if or in qualier to alter the prediction sample can change the estimate of the unobserved states in the period prior to beginning the dynamic predictions and hence alter the dynamic predictions. The initial states are estimated using e(b) and the prediction sample.
References
De Jong, P. 1988. The likelihood for a state space model. Biometrika 75: 165169. . 1989. Smoothing and interpolation with the state-space model. Journal of the American Statistical Association 84: 10851088. . 1991. The diffuse Kalman lter. Annals of Statistics 19: 10731083. Hamilton, J. D. 1994. Time Series Analysis. Princeton: Princeton University Press. Harvey, A. C. 1989. Forecasting, Structural Time Series Models and the Kalman Filter. Cambridge: Cambridge University Press. L tkepohl, H. 2005. New Introduction to Multiple Time Series Analysis. New York: Springer. u
Also see
[TS] sspace State-space models [TS] dfactor Dynamic-factor models [TS] dfactor postestimation Postestimation tools for dfactor [U] 20 Estimation and postestimation commands
Title
tsappend Add observations to a time-series dataset
Syntax
tsappend , options
add(#) | last(date | clock) tsfmt(string) Description
options
add(#) last(date | clock) tsfmt(string) panel(panel id)
add # observations add observations at date or clock use time-series function string with last(date | clock) add observations to panel panel id
Either add(#) is required, or last(date | clock) and tsfmt(string) are required. You must tsset your data before using tsappend; see [TS] tsset.
Menu
Statistics
>
Time series
>
Setup and utilities
>
Add observations to time-series dataset
Description
tsappend appends observations to a time-series dataset or to a panel dataset. tsappend uses and updates the information set by tsset.
Options
add(#) species the number of observations to add. last(date | clock) and tsfmt(string) must be specied together and are an alternative to add(). last(date | clock) species the date or the date and time of the last observation to add. tsfmt(string) species the name of the Stata time-series function to use in converting the date specied in last() to an integer. The function names are tc (clock), tC (Clock), td (daily), tw (weekly), tm (monthly), tq (quarterly), th (half-yearly), and ty (yearly). For clock times, the last time added (if any) will be earlier than the time requested in last(date | clock) if last() is not a multiple of delta units from the last time in the data. For instance, you might specify last(17may2007) tsfmt(td), last(2001m1) tsfmt(tm), or last(17may2007 15:30:00) tsfmt(tc). panel(panel id) species that observations be added only to panels with the ID specied in panel().
Remarks
Introduction Using tsappend with time-series data Using tsappend with panel data
381
382
Introduction
tsappend adds observations to a time-series dataset or to a panel dataset. You must tsset your data before using tsappend. tsappend simultaneously removes any gaps from the dataset. There are two ways to use tsappend: you can specify the add(#) option to request that # observations be added, or you can specify the last(date | clock) option to request that observations be appended until the date specied is reached. If you specify last(), you must also specify tsfmt(). tsfmt() species the Stata time-series date function that converts the date held in last() to an integer. tsappend works with time series of panel data. With panel data, tsappend adds the requested observations to all the panels, unless the panel() option is also specied.
Using tsappend with time-series data

tsappend can be useful for appending observations when dynamically predicting a time series. Consider an example in which tsappend adds the extra observations before dynamically predicting from an AR(1) regression:
. use http://www.stata-press.com/data/r12/tsappend1 . regress y l.y Source SS df MS Model Residual Total y y L1. _cons . mat b = e(b) . mat colnames b = L.xb one . tsset time variable: t2, 1960m2 to 2000m1 delta: 1 month . tsappend, add(12) . tsset time variable: t2, 1960m2 to 2001m1 delta: 1 month . predict xb if t2<=tm(2000m2) (option xb assumed; fitted values) (12 missing values generated) . gen one=1 . mat score xb=b if t2>=tm(2000m2), replace 115.349555 461.241577 576.591132 Coef. 1 477 478 115.349555 .966963473 1.2062576 t P>|t|
Number of obs F( 1, 477) Prob > F R-squared Adj R-squared Root MSE
= = = = = =
479 119.29 0.0000 0.2001 0.1984 .98334
Std. Err.
.4493507 11.11877
.0411417 .8314581
10.92 13.37
0.000 0.000
.3685093 9.484993
.5301921 12.75254
The calls to tsset before and after tsappend were unnecessary. Their output reveals that tsappend added another year of observations. We then used predict and matrix score to obtain the dynamic predictions, which allows us to produce the following graph:

. line y xb t2 if t2>=tm(1995m1), ytitle("") xtitle("time")
383
18 1995m1
19
20
21
22
23
1996m1
1997m1 y
1998m1 time
1999m1 Fitted values
2000m1
2001m1
In the call to tsappend, instead of saying that we wanted to add 12 observations, we could have specied that we wanted to ll in observations through the rst month of 2001:
. use http://www.stata-press.com/data/r12/tsappend1, clear . tsset time variable: t2, 1960m2 to 2000m1 delta: 1 month . tsappend, last(2001m1) tsfmt(tm) . tsset time variable: t2, 1960m2 to 2001m1 delta: 1 month
We specied the tm() function in the tsfmt() option. [D] functions contains a list of timeseries functions for translating date literals to integers. Because we have monthly data, and since [D] functions tells us that we want to use the tm() function, we specied the tsfmt(tm) option. The following table shows the most common types of time-series data, their formats, the appropriate translation functions, and the corresponding options for tsappend: Description time time daily weekly monthly quarterly half-yearly yearly Format %tc %tC %td %tw %tm %tq %th %ty Function tc() tC() td() tw() tm() tq() th() ty() Option tsfmt(tc) tsfmt(tC) tsfmt(td) tsfmt(tw) tsfmt(tm) tsfmt(tq) tsfmt(th) tsfmt(ty)
384
Using tsappend with panel data

tsappends actions on panel data are similar to its action on time-series data, except that tsappend performs those actions on each time series within the panels. If the end dates vary over panels, last() and add() will produce different results. add(#) always adds # observations to each panel. If the data end at different periods before tsappend, add() is used, the data will still end at different periods after tsappend, add(). In contrast, tsappend, last() tsfmt() will cause all the panels to end on the specied last date. If the beginning dates differ across panels, using tsappend, last() tsfmt() to provide a uniform ending date will not create balanced panels because the number of observations per panel will still differ. Consider the panel data summarized in the output below:
. use http://www.stata-press.com/data/r12/tsappend3, clear . xtdescribe id: 1, 2, ..., 3 n = t2: 1998m1, 1998m2, ..., 2000m1 T = Delta(t2) = 1 month Span(t2) = 25 periods (id*t2 uniquely identifies each observation) Distribution of T_i: min 5% 25% 50% 75% 13 13 13 20 24 Freq. Percent Cum. Pattern 1 1 1 33.33 33.33 33.33 33.33 66.67 100.00 ............1111111111111 1111.11111111111111111111 11111111111111111111..... XXXXXXXXXXXXXXXXXXXXXXXXX
3 25
95% 24
max 24
3 100.00 . by id: sum t2 -> id = 1 Variable t2 -> id = 2 Variable t2 -> id = 3 Variable t2 Obs 24 Obs 13
Mean 474
Std. Dev. 3.89444
Min 468
Max 480
Obs 20
Mean 465.5
Std. Dev. 5.91608
Min 456
Max 475
Mean 468.3333
Std. Dev. 7.322786
Min 456
Max 480
The output from xtdescribe and summarize on these data tells us that one panel starts later than the other, that another panel ends before the other two, and that the remaining panel has a gap in the time variable but otherwise spans the entire time frame.
385
Now consider the data after a call to tsappend, add(6):

. tsappend, add(6) . xtdescribe id: 1, 2, ..., 3 t2: 1998m1, 1998m2, ..., 2000m7 Delta(t2) = 1 month Span(t2) = 31 periods (id*t2 uniquely identifies each observation) Distribution of T_i: Freq. 1 1 1 3 Percent 33.33 33.33 33.33 100.00 min 19 Cum. 33.33 66.67 100.00 5% 25% 19 19 Pattern 50% 26
n = T =
3 31
75% 31
95% 31
max 31
............1111111111111111111 11111111111111111111111111..... 1111111111111111111111111111111 XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
. by id: sum t2 -> id = 1 Variable t2 -> id = 2 Variable t2 -> id = 3 Variable t2
Obs 19
Mean 477
Std. Dev. 5.627314
Min 468
Max 486
Obs 26
Mean 468.5
Std. Dev. 7.648529
Min 456
Max 481
Obs 31
Mean 471
Std. Dev. 9.092121
Min 456
Max 486
This output from xtdescribe and summarize after the call to tsappend shows that the call to tsappend, add(6) added 6 observations to each panel and lled in the gap in the time variable in the second panel. tsappend, add() did not cause a uniform end date over the panels. The following output illustrates the contrast between tsappend, add() and tsappend, last() tsfmt() with panel data that end at different dates. The output from xtdescribe and summarize shows that the call to tsappend, last() tsfmt() lled in the gap in t2 and caused all the panels to end at the specied end date. The output also shows that the panels remain unbalanced because one panel has a later entry date than the other two.
. use http://www.stata-press.com/data/r12/tsappend2, clear . tsappend, last(2000m7) tsfmt(tm)
386

. xtdescribe id: 1, 2, ..., 3 t2: 1998m1, 1998m2, ..., 2000m7 Delta(t2) = 1 month Span(t2) = 31 periods (id*t2 uniquely identifies each observation) Distribution of T_i: Freq. 2 1 Percent 66.67 33.33 min 19 Cum. 66.67 100.00 5% 25% 19 19 Pattern 50% 31
n = T =
3 31
75% 31
95% 31
max 31
1111111111111111111111111111111 ............1111111111111111111 XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
3 100.00 . by id: sum t2 -> id = 1 Variable t2 -> id = 2 Variable t2 -> id = 3 Variable t2
Obs 19
Mean 477
Std. Dev. 5.627314
Min 468
Max 486
Obs 31
Mean 471
Std. Dev. 9.092121
Min 456
Max 486
Obs 31
Mean 471
Std. Dev. 9.092121
Min 456
Max 486
Saved results
tsappend saves the following in r():
Scalars r(add) number of observations added

tsappend is implemented as an ado-le.
Also see
[TS] tsset Declare data to be time-series data [TS] tsll Fill in gaps in time variable
Title
tsll Fill in gaps in time variable
Syntax
tsfill , full
You must tsset your data before using tsfill; see [TS] tsset.
Menu
Statistics
>
Time series
>
Setup and utilities
>
Fill in gaps in time variable
Description
tsfill is used after tsset to ll in gaps in time-series data and gaps in panel data with new observations, which contain missing values. For instance, perhaps observations for timevar = 1, 3, 5, 6, . . . , 22 exist. tsfill would create observations for timevar = 2 and timevar = 4 containing all missing values. There is seldom reason to do this because Statas time-series operators consider timevar, not the observation number. Referring to L.gnp to obtain lagged gnp values would correctly produce a missing value for timevar = 3, even if the data were not lled in. Referring to L2.gnp would correctly return the value of gnp in the rst observation for timevar = 3, even if the data were not lled in.
Option
full is for use with panel data only. With panel data, tsfill by default lls in observations for each panel according to the minimum and maximum values of timevar for the panel. Thus if the rst panel spanned the times 520 and the second panel the times 115, after tsfill they would still span the same periods; observations would be created to ll in any missing times from 520 in the rst panel and from 115 in the second. If full is specied, observations are created so that both panels span the time 120, the overall minimum and maximum of timevar across panels.
Remarks
Using tsll with time-series data Using tsll with panel data
387
388
Using tsll with time-series data

You have monthly data, with gaps:
. use http://www.stata-press.com/data/r12/tsfillxmpl . tsset time variable: mdate, 1995m7 to 1996m3, but with gaps delta: 1 month . list mdate income mdate 1. 2. 3. 4. 5. 6. 1995m7 1995m8 1995m11 1995m12 1996m1 1996m3 income 1153 1181 1236 1297 1265 1282
You can ll in the gaps by interpolation easily with tsfill and ipolate. tsfill creates the missing observations:
. tsfill . list mdate income mdate 1. 2. 3. 4. 5. 6. 7. 8. 9. 1995m7 1995m8 1995m9 1995m10 1995m11 1995m12 1996m1 1996m2 1996m3 income 1153 1181 . . 1236 1297 1265 . 1282
new new
new
We can now use ipolate (see [D] ipolate) to ll them in:

. ipolate income mdate, gen(ipinc) . list mdate income ipinc mdate 1. 2. 3. 4. 5. 6. 7. 8. 9. 1995m7 1995m8 1995m9 1995m10 1995m11 1995m12 1996m1 1996m2 1996m3 income 1153 1181 . . 1236 1297 1265 . 1282 ipinc 1153 1181 1199.3333 1217.6667 1236 1297 1265 1273.5 1282
389
Using tsll with panel data

You have the following panel dataset:
. use http://www.stata-press.com/data/r12/tsfillxmpl2, clear . tsset panel variable: edlevel (unbalanced) time variable: year, 1988 to 1992, but with a gap delta: 1 unit . list edlevel year income edlevel 1. 2. 3. 4. 5. 6. 7. 1 1 1 1 2 2 2 year 1988 1989 1990 1991 1989 1990 1992 income 14500 14750 14950 15100 22100 22200 22800
Just as with nonpanel time-series datasets, you can use tsfill to ll in the gaps:
. tsfill . list edlevel year income edlevel 1. 2. 3. 4. 5. 6. 7. 8. 1 1 1 1 2 2 2 2 year 1988 1989 1990 1991 1989 1990 1991 1992 income 14500 14750 14950 15100 22100 22200 . 22800
new
You could instead use tsfill to produce fully balanced panels with the full option:
. tsfill, full . list edlevel year income, sep(0) edlevel 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 1 1 1 1 1 2 2 2 2 2 year 1988 1989 1990 1991 1992 1988 1989 1990 1991 1992 income 14500 14750 14950 15100 . . 22100 22200 . 22800
new new
new
390

tsfill is implemented as an ado-le.
Also see
[TS] tsset Declare data to be time-series data [TS] tsappend Add observations to a time-series dataset
Title
tslter Filter a time-series, keeping only selected periodicities
Syntax
Filter one variable tsfilter lter type newvar = varname if in , options
Filter multiple variables, unique names tsfilter lter type newvarlist = varlist if in , options
Filter multiple variables, common name stub tsfilter lter lter bk bw cf hp type stub* = varlist if in See [TS] [TS] [TS] [TS]
tslter tslter tslter tslter bk bw cf hp
, options
Name BaxterKing Butterworth ChristianoFitzgerald HodrickPrescott
You must tsset or xtset your data before using tsfilter; see [TS] tsset and [XT] xtset. varname and varlist may contain time-series operators; see [U] 11.4.4 Time-series varlists. options differ across the lters and are documented in each lters manual entry.
Description
tsfilter separates a time series into trend and cyclical components. The trend component may contain a deterministic or a stochastic trend. The stationary cyclical component is driven by stochastic cycles at the specied periods.
Remarks
The time-series lters implemented in tsfilter separate a time-series yt into trend and cyclical components: yt = t + ct where t is the trend component and ct is the cyclical component. t may be nonstationary; it may contain a deterministic or a stochastic trend, as discussed below. The primary objective of the methods implemented in tsfilter is to estimate ct , a stationary cyclical component that is driven by stochastic cycles within a specied range of periods. The trend component t is calculated by the difference t = yt ct . Although the lters implemented in tsfilter have been widely applied by macroeconomists, they are general time-series methods and may be of interest to other researchers.
391
392

An example dataset A baseline method: Symmetric moving-average (SMA) lters An overview of ltering in the frequency domain SMA revisited: The BaxterKing lter Filtering a random walk: The ChristianoFitzgerald lter A one-parameter high-pass lter: The HodrickPrescott lter A two-parameter high-pass lter: The Butterworth lter
An example dataset
Time series are frequently ltered to remove unwanted characteristics, such as trends and seasonal components, or to estimate components driven by stochastic cycles from a specic range of periods. Although the lters implemented in tsfilter can be used for both purposes, their primary purpose is the latter, and we restrict our discussion to that use. We explain the methods implemented in tsfilter by estimating the business-cycle component of a macroeconomic variable, because they are frequently used for this purpose. We estimate the business-cycle component of the natural log of an index of the industrial production of the United States, which is plotted below.
Example 1
. use http://www.stata-press.com/data/r12/ipq (Federal Reserve Economic Data, St. Louis Fed) . tsline ip_ln
5 1
1920q1 1930q1 1940q1 1950q1 1960q1 1970q1 1980q1 1990q1 2000q1 2010q1
log of industrial production 2 3 4
quarterly time variable
The above graph shows that ip ln contains a trend component. Time series may contain deterministic trends or stochastic trends. A polynomial function of time is the most common deterministic time trend. An integrated process is the most common stochastic trend. An integrated process is a random variable that must be differenced one or more times to be stationary; see Hamilton (1994) for a discussion. The different lters implemented in tsfilter allow for different orders of deterministic time trends or integrated processes.
393
We now illustrate the four methods implemented in tsfilter, each of which will remove the trend and estimate the business-cycle component. Burns and Mitchell (1946) dened oscillations in business data with recurring periods between 1.5 and 8 years to be business-cycle uctuations; we use their commonly accepted denition.
A baseline method: Symmetric moving-average (SMA) lters

Symmetric moving-average (SMA) lters form a baseline method for estimating a cyclical component because of their properties and simplicity. An SMA lter of a time series yt , t {1, . . . , T }, is the data transform dened by
q yt = j=q
j ytj
for each t {q + 1, . . . , T q}, where j = j for j {q, . . . , q}. Although the original series has T observations, the ltered series has only T 2q , where q is known as the order of the SMA lter.
SMA lters with weights that sum to zero remove deterministic and stochastic trends of order 2 or less, as shown by Fuller (1996) and Baxter and King (1999).
Example 2
This trend-removal property of SMA lters with coefcients that sum to zero may surprise some readers. For illustration purposes, we lter ip ln by the lter
0.2ip lnt2 0.2ip lnt1 + 0.8ip lnt 0.2ip lnt+1 0.2ip lnt+2
and plot the ltered series. We do not even need tsfilter to implement this second-order SMA lter; we can use generate.
. generate ip_sma = -.2*L2.ip_ln-.2*L.ip_ln+.8*ip_ln-.2*F.ip_ln-.2*F2.ip_ln (4 missing values generated) . tsline ip_sma
.2
1920q1 1930q1 1940q1 1950q1 1960q1 1970q1 1980q1 1990q1 2000q1 2010q1
.1
ip_sma 0
.1
.2
394
The lter has removed the trend.
There is no good reason why we chose that particular SMA lter. Baxter and King (1999) derived a class of SMA lters with coefcients that sum to zero and get as close as possible to keeping only the specied cyclical component.
An overview of ltering in the frequency domain

We need some concepts from the frequency-domain approach to time-series analysis to motivate how Baxter and King (1999) dened as close as possible. These concepts also motivate the other lters in tsfilter. The intuitive explanation presented here glosses over many technical details discussed by Priestley (1981), Hamilton (1994), Fuller (1996), and Wei (2006). As with much time-series analysis, the basic results are for covariance-stationary processes with additional results handling some nonstationary cases. We present some useful results for covariancestationary processes and discuss how to handle nonstationary series below. The autocovariances j , j {0, 1, . . . , }, of a covariance-stationary process yt specify its variance and dependence structure. In the frequency-domain approach to time-series analysis, yt and the autocovariances are specied in terms of independent stochastic cycles that occur at frequencies [, ]. The spectral density function fy () species the contribution of stochastic cycles at 2 each frequency relative to the variance of yt , which is denoted by y . The variance and the autocovariances can be expressed as an integral of the spectral density function. Formally,
j =
eij fy ()d
(1)
where i is the imaginary number i =
1.
Equation (1) can be manipulated to show what fraction of the variance of yt is attributable to stochastic cycles in a specied range of frequencies. Hamilton (1994, 156) discusses this point in more detail. Equation (1) implies that if fy () = 0 for [1 , 2 ], then stochastic cycles at these frequencies contribute zero to the variance and autocovariances of yt .
The goal of time-series lters is to transform the original series into a new series yt for which the spectral density function of the ltered series fy () is zero for unwanted frequencies and equal to fy () for desired frequencies.
A linear lter of yt can be written as

yt
=
j=
j ytj = (L)yt
where we let yt be an innitely long series as required by some of the results below. To see the impact of the lter on the components of yt at each frequency , we need an expression for fy () in terms of fy () and the lter weights j . Wei (2006, 282) shows that for each ,
fy () = |(ei )|2 fy ()
(2)
395
where |(ei )| is known as the gain of the lter. Equation (2) makes explicit that the squared gain function |a(ei )|2 converts the spectral density of the original series, fy (), into the spectral density of the ltered series, fy (). In particular, (2) says that for each frequency , the spectral density of the ltered series is the product of the square of the gain of the lter and the spectral density of the original series. As we will see in the examples below, the gain function provides a crucial interpretation of what a lter is doing. We want a lter for which fy () = 0 for unwanted frequencies and for which fy () = fy () for desired frequencies. So we seek a lter for which the gain is 0 for unwanted frequencies and for which the gain is 1 for desired frequencies. In practice, we cannot nd such an ideal lter exactly, because the constraints an ideal lter places on lter coefcients cannot be satised for time series with only a nite number of observations. The expansive literature on lters is a result of the trade-offs involved in designing implementable lters that approximate the ideal lter. Ideally, lters pass or block the stochastic cycles at specied frequencies by having a gain of 1 or 0. Band-pass lters, such as the BaxterKing (BK) and the ChristianoFitzgerald (CF) lters, pass through stochastic cycles in the specied range of frequencies and block all the other stochastic cycles. High-pass lters, such as the HodrickPrescott (HP) and Butterworth lters, only allow the stochastic cycles at or above a specied frequency to pass through and block the lower-frequency stochastic cycles. For band-pass lters, let [0 , 1 ] be the set of desired frequencies with all other frequencies being undesired. For high-pass lters, let 0 be the cutoff frequency with only those frequencies 0 being desired.
SMA revisited: The BaxterKing lter

We now return to the class of SMA lters with coefcients that sum to zero and get as close as possible to keeping only the specied cyclical component as derived by Baxter and King (1999). For an innitely long series, there is an ideal band-pass lter for which the gain function is 1 for [0 , 1 ] and 0 for all other frequencies. It just so happens that this ideal band-pass lter is an SMA lter with coefcients that sum to zero. Baxter and King (1999) derive the coefcients of this ideal band-pass lter and then dene the BK lter to be the SMA lter with 2q + 1 terms that are as close as possible to those of the ideal lter. There is a trade-off in choosing q : larger values of q cause the gain of the BK lter to be closer to the gain of the ideal lter, but larger values also increase the number of missing observations in the ltered series. Although the mathematics of the frequency-domain approach to time-series analysis is in terms of stochastic cycles at frequencies [, ], applied work is generally in terms of periods p, where p = 2/ . So the options for the tsfilter subcommands are in terms of periods.
Example 3
Below we use tsfilter bk, which implements the BK lter, to estimate the business-cycle component composed of stochastic cycles between 6 and 32 periods, and then we graph the estimated component.
396

. tsfilter bk ip_bk = ip_ln, minperiod(6) maxperiod(32) . tsline ip_bk
.3
1920q1 1930q1 1940q1 1950q1 1960q1 1970q1 1980q1 1990q1 2000q1 2010q1
ip_ln cyclical component from bk filter .2 .1 0 .1
.2
The above graph tells us what the estimated business-cycle component looks like, but it presents no evidence as to how well we have estimated the component. A periodogram is better for this purpose. A periodogram is an estimator of a transform of the spectral density function; see [TS] pergram for details. Below we plot the periodogram for the BK estimate of the business-cycle component. pergram displays the results in natural frequencies, which are the standard frequencies divided by 2 . We use the xline() option to draw vertical lines at the lower natural-frequency cutoff (1/32 = 0.03125) and the upper natural-frequency cutoff (1/6 0.16667).
. pergram ip_bk, xline(0.03125 0.16667)
ip_ln cyclical component from bk filter Log Periodogram 6.00 4.00 2.00 0.00 2.00 4.00 6.00
0.00
0.10
0.20 0.30 Frequency
0.40
0.50
If the lter completely removed the stochastic cycles corresponding to the unwanted frequencies, the periodogram would be a at line at the minimum value of 6 outside the range identied by the vertical lines. That the periodogram takes on values greater than 6 outside the specied range indicates the inability of the BK lter to pass through only stochastic cycles at frequencies inside the specied band.
6.00 4.00 2.00
0.00
2.00
4.00
6.00
397
We can also evaluate the BK lter by plotting its gain function against the gain function of an ideal lter. In the output below, we reestimate the business-cycle component to store the gain of the BK lter for the specied parameters. (The coefcients and the gain of the BK lter are completely determined by the specied minimum period, the maximum period, and the order of the SMA lter.) We label the variable bkgain for the graph below.
. drop ip_bk . tsfilter bk ip_bk = ip_ln, minperiod(6) maxperiod(32) gain(bkgain abk) . label variable bkgain "BK filter"
Below we generate ideal, the gain function of the ideal band-pass lter at the frequencies f. Then we plot the gain of the ideal lter and the gain of the BK lter.
. generate f = _pi*(_n-1)/_N . generate ideal = cond(f<_pi/16, 0, cond(f<_pi/3, 1,0)) . label variable ideal "Ideal filter" . twoway line ideal f || line bkgain abk
0 0
.5
1 Ideal filter
2 BK filter
The graph reveals that the gain of the BK lter deviates markedly from the square-wave gain of the ideal lter. Increasing the symmetric moving average via the smaorder() option will cause the gain of the BK lter to more closely approximate the gain of the ideal lter at the cost of lost observations in the ltered series.
Filtering a random walk: The ChristianoFitzgerald lter

Although Baxter and King (1999) minimized the error between the coefcients in their lter and the ideal band-pass lter, Christiano and Fitzgerald (2003) minimized the mean squared error between the estimated component and the true component, assuming that the raw series is a random-walk process. Christiano and Fitzgerald (2003) give three important reasons for using their lter: 1. The true dependence structure of the data affects which lter is optimal. 2. Many economic time series are well approximated by random-walk processes.
398
3. Their lter does a good job passing through stochastic cycles of desired frequencies and blocking stochastic cycles from unwanted frequencies on a range of processes that are close to being a random-walk process. The CF lter obtains its optimality properties at the cost of an additional parameter that must be estimated and a loss of robustness. The CF lter is optimal for a random-walk process. If the true process is a random walk with drift, then the drift term must be estimated and removed; see [TS] tslter cf for details. The CF lter is not symmetric, so it will not remove second-order deterministic or second-order integrated processes. tsfilter cf also implements another lter that Christiano and Fitzgerald (2003) derived that is an SMA lter with coefcients that sum to zero. This lter is designed to be as close as possible to the random-walk optimal lter under the constraint that it be an SMA lter with constraints that sum to zero; see [TS] tslter cf for details.
Technical note
A random-walk process is a rst-order integrated process; it must be differenced once to produce a stationary process. Formally, a random-walk process is given by yt = yt1 + t , where t is a zeromean stationary random variable. A random-walk-plus-drift process is given by yt = + yt1 + t , where t is a zero-mean stationary random variable.
Example 4
In this example, we use the CF lter to estimate the business-cycle component, and we plot the periodogram of the CF estimates. We specify the drift option because ip ln is well approximated by a random-walk-plus-drift process.
. tsfilter cf ip_cf = ip_ln, minperiod(6) maxperiod(32) drift . pergram ip_cf, xline(0.03125 0.16667)
ip_ln cyclical component from cf filter Log Periodogram 6.00 4.00 2.00 0.00 2.00 4.00 6.00
0.00
0.10
0.20 0.30 Frequency
0.40
0.50
The periodogram of the CF estimates of the business-cycle component indicates that the CF lter did a better job than the BK lter of passing through only the desired stochastic cycles. Given that ip ln is well approximated by a random-walk-plus-drift process, the relative performance of the CF lter is not surprising.
6.00 4.00 2.00
0.00
2.00
4.00
6.00
399
As with the BK lter, plotting the gain of the CF lter and the gain of the ideal lter gives an impression of how well the lter isolates the specied components. In the output below, we reestimate the business-cycle component, using the gain() option to store the gain of the CF lter, and we plot the gain functions.
. drop ip_cf . tsfilter cf ip_cf = ip_ln, minperiod(6) maxperiod(32) drift gain(cfgain acf) . label variable cfgain "CF filter" . twoway line ideal f || line cfgain acf
1.5 0 0 .5 1
1 Ideal filter
2 CF filter
Comparing this graph with the graph of the BK gain function reveals that the CF lter is closer to the gain of the ideal lter than is the BK lter. The graph also reveals that the gain of the CF lter oscillates above and below 1 for desired frequencies.
The choice between the BK or the CF lter is one between robustness or efciency. The BK lter handles a broader class of stochastic processes, but the CF lter produces a better estimate of ct if yt is close to a random-walk process or a random-walk-plus-drift process.
A one-parameter high-pass lter: The HodrickPrescott lter

Hodrick and Prescott (1997) motivated the HodrickPrescott (HP) lter as a trend-removal technique that could be applied to data that came from a wide class of data-generating processes. In their view, the technique specied a trend in the data, and the data were ltered by removing the trend. The smoothness of the trend depends on a parameter . The trend becomes smoother as . Hodrick and Prescott (1997) recommended setting to 1,600 for quarterly data. King and Rebelo (1993) showed that removing a trend estimated by the HP lter is equivalent to a high-pass lter. They derived the gain function of this high-pass lter and showed that the lter would make integrated processes of order 4 or less stationary, making the HP lter comparable with the band-pass lters discussed above.
400
Example 5
We begin by applying the HP high-pass lter to ip ln and plotting the periodogram of the estimated business-cycle component. We specify the gain() option because will use the gain of the lter in the next example.
. tsfilter hp ip_hp = ip_ln, gain(hpg1600 ahp1600) . label variable hpg1600 "HP(1600) filter" . pergram ip_hp, xline(0.03125)
ip_ln cyclical component from hp filter Log Periodogram 6.00 4.00 2.00 0.00 2.00 4.00 6.00
0.00
0.10
0.20 0.30 Frequency
0.40
0.50
Because the HP lter is a high-pass lter, the high-frequency stochastic cycles corresponding to those periods below 6 remain in the estimated component. Of more concern is the presence of the low-frequency stochastic cycles that the lter should remove. We address this issue in the example below.
Example 6
Hodrick and Prescott (1997) argued that the smoothing parameter should be 1,600 on the basis of a heuristic argument that specied values for the variance of the cyclical component and the variance of the second difference of the trend component, both recorded at quarterly frequencies. In this example, we choose the smoothing parameter to be 677.13, which sets the gain of the lter to 0.5 at the frequency corresponding to 32 periods, as explained in the technical note below. We then plot the periodogram of the ltered series.
6.00 4.00 2.00
0.00
2.00
4.00
6.00

. tsfilter hp ip_hp2 = ip_ln, smooth(677.13) . label variable hpg677 "HP(677) filter" . pergram ip_hp, xline(0.03125) gain(hpg677 ahp677)
401
ip_ln cyclical component from hp filter Log Periodogram 6.00 4.00 2.00 0.00 2.00 4.00 6.00
0.00
0.10
0.20 0.30 Frequency
0.40
0.50
Although the periodogram looks better than the periodogram with the default smoothing, the HP lter still did not zero out the low-frequency stochastic cycles as well as the CF lter did. We take another look at this issue by plotting the gain functions for these lters along with the gain function from the ideal band-pass lter.
. twoway line ideal f || line hpg677 ahp677
.2
.4
.6
.8
1 Ideal filter
2 HP(677) filter
Comparing the gain graphs reveals that the gain of the CF lter is closest to the gain of the ideal lter. Both the BK and the HP lters allow some low-frequency stochastic cycles to pass through. The plot also illustrates that the HP lter is a high-pass lter because its gain is 1 for those stochastic cycles at frequencies above 6 periods, whereas the other gain functions go to zero.
6.00 4.00 2.00
0.00
2.00
4.00
6.00
402
Technical note
Conventionally, economists have used = 1600, which Hodrick and Prescott (1997) recommended for quarterly data. Ravn and Uhlig (2002) derived values for at monthly and annual frequencies that are rescalings of the conventional = 1600 for quarterly data. These heuristic values are the default values; see [TS] tslter hp for details. In the lter literature, lter parameters are set as functions of the cutoff frequency; see Pollock (2000, 324), for instance. This method nds the lter parameter that sets the gain of the lter equal to 1/2 at the cutoff frequency. Applying this method to selecting at the cutoff frequency of 32 periods requires solving
1/2 =
4 {1 cos(2/32)}
2 2
1 + 4 {1 cos(2/32)}
for , which yields 677.13, which was used in the previous example. The gain function of the HP lter is a function of the parameter , and sets both the location of the cutoff frequency and the slope of the gain function. The graph below illustrates this dependence by plotting the gain function of the HP lter for set to 10, 677.13, and 1,600 along with the gain function for the ideal band-pass lter with cutoff periods of 32 periods and 6 periods.
1 0 0 .2 .4 .6 .8
1 Ideal filter HP(677) filter
2 HP(10) filter HP(1600) filter
A two-parameter high-pass lter: The Butterworth lter

Engineers have used Butterworth lters for a long time because they are maximally at. The gain functions of these lters are as close as possible to being a at line at 0 for the unwanted periods and a at line at 1 for the desired periods; see Butterworth (1930) and Bianchi and Sorrentino (2007, 1720). Pollock (2000) showed that Butterworth lters can be derived from some axioms that specify properties we would like a lter to have. Although the Butterworth and BK lters share the properties of symmetry and phase neutrality, the coefcients of Butterworth lters do not need to sum to zero. (Phase-neutral lters do not shift the signal forward or backward in time; see Pollock [1999].) Although the BK lter relies on the detrending properties of SMA lters with coefcients that sum to zero, Pollock (2000) shows that Butterworth lters have detrending properties that depend on the lters parameters.
403
tsfilter bw implements the high-pass Butterworth lter using the computational method that Pollock (2000) derived. This lter has two parameters: the cutoff period and the order of the lter denoted by m. The cutoff period sets the location where the gain function starts to lter out the high-period (low-frequency) stochastic cycles, and m sets the slope of the gain function for a given cutoff period. For a given cutoff period, the slope of the gain function at the cutoff period increases with m. For a given m, the slope of the gain function at the cutoff period increases with the cutoff period. We cannot obtain a vertical slope at the cutoff frequency, which is the ideal, because the computation becomes unstable; see Pollock (2000). The m for which the computation becomes unstable depends on the cutoff period. Pollock (2000) and G mez (1999) argue that the additional exibility produced by the additional o parameter makes the high-pass Butterworth lter a better lter than the HP lter for estimating the cyclical components. Pollock (2000) shows that the high-pass Butterworth lter can estimate the desired components of the dth difference of a dth-order integrated process as long as m d.
Example 7
Below we use tsfilter bw to estimate the components driven by stochastic cycles greater than 32 periods using Butterworth lters of order 2 and order 6. We also compute, label, and plot the gain functions for each lter.
. tsfilter bw ip_bw1 = ip_ln, gain(bwgain1 abw1) maxperiod(32) order(2) . label variable bwgain1 "BW 2" . tsfilter bw ip_bw6 = ip_ln, gain(bwgain6 abw6) maxperiod(32) order(6) . label variable bwgain6 "BW 6" . twoway line ideal f || line bwgain1 abw1 || line bwgain6 abw6
0 0
.2
.4
.6
.8
1 Ideal filter BW 6
2 BW 2
The graph illustrates that the slope of the gain function increases with the order of the lter. The graph below provides another perspective by plotting the gain function from the ideal band-pass lter on a graph with plots of the gain functions from the Butterworth lter of order 6, the CF lter, and the HP(677) lter.
404

. twoway line ideal f || line bwgain6 abw6 || line cfgain acf > || line hpg677 ahp677
1.25 0 0 .25 .5 .75 1
1 Ideal filter CF filter
2 BW 6 HP(677) filter
Although the slope of the gain function from the CF lter is closer to being vertical at the cutoff frequency, the gain function of the Butterworth lter does not oscillate above and below 1 after it rst reaches the value of 1. The atness of the Butterworth lter below and above the cutoff frequency is not an accident, it is one of the lters properties.
Example 8
In the previous example, we used the Butterworth lter of order 6 to remove low-frequency stochastic cycles, and we stored the results in ip bw6. The Butterworth lter did not address the high-frequency stochastic cycles below 6 periods because it is a high-pass lter. We remove those high-frequency stochastic cycles in this example by keeping the trend produced by reltering the previously ltered series. This example uses a common trick: keeping the trend produced by a high-pass lter turns that high-pass lter into a low-pass lter. Because we want to remove the high-frequency stochastic cycles still in the previously ltered series ip bw6, we need a low-pass lter. So we keep the trend produced by reltering the previously ltered series. In the output below, we apply a Butterworth lter of order 20 to the previously ltered series ip bw6. We explain why we used order 20 in the next example. We specify the trend() option to keep the low-frequency components from these lters. Then we compute and graph the periodogram for the trend variable.

. > . . tsfilter bw ip_bwu20 = ip_bw6, gain(bwg20 fbw20) maxperiod(6) order(20) trend(ip_bwb) label variable bwg20 "BW upper filter 20" pergram ip_bwb, xline(0.03125 0.16667)
405
ip_bw6 trend component from bw filter Log Periodogram 6.00 4.00 2.00 0.00 2.00 4.00 6.00
0.00
0.10
0.20 0.30 Frequency
0.40
0.50
The periodogram reveals that the two-pass process has passed the original series ip ln through a band-pass lter. It also reveals that the two-pass process did a reasonable job of ltering out the stochastic cycles corresponding to the unwanted frequencies.
Example 9
In the previous example, when the cutoff period was 6, we set the order of the Butterworth lter to 20. In contrast, in example 7, when the cutoff period was 32, we set the order of the Butterworth lter to 6. We had to increase lter order because the slope of the gain function of the Butterworth lter is increasing with the cutoff period. We needed a larger lter order to get an acceptable slope at the lower cutoff period. We illustrate this point in the output below. We apply Butterworth lters of orders 1 and 6 to the previously ltered series ip bw6, we compute the gain functions, we label the gain variables, and then we plot the gain functions from the ideal lter and the Butterworth lters.
6.00 4.00 2.00
0.00
2.00
4.00
6.00
406
. . . . .

tsfilter bw ip_bwu1 = ip_bw6, gain(bwg1 fbw1) maxperiod(6) order(2) label variable bwg1 "BW upper filter 2" tsfilter bw ip_bwu6 = ip_bw6, gain(bwg6 fbw6) maxperiod(6) order(6) label variable bwg6 "BW upper filter 6" twoway line ideal f || line bwg1 fbw1 || line bwg6 fbw6 || line bwg20 fbw20
1 0 0 .2 .4 .6 .8
1 Ideal filter BW upper filter 6
2 BW upper filter 2 BW upper filter 20
Because the cutoff period is 6, the gain functions for m = 2 and m = 6 are much atter than the gain functions for m = 2 and m = 6 in example 7 when the cutoff period was 32. The gain function for m = 20 is reasonably close to vertical, so we used it in example 8. We mentioned above that for any given cutoff period, the computation eventually becomes unstable for larger values of m. For instance, when the cutoff period is 32, m = 20 is not numerically feasible.
Example 10
As a conclusion, we plot the business-cycle components estimated by the CF lter and by the two passes of Butterworth lters. The shaded areas identify recessions. The two estimates are close but the differences could be important. Which estimate is better depends on whether the oscillations around 1 in the graph of the CF gain function (the second graph of example 7) cause more problems than the nonvertical slopes at the cutoff periods that occur in the BW6 gain function of that same graph and the BW upper lter 20 gain function graphed above.

.25
407
.25
1920q1 1930q1 1940q1 1950q1 1960q1 1970q1 1980q1 1990q1 2000q1 2010q1
quarterly time variable Butterworth filter CF filter
There is a long tradition in economics of using models to estimate components. Instead of comparing lters by their gain functions, some authors compare lters by nding underlying models for which the lter parameters are the model parameters. For instance, Harvey and Jaeger (1993), G mez (1999, o 2001), Pollock (2000, 2006), and Harvey and Trimbur (2003) derive models that correspond to the HP or the Butterworth lter. Some of these references also compare components estimated by lters with components estimated by making predictions from estimated models. In effect, these references point out that arima, dfactor, sspace, and ucm (see [TS] arima, [TS] dfactor, [TS] sspace, and [TS] ucm) implement alternative methods to component estimation.

All lters work with both time-series data and panel data when there are many observations on each panel. When used with panel data, the calculations are performed separately within each panel. For these lters, the default minimum and maximum periods of oscillation correspond to the boundaries used by economists (Burns and Mitchell 1946) for business cycles. Burns and Mitchell dened business cycles as oscillations in business data with recurring periods between 1.5 and 8 years. Their denition continues to be cited by economists investigating correlations between business cycles. If yt is a time series, then the cyclical component is
ct = B(L)yt =
j=
bj ytj
where bj are the coefcients of the impulseresponse sequence of some ideal lter. The impulse response sequence is the inverse Fourier transform of either a square wave or step function depending upon whether the lter is a band-pass or high-pass lter, respectively.
408
In nite sequences, it is necessary to approximate this calculation with a nite impulseresponse sequence bj :
n2
ct = Bt (L)yt =
j=n1
bj ytj
The innite-order impulseresponse sequence for the lters implemented in tsfilter are symmetric and time-invariant. In the frequency domain, the relationships between the true cyclical component and its nite estimates respectively are c() = B()y() and
c() = B()y()
where B() and B() are the frequency transfer functions of the lters B and B . The frequency transfer function for B() can be expressed in polar form as
B() = |B()|exp{i()}
where |B()| is the lters gain function and () is the lters phase function. The gain function determines whether the amplitude of the stochastic cycle is increased or decreased at a particular frequency. The phase function determines how a cycle at a particular frequency is shifted forward or backward in time. In this form, it can be shown that the spectrum of the cyclical component, fc (), is related to the spectrum of yt series by the squared gain:
fc () = |B()|2 fy ()
Each of the four lters in tsfilter has an option for returning an estimate of the gain function together with its associated scaled frequency a = / , where 0 . These are consistent estimates of |B()|, the gain from the ideal linear lter. The band-pass lters implemented in tsfilter, the BK and CF lters, use a square wave as the ideal transfer function: 1 if || [l , h ] B() = 0 if || [l , h ] / The high-pass lters, the HodrickPrescott and Butterworth lters, use a step function as the ideal transfer function: 1 if || h B() = 0 if || < h
Acknowledgments
We thank Christopher F. Baum, Boston College, for his previous implementations of these lters: BaxterKing (bking), ChristianoFitzgerald (cfitzrw), HodrickPrescott (hprescott), and Butterworth (butterworth).
409
We also thank D. S. G. Pollock for his helpful responses to our questions about Butterworth lters and the methods that he has developed.
References
Baxter, M., and R. G. King. 1999. Measuring business cycles: Approximate band-pass lters for economic time series. Review of Economics and Statistics 81: 575593. Bianchi, G., and R. Sorrentino. 2007. Electronic Filter Simulation and Design. New York: McGrawHill. Burns, A. F., and W. C. Mitchell. 1946. Measuring Business Cycles. New York: National Bureau of Economic Research. Butterworth, S. 1930. On the theory of lter ampliers. Experimental Wireless and the Wireless Engineer 7: 536541. Christiano, L. J., and T. J. Fitzgerald. 2003. The band pass lter. International Economic Review 44: 435465. Fuller, W. A. 1996. Introduction to Statistical Time Series. 2nd ed. New York: Wiley. G mez, V. 1999. Three equivalent methods for ltering nite nonstationary time series. Journal of Business and o Economic Statistics 17: 109116. . 2001. The use of Butterworth lters for trend and cycle estimation in economic time series. Journal of Business and Economic Statistics 19: 365373. Hamilton, J. D. 1994. Time Series Analysis. Princeton: Princeton University Press. Harvey, A. C., and A. Jaeger. 1993. Detrending, stylized facts and the business cycle. Journal of Applied Econometrics 8: 231247. Harvey, A. C., and T. M. Trimbur. 2003. General model-based lters for extracting cycles and trends in economic time series. The Review of Economics and Statistics 85: 244255. Hodrick, R. J., and E. C. Prescott. 1997. Postwar U.S. business cycles: An empirical investigation. Journal of Money, Credit, and Banking 29: 116. King, R. G., and S. T. Rebelo. 1993. Low frequency ltering and real business cycles. Journal of Economic Dynamics and Control 17: 207231. Leser, C. E. V. 1961. A simple method of trend construction. Journal of the Royal Statistical Society, Series B 23: 91107. Pollock, D. S. G. 1999. A Handbook of Time-Series Analysis, Signal Processing and Dynamics. London: Academic Press. . 2000. Trend estimation and de-trending via rational square-wave lters. Journal of Econometrics 99: 317334. . 2006. Econometric methods of signal extraction. Computational Statistics & Data Analysis 50: 22682292. Priestley, M. B. 1981. Spectral Analysis and Time Series. London: Academic Press. Ravn, M. O., and H. Uhlig. 2002. On adjusting the HodrickPrescott lter for the frequency of observations. Review of Economics and Statistics 84: 371376. Schmidt, T. J. 1994. sts5: Detrending with the HodrickPrescott lter. Stata Technical Bulletin 17: 2224. Reprinted in Stata Technical Bulletin Reprints, vol. 3, pp. 216219. College Station, TX: Stata Press. Wei, W. W. S. 2006. Time Series Analysis: Univariate and Multivariate Methods. 2nd ed. Boston: Pearson.
Also see
[TS] tsset Declare data to be time-series data [XT] xtset Declare data to be panel data [TS] tssmooth Smooth and forecast univariate time-series data
Title
tslter bk BaxterKing time-series lter
Syntax
Filter one variable tsfilter bk type newvar = varname if in , options
Filter multiple variables, unique names tsfilter bk type newvarlist = varlist if in , options
Filter multiple variables, common name stub tsfilter bk options

Main
type
stub* = varlist
if
in
, options
Description lter out stochastic cycles at periods smaller than # lter out stochastic cycles at periods larger than # number of observations in each direction that contribute to each ltered value use calculations for a stationary time series
minperiod(#) maxperiod(#) smaorder(#) stationary

Trend
trend(newvar | newvarlist | stub*) save the trend component(s) in new variable(s)

Gain
gain(gainvar anglevar)
save the gain and angular frequency
You must tsset or xtset your data before using tsfilter; see [TS] tsset and [XT] xtset. varname and varlist may contain time-series operators; see [U] 11.4.4 Time-series varlists.
Menu
Statistics
>
Time series
>
Filters for cyclical components
>
Baxter-King
Description
tsfilter bk uses the Baxter and King (1999) band-pass lter to separate a time series into trend and cyclical components. The trend component may contain a deterministic or a stochastic trend. The stationary cyclical component is driven by stochastic cycles at the specied periods. See [TS] tslter for an introduction to the methods implemented in tsfilter bk.
410
411
Options

Main
minperiod(#) lters out stochastic cycles at periods smaller than #, where # must be at least 2 and less than maxperiod(). By default, if the units of the time variable are set to daily, weekly, monthly, quarterly, or half-yearly, then # is set to the number of periods equivalent to 1.5 years; yearly data use minperiod(2); otherwise, the default value is minperiod(6). maxperiod(#) lters out stochastic cycles at periods larger than #, where # must be greater than minperiod(). By default, if the units of the time variable are set to daily, weekly, monthly, quarterly, half-yearly, or yearly, then # is set to the number of periods equivalent to 8 years; otherwise, the default value is maxperiod(32). smaorder(#) sets the order of the symmetric moving average, denoted by q . The order is an integer that species the number of observations in each direction used in calculating the symmetric moving average estimate of the cyclical component. This number must be an integer greater than zero and less than (T 1)/2. The estimate for the cyclical component for the tth observation, yt , is based upon the 2q + 1 values ytq , ytq+1 , . . . , yt , yt+1 , . . . , yt+q . By default, if the units of the time variable are set to daily, weekly, monthly, quarterly, half-yearly, or yearly, then # is set to the equivalent of 3 years; otherwise, the default value is smaorder(12). stationary modies the lter calculations to those appropriate for a stationary series. By default, the series is assumed nonstationary.
Trend
trend(newvar | newvarlist | stub*) saves the trend component(s) in the new variable(s) specied by newvar, newvarlist, or stub*.
Gain
gain(gainvar anglevar) saves the gain in gainvar and its associated angular frequency in anglevar. Gains are calculated at the N angular frequencies that uniformly partition the interval (0, ], where N is the sample size.
Remarks
We assume that you have already read [TS] tslter, which provides an introduction to ltering and the methods implemented in tsfilter bk, more examples using tsfilter bk, and a comparison of the four lters implemented by tsfilter. In particular, an understanding of gain functions as presented in [TS] tslter is required to understand these remarks. tsfilter bk uses the BaxterKing (BK) band-pass lter to separate a time-series yt into trend and cyclical components: yt = t + ct where t is the trend component and ct is the cyclical component. t may be nonstationary; it may contain a deterministic or a stochastic trend, as discussed below. The primary objective is to estimate ct , a stationary cyclical component that is driven by stochastic cycles within a specied range of periods. The trend component t is calculated by the difference t = yt ct . Although the BK band-pass lter implemented in tsfilter bk has been widely applied by macroeconomists, it is a general time-series method and may be of interest to other researchers.
412
Symmetric moving-average (SMA) lters with coefcients that sum to zero remove stochastic and deterministic trends of rst and second order; see Fuller (1996), Baxter and King (1995), and Baxter and King (1999). For an innitely long series, there is an ideal band-pass lter for which the gain function is 1 for [0 , 1 ] and 0 for all other frequencies; see [TS] tslter for an introduction to gain functions. It just so happens that this ideal band-pass lter is an SMA lter with coefcients that sum to zero. Baxter and King (1999) derive the coefcients of this ideal band-pass lter and then dene the BK lter to be the SMA lter with 2q + 1 terms that are as close as possible to those of the ideal lter. There is a trade-off in choosing q : larger values of q cause the gain of the BK lter to be closer to the gain of the ideal lter, but they also increase the number of missing observations in the ltered series. The smaorder() option species q . The default value of smaorder() is the number of periods equivalent to 3 years, following the Baxter and King (1999) recommendation. Although the mathematics of the frequency-domain approach to time-series analysis is in terms of stochastic cycles at frequencies [, ], applied work is generally in terms of periods p, where p = 2/ . So tsfilter bk has the minperiod() and maxperiod() options to specify the desired range of stochastic cycles. Among economists, the BK lter is commonly used for investigating business cycles. Burns and Mitchell (1946) dened business cycles as stochastic cycles in business data corresponding to periods between 1.5 and 8 years. The default values for minperiod() and maxperiod() are the Burns Mitchell values of 1.5 and 8 years, scaled to the frequency of the dataset. The calculations of the default values assume that the time variable is formatted as daily, weekly, monthly, quarterly, half-yearly, or yearly; see [D] format. For each variable, the band-pass BK lter estimate of ct is put in the corresponding new variable, and when the trend() option is specied, the estimate of t is put in the corresponding new variable. tsfilter bk automatically detects panel data from the information provided when the dataset was tsset or xtset. All calculations are done separately on each panel. Missing values at the beginning and end of the sample are excluded from the sample. The sample may not contain gaps. Baxter and King (1999) derived their method for nonstationary time series, but they noted that a small modication makes it applicable to stationary time series. Imposing the condition that the lter coefcients sum to zero is what makes their method applicable to nonstationary time series; dropping this condition yields a lter for stationary time series. Specifying the stationary option causes tsfilter bk to use coefcients calculated without the constraint that they sum to zero.
Example 1
In this and the subsequent examples, we use tsfilter bk to estimate the business-cycle component of the natural log of real gross domestic product (GDP) of the United States. Our sample of quarterly data goes from 1952q1 to 2010q4. Below we read in and plot the data.

. use http://www.stata-press.com/data/r12/gdp2 (Federal Reserve Economic Data, St. Louis Fed) . tsline gdp_ln
9.5
413
7.5
1950q1
natural log of real GDP 8.5 9
1960q1
1970q1
1980q1
1990q1
2000q1
2010q1
The series is nonstationary and is thus a candidate for the BK lter. Below we use tsfilter bk to lter gdp ln, and we use pergram (see [TS] pergram) to compute and to plot the periodogram of the estimated cyclical component.
. tsfilter bk gdp_bk = gdp_ln . pergram gdp_bk, xline(.03125 .16667)
gdp_ln cyclical component from bk filter Log Periodogram 6.00 4.00 2.00 0.00 2.00 4.00 6.00
0.00
0.10
0.20 0.30 Frequency
0.40
0.50
Because our sample is of quarterly data, tsfilter bk used the default values of minperiod(6), maxperiod(32), and smaorder(12). The minimum and maximum periods are the Burns and Mitchell (1946) business-cycle periods for quarterly data. The default of smaorder(12) was recommend by Baxter and King (1999) for quarterly data. In the periodogram, we added vertical lines at the natural frequencies corresponding to the conventional Burns and Mitchell (1946) values for business-cycle components. pergram displays the
6.00 4.00 2.00
0.00
2.00
4.00
6.00
414
results in natural frequencies, which are the standard frequencies divided by 2 . We use the xline() option to draw vertical lines at the lower natural-frequency cutoff (1/32 = 0.03125) and the upper natural-frequency cutoff (1/6 0.16667). If the lter completely removed the stochastic cycles at the unwanted frequencies, the periodogram would be a at line at the minimum value of 6 outside the range identied by the vertical lines. The periodogram reveals that the default value of smaorder(12) did not do a good job of ltering out the high-periodicity stochastic cycles, because there are too many points above 6.00 to the left of the left-hand vertical line. It also reveals that the lter did not remove enough low-periodicity stochastic cycles, because there are too many points above 6.00 to the right of the right-hand vertical line. We address these problems in the next example.
Example 2
In this example, we change the symmetric moving average of the lter via the smaorder() option so that it will remove more of the unwanted stochastic cycles. As mentioned, larger values of q cause the gain of the BK lter to be closer to the gain of the ideal lter, but larger values also increase the number of missing observations in the ltered series. In the output below, we estimate the business-cycle component and compute the gain functions when the SMA-order of the lter is 12 and when it is 20. We also generate ideal, the gain function of the ideal band-pass lter at the frequencies f. Then we plot the gain functions from all three lters.
. tsfilter bk gdp_bk12 = gdp_ln, gain(g12 a12) . label variable g12 "BK SMA-order 12" . tsfilter bk gdp_bk20 = gdp_ln, gain(g20 a20) smaorder(20) . label variable g20 "BK SMA-order 20" . generate f = _pi*(_n-1)/_N . generate ideal = cond(f<_pi/16, 0, cond(f<_pi/3, 1,0)) . label variable ideal "Ideal filter" . twoway line ideal f || line g12 a12 || line g20 a20
0 0
.5
1 Ideal filter BK SMAorder 20
2 BK SMAorder 12
415
As discussed in [TS] tslter, the gain function of the ideal lter is a square wave with a value of 0 at the frequencies corresponding to unwanted frequencies and a value of 1 at the desired frequencies. The vertical lines in the gain function of the ideal lter occur at the frequencies /16, corresponding to 32 periods, and at /3, corresponding to 6 periods. (Given that p = 2/ , where p is the period corresponding to the frequency , the frequency is given by 2/p.) The differences between the gain function of the lter with SMA-order 12 and the gain function of the ideal band-pass lter is the root of the issues mentioned at the end of example 1. The lter with SMA-order 20 is closer to the gain function of the ideal band-pass lter at the cost of 16 more missing values in the ltered series. Below we compute and graph the periodogram of the series ltered with SMA-order 20.
. pergram gdp_bk20, xline(.03125 .16667)
gdp_ln cyclical component from bk filter Log Periodogram 6.00 4.00 2.00 0.00 2.00 4.00 6.00
0.00
0.10
0.20 0.30 Frequency
0.40
0.50
The above periodogram indicates that the lter of SMA-order 20 removed more of the stochastic cycles at the unwanted periodicities than did the lter of SMA-order 12. Whether removing the stochastic cycles at the unwanted periodicities is worth losing more observations in the ltered series is a judgment call.
6.00 4.00 2.00
0.00
2.00
4.00
6.00
416
Below we plot the estimated business-cycle component with recessions identied by the shaded areas.
.04 .04 gdp_ln cyclical component from bk filter .02 0 .02
1957q3
1969q3
1981q3
1993q3
2005q3
quarterly time variable gdp_ln cyclical component from bk filter
Saved results
tsfilter bk saves the following in r():
Scalars r(smaorder) r(minperiod) r(maxperiod) Macros r(varlist) r(filterlist) r(trendlist) r(method) r(stationary) r(unit) Matrices r(filter) order of the symmetric moving average minimum period of stochastic cycles maximum period of stochastic cycles original time-series variables variables containing estimates of the cyclical components variables containing estimates of the trend components, if trend() was specied Baxter-King yes or no, indicating whether the calculations assumed the series was or was not stationary units of time variable set using tsset or xtset
(q+1)1 matrix of lter weights, where q is the order of the symmetric moving average

tsfilter bk is implemented as an ado-le. Baxter and King (1999) showed that there is an innite-order SMA lter with coefcients that sum to zero that can extract the specied components from a nonstationary time series. The components are specied in terms of the minimum and maximum periods of the stochastic cycles that drive these components in the frequency domain. This ideal lter is not feasible, because the constraints imposed on the lter can only be satised using an innite number of coefcients, so Baxter and King (1999) derived a nite approximation to this ideal lter. The innite-order, ideal band-pass lter obtains the cyclical component with the calculation
ct =
j=
bj ytj
417
Letting pl and ph be the minimum and maximum periods of the stochastic cycles of interest, the weights bj in this calculation are given by
bj =
1 (h l )
if j = 0
(j)1 {sin(jh ) sin(jl )} if j = 0
where l = 2/pl and h = 2/ph are the lower and higher cutoff frequencies, respectively. For the default case of nonstationary time series with nite length, the ideal band-pass lter cannot be used without modication. Baxter and King (1999) derived modied weights for a nite order SMA lter with coefcients that sum to zero. As a result, Baxter and King (1999) estimate ct by
+q
ct =
j=q
bj ytj
The coefcients bj in this calculation are equal to bj = bj bq , where bj = bj and bq is the mean of the ideal coefcients truncated at q :
q
bq = (2q + 1)1
j=q +q
bj
Note that j=q bj = 0 and that the rst and last q values of the cyclical component cannot be estimated using this lter. If the stationary option is set, the BK lter sets the coefcients to the ideal coefcients, that is, q bj = bj . For these weights, bj = bj , and although j= bj = 0, for small q , q bj = 0.
References
Baxter, M., and R. G. King. 1995. Measuring business cycles approximate band-pass lters for economic time series. NBER Working Paper No. 5022, National Bureau of Economic Research. http://www.nber.org/papers/w5022. . 1999. Measuring business cycles: Approximate band-pass lters for economic time series. Review of Economics and Statistics 81: 575593. Burns, A. F., and W. C. Mitchell. 1946. Measuring Business Cycles. New York: National Bureau of Economic Research. Fuller, W. A. 1996. Introduction to Statistical Time Series. 2nd ed. New York: Wiley. Pollock, D. S. G. 1999. A Handbook of Time-Series Analysis, Signal Processing and Dynamics. London: Academic Press. . 2006. Econometric methods of signal extraction. Computational Statistics & Data Analysis 50: 22682292.
Also see
[TS] tsset Declare data to be time-series data [XT] xtset Declare data to be panel data [TS] tslter Filter a time-series, keeping only selected periodicities [D] format Set variables output format [TS] tssmooth Smooth and forecast univariate time-series data
Title
tslter bw Butterworth time-series lter
Syntax
Filter one variable tsfilter bw type newvar = varname if in , options
Filter multiple variables, unique names tsfilter bw type newvarlist = varlist if in , options
Filter multiple variables, common name stub tsfilter bw options

Main
type
stub* = varlist
if
in
, options
Description lter out stochastic cycles at periods larger than # set the order of the lter; default is order(2)
maxperiod(#) order(#)
Trend
trend(newvar|newvarlist|stub*) save the trend component(s) in new variable(s)

Gain
Menu
Statistics
>
Time series
>
>
Butterworth
Description
tsfilter bw uses the Butterworth high-pass lter to separate a time series into trend and cyclical components. The trend component may contain a deterministic or a stochastic trend. The stationary cyclical component is driven by stochastic cycles at the specied periods. See [TS] tslter for an introduction to the methods implemented in tsfilter bw.
418
419
Options

Main
maxperiod(#) lters out stochastic cycles at periods larger than #, where # must be greater than 2. By default, if the units of the time variable are set to daily, weekly, monthly, quarterly, half-yearly, or yearly, then # is set to the number of periods equivalent to 8 years; otherwise, the default value is maxperiod(32). order(#) sets the order of the Butterworth lter, which must be an integer. The default is order(2).
Trend
Gain
Remarks
We assume that you have already read [TS] tslter, which provides an introduction to ltering and the methods implemented in tsfilter bw, more examples using tsfilter bw, and a comparison of the four lters implemented by tsfilter. In particular, an understanding of gain functions as presented in [TS] tslter is required to understand these remarks. tsfilter bw uses the Butterworth high-pass lter to separate a time-series yt into trend and cyclical components: yt = t + ct where t is the trend component and ct is the cyclical component. t may be nonstationary; it may contain a deterministic or a stochastic trend, as discussed below. The primary objective is to estimate ct , a stationary cyclical component that is driven by stochastic cycles within a specied range of periods. The trend component t is calculated by the difference t = yt ct . Although the Butterworth high-pass lter implemented in tsfilter bw has been widely applied by macroeconomists and engineers, it is a general time-series method and may be of interest to other researchers. Engineers have used Butterworth lters for a long time because they are maximally at. The gain functions of these lters are as close as possible to being a at line at 0 for the unwanted periods and a at line at 1 for the desired periods; see Butterworth (1930) and Bianchi and Sorrentino (2007, 1720). (See [TS] tslter for an introduction to gain functions.) The high-pass Butterworth lter is a two-parameter lter. The maxperiod() option species the maximum period; the stochastic cycles of all higher periodicities are ltered out. The maxperiod() option sets the location of the cutoff period in the gain function. The order() option species the order of the lter, which determines the slope of the gain function at the cutoff frequency. For a given cutoff period, the slope of the gain function at the cutoff period increases with lter order. For a given lter order, the slope of the gain function at the cutoff period increases with the cutoff period.
420
We cannot obtain a vertical slope at the cutoff frequency, which is the ideal, because the computation becomes unstable; see Pollock (2000). The lter order for which the computation becomes unstable depends on the cutoff period. Among economists, the high-pass Butterworth lter is commonly used for investigating business cycles. Burns and Mitchell (1946) dened business cycles as stochastic cycles in business data corresponding to periods between 1.5 and 8 years. For this reason, the default value for maxperiod() is the number of periods in 8 years, if the time variable is formatted as daily, weekly, monthly, quarterly, half-yearly, or yearly; see [D] format. The default value for maxperiod() is 32 for all other time formats. For each variable, the high-pass Butterworth lter estimate of ct is put in the corresponding new variable, and when the trend() option is specied, the estimate of t is put in the corresponding new variable. tsfilter bw automatically detects panel data from the information provided when the dataset was tsset or xtset. All calculations are done separately on each panel. Missing values at the beginning and end of the sample are excluded from the sample. The sample may not contain gaps.
Example 1
In this and the subsequent examples, we use tsfilter bw to estimate the business-cycle component of the natural log of the real gross domestic product (GDP) of the United States. Our sample of quarterly data goes from 1952q1 to 2010q4. Below we read in and plot the data.
9.5 7.5
1950q1
1960q1
1970q1
1980q1
1990q1
2000q1
2010q1
The series is nonstationary. Pollock (2000) shows that the high-pass Butterworth lter can estimate the components driven by the stochastic cycles at the specied frequencies when the original series is nonstationary. Below we use tsfilter bw to lter gdp ln and use pergram (see [TS] pergram) to compute and to plot the periodogram of the estimated cyclical component.
. tsfilter bw gdp_bw = gdp_ln . pergram gdp_bw, xline(.03125 .16667)
421
gdp_ln cyclical component from bw filter Log Periodogram 6.00 4.00 2.00 0.00 2.00 4.00 6.00
0.00
0.10
0.20 0.30 Frequency
0.40
0.50
tsfilter bw used the default value of maxperiod(32) because our sample is of quarterly data. In the periodogram, we added vertical lines at the natural frequencies corresponding to the conventional Burns and Mitchell (1946) values for business-cycle components. pergram displays the results in natural frequencies, which are the standard frequencies divided by 2 . We use option xline() to draw vertical lines at the lower natural-frequency cutoff (1/32 = 0.03125) and the upper natural-frequency cutoff (1/6 0.16667). If the lter completely removed the stochastic cycles at the unwanted frequencies, the periodogram would be a at line at the minimum value of 6 outside the range identied by the vertical lines. The periodogram reveals two issues. First, it indicates that the default value of order(2) did not do a good job of ltering out the high-periodicity stochastic cycles, because there are too many points above 6.00 to the left of the left-hand vertical line. Second, it reveals the high-pass nature of the lter, because none of the low-period (high-frequency) stochastic cycles have been ltered out. We cope with these two issues in the remaining examples.
Example 2
In this example, we change the order of the lter so that it will remove more of the unwanted low-frequency stochastic cycles. As previously mentioned, increasing the order of the lter increases the slope of the gain function at the cutoff period. For orders 2 and 8, we compute the ltered series, compute the gain functions, and label the gain variables. We also generate ideal, the gain function of the ideal band-pass lter at the frequencies f. Then we plot the gain function of the ideal band-pass lter and the gain functions of the high-pass Butterworth lters of orders 2 and 8.
. . . . . . . tsfilter bw gdp_bw2 = gdp_ln, gain(g1 a1) label variable g1 "BW order 2" tsfilter bw gdp_bw8 = gdp_ln, gain(g8 a8) order(8) label variable g8 "BW order 8" generate f = _pi*(_n-1)/_N generate ideal = cond(f<_pi/16, 0, cond(f<_pi/3, 1,0)) label variable ideal "Ideal filter"
6.00 4.00 2.00
0.00
2.00
4.00
6.00
422

. twoway line ideal f || line g1 a1 || line g8 a8
0 0
.2
.4
.6
.8
1 Ideal filter BW order 8
2 BW order 2
As discussed in [TS] tslter, the gain function of the ideal lter is a square wave with a value of 0 at the frequencies corresponding to unwanted frequencies and a value of 1 at the desired frequencies. The vertical lines in the gain function of the ideal lter occur at /16, corresponding to 32 periods, and at /3, corresponding to 6 periods. (Given that p = 2/ , where p is the period corresponding to frequency , the frequency is given by 2/p.) The distance between the gain function of the lter with order 2 and the gain function of the ideal band-pass lter at /16 is the root of the rst issue mentioned at the end of example 1. The lter with order 8 is much closer to the gain function of the ideal band-pass lter at /16 than is the lter with order 2. That both gain functions are 1 to the right of the vertical line at /3 reveals the high-pass nature of the lter.
Example 3
In this example, we use a common trick to resolve the second issue mentioned at the end of example 1. Keeping the trend produced by a high-pass lter turns that high-pass lter into a low-pass lter. Because we want to remove the high-frequency stochastic cycles still in the previously ltered series gdp bw8, we need to run gdp bw8 through a low-pass lter. So we keep the trend produced by reltering the previously ltered series. To determine an order for the lter, we run the lter with order(8), then with order(15), and then we plot the gain functions along with the gain function of the ideal lter.
. tsfilter bw gdp_bwn8 = gdp_bw8, gain(gc8 ac8) order(8) > maxperiod(6) trend(gdp_bwc8) . label variable gc8 "BW order 8" . tsfilter bw gdp_bwn15 = gdp_bw8, gain(gc15 ac15) order(15) > maxperiod(6) trend(gdp_bwc15) . label variable gc15 "BW order 15" . twoway line ideal f || line gc8 ac8 || line gc15 ac15
423
0 0
.2
.4
.6
.8
1 Ideal filter BW order 15
2 BW order 8
We specied much higher orders for the lter in this example because the cutoff period is 6 instead of 32. (As previously mentioned, holding the order of the lter constant, the slope of the gain function at the cutoff period decreases when the period decreases.) The above graph indicates that the lter with order(15) is reasonably close to the gain function of the ideal lter. Now we compute and plot the periodogram of the estimated business-cycle component.
. pergram gdp_bwc15, xline(.03125 .16667)
gdp_bw8 trend component from bw filter Log Periodogram 6.00 4.00 2.00 0.00 2.00 4.00 6.00
0.00
0.10
0.20 0.30 Frequency
0.40
0.50
The graph indicates that the above applications of the Butterworth lter did a reasonable job of ltering out the high-periodicity stochastic cycles but that the low-periodicity stochastic cycles have not been completely removed.
6.00 4.00 2.00
0.00
2.00
4.00
6.00
424
gdp_bw8 trend component from bw filter .02 0 .02 .04 .04
1950q1
1960q1
1970q1
1980q1
1990q1
2000q1
2010q1
quarterly time variable gdp_bw8 trend component from bw filter
Saved results
tsfilter bw saves the following in r():
Scalars r(order) r(maxperiod) Macros r(varlist) r(filterlist) r(trendlist) r(method) r(unit) order of the lter maximum period of stochastic cycles original time-series variables variables containing estimates of the cyclical components variables containing estimates of the trend components, if trend() was specied Butterworth units of time variable set using tsset or xtset

tsfilter bw is implemented as an ado-le. tsfilter bw uses the computational methods described in Pollock (2000) to implement the lter. Pollock (2000) shows that the gain of the Butterworth high-pass lter is given by
() = 1 +
tan(c /2) tan(/2)
2m 1
where m is the order of the lter, c = 2/ph is the cutoff frequency, and ph is the maximum period. Here is an outline of the computational procedure that Pollock (2000) derived. Pollock (2000) showed that the Butterworth lter corresponds to a particular model. Actually, his model is more general than the Butterworth lter, but tsfilter bw restricts the computations to the case in which the model corresponds to the Butterworth lter.
425
The model represents the series to be ltered, yt , in terms of zero mean, covariance stationary, and independent and identically distributed shocks t and t :
yt =
(1 + L)m t + t (1 L)m
From this model, Pollock (2000) shows that the optimal estimate for the cyclical component is given by c = Q(L + H )1 Q y
2 2 where Var{Q (y c)} = L and Var{Q c} = H . Here L and H are symmetric Toeplitz matrices with 2m + 1 nonzero diagonal bands and generating functions (1 + z)m (1 + z 1 )m and (1 z)m (1 z 1 )m , respectively.
The parameter in this expression is a function of ph (the maximum period of stochastic cycles ltered out) and the order of the lter:
= {tan(/ph )}2m
The matrix Q in this expression is a function of the coefcients in the polynomial (1 L)d = 1 + 1 L + + d Ld :
d . . . 0 Q = 0 . . . 0 0
. . . 1 . .. . . . . . . d ... 0 . . . ... ... 0 0
1 . . . d1 d . . . 0 0
... 0 . .. . . . ... 1 . . . 1 .. . . . . d ... 0
0 . . . 0 1 d1 d
... ... ... .. .
0 . . . 0 0 . . .
... 1 . . . 1
(T d)T 0 . . . 0 0 . . . 0 1
It can be shown that H = Q Q and L = |H |, which simplies the calculation of the cyclical component to c = Q{|Q Q| + (Q Q)}1 Q y
References
Bianchi, G., and R. Sorrentino. 2007. Electronic Filter Simulation and Design. New York: McGrawHill. Burns, A. F., and W. C. Mitchell. 1946. Measuring Business Cycles. New York: National Bureau of Economic Research. Butterworth, S. 1930. On the theory of lter ampliers. Experimental Wireless and the Wireless Engineer 7: 536541. Pollock, D. S. G. 1999. A Handbook of Time-Series Analysis, Signal Processing and Dynamics. London: Academic Press. . 2000. Trend estimation and de-trending via rational square-wave lters. Journal of Econometrics 99: 317334. . 2006. Econometric methods of signal extraction. Computational Statistics & Data Analysis 50: 22682292.
426
Also see
Title
tslter cf ChristianoFitzgerald time-series lter
Syntax
Filter one variable tsfilter cf type newvar = varname if in , options
Filter multiple variables, unique names tsfilter cf type newvarlist = varlist if in , options
Filter multiple variables, common name stub tsfilter cf options

Main
type
stub* = varlist
if
in
, options
Description lter out stochastic cycles at periods smaller than # lter out stochastic cycles at periods larger than # number of observations in each direction that contribute to each ltered value use calculations for a stationary time series remove drift from the time series
minperiod(#) maxperiod(#) smaorder(#) stationary drift

Trend

Gain
Menu
Statistics
>
Time series
>
>
Christiano-Fitzgerald
Description
tsfilter cf uses the Christiano and Fitzgerald (2003) band-pass lter to separate a time series into trend and cyclical components. The trend component may contain a deterministic or a stochastic trend. The stationary cyclical component is driven by stochastic cycles at the specied periods. See [TS] tslter for an introduction to the methods implemented in tsfilter cf.
427
428
Options

Main
minperiod(#) lters out stochastic cycles at periods smaller than #, where # must be at least 2 and less than maxperiod(). By default, if the units of the time variable are set to daily, weekly, monthly, quarterly, or half-yearly, then # is set to the number of periods equivalent to 1.5 years; yearly data use minperiod(2); otherwise, the default value is minperiod(6). maxperiod(#) lters out stochastic cycles at periods larger than #, where # must be greater than minperiod(). By default, if the units of the time variable are set to daily, weekly, monthly, quarterly, half-yearly, or yearly, then # is set to the number of periods equivalent to 8 years; otherwise, the default value is maxperiod(32). smaorder(#) sets the order of the symmetric moving average, denoted by q . By default, smaorder() is not set, which invokes the asymmetric calculations for the ChristianoFitzgerald lter. The order is an integer that species the number of observations in each direction used in calculating the symmetric moving average estimate of the cyclical component. This number must be an integer greater than zero and less than (T 1)/2. The estimate of the cyclical component for the tth observation, yt , is based upon the 2q + 1 values ytq , ytq+1 , . . . , yt , yt+1 , . . . , yt+q . stationary modies the lter calculations to those appropriate for a stationary series. By default, the series is assumed nonstationary. drift removes drift using the approach described in Christiano and Fitzgerald (2003). By default, drift is not removed.
Trend
Gain
Remarks
We assume that you have already read [TS] tslter, which provides an introduction to ltering and the methods implemented in tsfilter cf, more examples using tsfilter cf, and a comparison of the four lters implemented by tsfilter. In particular, an understanding of gain functions as presented in [TS] tslter is required to understand these remarks. tsfilter cf uses the ChristianoFitzgerald (CF) band-pass lter to separate a time-series yt into trend and cyclical components yt = t + ct where t is the trend component and ct is the cyclical component. t may be nonstationary; it may contain a deterministic or a stochastic trend, as discussed below. The primary objective is to estimate ct , a stationary cyclical component that is driven by stochastic cycles at a specied range of periods. The trend component t is calculated by the difference t = yt ct . Although the CF band-pass lter implemented in tsfilter cf has been widely applied by macroeconomists, it is a general time-series method and may be of interest to other researchers.
429
As discussed by Christiano and Fitzgerald (2003) and in [TS] tslter, if one had an innitely long series, one could apply an ideal band-pass lter that perfectly separates out cyclical components driven by stochastic cycles at the specied periodicities. In nite samples, it is not possible to exactly satisfy the conditions that a lter must fulll to perfectly separate out the specied stochastic cycles; the expansive lter literature reects the trade-offs involved in choosing a nite-length lter to separate out the specied stochastic cycles. Christiano and Fitzgerald (2003) derive a nite-length CF band-pass lter that minimizes the mean squared error between the ltered series and the series ltered by an ideal band-pass lter that perfectly separates out components driven by stochastic cycles at the specied periodicities. Christiano and Fitzgerald (2003) place two important restrictions on the mean squared error problem that their lter solves. First, the CF lter is restricted to be a linear lter. Second, yt is assumed to be a random-walk process; in other words, yt = yt1 + t , where t is independently and identically distributed with mean zero and nite variance. The CF lter is the best linear predictor of the series ltered by the ideal band-pass lter when yt is a random walk. Christiano and Fitzgerald (2003) make four points in support of the random-walk assumption. First, the mean squared error problem solved by their lter requires that the process for yt be specied. Second, they provide a method for removing drift so that their lter handles cases in which yt is a random walk with drift. Third, many economic time series are well approximated by a random-walk-plus-drift process. (We add that many time series encountered in applied statistics are well approximated by a random-walk-plus-drift process.) Fourth, they provide simulation evidence that their lter performs well when the process generating yt is not a random-walk-plus-drift process but is close to being a random-walk-plus-drift process. Comparing the CF lter with the BaxterKing (BK) lter provides some intuition and explains the smaorder() option in tsfilter cf. As discussed in [TS] tslter and Baxter and King (1999), symmetric moving-average (SMA) lters with coefcients that sum to zero can extract the components driven by stochastic cycles at specied periodicities when the series to be ltered has a deterministic or stochastic trend of order 1 or 2. The coefcients of the nite-length BK lter are as close as possible to the coefcients of an ideal SMA band-pass lter under the constraints that the BK coefcients are symmetric and sum to zero. The coefcients of the CF lter are not symmetric nor do they sum to zero, but the CF lter was designed to lter out the specied periodicities when yt has a rst-order stochastic trend. To be robust to second-order trends, Christiano and Fitzgerald (2003) derive a constrained version of the CF lter. The coefcients of the constrained lter are constrained to be symmetric and to sum to zero. Subject to these constraints, the coefcients of the constrained CF lter minimize the mean squared error between the ltered series and the series ltered by an ideal band-pass lter that perfectly separates out the components. Christiano and Fitzgerald (2003) note that the higher-order detrending properties of this constrained lter come at the cost of lost efciency. If the constraints are binding, the constrained lter cannot predict the series ltered by the ideal lter as well as the unconstrained lter can. Specifying the smaorder() option causes tsfilter cf to compute the SMA-constrained CF lter. The choice between the BK and the CF lters is one between robustness and efciency. The BK lter handles a broader class of stochastic processes than does the CF lter, but the CF lter produces a better estimate of ct if yt is close to a random-walk process or a random-walk-plus-drift process. Among economists, the CF lter is commonly used for investigating business cycles. Burns and Mitchell (1946) dened business cycles as stochastic cycles in business data corresponding to periods between 1.5 and 8 years. The default values for minperiod() and maxperiod() are the Burns Mitchell values of 1.5 and 8 years scaled to the frequency of the dataset. The calculations of the default
430
values assume that the time variable is formatted as daily, weekly, monthly, quarterly, half-yearly, or yearly; see [D] format. When yt is assumed to be a random-walk-plus-drift process instead of a random-walk process, specify the drift option, which removes the linear drift in the series before applying the lter. Drift is removed by transforming the original series to a new series by using the calculation
zt = yt
(t 1)(yT y1 ) T 1
The cyclical component ct is calculated from drift-adjusted series zt . The trend component t is calculated by t = yt ct . By default, the CF lter assumes the series is nonstationary. If the series is stationary, the stationary option is used to change the calculations to those appropriate for a stationary series. For each variable, the CF lter estimate of ct is put in the corresponding new variable, and when the trend() option is specied, the estimate of t is put in the corresponding new variable. tsfilter cf automatically detects panel data from the information provided when the dataset was tsset or xtset. All calculations are done separately on each panel. Missing values at the beginning and end of the sample are excluded from the sample. The sample may not contain gaps.
Example 1
In this and the subsequent examples, we use tsfilter cf to estimate the business-cycle component of the natural log of real gross domestic product (GDP) of the United States. Our sample of quarterly data goes from 1952q1 to 2010q4. Below we read in and plot the data.
9.5 7.5
1950q1
1960q1
1970q1
1980q1
1990q1
2000q1
2010q1
The series looks like it might be generated by a random-walk-plus-drift process and is thus a candidate for the CF lter. Below we use tsfilter cf to lter gdp ln, and we use pergram (see [TS] pergram) to compute and to plot the periodogram of the estimated cyclical component.

. tsfilter cf gdp_cf = gdp_ln . pergram gdp_cf, xline(.03125 .16667)
431
gdp_ln cyclical component from cf filter Log Periodogram 6.00 4.00 2.00 0.00 2.00 4.00 6.00
0.00
0.10
0.20 0.30 Frequency
0.40
0.50
Because our sample is of quarterly data, tsfilter cf used the default values of minperiod(6) and maxperiod(32). The minimum and maximum periods are the Burns and Mitchell (1946) business-cycle periods for quarterly data. In the periodogram, we added vertical lines at the natural frequencies corresponding to the conventional Burns and Mitchell (1946) values for business-cycle components. pergram displays the results in natural frequencies, which are the standard frequencies divided by 2 . We use the xline() option to draw vertical lines at the lower natural-frequency cutoff (1/32 = 0.03125) and the upper natural-frequency cutoff (1/6 0.16667). If the lter completely removed the stochastic cycles at the unwanted frequencies, the periodogram would be a at line at the minimum value of 6 outside the range identied by the vertical lines. The periodogram reveals that the CF did a reasonable job of ltering out the unwanted stochastic cycles.
6.00 4.00 2.00
0.00
2.00
4.00
6.00
432
gdp_ln cyclical component from cf filter .02 0 .02 .04 .04
1950q1
1960q1
1970q1
1980q1
1990q1
2000q1
2010q1
quarterly time variable gdp_ln cyclical component from cf filter
Saved results
tsfilter cf saves the following in r():
Scalars r(smaorder) r(minperiod) r(maxperiod) Macros r(varlist) r(filterlist) r(trendlist) r(method) r(symmetric) r(drift) r(stationary) r(unit) Matrices r(filter) order of the symmetric moving average, if specied minimum period of stochastic cycles maximum period of stochastic cycles original time-series variables variables containing estimates of the cyclical components variables containing estimates of the trend components, if trend() was specied Christiano-Fitzgerald yes or no, indicating whether the symmetric version of the lter was or was not used yes or no, indicating whether drift was or was not removed before ltering yes or no, indicating whether the calculations assumed the series was or was not stationary units of time variable set using tsset or xtset
(q+1)1 matrix of weights (b0 ,b1 ,...,bq ) , where q is the order of the symmetric moving
average, and the weights are the ChristianoFitzgerald coefcients; only returned when smaorder() is used to set q

tsfilter cf is implemented as an ado-le. For an innitely long series, there is an ideal band-pass lter that extracts the cyclical component by using the calculation
ct =
j=
bj ytj
If pl and ph are the minimum and maximum periods of the stochastic cycles of interest, the weights bj in the ideal band-pass lter are given by
433
bj =
1 (h l )
if j = 0
(j)1 {sin(jh ) sin(jl )} if j = 0
where l = 2/pl and h = 2/ph are the lower and higher cutoff frequencies, respectively. Because our time series has nite length, the ideal band-pass lter cannot be computed exactly. Christiano and Fitzgerald (2003) derive the nite-length CF band-pass lter that minimizes the mean squared error between the ltered series and the series ltered by an ideal band-pass lter that perfectly separates out the components. This lter is not symmetric nor do the coefcients sum to zero. The formula for calculating the value of cyclical component ct for t = 2, 3, . . . , T 1 using the asymmetric version of the CF lter can be expressed as
T t1 t2
ct = b0 yt +
j=1
bj yt+j + bT t yT +
j=1
bj ytj + bt1 y1
where b0 , b1 , . . . are the weights used by the ideal band-pass lter. bT t and bt1 are linear functions of the ideal weights used in this calculation. The CF lter uses two different calculations for bt depending upon whether the series is assumed to be stationary or nonstationary. For the default nonstationary case with 1 < t < T , Christiano and Fitzgerald (2003) set bT t and bt1 to T t1 t2 1 1 bj and bt1 = b0 bj bT t = b0 2 2 j=1 j=1 which forces the weights to sum to zero. For the nonstationary case, when t = 1 or t = T , the two endpoints (c1 and cT ) use only one modied weight, bT 1 :
c1 =
1 b0 y1 + 2
T 2
bj yj+1 + bT 1 yT
j=1
and
cT =
1 b0 y T + 2
T 2
bj yT j + bT 1 y1
j=1
When the stationary option is used to invoke the stationary calculations, all weights are set to the ideal lter weight, that is, bj = bj . If the smaorder() option is set, the symmetric version of the CF lter is used. This option species the length of the symmetric moving average denoted by q . The symmetric calculations for ct are similar to those used by the BK lter:
q1
ct = bq {Lq (yt ) + Lq (yt )} +

j=q+1
bj Lj (yt )
q1 j=1 bj .
where, for the default nonstationary calculations, bq = (1/2)b0
If the smaorder()
and stationary options are set, then bq is set equal to the ideal weight bq .
434
References
Baxter, M., and R. G. King. 1999. Measuring business cycles: Approximate band-pass lters for economic time series. Review of Economics and Statistics 81: 575593. Burns, A. F., and W. C. Mitchell. 1946. Measuring Business Cycles. New York: National Bureau of Economic Research. Christiano, L. J., and T. J. Fitzgerald. 2003. The band pass lter. International Economic Review 44: 435465. Pollock, D. S. G. 1999. A Handbook of Time-Series Analysis, Signal Processing and Dynamics. London: Academic Press. . 2006. Econometric methods of signal extraction. Computational Statistics & Data Analysis 50: 22682292.
Also see
Title
tslter hp HodrickPrescott time-series lter
Syntax
Filter one variable tsfilter hp type newvar = varname if in , options
Filter multiple variables, unique names tsfilter hp type newvarlist = varlist if in , options
Filter multiple variables, common name stub tsfilter hp options

Main
type
stub* = varlist
if
in
, options
Description smoothing parameter for the HodrickPrescott lter
smooth(#)
Trend

Gain
Menu
Statistics
>
Time series
>
>
Hodrick-Prescott
Description
tsfilter hp uses the HodrickPrescott high-pass lter to separate a time series into trend and cyclical components. The trend component may contain a deterministic or a stochastic trend. The smoothing parameter determines the periods of the stochastic cycles that drive the stationary cyclical component. See [TS] tslter for an introduction to the methods implemented in tsfilter hp.
435
436
Options

Main
smooth(#) sets the smoothing parameter for the HodrickPrescott lter. By default if the units of the time variable are set to daily, weekly, monthly, quarterly, half-yearly, or yearly, then the RavnUhlig rule is used to set the smoothing parameter; otherwise, the default value is smooth(1600). The RavnUhlig rule sets # to 1600p4 , where pq is the number of periods per quarter. The smoothing q parameter must be greater than 0.
Trend
Gain
Remarks
We assume that you have already read [TS] tslter, which provides an introduction to ltering and the methods implemented in tsfilter hp, more examples using tsfilter hp, and a comparison of the four lters implemented by tsfilter. In particular, an understanding of gain functions as presented in [TS] tslter is required to understand these remarks. tsfilter hp uses the HodrickPrescott (HP) high-pass lter to separate a time-series yt into trend and cyclical components yt = t + ct where t is the trend component and ct is the cyclical component. t may be nonstationary; it may contain a deterministic or a stochastic trend, as discussed below. The primary objective is to estimate ct , a stationary cyclical component that is driven by stochastic cycles at a range of periods. The trend component t is calculated by the difference t = yt ct . Although the HP high-pass lter implemented in tsfilter hp has been widely applied by macroeconomists, it is a general time-series method and may be of interest to other researchers. Hodrick and Prescott (1997) motivated the HP lter as a trend-removal technique that could be applied to data that came from a wide class of data-generating processes. In their view, the technique specied a trend in the data and the data was ltered by removing the trend. The smoothness of the trend depends on a parameter . The trend becomes smoother as , and Hodrick and Prescott (1997) recommended setting to 1,600 for quarterly data. King and Rebelo (1993) showed that removing a trend estimated by the HP lter is equivalent to a high-pass lter. They derived the gain function of this high-pass lter and showed that the lter would make integrated processes of order 4 or less stationary, making the HP lter comparable to the other lters implemented in tsfilter.
437
Example 1
In this and the subsequent examples, we use tsfilter hp to estimate the business-cycle component of the natural log of real gross domestic product (GDP) of the United States. Our sample of quarterly data goes from 1952q1 to 2010q4. Below we read in and plot the data.
9.5 7.5
1950q1
1960q1
1970q1
1980q1
1990q1
2000q1
2010q1
The series is nonstationary and is thus a candidate for the HP lter. Below we use tsfilter hp to lter gdp ln, and we use pergram (see [TS] pergram) to compute and to plot the periodogram of the estimated cyclical component.
. tsfilter hp gdp_hp = gdp_ln . pergram gdp_hp, xline(.03125 .16667)
Because our sample is of quarterly data, tsfilter hp used the default value for the smoothing parameter of 1,600. In the periodogram, we added vertical lines at the natural frequencies corresponding to the conventional Burns and Mitchell (1946) values for business-cycle components of 32 periods and 6 periods. pergram displays the results in natural frequencies, which are the standard frequencies divided by 2 . We use the xline() option to draw vertical lines at the lower natural-frequency cutoff (1/32 = 0.03125) and the upper natural-frequency cutoff (1/6 0.16667). If the lter completely removed the stochastic cycles at the unwanted frequencies, the periodogram would be a at line at the minimum value of 6 outside the range identied by the vertical lines.
438
gdp_ln cyclical component from hp filter Log Periodogram 6.00 4.00 2.00 0.00 2.00 4.00 6.00
0.00
0.10
0.20 0.30 Frequency
0.40
0.50
The periodogram reveals a high-periodicity issue and a low-periodicity issue. The points above 6.00 to the left of the left-hand vertical line in the periodogram reveal that the lter did not do a good job of ltering out the high-periodicity stochastic cycles with the default value smoothing parameter of 1,600. That there is no tendency of the points to the right of the right-hand vertical line to be smoothed toward 6.00 reveals that the HP lter did not remove any of the low-periodicity stochastic cycles. This result is not surprising, because the HP lter is a high-pass lter. In the next example, we address the high-periodicity issue. See [TS] tslter and [TS] tslter bw for how to turn a high-pass lter into a band-pass lter.
Example 2
In the lter literature, lter parameters are set as functions of the cutoff frequency; see Pollock (2000, 324), for instance. This method nds the lter parameter that sets the gain of the lter equal to 1/2 at the cutoff frequency. In a technical note in [TS] tslter, we showed that applying this method to selecting at the cutoff frequency of 32 periods suggests setting 677.13. In the output below, we estimate the business-cycle component using this value for the smoothing parameter, and we compute and plot the periodogram of the estimated business-cycle component.
6.00 4.00 2.00
0.00
2.00
4.00
6.00

. tsfilter hp gdp_hp677 = gdp_ln, smooth(677.13) . pergram gdp_hp677, xline(.03125 .16667)
439
gdp_ln cyclical component from hp filter Log Periodogram 6.00 4.00 2.00 0.00 2.00 4.00 6.00
0.00
0.10
0.20 0.30 Frequency
0.40
0.50
A comparison of the two periodograms reveals that setting the smoothing parameter to 677.13 removes more of the high-periodicity stochastic cycles than does the default 1,600. In [TS] tslter, we found that the HP lter was not as good at removing the high-periodicity stochastic cycles as was the ChristianoFitzgerald lter implemented in tsfilter cf or as was the Butterworth lter implemented in tsfilter bw. Below we plot the estimated business-cycle component with recessions identied by the shaded areas.
.04
.04
gdp_ln cyclical component from hp filter .02 0 .02
1950q1
1960q1
1970q1
1980q1
1990q1
2000q1
2010q1
quarterly time variable gdp_ln cyclical component from hp filter
tsfilter hp automatically detects panel data from the information provided when the dataset was tsset or xtset. All calculations are done separately on each panel. Missing values at the beginning and end of the sample are excluded from the sample. The sample may not contain gaps.
6.00 4.00 2.00
0.00
2.00
4.00
6.00
440
Saved results
tsfilter hp saves the following in r():
Scalars r(smooth) Macros r(varlist) r(filterlist) r(trendlist) r(method) r(unit) smoothing parameter original time-series variables variables containing estimates of the cyclical components variables containing estimates of the trend components, if trend() was specied Hodrick-Prescott units of time variable set using tsset or xtset

tsfilter hp is implemented as an ado-le. Formally, the lter is dened as the solution to the following optimization problem for t
T T 1
min
t t=1
(yt t )2 +
t=2
{(t+1 t ) (t t1 )}
where the smoothing parameter is set xed to a value. If = 0, the solution degenerates to t = yt , in which case the lter excludes all frequencies, that is, ct = 0. On the other extreme, as , the solution approaches the least-squares t to the line t = 0 + 1 t; see Hodrick and Prescott (1997) for a discussion. For a xed , it can be shown that the cyclical component c = (c1 , c2 , . . . , cT ) is calculated by
c = (IT M1 )y
where y is the column vector y = (y1 , y2 , . . . , yT ), IT is the T T identity matrix, and M is the T T matrix:
(1 + ) 2 0 0 0 ... 0 2 (1 + 5) 4 0 0 ... 0 4 (1 + 6) 4 0 ... 0 0 4 (1 + 6) 4 ... 0 . . .. .. .. .. .. . . . . . . . . ... . M= . .. .. .. .. .. . . . . . . . 0 0 0 ... 4 (1 + 6) 4 0 0 ... 0 4 (1 + 6) 4 0 ... 0 0 4 (1 + 5) 2 0 ... 0 0 0 2 (1 + )

The gain of the HP lter is given by (see King and Rebelo [1993], Maravall and del Rio [2007], or Harvey and Trimbur [2008])
() =
4{1 cos()}2 1 + 4{1 cos()}2
441
As discussed in [TS] tslter, there are two approaches to selecting . One method, based on the heuristic argument of Hodrick and Prescott (1997), is used to compute the default values for . The method sets to 1,600 for quarterly data and to the rescaled values worked out by Ravn and Uhlig (2002). The rescaled default values for are 6.25 for yearly data, 100 for half-yearly data, 129,600 for monthly data, 1600 124 for weekly data, and 1600 (365/4)4 for daily data. The second method for selecting uses the recommendations of Pollock (2000, 324), who uses the gain function of the lter to identify a value for . Additional literature critiques the HP lter by pointing out that the HP lter corresponds to a specic model. Harvey and Trimbur (2008) show that the cyclical component estimated by the HP lter is equivalent to one estimated by a particular unobserved-components model. Harvey and Jaeger (1993), G mez (1999), Pollock (2000), and G mez (2001) also show this result and provide interesting o o comparisons of estimating ct by ltering and model-based methods.
References
Burns, A. F., and W. C. Mitchell. 1946. Measuring Business Cycles. New York: National Bureau of Economic Research. G mez, V. 1999. Three equivalent methods for ltering nite nonstationary time series. Journal of Business and o Economic Statistics 17: 109116. . 2001. The use of Butterworth lters for trend and cycle estimation in economic time series. Journal of Business and Economic Statistics 19: 365373. Harvey, A. C., and A. Jaeger. 1993. Detrending, stylized facts and the business cycle. Journal of Applied Econometrics 8: 231247. Harvey, A. C., and T. M. Trimbur. 2008. Trend estimation and the HodrickPrescott lter. Journal of the Japanese Statistical Society 38: 4149. Hodrick, R. J., and E. C. Prescott. 1997. Postwar U.S. business cycles: An empirical investigation. Journal of Money, Credit, and Banking 29: 116. King, R. G., and S. T. Rebelo. 1993. Low frequency ltering and real business cycles. Journal of Economic Dynamics and Control 17: 207231. Leser, C. E. V. 1961. A simple method of trend construction. Journal of the Royal Statistical Society, Series B 23: 91107. Maravall, A., and A. del Rio. 2007. Temporal aggregation, systematic sampling, and the HodrickPrescott lter. Working Paper No. 0728, Banco de Espa a. n http://www.bde.es/webbde/Secciones/Publicaciones/PublicacionesSeriadas/DocumentosTrabajo/07/Fic/dt0728e.pdf. Pollock, D. S. G. 1999. A Handbook of Time-Series Analysis, Signal Processing and Dynamics. London: Academic Press. . 2000. Trend estimation and de-trending via rational square-wave lters. Journal of Econometrics 99: 317334. . 2006. Econometric methods of signal extraction. Computational Statistics & Data Analysis 50: 22682292. Ravn, M. O., and H. Uhlig. 2002. On adjusting the HodrickPrescott lter for the frequency of observations. Review of Economics and Statistics 84: 371376.
Also see
Title
tsline Plot time-series data
Syntax
Time-series line plot twoway tsline varlist if in , tsline options
Time-series range plot with lines twoway tsrline y1 y2 if in , tsrline options
where the time variable is assumed set by tsset (see [TS] tsset), varlist has the interpretation y1 y2 . . . yk . tsline options
Plots
Description any of the options documented in [G-2] graph twoway scatter with the exception of marker options, marker placement options, and marker label options, which will be ignored if specied any options documented in [G-3] twoway options
scatter options
Y axis, Time axis, Titles, Legend, Overall, By
twoway options
tsrline options
Plots
Description any of the options documented in [G-2] graph twoway rline any options documented in [G-3] twoway options
rline options twoway options
Menu
Statistics
>
Time series
>
Graphs
>
Line plots
Description
tsline draws line plots for time-series data. tsrline draws a range plot with lines for time-series data. tsline and tsrline are both commands and plottypes as dened in [G-2] graph twoway. Thus the syntax for tsline is
. graph twoway tsline ...
442

. twoway tsline ... . tsline ...
443
and similarly for tsrline. Being plot types, these commands may be combined with other plot types in the twoway family, as in,
. twoway (tsrline
. . . ) (tsline . . . ) (lfit . . . ) . . .
which can equivalently be written

. tsrline
. . . || tsline . . . || lfit . . . || . . .
Options

Plots
scatter options are any of the options allowed by the graph twoway scatter command except that marker options, marker placement option, and marker label options will be ignored if specied; see [G-2] graph twoway scatter. rline options are any of the options allowed by the graph twoway rline command; see [G-2] graph twoway rline.
twoway options are any of the options documented in [G-3] twoway options. These include options for titling the graph (see [G-3] title options), for saving the graph to disk (see [G-3] saving option), and the by() option, which will allow you to simultaneously plot different subsets of the data (see [G-3] by option). Also see the recast() option discussed in [G-3] advanced options for information on how to plot spikes, bars, etc., instead of lines.
Remarks
Example 1
We simulated two separate time series (each of 200 observations) and placed them in a Stata dataset, tsline1.dta. The rst series simulates an AR(2) process with 1 = 0.8 and 2 = 0.2; the second series simulates an MA(2) process with 1 = 0.8 and 2 = 0.2. We use tsline to graph these two series.
. use http://www.stata-press.com/data/r12/tsline1 . tsset lags time variable: lags, 0 to 199 delta: 1 unit . tsline ar ma
444
1 0
50 Simulated AR(.8,.2)
100 lags
150 Simulated MA(.8,.2)
200
Example 2
Suppose that we kept a calorie log for an entire calendar year. At the end of the year, we would have a dataset (for example, tsline2.dta) that contains the number of calories consumed for 365 days. We could then use tsset to identify the date variable and tsline to plot calories versus time. Knowing that we tend to eat a little more food on Thanksgiving and Christmas day, we use the ttick() and ttext() options to point these days out on the time axis.
. use http://www.stata-press.com/data/r12/tsline2 . tsset day time variable: day, 01jan2002 to 31dec2002 delta: 1 day . tsline calories, ttick(28nov2002 25dec2002, tpos(in)) > ttext(3470 28nov2002 "thanks" 3470 25dec2002 "x-mas", orient(vert))
4400 3600 Calories consumed 3800 4000 4200
3400
01jan2002
01apr2002
01jul2002 Date
01oct2002
01jan2003
We were uncertain of the exact values we logged, so we also gave a range for each day. Here is a plot of the summer months.
xmas
thanks

. tsrline lcalories ucalories if tin(1may2002,31aug2002) || tsline cal || > if tin(1may2002,31aug2002), ytitle(Calories)
445
3300 01may2002
3400
Calories 3500 3600
3700
3800
01jun2002
01jul2002 Date
01aug2002 Calories consumed
01sep2002
Calorie range
Options associated with the time axis allow dates (and times) to be specied in place of numeric date (and time) values. For instance, we used
ttick(28nov2002 25dec2002, tpos(in))
to place tick marks at the specied dates. This works similarly for tlabel, tmlabel, and tmtick. Suppose that we wanted to place vertical lines for the previously mentioned holidays. We could specify the dates in the tline() option as follows:
. tsline calories, tline(28nov2002 25dec2002)
4400 3400 01jan2002 3600 Calories consumed 3800 4000 4200
01apr2002
01jul2002 Date
01oct2002
01jan2003
446
We could also modify the format of the time axis so that only the day in the year is displayed in the labeled ticks:
. tsline calories, tlabel(, format(%tdmd)) ttitle("Date (2002)")
4400 3400 Jan1 3600 Calories consumed 3800 4000 4200
Apr1
Jul1 Date (2002)
Oct1
Jan1

tsline and tsrline are implemented as ado-les.
Reference
Cox, N. J. 2009. Stata tip 76: Separating seasonal time series. Stata Journal 9: 321326.
Also see
[TS] tsset Declare data to be time-series data [G-2] graph twoway Twoway graphs [XT] xtline Panel-data line plots
Title
tsreport Report time-series aspects of a dataset or estimation sample
Syntax
tsreport options
Main
if
in
, options
Description display number of gaps in the time series display count of gaps even if no gaps display gaps data in tabular list ignore panel changes
report report0 list panel
tsreport typed without options produces no output but provides its standard saved results.
Menu
Statistics
>
Time series
>
Setup and utilities
>
Report time-series aspects of dataset
Description
tsreport reports time gaps in a sample of observations. The report option displays a one-line statement showing the count of the gaps, and list displays a list of records that follow gaps. A return value r(N gaps) indicates the number of gaps in the sample.
Options

Main
report species that the number of gaps in the time series be reported, if any gaps exist. report0 species that the count of gaps be reported, even if there are no gaps. list species that a tabular list of gaps be displayed. panel species that panel changes not be counted as gaps. Whether panel changes are counted as gaps usually depends on how the calling command handles panels.
Remarks
Time-series commands sometimes require that observations be on a xed time interval with no gaps, or the command may not function properly. tsreport provides a tool for reporting the gaps in a sample.
447
448
Example 1
The following monthly panel data have two panels and a missing month (March) in the second panel:
. use http://www.stata-press.com/data/r12/tsrptxmpl . list edlevel month income in 1/6, sep(0) edlevel 1. 2. 3. 4. 5. 6. 1 1 1 2 2 2 month 1998m1 1998m2 1998m3 1998m1 1998m2 1998m4 income 687 783 790 1435 1522 1532
Invoking tsreport without the panel option gives us the following report:
. tsreport, report Number of gaps in sample: 2 (gap count includes panel changes)
We could get a list of gaps and better see what has been counted as a gap by using the list option:
. tsreport, report list Number of gaps in sample: 2 (gap count includes panel changes)
Observations with preceding time gaps (gaps include panel changes) Record 4 6 edlevel 2 2 month 1998m1 1998m4
We now see why tsreport is reporting two gaps. It is counting the known gap in March of the second panel and counting the change from the rst to the second panel. (If we are programmers writing a procedure that does not account for panels, a change from one panel to the next represents a break in the time series just as a gap in the data does.) We may prefer that the changes in panels not be counted as gaps. We can obtain a count without the panel change by using the panel option:
. tsreport, report panel Number of gaps in sample: 1
To obtain a fuller report, we type

. tsreport, report list panel Number of gaps in sample: 1 Observations with preceding time gaps Record 6 edlevel 2 month 1998m4
449
Saved results
tsreport saves the following in r():
Scalars r(N gaps) number of gaps in sample

tsreport is implemented as an ado-le.
Also see
[TS] tsset Declare data to be time-series data
Title
tsrevar Time-series operator programming command
Syntax
tsrevar varlist if in , substitute list
You must tsset your data before using tsrevar; see [TS] tsset.
Description
tsrevar, substitute takes a varlist that might contain op.varname combinations and substitutes equivalent temporary variables for the combinations. tsrevar, list creates no new variables. It returns in r(varlist) the list of base variables corresponding to varlist.
Options
substitute species that tsrevar resolve op.varname combinations by creating temporary variables as described above. substitute is the default action taken by tsrevar; you do not need to specify the option. list species that tsrevar return a list of base variable names.
Remarks
tsrevar substitutes temporary variables for any op.varname combinations in a variable list. For instance, the original varlist might be gnp L.gnp r, and tsrevar, substitute would create newvar = L.gnp and create the equivalent varlist gnp newvar r. This new varlist could then be used with commands that do not otherwise support time-series operators, or it could be used in a program to make execution faster at the expense of using more memory. tsrevar, substitute might create no new variables, one new variable, or many new variables, depending on the number of op.varname combinations appearing in varlist. Any new variables created are temporary. The new, equivalent varlist is returned in r(varlist). The new varlist corresponds one to one with the original varlist. tsrevar, list returns in r(varlist) the list of base variable names of varlist with the timeseries operators removed. tsrevar, list creates no new variables. For instance, if the original varlist were gnp l.gnp l2.gnp r l.cd, then r(varlist) would contain gnp r cd. This is useful for programmers who might want to create programs to keep only the variables corresponding to varlist.
Example 1
. use http://www.stata-press.com/data/r12/tsrevarex . tsrevar l.gnp d.gnp r
creates two temporary variables containing the values for l.gnp and d.gnp. The variable r appears in the new variable list but does not require a temporary variable.
450
451
The resulting variable list is

. display "r(varlist)" 00014P 00014Q r
(Your temporary variable names may be different, but that is of no consequence.) We can see the results by listing the new variables alongside the original value of gnp.
. list gnp r(varlist) gnp 1. 2. 3. 4. 5. 128 135 132 138 145 __00014P . 128 135 132 138 in 1/5 __00014Q . 7 -3 6 7 r 3.2 3.8 2.6 3.9 4.2
Temporary variables automatically vanish when the program concludes. If we had needed only the base variable names, we could have specied
. tsrevar l.gnp d.gnp r, list . display "r(varlist)" gnp r
The order of the list will probably differ from that of the original list; base variables are listed only once and are listed in the order that they appear in the dataset.
Technical note
tsrevar, substitute avoids creating duplicate variables. Consider
. tsrevar gnp l.gnp r cd l.cd l.gnp
l.gnp appears twice in the varlist. tsrevar will create only one new variable for l.gnp and use that new variable twice in the resulting r(varlist). Moreover, tsrevar will even do this across multiple calls:
. tsrevar gnp l.gnp cd l.cd . tsrevar cpi l.gnp
l.gnp appears in two separate calls. At the rst call, tsrevar creates a temporary variable corresponding to l.gnp. At the second call, tsrevar remembers what it has done and uses that same temporary variable for l.gnp again.
Saved results
tsrevar saves the following in r():
Macros r(varlist) the modied variable list or list of base variable names
452
Also see
[P] syntax Parse Stata syntax [P] unab Unabbreviate variable list [U] 11 Language syntax [U] 11.4.4 Time-series varlists [U] 18 Programming Stata
Title
tsset Declare data to be time-series data
Syntax
Declare data to be time series tsset timevar , options , options
tsset panelvar timevar
Display how data are currently tsset tsset Clear time-series settings tsset, clear In the declare syntax, panelvar identies the panels and timevar identies the times. options
Main
Description specify units of timevar specify period of timevar suppress summary calculations and output
unitoptions
Delta
deltaoption noquery
noquery is not shown in the dialog box.
unitoptions (default) clocktime daily weekly monthly quarterly halfyearly yearly generic format(% fmt)
Description timevars units to be obtained from timevars display format timevar is %tc: 0 = 1jan1960 00:00:00.000, 1 = 1jan1960 00:00:00.001, . . . timevar is %td: 0 = 1jan1960, 1 = 2jan1960, . . . timevar is %tw: 0 = 1960w1, 1 = 1960w2, . . . timevar is %tm: 0 = 1960m1, 1 = 1960m2, . . . timevar is %tq: 0 = 1960q1, 1 = 1960q2,. . . timevar is %th: 0 = 1960h1, 1 = 1960h2,. . . timevar is %ty: 1960 = 1960, 1961 = 1961, . . . timevar is %tg: 0 = ?, 1 = ?, . . . specify timevars format and then apply default rule
In all cases, negative timevar values are allowed.
453
454
deltaoption species the period between observations in timevar units and may be specied as deltaoption delta(#) delta((exp)) delta(# units) delta((exp) units) Example delta(1) or delta(2) delta((7*24)) delta(7 days) or delta(15 minutes) or delta(7 days 15 minutes) delta((2+3) weeks)
Allowed units for %tc and %tC timevars are seconds minutes hours days weeks second minute hour day week secs mins sec min
and for all other %t timevars, units specied must match the frequency of the data; for example, for %ty, units must be year or years.
Menu
Statistics
>
Time series
>
Setup and utilities
>
Declare dataset to be time-series data
Description
tsset declares the data in memory to be a time series. tssetting the data is what makes Statas time-series operators such as L. and F. (lag and lead) work; the operators are discussed under Remarks below. Also, before using the other ts commands, you must tsset the data rst. If you save the data after tsset, the data will be remembered to be time series and you will not have to tsset again. There are two syntaxes for setting the data: tsset timevar tsset panelvar timevar In the rst syntaxtsset timevarthe data are set to be a straight time series. In the second syntaxtsset panelvar timevarthe data are set to be a collection of time series, one for each value of panelvar, also known as panel data, cross-sectional time-series data, and xt data. Such datasets can be analyzed by xt commands as well as ts commands. If you tsset panelvar timevar, you do not need to xtset panelvar timevar to use the xt commands. tsset without argumentstssetdisplays how the data are currently tsset and sorts the data on timevar or panelvar timevar if they are sorted differently from that. tsset, clear is a rarely used programmers command to declare that the data are no longer a time series.
455
Options

Main
unitoptions clocktime, daily, weekly, monthly, quarterly, halfyearly, yearly, generic, and format(% fmt) specify the units in which timevar is recorded. timevar will usually be a %t variable; see [D] datetime. If timevar already has a %t display format assigned to it, you do not need to specify a unitoption; tsset will obtain the units from the format. If you have not yet bothered to assign the appropriate %t format, however, you can use the unitoptions to tell tsset the units. Then tsset will set timevars display format for you. Thus, the unitoptions are convenience options; they allow you to skip formatting the time variable. The following all have the same net result:
Alternative 1 Alternative 2 (t not formatted) Alternative 3 (t not formatted)
format t %td tsset t
tsset t, daily
tsset t, format(%td)
timevar is not required to be a %t variable; it can be any variable of your own concocting so long as it takes on only integer values. In such cases, it is called generic and considered to be %tg. Specifying the unitoption generic or attaching a special format to timevar, however, is not necessary because tsset will assume that the variable is generic if it has any numerical format other than a %t format (or if it has a %tg format). clearused in tsset, clearmakes Stata forget that the data ever were tsset. This is a rarely used programmers option.
Delta
delta() species the period of timevar and is commonly used when timevar is %tc. delta() is only sometimes used with the other %t formats or with generic time variables. If delta() is not specied, delta(1) is assumed. This means that at timevar = 5, the previous time is timevar = 5 1 = 4 and the next time would be timevar = 5 + 1 = 6. Lag and lead operators, for instance, would work this way. This would be assumed regardless of the units of timevar. If you specied delta(2), then at timevar = 5, the previous time would be timevar = 5 2 = 3 and the next time would be timevar = 5 + 2 = 7. Lag and lead operators would work this way. In the observation with timevar = 5, L.price would be the value of price in the observation for which timevar = 3 and F.price would be the value of price in the observation for which timevar = 7. If you then add an observation with timevar = 4, the operators will still work appropriately; that is, at timevar = 5, L.price will still have the value of price at timevar = 3. There are two aspects of timevar: its units and its periodicity. The unitoptions set the units. delta() sets the periodicity. We mentioned that delta() is commonly used with %tc timevars because Statas %tc variables have units of milliseconds. If delta() is not specied and in some model you refer to L.price, you will be referring to the value of price 1 ms ago. Few people have data with periodicity of a millisecond. Perhaps your data are hourly. You could specify delta(3600000). Or you could specify delta((60*60*1000)), because delta() will allow expressions if you include an extra pair of parentheses. Or you could specify delta(1 hour). They all mean the same thing: timevar has periodicity of 3,600,000 ms. In an observation for which timevar = 1,489,572,000,000 (corresponding to 15mar2007 10:00:00), L.price would be the observation for which timevar = 1,489,572,000,000 3,600,000 = 1,489,568,400,000 (corresponding to 15mar2007 9:00:00).
456
When you tsset the data and specify delta(), tsset veries that all the observations follow the specied periodicity. For instance, if you specied delta(2), then timevar could contain any subset of {. . . , 4, 2, 0, 2, 4, . . . } or it could contain any subset of {. . . , 3, 1, 1, 3, . . . }. If timevar contained a mix of values, tsset would issue an error message. If you also specify a panelvaryou type tsset panelvar timevar, delta(2)the check is made on each panel independently. One panel might contain timevar values from one set and the next, another, and that would be ne. The following option is available with tsset but is not shown in the dialog box: noquery prevents tsset from performing most of its summary calculations and suppresses output. With this option, only the following results are posted: r(tdelta) r(panelvar) r(timevar) r(tsfmt) r(unit) r(unit1)
Remarks
tsset sets timevar so that Statas time-series operators are understood in varlists and expressions. The time-series operators are Operator L. L2. ... F. F2. ... D. D2. ... S. S2. ... Meaning lag xt1 2-period lag xt2 lead xt+1 2-period lead xt+2 difference xt xt1 difference of difference xt xt1 (xt1 xt2 ) = xt 2xt1 + xt2 seasonal difference xt xt1 lag-2 (seasonal) difference xt xt2
Time-series operators may be repeated and combined. L3.gnp refers to the third lag of variable gnp, as do LLL.gnp, LL2.gnp, and L2L.gnp. LF.gnp is the same as gnp. DS12.gnp refers to the one-period difference of the 12-period difference. LDS12.gnp refers to the same concept, lagged once. D1. = S1., but D2. = S2., D3. = S3., and so on. D2. refers to the difference of the difference. S2. refers to the two-period difference. If you wanted the difference of the difference of the 12-period difference of gnp, you would write D2S12.gnp. Operators may be typed in uppercase or lowercase. Most users would type d2s12.gnp instead of D2S12.gnp. You may type operators however you wish; Stata internally converts operators to their canonical form. If you typed ld2ls12d.gnp, Stata would present the operated variable as L2D3S12.gnp. Stata also understands operator(numlist). to mean a set of operated variables. For instance, typing L(1/3).gnp in a varlist is the same as typing L.gnp L2.gnp L3.gnp. The operators can also be applied to a list of variables by enclosing the variables in parentheses; for example,

. list year L(1/3).(gnp cpi) year 1. 2. 3. 4. 8. 1989 1990 1991 1992 1996 L.gnp . 5452.8 5764.9 5932.4 7330.1 L2.gnp L3.gnp L.cpi . 100 105 108 122 L2.cpi . . 100 105 119 L3.cpi . . . 100 112
457
. . . . 5452.8 . 5764.9 5452.8 (output omitted ) 6892.2 6519.1
In operator#., making # zero returns the variable itself. L0.gnp is gnp. Thus, you can type list year l(0/3).gnp to mean list year gnp L.gnp L2.gnp L3.gnp. The parenthetical notation may be used with any operator. Typing D(1/3).gnp would return the rst through third differences. The parenthetical notation may be used in operator lists with multiple operators, such as L(0/3)D2S12.gnp. Operator lists may include up to one set of parentheses, and the parentheses may enclose a numlist; see [U] 11.1.8 numlist. Before you can use these time-series operators, however, the dataset must satisfy two requirements: 1. the dataset must be tsset and 2. the dataset must be sorted by timevar or, if it is a cross-sectional time-series dataset, by panelvar timevar. tsset handles both requirements. As you use Stata, however, you may later use a command that re-sorts that data, and if you do, the time-series operators will not work:
. tsset time (output omitted ) . regress y x l.x (output omitted ) . (you continue to use Stata and, sometime later:) . regress y x l.x not sorted r(5);
Then typing tsset without arguments will reestablish the sort order:
. tsset (output omitted ) . regress y x l.x (output omitted )
Here typing tsset is the same as typing sort time. Had we previously tsset country time, however, typing tsset would be the same as typing sort country time. You can type the sort command or type tsset without arguments; it makes no difference. There are two syntaxes for setting your data: tsset timevar tsset panelvar timevar In both, timevar must contain integer values. If panelvar is specied, it too must contain integer values, and the dataset is declared to be a cross-section of time series, such as a collection of time series for different countries.
458
Example 1: Numeric time variable

You have monthly data on personal income. Variable t records the time of an observation, but there is nothing special about the name of the variable. There is nothing special about the values of the variable, either. t is not required to be %tm variableperhaps you do not even know what that means. t is just a numeric variable containing integer values that represent the month, and we will imagine that t takes on the values 1, 2, . . . , 9, although it could just as well be 3, 2 . . . , 5, or 1,023, 1,024, . . . , 1,031. What is important is that the values are dense: adjacent months have a time value that differs by 1.
. use http://www.stata-press.com/data/r12/tssetxmpl . list t income t 1. 2. income
1 1153 2 1181 (output omitted ) 9. 9 1282 . tsset t time variable: t, 1 to 9 delta: 1 unit . regress income l.income (output omitted )
Example 2: Adjusting the starting date

In the example above, that t started at 1 was not important. As we said, the t variable could just as well be recorded 3, 2 . . . , 5, or 1,023, 1,024, . . . , 1,031. What is important is that the difference in t between observations be delta() when there are no gaps. Although how time is measured makes no difference, Stata has formats to display time nicely if it is recorded in certain ways; you can learn about the formats by seeing [D] datetime. Stata likes time variables in which 1jan1960 is recorded as 0. In our previous example, if t = 1 corresponds to July 1995, then we could make a variable that ts Statas preference by typing
. generate newt = tm(1995m7) + t - 1
tm() is the function that returns a month equivalent; tm(1995m7) evaluates to the constant 426, meaning 426 months after January 1960. We now have variable newt containing
. list t newt income t 1. 2. 3. 9. 1 2 3 newt income
426 1153 427 1181 428 1208 (output omitted ) 9 434 1282
If we put a %tm format on newt, it will display more cleanly:

. format newt %tm . list t newt income t 1. 2. 3. 9. 1 2 3 9 newt income
459
1995m7 1153 1995m8 1181 1995m9 1208 (output omitted ) 1996m3 1282
We could now tsset newt rather than t:

. tsset newt time variable: delta: newt, 1995m7 to 1996m3 1 month
Technical note
In addition to monthly, Stata understands clock times (to the millisecond level) as well as daily, weekly, quarterly, half-yearly, and yearly data. See [D] datetime for a description of these capabilities. Lets reconsider the previous example, but rather than monthly, lets assume the data are daily, weekly, etc. The only thing to know is that, corresponding to function tm(), there are functions td(), tw(), tq(), th(), and ty() and that, corresponding to format %tm, there are formats %td, %tw, %tq, %th, and %ty. Here is what we would have typed had our data been on a different time scale:
Daily: if your t variable had t=1 corresponding to 15mar1993 . gen newt = td(15mar1993) + t - 1 . tsset newt, daily if your t variable had t=1 corresponding to 1994w1: . gen newt = tw(1994w1) + t - 1 . tsset newt, weekly if your t variable had t=1 corresponding to 2004m7: . gen newt = tm(2004m7) + t - 1 . tsset newt, monthly if your t variable had t=1 corresponding to 1994q1: . gen newt = tq(1994q1) + t - 1 . tsset newt, quarterly
Weekly:
Monthly:
Quarterly:
Half-yearly: if your t variable had t=1 corresponding to 1921h2: . gen newt = th(1921h2) + t - 1 . tsset newt, halfyearly Yearly: if your t variable had t=1 corresponding to 1842: . gen newt = 1842 + t - 1 . tsset newt, yearly
In each example above, we subtracted one from our time variable in constructing the new time variable newt because we assumed that our starting time value was 1. For the quarterly example, if our starting time value were 5 and that corresponded to 1994q1, we would type
. generate newt = tq(1994q1) + t - 5
Had our initial time value been t = 742 and that corresponded to 1994q1, we would have typed
. generate newt = tq(1994q1) + t - 742
460
Example 3: Time-series data but no time variable

Perhaps we have the same time-series data but no time variable:
. use http://www.stata-press.com/data/r12/tssetxmpl2, clear . list income income 1. 2. 3. 4. 5. 6. 7. 8. 9. 1153 1181 1208 1272 1236 1297 1265 1230 1282
Say that we know that the rst observation corresponds to July 1995 and continues without gaps. We can create a monthly time variable and format it by typing
. generate t = tm(1995m7) + _n - 1 . format t %tm
We can now tsset our dataset and list it:

. tsset t time variable: delta: . list t income t 1. 2. 3. 9. income t, 1995m7 to 1996m3 1 month
1995m7 1153 1995m8 1181 1995m9 1208 (output omitted ) 1996m3 1282
Example 4: Time variable as a string

Your data might include a time variable that is encoded into a string. In the example below each monthly observation is identied by string variable yrmo containing the month and year of the observation, sometimes with punctuation between:

. use http://www.stata-press.com/data/r12/tssetxmpl, clear . list yrmo income yrmo 1. 2. 3. 4. 5. 6. 7. 8. 9. 7/1995 8/1995 9-1995 10,1995 11 1995 12 1995 1/1996 2.1996 3- 1996 income 1153 1181 1208 1272 1236 1297 1265 1230 1282
461
The rst step is to convert the string to a numeric representation. Doing so is easy using the monthly() function; see [D] datetime.
. gen mdate = monthly(yrmo, "MY") . list yrmo mdate income yrmo 1. 2. 3. 9. mdate income 1153 1181 1208 1282
7/1995 426 8/1995 427 9-1995 428 (output omitted ) 3- 1996 434
Our new variable, mdate, contains the number of months from January 1960. Now that we have numeric variable mdate, we can tsset the data:
. format mdate %tm . tsset mdate time variable: delta:
mdate, 1995m7 to 1996m3 1 month
In fact, we can combine the two and type

. tsset mdate, format(%tm) time variable: mdate, 1995m7 to 1996m3 delta: 1 month
or type
. tsset mdate, monthly time variable: delta: mdate, 1995m7 to 1996m3 1 month
462
In all cases, we obtain

. list yrmo mdate income yrmo 1. 2. 3. 4. 5. 6. 7. 8. 9. 7/1995 8/1995 9-1995 10,1995 11 1995 12 1995 1/1996 2.1996 3- 1996 mdate 1995m7 1995m8 1995m9 1995m10 1995m11 1995m12 1996m1 1996m2 1996m3 income 1153 1181 1208 1272 1236 1297 1265 1230 1282
Stata can translate many different date formats, including strings like 12jan2009; January 12, 2009; 12-01-2009; 01/12/2009; 01/12/09; 12jan2009 8:14; 12-01-2009 13:12; 01/12/09 1:12 pm; Wed Jan 31 13:03:25 CST 2009; 1998q1; and more. See [D] datetime.
Example 5: Time-series data with gaps

Gaps in the time series cause no difculties:
. use http://www.stata-press.com/data/r12/tssetxmpl3, clear . list yrmo income yrmo 1. 2. 3. 4. 5. 6. 7/1995 8/1995 11 1995 12 1995 1/1996 3- 1996 income 1153 1181 1236 1297 1265 1282
. gen mdate = monthly(yrmo, "MY") . tsset mdate, monthly time variable: mdate, 1995m7 to 1996m3, but with gaps delta: 1 month
Once the dataset has been tsset, we can use the time-series operators. The D operator species rst differences:
. list mdate income d.income mdate 1. 2. 3. 4. 5. 6. 1995m7 1995m8 1995m11 1995m12 1996m1 1996m3 income 1153 1181 1236 1297 1265 1282 D.income . 28 . 61 -32 .
463
We can use the operators in an expression or varlist context; we do not have to create a new variable to hold D.income. We can use D.income with the list command, with regress or any other Stata command that allows time-series varlists.
Example 6: Clock times

We have data from a large hotel in Las Vegas that changes the reservation prices for its rooms hourly. A piece of the data looks like
. use http://www.stata-press.com/data/r12/tssetxmpl4, clear . list in 1/5 time 1. 2. 3. 4. 5. 02.13.2007 02.13.2007 02.13.2007 02.13.2007 02.13.2007 08:00 09:00 10:00 11:00 12:00 price 140 155 160 155 160
Variable time is a string variable. The rst step in making this dataset a time-series dataset is to translate the string to a numeric variable:
. generate double t = clock(time, "MDY hm") . list in 1/5 time 1. 2. 3. 4. 5. 02.13.2007 02.13.2007 02.13.2007 02.13.2007 02.13.2007 08:00 09:00 10:00 11:00 12:00 price 140 155 160 155 160 t 1.487e+12 1.487e+12 1.487e+12 1.487e+12 1.487e+12
See [D] datetime for an explanation of what is going on here. clock() is the function that converts strings to datetime (%tc) values. We typed clock(time, "MDY hm") to convert string variable time, and we told clock() that the values in time were in the order month, day, year, hour, and minute. We stored new variable t as a double because time values are large, and doing so is required to prevent rounding. Even so, the resulting values 1.487e+12 look rounded, but that is only because of the default display format for new variables. We can see the values better if we change the format:
. format t %20.0gc . list in 1/5 time 1. 2. 3. 4. 5. 02.13.2007 02.13.2007 02.13.2007 02.13.2007 02.13.2007 08:00 09:00 10:00 11:00 12:00 price 140 155 160 155 160 t 1,486,972,800,000 1,486,976,400,000 1,486,980,000,000 1,486,983,600,000 1,486,987,200,000
464
Even better would be to change the format to %tcStatas clock-time format:

. format t %tc . list in 1/5 time 1. 2. 3. 4. 5. 02.13.2007 02.13.2007 02.13.2007 02.13.2007 02.13.2007 08:00 09:00 10:00 11:00 12:00 price 140 155 160 155 160 13feb2007 13feb2007 13feb2007 13feb2007 13feb2007 t 08:00:00 09:00:00 10:00:00 11:00:00 12:00:00
We could drop variable time. New variable t contains the same information as time and t is better because it is a Stata time variable, the most important property of which being that it is numeric rather than string. We can tsset it. Here, however, we also need to specify the period with tssets delta() option. Statas time variables are numeric, but they record milliseconds since 01jan1960 00:00:00. By default, tsset uses delta(1), and that means the time-series operators would not work as we want them to work. For instance, L.price would look back only 1 ms (and nd nothing). We want L.price to look back 1 hour (3,600,000 ms):
. tsset t, delta(1 hour) time variable: t, 13feb2007 08:00:00.000 to 13feb2007 14:00:00.000 delta: 1 hour . list t price l.price in 1/5 t 1. 2. 3. 4. 5. 13feb2007 13feb2007 13feb2007 13feb2007 13feb2007 08:00:00 09:00:00 10:00:00 11:00:00 12:00:00 price 140 155 160 155 160 L.price . 140 155 160 155
Example 7: Clock times must be double

In the previous example, it was of vital importance that when we generated the %tc variable t, . generate double t = clock(time, "MDY hm") we generated it as a double. Lets see what would have happened had we forgotten and just typed generate t = clock(time, "MDY hm"). Lets go back and start with the same original data:
. use http://www.stata-press.com/data/r12/tssetxmpl4, clear . list in 1/5 time 1. 2. 3. 4. 5. 02.13.2007 02.13.2007 02.13.2007 02.13.2007 02.13.2007 08:00 09:00 10:00 11:00 12:00 price 140 155 160 155 160
465
Remember, variable time is a string variable, and we need to translate it to numeric. So we translate, but this time we forget to make the new variable a double:
. generate t = clock(time, "MDY hm") . list in 1/5 time 1. 2. 3. 4. 5. 02.13.2007 02.13.2007 02.13.2007 02.13.2007 02.13.2007 08:00 09:00 10:00 11:00 12:00 price 140 155 160 155 160 t 1.49e+12 1.49e+12 1.49e+12 1.49e+12 1.49e+12
We see the rst differencet now lists as 1.49e+12 rather than 1.487e+12 as it did previouslybut this is nothing that would catch our attention. We would not even know that the value is different. Lets continue. We next put a %20.0gc format on t to better see the numerical values. In fact, that is not something we would usually do in an analysis. We did that in the example to emphasize to you that the t values were really big numbers. We will repeat the exercise just to be complete, but in real analysis, we would not bother.
. format t %20.0gc . list in 1/5 time 1. 2. 3. 4. 5. 02.13.2007 02.13.2007 02.13.2007 02.13.2007 02.13.2007 08:00 09:00 10:00 11:00 12:00 price 140 155 160 155 160 t 1,486,972,780,544 1,486,976,450,560 1,486,979,989,504 1,486,983,659,520 1,486,987,198,464
Okay, we see big numbers in t. Lets continue. Next we put a %tc format on t, and that is something we would usually do, and you should always do. You should also list a bit of the data, as we did:
. format t %tc . list in 1/5 time 1. 2. 3. 4. 5. 02.13.2007 02.13.2007 02.13.2007 02.13.2007 02.13.2007 08:00 09:00 10:00 11:00 12:00 price 140 155 160 155 160 13feb2007 13feb2007 13feb2007 13feb2007 13feb2007 t 07:59:40 09:00:50 09:59:49 11:00:59 11:59:58
By now, you should see a problem: the translated datetime values are off by a second or two. That was caused by rounding. Dates and times should be the same, not approximately the same, and when you see a difference like this, you should say to yourself, The translation is off a little. Why is that? and then you should think, Of course, rounding. I bet that I did not create t as a double.
466
Let us assume, however, that you do not do this. You instead plow ahead:
. tsset t, delta(1 hour) time values with period less than delta() found r(451);
And that is what will happen when you forget to create t as a double. The rounding will cause uneven period, and tsset will complain. By the way, it is only important that clock times (%tc and %tC variables) be stored as doubles. The other date values %td, %tw, %tm, %tq, %th, and %ty are small enough that they can safely be stored as floats, although forgetting and storing them as doubles does no harm.
Technical note
Stata provides two clock-time formats, %tc and %tC. %tC provides a clock with leap seconds. Leap seconds are occasionally inserted to account for randomness of the earths rotation, which gradually slows. Unlike the extra day inserted in leap years, the timing of when leap seconds will be inserted cannot be foretold. The authorities in charge of such matters announce a leap second approximately 6 months before insertion. Leap seconds are inserted at the end of the day, and the leap second is called 23:59:60 (that is, 11:59:60 pm), which is then followed by the usual 00:00:00 (12:00:00 am). Most nonastronomers nd these leap seconds vexing. The added seconds cause problems because of their lack of predictabilityknowing how many seconds there will be between 01jan2012 and 01jan2013 is not possibleand because there are not necessarily 24 hours in a day. If you use a leap second adjustedclock, most days have 24 hours, but a few have 24 hours and 1 second. You must look at a table to nd out. From a time-series analysis point of view, the nonconstant day causes the most problems. Lets say that you have data on blood pressure, taken hourly at 1:00, 2:00, . . . , and that you have tsset your data with delta(1 hour). On most days, L24.bp would be blood pressure at the same time yesterday. If the previous day had a leap second, however, and your data were recorded using a leap second adjustedclock, there would be no observation L24.bp because 86,400 seconds before the current reading does not correspond to an on-the-hour time; 86,401 seconds before the current reading corresponds to yesterdays time. Thus, whenever possible, using Statas %tc encoding rather than %tC is better. When times are recorded by computers using leap secondadjusted clocks, however, avoiding %tC is not possible. For performing most time-series analysis, the recommended procedure is to map the %tC values to %tc and then tsset those. You must ask yourself whether the process you are studying is based on the clockthe nurse does something at 2 oclock every dayor the true passage of timethe emitter spits out an electron every 86,400,000 ms. When dealing with computer-recorded times, rst nd out whether the computer (and its timerecording software) use a leap secondadjusted clock. If it does, translate that to a %tC value. Then use function cofC() to convert to a %tc value and tsset that. If variable T contains the %tC value,
. gen double t = cofC(T) . format t %tc . tsset t, delta(. . . )
Function cofC() moves leap seconds forward: 23:59:60 becomes 00:00:00 of the next day.
467
Panel data Example 8: Time-series data for multiple groups

Assume that we have a time series on average annual income and that we have the series for two groups: individuals who have not completed high school (edlevel = 1) and individuals who have (edlevel = 2).
. use http://www.stata-press.com/data/r12/tssetxmpl5, clear . list edlevel year income, sep(0) edlevel 1. 2. 3. 4. 5. 6. 7. 1 1 1 1 2 2 2 year 1988 1989 1990 1991 1989 1990 1992 income 14500 14750 14950 15100 22100 22200 22800
We declare the data to be a panel by typing

. tsset edlevel year, yearly panel variable: edlevel, (unbalanced) time variable: year, 1988 to 1992, but with a gap delta: 1 year
Having tsset the data, we can now use time-series operators. The difference operator, for example, can be used to list annual changes in income:
. list edlevel year income d.income, sep(0) edlevel 1. 2. 3. 4. 5. 6. 7. 1 1 1 1 2 2 2 year 1988 1989 1990 1991 1989 1990 1992 income 14500 14750 14950 15100 22100 22200 22800 D.income . 250 200 150 . 100 .
We see that in addition to producing missing values due to missing times, the difference operator correctly produced a missing value at the start of each panel. Once we have tsset our panel data, we can use time-series operators and be assured that they will handle missing time periods and panel changes correctly.
468
Saved results
tsset saves the following in r():
Scalars r(imin) r(imax) r(tmin) r(tmax) r(tdelta) Macros r(panelvar) r(timevar) r(tdeltas) r(tmins) r(tmaxs) r(tsfmt) r(unit) r(unit1) r(balanced) minimum panel ID maximum panel ID minimum time maximum time delta name of panel variable name of time variable formatted delta formatted minimum time formatted maximum time %fmt of time variable units of time variable: Clock, clock, daily, weekly, monthly, quarterly, halfyearly, yearly, or generic units of time variable: C, c, d, w, m, q, h, y, or "" unbalanced, weakly balanced, or strongly balanced; a set of panels are strongly balanced if they all have the same time values, otherwise balanced if same number of time values, otherwise unbalanced

tsset is implemented as an ado-le.
Reference
Baum, C. F. 2000. sts17: Compacting time series data. Stata Technical Bulletin 57: 4445. Reprinted in Stata Technical Bulletin Reprints, vol. 10, pp. 369370. College Station, TX: Stata Press.
Also see
[TS] tsll Fill in gaps in time variable
Title
tssmooth Smooth and forecast univariate time-series data
Syntax
tssmooth smoother type newvar = exp if in , ...
Smoother category Moving average with uniform weights with specied weights Recursive exponential double exponential nonseasonal HoltWinters seasonal HoltWinters Nonlinear lter
smoother ma ma exponential dexponential hwinters shwinters nl
See [TS] tssmooth ma, [TS] tssmooth exponential, [TS] tssmooth dexponential, [TS] tssmooth hwinters, [TS] tssmooth shwinters, and [TS] tssmooth nl.
Description
tssmooth creates new variable newvar and lls it in by passing the specied expression (usually a variable name) through the requested smoother.
Remarks
The recursive smoothers may also be used for forecasting univariate time series; indeed, the HoltWinters methods are used almost exclusively for this. All can perform dynamic out-of-sample forecasts, and the smoothing parameters may be chosen to minimize the in-sample sum-of-squared prediction errors. The moving-average and nonlinear smoothers are generally used to extract the trendor signal from a time series while omitting the high-frequency or noise components. All smoothers work both with time-series data and panel data. When used with panel data, the calculation is performed separately within panel. Several texts provide good introductions to the methods available in tssmooth. Chateld (2004) discusses how these methods t into time-series analysis in general. Abraham and Ledolter (1983); Montgomery, Johnson, and Gardiner (1990); Bowerman, OConnell, and Koehler (2005); and Chateld (2001) discuss using these methods for modern time-series forecasting.
469
470
tssmooth Smooth and forecast univariate time-series data
References
Abraham, B., and J. Ledolter. 1983. Statistical Methods for Forecasting. New York: Wiley. Bowerman, B. L., R. T. OConnell, and A. B. Koehler. 2005. Forecasting, Time Series, and Regression: An Applied Approach. 4th ed. Pacic Grove, CA: Brooks/Cole. Chateld, C. 2001. Time-Series Forecasting. London: Chapman & Hall/CRC. . 2004. The Analysis of Time Series: An Introduction. 6th ed. Boca Raton, FL: Chapman & Hall/CRC. Chateld, C., and M. Yar. 1988. Holt-Winters forecasting: Some practical issues. Statistician 37: 129140. Holt, C. C. 2004. Forecasting seasonals and trends by exponentially weighted moving averages. International Journal of Forecasting 20: 510. Montgomery, D. C., L. A. Johnson, and J. S. Gardiner. 1990. Forecasting and Time Series Analysis. 2nd ed. New York: McGrawHill. Winters, P. R. 1960. Forecasting sales by exponentially weighted moving averages. Management Science 6: 324342.
Also see
[TS] tsset Declare data to be time-series data [TS] arima ARIMA, ARMAX, and other dynamic regression models [TS] sspace State-space models [TS] tslter Filter a time-series, keeping only selected periodicities [R] smooth Robust nonlinear smoother
Title
tssmooth dexponential Double-exponential smoothing
Syntax
tssmooth dexponential options
Main
type
newvar = exp
if
in
, options
Description replace newvar if it already exists use # as smoothing parameter use # observations to obtain initial values for recursions use #1 and #2 as initial values for recursions use # periods for the out-of-sample forecast
replace parms(# ) samp0(#) s0(#1 #2 ) forecast(#)
You must tsset your data before using tssmooth dexponential; see [TS] tsset. exp may contain time-series operators; see [U] 11.4.4 Time-series varlists.
Menu
Statistics
>
Time series
>
Smoothers/univariate forecasters
>
Double-exponential smoothing
Description
tssmooth dexponential models the trend of a variable whose difference between changes from the previous values is serially correlated. More precisely, it models a variable whose second difference follows a low-order, moving-average process.
Options

Main
replace replaces newvar if it already exists. parms(# ) species the parameter for the double-exponential smoothers; 0 < # < 1. If parms(# ) is not specied, the smoothing parameter is chosen to minimize the in-sample sum-of-squared forecast errors. samp0(#) and s0(#1 #2 ) are mutually exclusive ways of specifying the initial values for the recursion. By default, initial values are obtained by tting a linear regression with a time trend, using the rst half of the observations in the dataset; see Remarks. samp0(#) species that the rst # be used in that regression. s0(#1 #2 ) species that #1 #2 be used as initial values. forecast(#) species the number of periods for the out-of-sample prediction; 0 # 500. The default is forecast(0), which is equivalent to not performing an out-of-sample forecast.
471
472
Remarks
The double-exponential smoothing procedure is designed for series that can be locally approximated as
xt = mt + bt t
where xt is the smoothed or predicted value of the series x, and the terms mt and bt change over time. Abraham and Ledolter (1983), Bowerman, OConnell, and Koehler (2005), and Montgomery, Johnson, and Gardiner (1990) all provide good introductions to double-exponential smoothing. Chateld (2001, 2004) provides helpful discussions of how double-exponential smoothing relates to modern time-series methods. The double-exponential method has been used both as a smoother and as a prediction method. Recall from [TS] tssmooth exponential that the single-exponential smoothed series is given by
St = xt + (1 )St1
where is the smoothing constant and xt is the original series. The double-exponential smoother is obtained by smoothing the smoothed series,
St = St + (1 )St1
Values of S0 and S0 are necessary to begin the process. Per Montgomery, Johnson, and Gar[2] diner (1990), the default method is to obtain S0 and S0 from a regression of the rst Npre values of xt on t = (1, . . . , Npre t0 ) . By default, Npre is equal to one-half the number of observations in the sample. Npre can be specied using the samp0() option. The values of S0 and S0 can also be specied using the option s0().
[2] [2]
[2]
[2]
Example 1
Suppose that we had some data on the monthly sales of a book and that we wanted to smooth this series. The graph below illustrates that this series is locally trending over time, so we would not want to use single-exponential smoothing.
Monthly book sales
140 Sales 90 0 100 110 120 130
20 Time
40
60
473
The following example illustrates that double-exponential smoothing is simply smoothing the smoothed series. Because the starting values are treated as time-zero values, we actually lose 2 observations when smoothing the smoothed series.
. use http://www.stata-press.com/data/r12/sales2 . tssmooth exponential double sm1=sales, p(.7) s0(1031) exponential coefficient = sum-of-squared residuals = root mean squared error = exponential coefficient = sum-of-squared residuals = root mean squared error = double-exponential coefficient sum-of-squared residuals root mean squared error . generate double sm2c = f2.sm2 (2 missing values generated) . list sm2b sm2c in 1/10 sm2b 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 1031 1028.3834 1030.6306 1017.8182 1022.938 1026.0752 1041.8587 1042.8341 1035.9571 1030.6651 sm2c 1031 1028.3834 1030.6306 1017.8182 1022.938 1026.0752 1041.8587 1042.8341 1035.9571 1030.6651 0.7000 13923 13.192 0.7000 7698.6 9.8098 = = = 0.7000 3724.4 6.8231
. tssmooth exponential double sm2=sm1, p(.7) s0(1031)
. tssmooth dexponential double sm2b=sales, p(.7) s0(1031 1031)
The double-exponential method can also be viewed as a forecasting mechanism. The exponential forecast method is a constrained version of the HoltWinters method implemented in [TS] tssmooth hwinters (as discussed by Gardner [1985] and Chateld [2001]). Chateld (2001) also notes that the double-exponential method arises when the underlying model is an ARIMA(0,2,2) with equal roots. This method produces predictions xt for t = t1 , . . . , T + forecast(). These predictions are obtained as a function of the smoothed series and the smoothed-smoothed series. For t [t0 , T ],
xt = 2 +
where St and St
[2]
[2] St 1 + S 1 1 t
are as given above.
The out-of-sample predictions are obtained as a function of the constant term, the linear term of the [2] smoothed series at the last observation in the sample, and time. The constant term is aT = 2ST ST , [2] and the linear term is bT = 1 (ST ST ). The th-step-ahead out-of-sample prediction is given by xt = a t + b T
474
Example 2
Specifying the forecast option puts the double-exponential forecast into the new variable instead of the double-exponential smoothed series. The code given below uses the smoothed series sm1 and sm2 that were generated above to illustrate how the double-exponential forecasts are computed.
. tssmooth dexponential double f1=sales, p(.7) s0(1031 1031) forecast(4) double-exponential coefficient = 0.7000 sum-of-squared residuals = 20737 root mean squared error = 16.1 . generate double xhat = (2 + .7/.3) * sm1 - (1 + .7/.3) * f.sm2 (5 missing values generated) . list xhat f1 in 1/10 xhat 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 1031 1031 1023.524 1034.8039 994.0237 1032.4463 1031.9015 1071.1709 1044.6454 1023.1855 f1 1031 1031 1023.524 1034.8039 994.0237 1032.4463 1031.9015 1071.1709 1044.6454 1023.1855
Example 3
Generally, when you are forecasting, you do not know the smoothing parameter. tssmooth dexponential computes the double-exponential forecasts of a series and obtains the optimal smoothing parameter by nding the smoothing parameter that minimizes the in-sample sum-of-squared forecast errors.
. tssmooth dexponential f2=sales, forecast(4) computing optimal double-exponential coefficient (0,1) optimal double-exponential coefficient = 0.3631 sum-of-squared residuals = 16075.805 root mean squared error = 14.175598
The following graph describes the t that we obtained by applying the double-exponential forecast method to our sales data. The out-of-sample dynamic predictions are not constant, as in the singleexponential case.

. line f2 sales t, title("Double exponential forecast with optimal alpha") > ytitle(Sales) xtitle(time)
475
Sales
950
1000
1050
1100
Double exponential forecast with optimal alpha
20
40 time dexpc(0.3631) = sales
60 sales
80
tssmooth dexponential automatically detects panel data from the information provided when the dataset was tsset. The starting values are chosen separately for each series. If the smoothing parameter is chosen to minimize the sum-of-squared prediction errors, the optimization is performed separately on each panel. The saved results contain the results from the last panel. Missing values at the beginning of the sample are excluded from the sample. After at least one value has been found, missing values are lled in using the one-step-ahead predictions from the previous period.
Saved results
tssmooth dexponential saves the following in r():
Scalars r(N) r(alpha) r(rss) r(rmse) r(N pre) r(s2 0) r(s1 0) r(linear) r(constant) r(period) Macros r(method) r(exp) r(timevar) r(panelvar) number of observations smoothing parameter sum-of-squared errors root mean squared error number of observations used in calculating starting values, if starting values calculated [2] initial value for linear term, i.e., S0 initial value for constant term, i.e., S0 nal value of linear term nal value of constant term period, if lter is seasonal smoothing method expression specied time variable specied in tsset panel variable specied in tsset

tssmooth dexponential is implemented as an ado-le.
476
A truncated description of the specied double-exponential lter is used to label the new variable. See [D] label for more information on labels. An untruncated description of the specied double-exponential lter is saved in the characteristic tssmooth for the new variable. See [P] char for more information on characteristics. The updating equations for the smoothing and forecasting versions are as given previously. The starting values for both the smoothing and forecasting versions of double-exponential are obtained using the same method, which begins with the model
xt = 0 + 1 t
where xt is the series to be smoothed and t is a time variable that has been normalized to equal 1 in the rst period included in the sample. The regression coefcient estimates 0 and 1 are obtained via OLS. The sample is determined by the option samp0(). By default, samp0() includes the rst half of the observations. Given the estimates 0 and 1 , the starting values are
S0 = 0 {(1 )/}1 S0 = 0 2{(1 )/}1

[2]
References
Abraham, B., and J. Ledolter. 1983. Statistical Methods for Forecasting. New York: Wiley. Bowerman, B. L., R. T. OConnell, and A. B. Koehler. 2005. Forecasting, Time Series, and Regression: An Applied Approach. 4th ed. Pacic Grove, CA: Brooks/Cole. Chateld, C. 2001. Time-Series Forecasting. London: Chapman & Hall/CRC. . 2004. The Analysis of Time Series: An Introduction. 6th ed. Boca Raton, FL: Chapman & Hall/CRC. Chateld, C., and M. Yar. 1988. Holt-Winters forecasting: Some practical issues. Statistician 37: 129140. Gardner, E. S., Jr. 1985. Exponential smoothing: The state of the art. Journal of Forecasting 4: 128. Holt, C. C. 2004. Forecasting seasonals and trends by exponentially weighted moving averages. International Journal of Forecasting 20: 510. Montgomery, D. C., L. A. Johnson, and J. S. Gardiner. 1990. Forecasting and Time Series Analysis. 2nd ed. New York: McGrawHill. Winters, P. R. 1960. Forecasting sales by exponentially weighted moving averages. Management Science 6: 324342.
Also see
[TS] tsset Declare data to be time-series data [TS] tssmooth Smooth and forecast univariate time-series data
Title
tssmooth exponential Single-exponential smoothing
Syntax
tssmooth exponential options
Main
type
newvar = exp
if
in
, options
Description replace newvar if it already exists use # as smoothing parameter use # observations to obtain initial value for recursion use # as initial value for recursion use # periods for the out-of-sample forecast
replace parms(# ) samp0(#) s0(#) forecast(#)
You must tsset your data before using tssmooth exponential; see [TS] tsset. exp may contain time-series operators; see [U] 11.4.4 Time-series varlists.
Menu
Statistics
>
Time series
>
>
Single-exponential smoothing
Description
tssmooth exponential models the trend of a variable whose change from the previous value is serially correlated. More precisely, it models a variable whose rst difference follows a low-order, moving-average process.
Options

Main
replace replaces newvar if it already exists. parms(# ) species the parameter for the exponential smoother; 0 < # < 1. If parms(# ) is not specied, the smoothing parameter is chosen to minimize the in-sample sum-of-squared forecast errors. samp0(#) and s0(#) are mutually exclusive ways of specifying the initial value for the recursion. samp0(#) species that the initial value be obtained by calculating the mean over the rst # observations of the sample. s0(#) species the initial value to be used. If neither option is specied, the default is to use the mean calculated over the rst half of the sample. forecast(#) gives the number of observations for the out-of-sample prediction; 0 # 500. The default value is forecast(0) and is equivalent to not forecasting out of sample.
477
478
Remarks
Introduction Examples Treatment of missing values
Introduction
Exponential smoothing can be viewed either as an adaptive-forecasting algorithm or, equivalently, as a geometrically weighted moving-average lter. Exponential smoothing is most appropriate when used with time-series data that exhibit no linear or higher-order trends but that do exhibit lowvelocity, aperiodic variation in the mean. Abraham and Ledolter (1983), Bowerman, OConnell, and Koehler (2005), and Montgomery, Johnson, and Gardiner (1990) all provide good introductions to single-exponential smoothing. Chateld (2001, 2004) discusses how single-exponential smoothing relates to modern time-series methods. For example, simple exponential smoothing produces optimal forecasts for several underlying models, including ARIMA(0,1,1) and the random-walk-plus-noise state-space model. (See Chateld [2001, sec. 4.3.1].) The exponential lter with smoothing parameter creates the series St , where
St = Xt + (1 )St1
for t = 1, . . . , T
and S0 is the initial value. This is the adaptive forecast-updating form of the exponential smoother. This implies that
T 1
St =
k=0
(1 )K XT k + (1 )T S0
which is the weighted moving-average representation, with geometrically declining weights. The choice of the smoothing constant determines how quickly the smoothed series or forecast will adjust to changes in the mean of the unltered series. For small values of , the response will be slow because more weight is placed on the previous estimate of the mean of the unltered series, whereas larger values of will put more emphasis on the most recently observed value of the unltered series.
Examples Example 1
Lets consider some examples using sales data. Here we forecast sales for three periods with a smoothing parameter of 0.4:
. use http://www.stata-press.com/data/r12/sales1 . tssmooth exponential sm1=sales, parms(.4) forecast(3) exponential coefficient = 0.4000 sum-of-squared residuals = 8345 root mean squared error = 12.919
To compare our forecast with the actual data, we graph the series and the forecasted series over time.

. line sm1 sales t, title("Single exponential forecast") ytitle(Sales) xtitle(Time)
479
1000
1020
Sales 1040 1060
1080
1100
Single exponential forecast
10
20 Time
30
40 sales
50
exp parms(0.4000) = sales
The graph indicates that our forecasted series may not be adjusting rapidly enough to the changes in the actual series. The smoothing parameter controls the rate at which the forecast adjusts. Smaller values of adjust the forecasts more slowly. Thus we suspect that our chosen value of 0.4 is too small. One way to investigate this suspicion is to ask tssmooth exponential to choose the smoothing parameter that minimizes the sum-of-squared forecast errors.
. tssmooth exponential sm2=sales, forecast(3) computing optimal exponential coefficient (0,1) 0.7815 6727.7056 11.599746 optimal exponential coefficient = sum-of-squared residuals = root mean squared error =
The output suggests that the value of = 0.4 is too small. The graph below indicates that the new forecast tracks the series much more closely than the previous forecast.
. line sm2 sales t, title("Single exponential forecast with optimal alpha") > ytitle(sales) xtitle(Time)
1000
1020
Sales 1040 1060
1080
1100
Single exponential forecast with optimal alpha
10
20 Time
30
40 sales
50
parms(0.7815) = sales
480
We noted above that simple exponential forecasts are optimal for an ARIMA (0,1,1) model. (See [TS] arima for tting ARIMA models in Stata.) Chateld (2001, 90) gives the following useful derivation that relates the MA coefcient in an ARIMA (0,1,1) model to the smoothing parameter in single-exponential smoothing. An ARIMA (0,1,1) is given by
xt xt1 =
t1
where t is an identically and independently distributed white-noise error term. Thus given , an estimate of , an optimal one-step prediction of xt+1 is xt+1 = xt + t . Because t is not observable, it can be replaced by t = xt xt1 yielding
xt+1 = xt + (xt xt1 )

Letting = 1 + and doing more rearranging implies that
xt+1 = (1 + )xt xt1 xt+1 = xt (1 )xt1
Example 2
Lets compare the estimate of the optimal smoothing parameter of 0.7815 with the one we could obtain using [TS] arima. Below we t an ARIMA(0,1,1) to the sales data and then remove the estimate of . The two estimates of are quite close, given the large estimated standard error of .
. arima sales, arima(0,1,1) (setting optimization to BHHH) Iteration 0: log likelihood = -189.91037 Iteration 1: log likelihood = -189.62405 Iteration 2: log likelihood = -189.60468 Iteration 3: log likelihood = -189.60352 Iteration 4: log likelihood = -189.60343 (switching optimization to BFGS) Iteration 5: log likelihood = -189.60342 ARIMA regression Sample: 2 - 50 Log likelihood = -189.6034 OPG Std. Err.
= = =
49 1.41 0.2347
D.sales sales _cons ARMA ma L1. /sigma
Coef.
P>|z|
.5025469
1.382727
0.36
0.716
-2.207548
3.212641
-.1986561 11.58992
.1671699 1.240607
-1.19 9.34
0.235 0.000
-.5263031 9.158378
.1289908 14.02147
Note: The test of the variance against zero is one sided, and the two-sided confidence interval is truncated at zero. . di 1 + _b[ARMA:L.ma] .80134387
481
Example 3
tssmooth exponential automatically detects panel data. Suppose that we had sales gures for ve companies in long form. Running tssmooth exponential on the variable that contains all ve series puts the smoothed series and the predictions in one variable in long form. When the smoothing parameter is chosen to minimize the squared prediction error, an optimal value for the smoothing parameter is chosen separately for each panel.
. use http://www.stata-press.com/data/r12/sales_cert, clear . tsset panel variable: id (strongly balanced) time variable: t, 1 to 100 delta: 1 unit . tssmooth exponential sm5=sales, forecast(3) -> id = 1 computing optimal exponential
coefficient (0,1) 0.8702 16070.567 12.676974
optimal exponential coefficient = sum-of-squared residuals = root mean squared error =
-> id = 2 computing optimal exponential coefficient (0,1) optimal exponential coefficient = 0.7003 sum-of-squared residuals = 20792.393 root mean squared error = 14.419568 -> id = 3 computing optimal exponential coefficient (0,1) optimal exponential coefficient = 0.6927 sum-of-squared residuals = 21629 root mean squared error = 14.706801 -> id = 4 computing optimal exponential coefficient (0,1) optimal exponential coefficient = 0.3866 sum-of-squared residuals = 22321.334 root mean squared error = 14.940326 -> id = 5 computing optimal exponential coefficient (0,1) optimal exponential coefficient = 0.4540 sum-of-squared residuals = 20714.095 root mean squared error = 14.392392
tssmooth exponential computed starting values and chose an optimal for each panel individually.
482
Treatment of missing values

Missing values in the middle of the data are lled in with the one-step-ahead prediction using the previous values. Missing values at the beginning or end of the data are treated as if the observations were not there. tssmooth exponential treats observations excluded from the sample by if and in just as if they were missing.
Example 4
Here the 28th observation is missing. The prediction for the 29th observation is repeated in the new series.
. use http://www.stata-press.com/data/r12/sales1, clear . tssmooth exponential sm1=sales, parms(.7) forecast(3) (output omitted ) . generate sales2=sales if t!=28 (4 missing values generated) . tssmooth exponential sm3=sales2, parms(.7) forecast(3) exponential coefficient = 0.7000 sum-of-squared residuals = 6842.4 root mean squared error = 11.817 . list t sales2 sm3 if t>25 & t<31 t 26. 27. 28. 29. 30. 26 27 28 29 30 sales2 1011.5 1028.3 . 1028.4 1054.8 sm3 1007.5 1010.3 1022.9 1022.9 1026.75
Because the data for t = 28 are missing, the prediction for period 28 has been used in its place. This implies that the updating equation for period 29 is
S29 = S28 + (1 )S28 = S28

which explains why the prediction for t = 28 is repeated. Because this is a single-exponential procedure, the loss of that one observation will not be noticed several periods later.

. generate diff = sm3-sm1 if t>28 (28 missing values generated) . list t diff if t>28 & t<39 t 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 29 30 31 32 33 34 35 36 37 38 diff -3.5 -1.050049 -.3150635 -.0946045 -.0283203 -.0085449 -.0025635 -.0008545 -.0003662 -.0001221
483
Example 5
Now consider an example in which there are data missing at the beginning and end of the sample.
. generate sales3=sales if t>2 & t<49 (7 missing values generated) . tssmooth exponential sm4=sales3, parms(.7) forecast(3) exponential coefficient = sum-of-squared residuals = root mean squared error = 0.7000 6215.3 11.624
. list t sales sales3 sm4 if t<5 | t>45 t 1. 2. 3. 4. 46. 47. 48. 49. 50. 51. 52. 53. 1 2 3 4 46 47 48 49 50 51 52 53 sales 1031 1022.1 1005.6 1025 1055.2 1056.8 1034.5 1041.1 1056.1 . . . sales3 . . 1005.6 1025 1055.2 1056.8 1034.5 . . . . . sm4 . . 1016.787 1008.956 1057.2 1055.8 1056.5 1041.1 1041.1 1041.1 1041.1 1041.1
The output above illustrates that missing values at the beginning or end of the sample cause the sample to be truncated. The new series begins with nonmissing data and begins predicting immediately after it stops. One period after the actual data concludes, the exponential forecast becomes a constant. After the actual end of the data, the forecast at period t is substituted for the missing data. This also illustrates why the forecasted series is a constant.
484
Saved results
tssmooth exponential saves the following in r():
Scalars r(N) r(alpha) r(rss) r(rmse) r(N pre) r(s1 0) Macros r(method) r(exp) r(timevar) r(panelvar) number of observations smoothing parameter sum-of-squared prediction errors root mean squared error number of observations used in calculating starting values initial value for St smoothing method expression specied time variable specied in tsset panel variable specied in tsset

tssmooth exponential is implemented as an ado-le. The formulas for deriving smoothed series are as given in the text. When the value of is not specied, an optimal value is found that minimizes the mean squared forecast error. A method of bisection is used to nd the solution to this optimization problem. A truncated description of the specied exponential lter is used to label the new variable. See [D] label for more information about labels. An untruncated description of the specied exponential lter is saved in the characteristic tssmooth for the new variable. See [P] char for more information about characteristics.
References
Also see
Title
tssmooth hwinters HoltWinters nonseasonal smoothing
Syntax
tssmooth hwinters options
Main
type
newvar = exp
if
in
, options
Description replace newvar if it already exists use # and # as smoothing parameters use # observations to obtain initial values for recursion use #cons and #lt as initial values for recursion use # periods for the out-of-sample forecast alternative initial-value specication; see Options control the maximization process; seldom used use # and # as starting values for the parameters
replace parms(# # ) samp0(#) s0(#cons #lt ) forecast(#)

Options
diff
Maximization
maximize options from(# # )
You must tsset your data before using tssmooth hwinters; see [TS] tsset. exp may contain time-series operators; see [U] 11.4.4 Time-series varlists.
Menu
Statistics
>
Time series
>
>
Holt-Winters nonseasonal smoothing
Description
tssmooth hwinters is used in smoothing or forecasting a series that can be modeled as a linear trend in which the intercept and the coefcient on time vary over time.
Options

Main
replace replaces newvar if it already exists. parms(# # ), 0 # 1 and 0 # 1, species the parameters. If parms() is not specied, the values are chosen by an iterative process to minimize the in-sample sum-of-squared prediction errors. If you experience difculty converging (many iterations and not concave messages), try using from() to provide better starting values. samp0(#) and s0(#cons #lt ) specify how the initial values #cons and #lt for the recursion are obtained.
485
486
By default, initial values are obtained by tting a linear regression with a time trend using the rst half of the observations in the dataset. samp0(#) species that the rst # observations be used in that regression. s0(#cons #lt ) species that #cons and #lt be used as initial values. forecast(#) species the number of periods for the out-of-sample prediction; 0 # 500. The default is forecast(0), which is equivalent to not performing an out-of-sample forecast.
Options
diff species that the linear term is obtained by averaging the rst difference of expt and the intercept is obtained as the difference of exp in the rst observation and the mean of D.expt . If the diff option is not specied, a linear regression of expt on a constant and t is t.
Maximization
maximize options controls the process for solving for the optimal and when parms() is not specied. maximize options: nodifficult, technique(algorithm spec), iterate(#), no log, trace, gradient, showstep, hessian, showtolerance, tolerance(#), ltolerance(#), nrtolerance(#), and nonrtolerance; see [R] maximize. These options are seldom used. from(# # ), 0 < # < 1 and 0 < # < 1, species starting values from which the optimal values of and will be obtained. If from() is not specied, from(.5 .5) is used.
Remarks
The HoltWinters method forecasts series of the form
xt+1 = at + bt t
where xt is the forecast of the original series xt , at is a mean that drifts over time, and bt is a coefcient on time that also drifts. In fact, as Gardner (1985) has noted, the HoltWinters method produces optimal forecasts for an ARIMA(0,2,2) model and some local linear models. See [TS] arima and the references in that entry for ARIMA models, and see Harvey (1989) for a discussion of the local linear model and its relationship to the HoltWinters method. Abraham and Ledolter (1983), Bowerman, OConnell, and Koehler (2005), and Montgomery, Johnson, and Gardiner (1990) all provide good introductions to the HoltWinters method. Chateld (2001, 2004) provides helpful discussions of how this method relates to modern time-series analysis. The HoltWinters method can be viewed as an extension of double-exponential smoothing with two parameters, which may be explicitly set or chosen to minimize the in-sample sum-of-squared forecast errors. In the latter case, as discussed in Methods and formulas, the smoothing parameters are chosen to minimize the in-sample sum-of-squared forecast errors plus a penalty term that helps to achieve convergence when one of the parameters is too close to the boundary. Given the series xt , the smoothing parameters and , and the starting values a0 and b0 , the updating equations are at = xt + (1 ) (at1 + bt1 )
bt = (at at1 ) + (1 ) bt1
487
After computing the series of constant and linear terms, at and bt , respectively, the -step-ahead prediction of xt is given by
xt+ = at + bt
Example 1
Below we show how to use tssmooth hwinters with specied smoothing parameters. This example also shows that the HoltWinters method can closely follow a series in which both the mean and the time coefcient drift over time. Suppose that we have data on the monthly sales of a book and that we want to forecast this series with the HoltWinters method.
. use http://www.stata-press.com/data/r12/bsales . tssmooth hwinters hw1=sales, parms(.7 .3) forecast(3) Specified weights: alpha = 0.7000 beta = 0.3000 sum-of-squared residuals = 2301.046 root mean squared error = 6.192799 . line sales hw1 t, title("Holt-Winters Forecast with alpha=.7 > ytitle(Sales) xtitle(Time)
and beta=.3")
HoltWinters forecast with alpha=.7 and beta = .3

140 90 0 sales 100 Sales 110 120 130
20 Time
40 hw parms(0.700 0.300) = sales
60
The graph indicates that the forecasts are for linearly decreasing sales. Given aT and bT , the out-ofsample predictions are linear functions of time. In this example, the slope appears to be too steep, probably because our choice of and .
Example 2
The graph in the previous example illustrates that the starting values for the linear and constant series can affect the in-sample t of the predicted series for the rst few observations. The previous example used the default method for obtaining the initial values for the recursion. The output below illustrates that, for some problems, the differenced-based initial values provide a better in-sample t
488
for the rst few observations. However, the differenced-based initial values do not always outperform the regression-based initial values. Furthermore, as shown in the output below, for series of reasonable length, the predictions produced are nearly identical.
. tssmooth hwinters hw2=sales, parms(.7 .3) forecast(3) diff Specified weights: alpha beta sum-of-squared residuals root mean squared error = = = = 0.7000 0.3000 2261.173 6.13891
. list hw1 hw2 if _n<6 | _n>57 hw1 1. 2. 3. 4. 5. 58. 59. 60. 61. 62. 63. 93.31973 98.40002 100.8845 98.50404 93.62408 116.5771 119.2146 119.2608 111.0299 109.2815 107.5331 hw2 97.80807 98.11447 99.2267 96.78276 92.2452 116.5771 119.2146 119.2608 111.0299 109.2815 107.5331
When the smoothing parameters are chosen to minimize the in-sample sum-of-squared forecast errors, changing the initial values can affect the choice of the optimal and . When changing the initial values results in different optimal values for and , the predictions will also differ.
When the HoltWinters model ts the data well, nding the optimal smoothing parameters generally proceeds well. When the model ts poorly, nding the and that minimize the in-sample sum-of-squared forecast errors can be difcult.
Example 3
In this example, we forecast the book sales data using the and that minimize the in-sample squared forecast errors.
. tssmooth hwinters hw3=sales, forecast(3) computing optimal weights Iteration Iteration Iteration Iteration Iteration Iteration 0: 1: 2: 3: 4: 5: penalized penalized penalized penalized penalized penalized RSS RSS RSS RSS RSS RSS = = = = = = -2632.2073 -1982.8431 -1976.4236 -1975.9175 -1975.9036 -1975.9036 = = = = = (not concave)
Optimal weights: alpha beta penalized sum of squared residuals sum of squared residuals root mean-squared error 0.8209 0.0067 1975.904 1975.904 5.738617
489
The following graph contains the data and the forecast using the optimal and . Comparing this graph with the one above illustrates how different choices of and can lead to very different forecasts. Instead of linearly decreasing sales, the new forecast is for linearly increasing sales.
. line sales hw3 t, title("Holt-Winters Forecast with optimal alpha and beta") > ytitle(Sales) xtitle(Time)
HoltWinters forecast with optimal alpha and beta

140 90 0 sales 100 Sales 110 120 130
20 Time
40 hw parms(0.821 0.007) = sales
60
Saved results
tssmooth hwinters saves the following in r():
Scalars r(N) r(alpha) r(beta) r(rss) r(prss) r(rmse) Macros r(method) r(exp) number of observations smoothing parameter smoothing parameter sum-of-squared errors penalized sum-of-squared errors, if parms() not specied root mean squared error smoothing method expression specied r(N pre) r(s2 0) r(s1 0) r(linear) r(constant) number of observations used in calculating starting values initial value for linear term initial value for constant term nal value of linear term nal value of constant term
r(timevar) r(panelvar)
time variables specied in tsset panel variables specied in tsset

tssmooth hwinters is implemented as an ado-le. A truncated description of the specied HoltWinters lter is used to label the new variable. See [D] label for more information on labels. An untruncated description of the specied HoltWinters lter is saved in the characteristic named tssmooth for the new variable. See [P] char for more information on characteristics.
490
Given the series, xt ; the smoothing parameters, and ; and the starting values, a0 and b0 , the updating equations are
at = xt + (1 ) (at1 + bt1 ) bt = (at at1 ) + (1 ) bt1

By default, the initial values are found by tting a linear regression with a time trend. The time variable in this regression is normalized to equal one in the rst period included in the sample. By default, one-half of the data is used in this regression, but this sample can be changed using samp0(). a0 is then set to the estimate of the constant, and b0 is set to the estimate of the coefcient on the time trend. Specifying the diff option sets b0 to the mean of D.x and a0 to x1 b0 . s0() can also be used to specify the initial values directly. Sometimes, one or both of the optimal parameters may lie on the boundary of [ 0, 1 ]. To keep the estimates inside [ 0, 1 ], tssmooth hwinters parameterizes the objective function in terms of their inverse logits, that is, in terms of exp()/{1 + exp()} and exp()/{1 + exp()}. When one of these parameters is actually on the boundary, this can complicate the optimization. For this reason, tssmooth hwinters optimizes a penalized sum-of-squared forecast errors. Let xt (, ) be the forecast for the series xt , given the choices of and . Then the in-sample penalized sum-of-squared prediction errors is
T
P =
t=1
{xt xt (, )}2 + I|f ()|>12) (|f ()| 12)2 + I|f ()|>12) (|f ()| 12)2
where f (x) = ln {x(1 x)}. The penalty term is zero unless one of the parameters is close to the boundary. When one of the parameters is close to the boundary, the penalty term will help to obtain convergence.
Acknowledgment
We thank Nicholas J. Cox of Durham University for his helpful comments.
References
Abraham, B., and J. Ledolter. 1983. Statistical Methods for Forecasting. New York: Wiley. Bowerman, B. L., R. T. OConnell, and A. B. Koehler. 2005. Forecasting, Time Series, and Regression: An Applied Approach. 4th ed. Pacic Grove, CA: Brooks/Cole. Chateld, C. 2001. Time-Series Forecasting. London: Chapman & Hall/CRC. . 2004. The Analysis of Time Series: An Introduction. 6th ed. Boca Raton, FL: Chapman & Hall/CRC. Chateld, C., and M. Yar. 1988. Holt-Winters forecasting: Some practical issues. Statistician 37: 129140. Gardner, E. S., Jr. 1985. Exponential smoothing: The state of the art. Journal of Forecasting 4: 128. Harvey, A. C. 1989. Forecasting, Structural Time Series Models and the Kalman Filter. Cambridge: Cambridge University Press. Holt, C. C. 2004. Forecasting seasonals and trends by exponentially weighted moving averages. International Journal of Forecasting 20: 510. Montgomery, D. C., L. A. Johnson, and J. S. Gardiner. 1990. Forecasting and Time Series Analysis. 2nd ed. New York: McGrawHill. Winters, P. R. 1960. Forecasting sales by exponentially weighted moving averages. Management Science 6: 324342.
491
Also see
Title
tssmooth ma Moving-average lter
Syntax
Moving average with uniform weights tssmooth ma type newvar = exp if in , window(#l #c #f ) replace
Moving average with specied weights tssmooth ma replace

You must tsset your data before using tssmooth ma; see [TS] tsset. exp may contain time-series operators; see [U] 11.4.4 Time-series varlists.
type
newvar = exp
if
in , weights( numlistl
<#c > numlistf )
Menu
Statistics
>
Time series
>
>
Moving-average lter
Description
tssmooth ma creates a new series in which each observation is an average of nearby observations in the original series. In the rst syntax, window() is required and species the span of the lter. tssmooth ma constructs a uniformly weighted moving average of the expression. In the second syntax, weights() is required and species the weights to be used. tssmooth ma then applies the specied weights to construct a weighted moving average of the expression.
Options
window(#l #c #f ) describes the span of the uniformly weighted moving average. #l species the number of lagged terms to be included, 0 #l one-half the number of observations in the sample. #c is optional and species whether to include the current observation in the lter. A 0 indicates exclusion and 1, inclusion. The current observation is excluded by default. #f is optional and species the number of forward terms to be included, 0 #f one-half the number of observations in the sample. weights( numlistl <#c > numlistf ) is required for the weighted moving average and describes the span of the moving average, as well as the weights to be applied to each term in the average. The middle term literally is surrounded by < and >, so you might type weights(1/2 <3> 2/1).
492
493
numlistl is optional and species the weights to be applied to the lagged terms when computing the moving average. #c is required and species the weight to be applied to the current term. numlistf is optional and species the weights to be applied to the forward terms when computing the moving average. The number of elements in each numlist is limited to one-half the number of observations in the sample. replace replaces newvar if it already exists.
Remarks
Moving averages are simple linear lters of the form
f i=l wi xt+i f i=l wi
xt =
where
xt xt wi f
is the moving average is the variable or expression to be smoothed are the weights being applied to the terms in the lter is the longest lead in the span of the lter
l is the longest lag in the span of the lter
Moving averages are used primarily to reduce noise in time-series data. Using moving averages to isolate signals is problematic, however, because the moving averages themselves are serially correlated, even when the underlying data series is not. Still, Chateld (2004) discusses moving-average lters and provides several specic moving-average lters for extracting certain trends.
Example 1
Suppose that we have a time series of sales data, and we want to separate the data into two components: signal and noise. To eliminate the noise, we apply a moving-average lter. In this example, we use a symmetric moving average with a span of 5. This means that we will average the rst two lagged values, the current value, and the rst two forward terms of the series, with each term in the average receiving a weight of 1.
. use http://www.stata-press.com/data/r12/sales1 . tsset time variable: t, 1 to 50 delta: 1 unit . tssmooth ma sm1 = sales, window(2 1 2) The smoother applied was (1/5)*[x(t-2) + x(t-1) + 1*x(t) + x(t+1) + x(t+2)]; x(t)= sales
We would like to smooth our series so that there is no autocorrelation in the noise. Below we compute the noise as the difference between the smoothed series and the series itself. Then we use ac (see [TS] corrgram) to check for autocorrelation in the noise.
494

. generate noise = sales-sm1 . ac noise
0.40 0.40 Autocorrelations of noise 0.20 0.00 0.20 0
10 Lag
15
20
25
Example 2
In the previous example, there is some evidence of negative second-order autocorrelation, possibly due to the uniform weighting or the length of the lter. We are going to specify a shorter lter in which the weights decline as the observations get farther away from the current observation. The weighted moving-average lter requires that we supply the weights to apply to each element with the weights() option. In specifying the weights, we implicitly specify the span of the lter. Below we use the lter
xt = (1/9)(1xt2 + 2xt1 + 3xt + 2xt+1 + 1xt+2 )

In what follows, 1/2 does not mean one-half, it means the numlist 1 2:
. tssmooth ma sm2 = sales, weights( 1/2 <3> 2/1) The smoother applied was (1/9)*[1*x(t-2) + 2*x(t-1) + 3*x(t) + 2*x(t+1) + 1*x(t+2)]; x(t)= sales . generate noise2 = sales-sm2
We compute the noise and use ac to check for autocorrelation.

. ac noise2
0.40
495
0.40
Autocorrelations of noise2 0.20 0.00 0.20 0
10 Lag
15
20
25
The graph shows no signicant evidence of autocorrelation in the noise from the second lter.
Technical note
tssmooth ma gives any missing observations a coefcient of zero in both the uniformly weighted and weighted moving-average lters. This simply means that missing values or missing periods are excluded from the moving average. Sample restrictions, via if and in, cause the expression smoothed by tssmooth ma to be missing for the excluded observations. Thus sample restrictions have the same effect as missing values in a variable that is ltered in the expression. Also, gaps in the data that are longer than the span of the lter will generate missing values in the ltered series. Because the rst l observations and the last f observations will be outside the span of the lter, those observations will be set to missing in the moving-average series.
Saved results
tssmooth ma saves the following in r():
Scalars r(N) r(w0) r(wlead#) r(wlag#) Macros r(method) r(exp) r(timevar) r(panelvar) number of observations weight on the current observation weight on lead #, if leads are specied weight on lag #, if lags are specied smoothing method expression specied time variable specied in tsset panel variable specied in tsset
496

tssmooth ma is implemented as an ado-le. The formula for moving averages is the same as previously given. A truncated description of the specied moving-average lter labels the new variable. See [D] label for more information on labels. An untruncated description of the specied moving-average lter is saved in the characteristic tssmooth for the new variable. See [P] char for more information on characteristics.
Reference
Chateld, C. 2004. The Analysis of Time Series: An Introduction. 6th ed. Boca Raton, FL: Chapman & Hall/CRC.
Also see
Title
tssmooth nl Nonlinear lter
Syntax
tssmooth nl replace where smoother is specied as Sm Sm . . . and Sm is one of R type newvar = exp if in , smoother(smoother , twice )
1|2|3|4|5|6|7|8| 9 3 R S S|R S|R ... E H
The numbers specied in smoother represent the span of a running median smoother. For example, a number 3 species that each value be replaced by the median of the point and the two adjacent data values. The letter H indicates that a Hanning linear smoother, which is a span-3 smoother with binomial weights, be applied. The letters E, S, and R are three renements that can be combined with the running median and Hanning smoothers. First, the end points of a smooth can be given special treatment. This is specied by the E operator. Second, smoothing by 3, the span-3 running median, tends to produce at-topped hills and valleys. The splitting operator, S, splits these repeated values, applies the end-point operator to them, and then rejoins the series. Third, it is sometimes useful to repeat an odd-span median smoother or the splitting operator until the smooth no longer changes. Following a digit or an S with an R species this type of repetition. Finally, the twice operator species that after smoothing, the smoother be reapplied to the resulting rough, and any recovered signal be added back to the original smooth. Letters may be specied in lowercase, if preferred. Examples of smoother , twice include 3RSSH 3rssh 3RSSH,twice 3rssh,twice 4253H 4253h 4253H,twice 4253h,twice 43RSR2H,twice 43rsr2h,twice
You must tsset your data before using tssmooth nl; see [TS] tsset. exp may contain time-series operators; see [U] 11.4.4 Time-series varlists.
Menu
Statistics
>
Time series
>
>
Nonlinear lter
497
498
tssmooth nl Nonlinear lter
Description
tssmooth nl uses nonlinear smoothers to identify the underlying trend in a series.
Options

Main
smoother(smoother , twice ) species the nonlinear smoother to be used. replace replaces newvar if it already exists.
Remarks
tssmooth nl works as a front end to smooth. See [R] smooth for details.
Saved results
tssmooth nl saves the following in r():
Scalars r(N) Macros r(method) r(smoother) r(timevar) r(panelvar) number of observations nl specied smoother time variable specied in tsset panel variable specied in tsset

tssmooth nl is implemented as an ado-le. The methods are documented in [R] smooth. A truncated description of the specied nonlinear lter labels the new variable. See [D] label for more information on labels. An untruncated description of the specied nonlinear lter is saved in the characteristic tssmooth for the new variable. See [P] char for more information on characteristics.
Also see
Title
tssmooth shwinters HoltWinters seasonal smoothing
Syntax
tssmooth shwinters options
Main
type
newvar = exp
if
in
, options
Description replace newvar if it already exists use # , # , and # as smoothing parameters use # observations to obtain initial values for recursion use #cons and #lt as initial values for recursion use # periods for the out-of-sample forecast use # for period of the seasonality use additive seasonal HoltWinters method use initial seasonal values in varname store estimated initial values for seasonal terms in newvar store nal years estimated seasonal terms in newvar normalize seasonal values use alternative method for computing the starting values control the maximization process; seldom used use # , # , and # as starting values for the parameters
replace parms(# # # ) samp0(#) s0(#cons #lt ) forecast(#) period(#) additive

Options
sn0 0(varname) sn0 v(newvar) snt v(newvar) normalize altstarts

Maximization
maximize options from(# # # )
You must tsset your data before using tssmooth shwinters; see [TS] tsset. exp may contain time-series operators; see [U] 11.4.4 Time-series varlists.
Menu
Statistics
>
Time series
>
>
Holt-Winters seasonal smoothing
Description
tssmooth shwinters performs the seasonal HoltWinters method on a user-specied expression, which is usually just a variable name, and generates a new variable containing the forecasted series.
Options

Main
replace replaces newvar if it already exists.

499
500
parms(# # # ), 0 # 1, 0 # 1, and 0 # 1, species the parameters. If parms() is not specied, the values are chosen by an iterative process to minimize the in-sample sum-of-squared prediction errors. If you experience difculty converging (many iterations and not concave messages), try using from() to provide better starting values. samp0(#) and s0(#cons #lt ) have to do with how the initial values #cons and #lt for the recursion are obtained. s0(#cons #lt ) species the initial values to be used. samp0(#) species that the initial values be obtained using the rst # observations of the sample. This calculation is described under Methods and formulas and depends on whether the altstart and additive options are also specied. If neither option is specied, the rst half of the sample is used to obtain initial values. forecast(#) species the number of periods for the out-of-sample prediction; 0 # 500. The default is forecast(0), which is equivalent to not performing an out-of-sample forecast. period(#) species the period of the seasonality. If period() is not specied, the seasonality is obtained from the tsset options daily, weekly, . . . , yearly; see [TS] tsset. If you did not specify one of those options when you tsset the data, you must specify the period() option. For instance, if your data are quarterly and you did not specify tssets quarterly option, you must now specify period(4). By default, seasonal values are calculated, but you may specify the initial seasonal values to be used via the sn0 0(varname) option. The rst period() observations of varname are to contain the initial seasonal values. additive uses the additive seasonal HoltWinters method instead of the default multiplicative seasonal HoltWinters method.
Options
sn0 0(varname) species the initial seasonal values to use. varname must contain a complete years worth of seasonal values, beginning with the rst observation in the estimation sample. For example, if you have monthly data, the rst 12 observations of varname must contain nonmissing data. sn0 0() cannot be used with sn0 v(). sn0 v(newvar) stores in newvar the initial seasonal values after they have been estimated. sn0 v() cannot be used with sn0 0(). snt v(newvar) stores in newvar the seasonal values for the nal years worth of data. normalize species that the seasonal values be normalized. In the multiplicative model, they are normalized to sum to one. In the additive model, the seasonal values are normalized to sum to zero. altstarts uses an alternative method to compute the starting values for the constant, the linear, and the seasonal terms. The default and the alternative methods are described in Methods and formulas. altstarts may not be specied with s0().
Maximization
maximize options controls the process for solving for the optimal , , and when the parms() option is not specied.
501
maximize options: nodifficult, technique(algorithm spec), iterate(#), no log, trace, gradient, showstep, hessian, showtolerance, tolerance(#), ltolerance(#), nrtolerance(#), and nonrtolerance; see [R] maximize. These options are seldom used. from(# # # ), 0 < # < 1, 0 < # < 1, and 0 < # < 1, species starting values from which the optimal values of , , and will be obtained. If from() is not specied, from(.5 .5 .5) is used.
Remarks
Introduction HoltWinters seasonal multiplicative method HoltWinters seasonal additive method
Introduction
The seasonal HoltWinters methods forecast univariate series that have a seasonal component. If the amplitude of the seasonal component grows with the series, the HoltWinters multiplicative method should be used. If the amplitude of the seasonal component is not growing with the series, the HoltWinters additive method should be used. Abraham and Ledolter (1983), Bowerman, OConnell, and Koehler (2005), and Montgomery, Johnson, and Gardiner (1990) provide good introductions to the HoltWinters methods in recursive univariate forecasting methods. Chateld (2001, 2004) provides introductions in the broader context of modern time-series analysis. Like the other recursive methods in tssmooth, tssmooth shwinters uses the information stored by tsset to detect panel data. When applied to panel data, each series is smoothed separately, and the starting values are computed separately for each panel. If the smoothing parameters are chosen to minimize the in-sample sum-of-squared forecast errors, the optimization is performed separately on each panel. When there are missing values at the beginning of the series, the sample begins with the rst nonmissing observation. Missing values after the rst nonmissing observation are lled in with forecasted values.
HoltWinters seasonal multiplicative method

This method forecasts seasonal time series in which the amplitude of the seasonal component grows with the series. Chateld (2001) notes that there are some nonlinear state-space models whose optimal prediction equations correspond to the multiplicative HoltWinters method. This procedure is best applied to data that could be described by
xt+j = (t + j)St+j +
t+j
where xt is the series, t is the time-varying mean at time t, is a parameter, St is the seasonal component at time t, and t is an idiosyncratic error. See Methods and formulas for the updating equations.
502
Example 1
We have quarterly data on turkey sales by a new producer in the 1990s. The data have a strong seasonal component and an upward trend. We use the multiplicative HoltWinters method to forecast sales for the year 2000. Because we have already tsset our data to the quarterly format, we do not need to specify the period() option.
. use http://www.stata-press.com/data/r12/turksales . tssmooth shwinters shw1 = sales, forecast(4) computing optimal weights Iteration 0: penalized RSS = -189.34609 (not concave) Iteration 1: penalized RSS = -108.68038 (not concave) Iteration 2: penalized RSS = -106.24548 Iteration 3: penalized RSS = -106.141 Iteration 4: penalized RSS = -106.14093 Iteration 5: penalized RSS = -106.14093 Optimal weights: alpha = 0.1310 beta = 0.1428 gamma = 0.2999 penalized sum-of-squared residuals = 106.1409 sum-of-squared residuals = 106.1409 root mean squared error = 1.628964
The graph below describes the t and the forecast that was obtained.
. line sales shw1 t, title("Multiplicative Holt-Winters forecast") > xtitle(Time) ytitle(Sales)
Multiplicative HoltWinters forecast

115 95 1990q1 100 Sales 105 110
1992q1 sales
1994q1
1996q1 Time
1998q1
2000q1
shw parms(0.131 0.143 0.300) = sales
HoltWinters seasonal additive method

This method is similar to the previous one, but the seasonal effect is assumed to be additive rather than multiplicative. This method forecasts series that can be described by the equation
xt+j = (t + j) + St+j +
See Methods and formulas for the updating equations.
t+j
503
Example 2
In this example, we t the data from the previous example to the additive model to forecast sales in the coming year. We use the snt v() option to save the last years seasonal terms in the new variable seas.
. tssmooth shwinters shwa = sales, forecast(4) snt_v(seas) normalize additive computing optimal weights Iteration 0: penalized Iteration 1: penalized Iteration 2: penalized Iteration 3: penalized Iteration 4: penalized Iteration 5: penalized Iteration 6: penalized Optimal weights: RSS RSS RSS RSS RSS RSS RSS = = = = = = = -190.90242 -108.8357 -108.25359 -107.68187 -107.66444 -107.66442 -107.66442 = = = = = = (not concave)
alpha beta gamma penalized sum-of-squared residuals sum-of-squared residuals root mean squared error
0.1219 0.1580 0.3340 107.6644 107.6644 1.640613
The output reveals that the multiplicative model has a better in-sample t, and the graph below shows that the forecast from the multiplicative model is higher than that of the additive model.
. line shw1 shwa t if t>=tq(2000q1), title("Multiplicative and additive" > "Holt-Winters forecasts") xtitle("Time") ytitle("Sales") legend(cols(1))
Multiplicative and additive HoltWinters forecasts

113 108 2000q1 109 Sales 110 111 112
2000q2
2000q3 Time
2000q4
2001q1
shw parms(0.131 0.143 0.300) = sales shwadd parms(0.122 0.158 0.334) = sales
To check whether the estimated seasonal components are intuitively sound, we list the last years seasonal components.
504

. list t seas if seas < . t 37. 38. 39. 40. 1999q1 1999q2 1999q3 1999q4 seas -2.7533393 -.91752567 1.8082417 1.8626233
The output indicates that the signs of the estimated seasonal components agree with our intuition.
Saved results
tssmooth shwinters saves the following in r():
Scalars r(N) r(alpha) r(beta) r(gamma) r(prss) r(rss) r(rmse) Macros r(method) number of observations
smoothing parameter smoothing parameter smoothing parameter
r(N pre) r(s2 0) r(s1 0) r(linear) r(constant) r(period) r(exp) r(timevar) r(panelvar)
penalized sum-of-squared errors sum-of-squared errors root mean squared error
number of seasons used in calculating starting values initial value for linear term initial value for constant term nal value of linear term nal value of constant term period, if lter is seasonal expression specied time variable specied in tsset panel variable specied in tsset
shwinters, additive or shwinters, multiplicative r(normalize) normalize, if specied

tssmooth shwinters is implemented as an ado-le. A truncated description of the specied seasonal HoltWinters lter labels the new variable. See [D] label for more information on labels. An untruncated description of the specied seasonal HoltWinters lter is saved in the characteristic named tssmooth for the new variable. See [P] char for more information on characteristics. When the parms() option is not specied, the smoothing parameters are chosen to minimize the in-sample sum of penalized squared-forecast errors. Sometimes, one or more of the three optimal parameters lies on the boundary [ 0, 1 ]. To keep the estimates inside [ 0, 1 ], tssmooth shwinters parameterizes the objective function in terms of their inverse logits, that is, exp()/{1 + exp()}, exp()/{1 + exp()}, and exp()/{1 + exp()}. When one of these parameters is actually on the boundary, this can complicate the optimization. For this reason, tssmooth shwinters optimizes a penalized sum-of-squared forecast errors. Let xt (, , ) be the forecast for the series xt given the choices of , , and . Then the in-sample penalized sum-of-squared prediction errors is
T
P =
t=1
{xt xt (, , )}2 + I|f ()|>12) (|f ()| 12)2 + I|f ()|>12) (|f ()| 12)2 +I|f ()|>12) (|f ()| 12)2
505
x where f (x) = ln 1x . The penalty term is zero unless one of the parameters is close to the boundary. When one of the parameters is close to the boundary, the penalty term will help to obtain convergence.
HoltWinters seasonal multiplicative procedure As with the other recursive methods in tssmooth, there are three aspects to implementing the HoltWinters seasonal multiplicative procedure: the forecasting equation, the initial values, and the updating equations. Unlike in the other methods, the data are now assumed to be seasonal with period L. Given the estimates a(t), b(t), and s(t + L), a step-ahead point forecast of xt , denoted by yt+ , is yt+ = {a(t) + b(t) } s(t + L) Given the smoothing parameters , , and , the updating equations are
a(t) =
xt + (1 ) {a(t 1) + b(t 1)} s(t L)
b(t) = {a(t) a(t 1)} + (1 ) b(t 1)

and
s(t) =
xt a(t)
+ (1 )s(t L)
To restrict the seasonal terms to sum to 1 over each year, specify the normalize option. The updating equations require the L + 2 initial values a(0), b(0), s(1 L), s(2 L), . . . , s(0). Two methods calculate the initial values with the rst m years, each of which contains L seasons. By default, m is set to the number of seasons in half the sample. The initial value of the trend component, b(0), can be estimated by
b(0) =
xm x1 (m 1)L
where xm is the average level of xt in year m and x1 is the average level of xt in the rst year. The initial value for the linear term, a(0), is then calculated as
a(0) = x1
L b(0) 2
To calculate the initial values for the seasons 1, 2, . . . , L, we rst calculate the deviation-adjusted values, xt S(t) = (L+1) xi j b(0) 2 where i is the year that corresponds to time t, j is the season that corresponds to time t, and xi is the average level of xt in year i.
506
Next, for each season l = 1, 2, . . . , L, we dene sl as the average St over the years. That is,
sl =
1 m
m1
Sl+kL
k=0
for l = 1, 2, . . . , L
Then the initial seasonal estimates are
s0l = sl
L
L l=1 sl
for l = 1, 2, . . . , L
and these values are used to ll in s(1 L), . . . , s(0). If the altstarts option is specied, the starting values are computed based on a regression with seasonal indicator variables. Specically, the series xt is regressed on a time variable normalized to equal one in the rst period in the sample and on a constant. Then b(0) is set to the estimated coefcient on the time variable, and a(0) is set to the estimated constant term. To calculate the seasonal starting values, xt is regressed on a set of L seasonal dummy variables. The lth seasonal 1 starting value is set to ( )l , where is the mean of xt and l is the estimated coefcient on the lth seasonal dummy variable. The sample used in both regressions and the mean computation is restricted to include the rst samp0() years. By default, samp0() includes half the data.
Technical note
If there are missing values in the rst few years, a small value of m can cause the starting value methods for seasonal term to fail. Here you should either specify a larger value of m by using samp0() or directly specify the seasonal starting values by using the snt0 0() option.
HoltWinters seasonal additive procedure This procedure is similar to the previous one, except that the data are assumed to be described by
xt = (0 + 1 t) + st +
As in the multiplicative case, there are three smoothing parameters, , , and , which can either be set or chosen to minimize the in-sample sum-of-squared forecast errors. The updating equations are
a(t) = {xt s(t L)} + (1 ) {a(t 1) + b(t 1)}
b(t) = {a(t) a(t 1)} + (1 )b(t 1)

and
s(t) = {xt a(t)} + (1 )s(t L)

To restrict the seasonal terms to sum to 0 over each year, specify the normalize option. A -step-ahead forecast, denoted by yt+ , is given by
xt+ = a(t) + b(t) + s(t + L)
507
As in the multiplicative case, there are two methods for setting the initial values. The default method is to obtain the initial values for a(0), b(0), s(1 L), . . . , s(0) from the regression xt = a(0) + b(0)t + s,1L D1 + s,2L D2 + + s,0 DL + et where the D1 , . . . , DL are dummy variables with
Di =
1 0
if t corresponds to season i otherwise
When altstarts is specied, an alternative method is used that regresses the xt series on a time variable that has been normalized to equal one in the rst period in the sample and on a constant term. b(0) is set to the estimated coefcient on the time variable, and a(0) is set to the estimated constant term. Then the demeaned series xt = xt is created, where is the mean of the xt . The xt are regressed on L seasonal dummy variables. The lth seasonal starting value is then set to l , where l is the estimated coefcient on the lth seasonal dummy variable. The sample in both the regression and the mean calculation is restricted to include the rst samp0 years, where, by default, samp0() includes half the data.
Acknowledgment
We thank Nicholas J. Cox of Durham University for his helpful comments.
References
Also see
Title
ucm Unobserved-components model
Syntax
ucm depvar options
Model
indepvars
if
in
, options
Description
model(model) specify trend and idiosyncratic components seasonal(#) include a seasonal component with a period of # time units cycle(# , frequency(#f ) ) include a cycle of order # and optionally set initial frequency to #f , 0 < #f < ; cycle() may be specied up to three times constraints(constraints) apply specied linear constraints collinear keep collinear variables
SE/Robust
vce(vcetype)
Reporting
vcetype may be oim or robust set condence level; default is level(95) do not display constraints control column formats, row spacing, and display of omitted variables and base and empty cells control the maximization process display legend instead of statistics

Maximization
model rwalk none ntrend dconstant llevel dtrend lldtrend rwdrift lltrend strend rtrend
Description random-walk model; the default no trend or idiosyncratic component no trend component but include idiosyncratic component deterministic constant with idiosyncratic component local-level model deterministic-trend model with idiosyncratic component local-level model with deterministic trend random-walk-with-drift model local-linear-trend model smooth-trend model random-trend model
508

You must tsset your data before using ucm; see [TS] tsset. indepvars may contain factor variables; see [U] 11.4.3 Factor variables. indepvars and depvar may contain time-series operators; see [U] 11.4.4 Time-series varlists. by, statsby, and rolling are allowed; see [U] 11.1.10 Prex commands. coeflegend does not appear in the dialog box. See [U] 20 Estimation and postestimation commands for more capabilities of estimation commands.
509
Menu
Statistics
>
Time series
>
Unobserved-components model
Description
Unobserved-components models (UCMs) decompose a time series into trend, seasonal, cyclical, and idiosyncratic components and allow for exogenous variables. ucm estimates the parameters of UCMs by maximum likelihood. All the components are optional. The trend component may be rst-order deterministic or it may be rst-order or second-order stochastic. The seasonal component is stochastic; the seasonal effects at each time period sum to a zero-mean nite-variance random variable. The cyclical component is modeled by the stochastic-cycle model derived by Harvey (1989).
Options

Model
model(model) species the trend and idiosyncratic components. The default is model(rwalk). The available models are listed in Syntax and discussed in detail in Models for the trend and idiosyncratic components under Remarks below. seasonal(#) adds a stochastic-seasonal component to the model. # is the period of the season, that is, the number of time-series observations required for the period to complete. cycle(#) adds a stochastic-cycle component of order # to the model. The order # must be 1, 2, or 3. Multiple cycles are added by repeating the cycle(#) option with up to three cycles allowed. cycle(#, frequency(#f )) species #f as the initial value for the central-frequency parameter in the stochastic-cycle component of order #. #f must be in the interval (0, ). constraints(constraints), collinear; see [R] estimation options.
SE/Robust
vce(vcetype) species the estimator for the variancecovariance matrix of the estimator. vce(oim), the default, causes ucm to use the observed information matrix estimator. vce(robust) causes ucm to use the Huber/White/sandwich estimator.
Reporting
level(#), nocnsreport; see [R] estimation options. display options: noomitted, vsquish, noemptycells, baselevels, allbaselevels, cformat(% fmt), pformat(% fmt), and sformat(% fmt); see [R] estimation options.
510
Maximization
maximize options: difficult, technique(algorithm spec), iterate(#), no log, trace, gradient, showstep, hessian, showtolerance, tolerance(#), ltolerance(#), nrtolerance(#), and from(matname); see [R] maximize for all options except from(), and see below for information on from(). from(matname) species initial values for the maximization process. from(b0) causes ucm to begin the maximization algorithm with the values in b0. b0 must be a row vector; the number of columns must equal the number of parameters in the model; and the values in b0 must be in the same order as the parameters in e(b). If you model fails to converge, try using the difficult option. Also see the technical note below example 5. The following option is available with ucm but is not shown in the dialog box: coeflegend; see [R] estimation options.
Remarks
An introduction to UCMs A random-walk-model example Frequency-domain concepts used in the stochastic-cycle model Another random-walk-model example Comparing UCM and ARIMA A local-level-model example Comparing UCM and ARIMA, revisited Models for the trend and idiosyncratic components Seasonal component
An introduction to UCMs
UCMs decompose a time series into trend, seasonal, cyclical, and idiosyncratic components and allow for exogenous variables. Formally, UCMs can be written as
yt = t + t + t + xt +
(1)
where yt is the dependent variable, t is the trend component, t is the seasonal component, t is the cyclical component, is a vector of xed parameters, xt is a vector of exogenous variables, and t is the idiosyncratic component. By placing restrictions on t and t , Harvey (1989) derived a series of models for the trend and the idiosyncratic components. These models are briey described in Syntax and are further discussed in Models for the trend and idiosyncratic components. To these models, Harvey (1989) added models for the seasonal and cyclical components, and he also allowed for the presence of exogenous variables. It is rare that a UCM contains all the allowed components. For instance, the seasonal component is rarely needed when modeling deseasonalized data. Harvey (1989) and Durbin and Koopman (2001) show that UCMs can be written as state-space models that allow the parameters of a UCM to be estimated by maximum likelihood. In fact, ucm uses sspace (see [TS] sspace) to perform the estimation calculations; see Methods and formulas for details.
511
After estimating the parameters, predict can produce in-sample predictions or out-of-sample forecasts; see [TS] ucm postestimation. After estimating the parameters of a UCM that contains a cyclical component, estat period converts the estimated central frequency to an estimated central period and psdensity estimates the spectral density implied by the model; see [TS] ucm postestimation and the examples below. We illustrate the basic approach of analyzing data with UCMs, and then we discuss the details of the different trend models in Models for the trend and idiosyncratic components. Although the methods implemented in ucm have been widely applied by economists, they are general time-series techniques and may be of interest to researchers from other disciplines. In example 8, we analyze monthly data on the reported cases of mumps in New York City.
A random-walk-model example Example 1

We begin by plotting monthly data on the U.S. civilian unemployment rate.
. use http://www.stata-press.com/data/r12/unrate . tsline unrate, name(unrate)
2 1950m1
Civilian Unemployment Rate 4 6 8
10
1960m1
1970m1
1980m1 Month
1990m1
2000m1
2010m1
This series looks like it might be well approximated by a random-walk model. Formally, a random-walk model is given by
yt = t t = t1 + t
The random-walk is so frequently applied, at least as a starting model, that it is the default model for ucm. In the output below, we t the random-walk model to the unemployment data.
512

. ucm unrate searching for initial values .......... (setting technique to bhhh) Iteration 0: log likelihood = 84.272992 Iteration 1: log likelihood = 84.394942 Iteration 2: log likelihood = 84.400923 Iteration 3: log likelihood = 84.401282 Iteration 4: log likelihood = 84.401305 (switching technique to nr) Iteration 5: log likelihood = 84.401306 Refining estimates: Iteration 0: log likelihood = 84.401306 Iteration 1: log likelihood = 84.401307 Unobserved-components model Components: random walk Sample: 1948m1 - 2011m1 Log likelihood = 84.401307 OIM Std. Err.
Number of obs
757
unrate Variance level
Coef.
P>|z|
.0467196
.002403
19.44
0.000
.0420098
.0514294
The output indicates that the model is nonstationary, as all random-walk models are. We consider a richer model in the next example.
Example 2
We suspect that there should be a stationary cyclical component that produces serially correlated shocks around the random-walk trend. Harvey (1989) derived a stochastic-cycle model for these stationary cyclical components. The stochastic-cycle model has three parameters: the frequency at which the random components are centered, a damping factor that parameterizes the dispersion of the random components around the central frequency, and the variance of the stochastic-cycle process that acts as a scale factor.
513
Fitting this model to unemployment data yields

. ucm unrate, cycle(1) searching for initial values .................... (setting technique to bhhh) Iteration 0: log likelihood = 84.273579 Iteration 1: log likelihood = 87.852115 Iteration 2: log likelihood = 88.253422 Iteration 3: log likelihood = 89.191311 Iteration 4: log likelihood = 94.675898 (switching technique to nr) Iteration 5: log likelihood = 98.394691 (not concave) Iteration 6: log likelihood = 98.983092 Iteration 7: log likelihood = 99.983642 Iteration 8: log likelihood = 104.8308 Iteration 9: log likelihood = 114.27184 Iteration 10: log likelihood = 116.47396 Iteration 11: log likelihood = 118.45803 Iteration 12: log likelihood = 118.88055 Iteration 13: log likelihood = 118.88421 Iteration 14: log likelihood = 118.88421 Refining estimates: Iteration 0: log likelihood = 118.88421 Iteration 1: log likelihood = 118.88421 Unobserved-components model Components: random walk, order 1 cycle Sample: 1948m1 - 2011m1 Log likelihood = 118.88421 OIM Std. Err. .0103609 .0061121 Number of obs Wald chi2(2) Prob > chi2 = = = 757 26650.81 0.0000
unrate frequency damping Variance level cycle1
Coef. .0933466 .9820003
z 9.01 160.66
P>|z| 0.000 0.000
[95% Conf. Interval] .0730397 .9700207 .1136535 .9939798
.0143786 .0270339
.0051392 .0054343
2.80 4.97
0.003 0.000
.004306 .0163829
.0244511 .0376848
The estimated central frequency for the cyclical component is small, implying that the cyclical component is centered around low-frequency components. The high-damping factor indicates that all the components from this cyclical component are close to the estimated central frequency. The estimated variance of the stochastic-cycle process is small but signicant. We use estat period to convert the estimate of the central frequency to an estimated central period.
. estat period cycle1 period frequency damping Coef. 67.31029 .0933466 .9820003 Std. Err. 7.471004 .0103609 .0061121 [95% Conf. Interval] 52.66739 .0730397 .9700207 81.95319 .1136535 .9939798
Note: Cycle time unit is monthly.
514
Because we have monthly data, the estimated central period of 67.31 implies that the cyclical component is composed of random components that occur around a central periodicity of around 5.61 years. This estimate falls within the conventional Burns and Mitchell (1946) denition of business-cycle shocks occurring between 1.5 and 8 years. We can convert the estimated parameters of the cyclical component to an estimated spectral density of the cyclical component, as described by Harvey (1989). The spectral density of the cyclical component describes the relative importance of the random components at different frequencies; see Frequency-domain concepts used in the stochastic-cycle model for details. We use psdensity (see [TS] psdensity) to obtain the spectral density of the cyclical component implied by the estimated parameters, and we use twoway line (see [G-2] graph twoway line) to plot the estimated spectral density.
. psdensity sdensity omega . line sdensity omega
0 0
UCM cycle 1 spectral density 2 4 6
1 Frequency
The estimated spectral density shows that the cyclical component is composed of random components that are tightly distributed around the low-frequency peak.
Frequency-domain concepts used in the stochastic-cycle model

The parameters of the stochastic-cycle model are easiest to interpret in the frequency domain. We now provide a review of the useful concepts from the frequency domain. Crucial to understanding the stochastic-cycle model is the frequency-domain concept that a stationary process can be decomposed into random components that occur at the frequencies in the interval [0, ]. We need some concepts from the frequency-domain approach to interpret the parameters in the stochastic-cycle model of the cyclical component. Here we provide a simple, intuitive explanation. More technical presentations can be found in Priestley (1981), Harvey (1989, 1993), Hamilton (1994), Fuller (1996), and Wei (2006). As with much time-series analysis, the basic results are for covariance-stationary processes with additional results handling some nonstationary cases. We present some useful results for covariancestationary processes. These results provide what we need to interpret the stochastic-cycle model for the stationary cyclical component.
515
The autocovariances j , j {0, 1, . . . , }, of a covariance-stationary process yt specify its variance and dependence structure. In the frequency-domain approach to time-series analysis, the spectral density describes the importance of the random components that occur at frequency relative to the components that occur at other frequencies. The frequency-domain approach focuses on the relative contributions of random components that occur at the frequencies [0, ]. The spectral density can be written as a weighted average of the autocorrelations of yt . Like autocorrelations, the spectral density is normalized by 0 , the variance of yt . Multiplying the spectral density by 0 yields the power-spectrum of yt . In an independent and identically distributed (i.i.d.) process, the components at all frequencies are equally important, so the spectral density is a at line. In common parlance, we speak of high-frequency noise making a series look more jagged and of low-frequency components causing smoother plots. More formally, we say that a process composed primarily of high-frequency components will have fewer runs above or below the mean than an i.i.d. process and that a process composed primarily of low-frequency components will have more runs above or below the mean than an i.i.d. process. To further formalize these ideas, consider the rst-order autoregressive (AR(1)) process given by
yt = yt1 +
where t is a zero-mean, covariance-stationary process with nite variance 2 , and || < 1 so that yt is covariance stationary. The rst-order autocorrelation of this AR(1) process is . Below are plots of simulated data when is set to 0, 0.8, and 0.8. When = 0, the data are i.i.d. When = 0.8, the value today is strongly negatively correlated with the value yesterday, so this case should be a prototypical high-frequency noise example. When = 0.8, the value today is strongly positively correlated with the value yesterday, so this case should be a prototypical low-frequency shock example.
=0.8 321 0 1 2
=0.8 21 0 1 2 3
=0 2 1 0 1 2
Time
The plots above conrm our conjectures. The plot when = 0.8 contains fewer runs above or below the mean, and it is more jagged than the i.i.d. plot. The plot when = 0.8 contains more runs above or below the mean, and it is smoother than the i.i.d. plot.
516
Below we plot the spectral densities for the AR(1) model with = 0, = 0.8, and = 0.8.
25 0 0 5 Spectral density 10 15 20
1 Frequency =0 =0.8
2 =0.8
The high-frequency components are much more important to the AR(1) process with = 0.8 than to the i.i.d. process with = 0. The low-frequency components are much more important to the AR(1) process with = 0.8 than to the i.i.d. process.
Technical note
Autoregressive moving-average (ARMA) models parameterize the autocorrelation in a time series by allowing todays value to be a weighted average of past values and a weighted average of past i.i.d. shocks; see Hamilton (1994), Wei (2006), and [TS] arima for introductions and a Stata implementation. The intuitive ARMA parameterization has many nice features, including that one can easily rewrite the ARMA model as a weighted average of past i.i.d. shocks to trace how a shock feeds through the system. Although it is easy to obtain the spectral density of an ARMA process, the parameters themselves provide limited information about the underlying spectral density. In contrast, the parameters of the stochastic-cycle parameterization of autocorrelation in a time series directly provide information about the underlying spectral density. The parameter 0 is the central frequency around which the random components are clustered. If 0 is small, then the model is centered around low-frequency components. If 0 is close to , then the model is centered around high-frequency components. The parameter is the damping factor that indicates how tightly clustered the random components are around the central frequency 0 . If is close to 0, there is no clustering of the random components. If is close to 1, the random components are tightly distributed around the central frequency 0 . In the graph below, we draw the spectral densities implied by stochastic-cycle models with four sets of parameters: 0 = /4, = 0.8; 0 = /4, = 0.9; 0 = 4/5, = 0.8; and 0 = 4/5, = 0.9. The graph below illustrates that 0 is the central frequency around which the other important random components are distributed. It also illustrates that the damping parameter controls the dispersion of the important components around the central frequency.
517
50
0 = 4 40 = 0.9
0 = 5 = 0.9
30
20
0 = 4 = 0.8
0 = 5 = 0.8
10
/4
/2
3/4
Another random-walk-model example Example 3

Now lets reconsider example 2. Although we might be happy with how our model has identied a stationary cyclical component that we could interpret in business-cycle terms, we suspect that there should also be a high-frequency cyclical component. It is difcult to estimate the parameters of a UCM with two or more stochastic-cycle models. Providing starting values for the central frequencies can be a crucial help to the optimization procedure. Below we estimate a UCM with two cyclical components. We use the frequency() suboption to provide starting values for the central frequencies; we specied the values below because we suspect one model will pick up the low-frequency components and the other will pick up the high-frequency components. We specied the low-frequency model to be order 2 to make it less peaked for any given damping factor. (Trimbur [2006] provides a nice introduction and some formal results for higher-order stochastic-cycle models.)
. ucm unrate, cycle(1, frequency(2.9)) cycle(2, frequency(.09)) searching for initial values .................... (setting technique to bhhh) Iteration 0: log likelihood = 115.98563 Iteration 1: log likelihood = 125.04043 Iteration 2: log likelihood = 127.69387 Iteration 3: log likelihood = 134.50864 Iteration 4: log likelihood = 136.91353 (switching technique to nr) Iteration 5: log likelihood = 138.5091 Iteration 6: log likelihood = 146.09273 Iteration 7: log likelihood = 146.28132 Iteration 8: log likelihood = 146.28326 Iteration 9: log likelihood = 146.28326 Refining estimates: Iteration 0: log likelihood = 146.28326 Iteration 1: log likelihood = 146.28326
518

Unobserved-components model Components: random walk, 2 cycles of order 1 2 Sample: 1948m1 - 2011m1 Log likelihood = 146.28326 OIM Std. Err. Number of obs Wald chi2(4) Prob > chi2 = = = 757 7681.33 0.0000
unrate cycle1 frequency damping cycle2 frequency damping Variance level cycle1 cycle2
Coef.
P>|z|
2.882382 .7004295
.0668017 .1251571
43.15 5.60
0.000 0.000
2.751453 .4551262
3.013311 .9457328
.0667929 .9074708
.0206849 .0142273
3.23 63.78
0.001 0.000
.0262513 .8795858
.1073345 .9353559
.0207704 .0027886 .002714
.0039669 .0014363 .001028
5.24 1.94 2.64
0.000 0.026 0.004
.0129953 0 .0006991
.0285454 .0056037 .0047289
The output provides some support for the existence of a second, high-frequency cycle. The high-frequency components are centered around 2.88, whereas the low-frequency components are centered around 0.067. That the estimated damping factor is 0.70 for the high-frequency cycle whereas the estimated damping factor for the low-frequency cycle is 0.91 indicates that the high-frequency components are more diffusely distributed around 2.88 than the low-frequency components are around 0.067. We obtain and plot the estimated spectral densities to get another look at these results.
. psdensity sdensity2a omega2a . psdensity sdensity2b omega2b, cycle(2) . line sdensity2a sdensity2b omega2a, legend(col(1))
0 0
1 Frequency
2 UCM cycle 1 spectral density UCM cycle 2 spectral density
The estimated spectral densities indicate that we have found two distinct cyclical components.
519
It does not matter whether we specify omega2a or omega2b to be the x-axis variable, because they are equal to each other.
Technical note
That the estimated spectral densities in the previous example do not overlap is important for parameter identication. Although the parameters are identied in large-sample theory, we have found it difcult to estimate the parameters of two cyclical components when the spectral densities overlap. When the spectral densities of two cyclical components overlap, the parameters may not be well identied and the optimization procedure may not converge.
Comparing UCM and ARIMA Example 4

This example provides some insight for readers familiar with autoregressive integrated movingaverage (ARIMA) models but not with UCMs. If you are not familiar with ARIMA models, you may wish to skip this example. See [TS] arima for an introduction to ARIMA models in Stata.
UCMs provide an alternative to ARIMA models implemented in [TS] arima. Neither set of models is nested within the other, but there are some cases in which instructive comparisons can be made.
The random-walk model corresponds to an ARIMA model that is rst-order integrated and has an i.i.d. error term. In other words, the random-walk UCM and the ARIMA(0,1,0) are asymptotically equivalent. Thus
ucm unrate
and
arima unrate, arima(0,1,0) noconstant
produce asymptotically equivalent results. The stochastic-cycle model for the stationary cyclical component is an alternative functional form for stationary processes to stationary autoregressive moving-average (ARMA) models. Which model is preferred depends on the application and which parameters a researchers wants to interpret. Both the functional forms and the parameter interpretations differ between the stochastic-cycle model and the ARMA model. See Trimbur (2006, eq. 25) for some formal comparisons of the two models. That both models can be used to estimate the stationary cyclical components for the random-walk model implies that we can compare the results in this case by comparing their estimated spectral densities. Below we estimate the parameters of an ARIMA(2,1,1) model and plot the estimated spectral density of the stationary component.
520

. arima unrate, noconstant arima(2,1,1) (setting optimization to BHHH) Iteration 0: log likelihood = Iteration 1: log likelihood = Iteration 2: log likelihood = Iteration 3: log likelihood = Iteration 4: log likelihood = (switching optimization to BFGS) Iteration 5: log likelihood = Iteration 6: log likelihood = Iteration 7: log likelihood = Iteration 8: log likelihood = Iteration 9: log likelihood = Iteration 10: log likelihood = ARIMA regression Sample: 1948m2 - 2011m1 138.8364 OPG Std. Err. Number of obs Wald chi2(3) Prob > chi2 = = = 756 683.34 0.0000 129.8801 134.61953 137.04909 137.71386 138.25255 138.51924 138.81638 138.83615 138.8364 138.83642 138.83642
Log likelihood =
D.unrate ARMA ar L1. L2. ma L1. /sigma
Coef.
P>|z|
.5398016 .2468148
.0586304 .0359396
9.21 6.87
0.000 0.000
.4248882 .1763744
.6547151 .3172551
-.5146506 .2013332
.0632838 .0032644
-8.13 61.68
0.000 0.000
-.6386845 .1949351
-.3906167 .2077313
Note: The test of the variance against zero is one sided, and the two-sided confidence interval is truncated at zero. . psdensity sdensity_arma omega_arma . line sdensity_arma omega_arma
0 0
.2
ARMA spectral density .4 .6
.8
1 Frequency
The estimated spectral density from the ARIMA(2,1,1) has a similar shape to the plot obtained by combining the two spectral densities estimated from the stochastic-cycle model in example 3. For
521
this particular application, the estimated central frequencies of the two cyclical components from the stochastic-cycle model provide information about the business-cycle component and the high-frequency component that is not easily obtained from the ARIMA(2,1,1) model. On the other hand, it is easier to work out the impulseresponse function for the ARMA model than for the stochastic-cycle model, implying that the ARMA model is easier to use when tracing the effect of a shock feeding through the system.
A local-level-model example
We now consider the weekly series of initial claims for unemployment insurance in the United States, which is plotted below.
Example 5
. use http://www.stata-press.com/data/r12/icsa1, clear . tsline icsa
700 200 01jan1970 Change in initial claims 300 400 500 600
01jan1980
01jan1990 Date
01jan2000
01jan2010
This series looks like it was generated by a random walk with extra noise, so we want to use a random-walk model that includes an additional random term. This structure causes the model to be occasionally known as the random-walk-plus-noise model, but it is more commonly known as the local-level model in the UCM literature. The local-level model models the trend as a random walk and models the idiosyncratic components as independent and identically distributed components. Formally, the local-level model species the observed time-series yt , for t = 1, . . . , T , as
yt = t +
t = t1 + t
where
t 2 i.i.d. N (0, 2 ) and t i.i.d. N (0, ) and are mutually independent.
522
We t the local-level model in the output below:

. ucm icsa, model(llevel) searching for initial values .......... (setting technique to bhhh) Iteration 0: log likelihood = -9982.7798 Iteration 1: log likelihood = -9913.2745 Iteration 2: log likelihood = -9894.9925 Iteration 3: log likelihood = -9893.7191 Iteration 4: log likelihood = -9893.2876 (switching technique to nr) Iteration 5: log likelihood = -9893.2614 Iteration 6: log likelihood = -9893.2469 Iteration 7: log likelihood = -9893.2469 Refining estimates: Iteration 0: log likelihood = -9893.2469 Iteration 1: log likelihood = -9893.2469 Unobserved-components model Components: local level Sample: 07jan1967 - 19feb2011 Log likelihood = -9893.2469 OIM Std. Err. Number of obs = 2303
icsa Variance level icsa
Coef.
P>|z|
116.558 124.2715
8.806587 7.615506
13.24 16.32
0.000 0.000
99.29745 109.3454
133.8186 139.1976
Note: Model is not stationary. Note: Tests of variances against zero are one sided, and the two-sided confidence intervals are truncated at zero. Note: Time units are in 7 days.
The output indicates that both components are statistically signicant.
Technical note
The estimation procedure will not always converge when estimating the parameters of the local-level model. If the series does not vary enough in the random level, modeled by the random walk, and in the stationary shocks around the random level, the estimation procedure will not converge because it will be unable to set the variance of one of the two components to 0. Take another look at the graphs of unrate and icsa. The extra noise around the random level that can be seen in the graph of icsa allows us to estimate both variances. A closely related point is that it is difcult to estimate the parameters of a local-level model with a stochastic-cycle component because the series must have enough variation to identify the variance of the random-walk component, the variance of the idiosyncratic term, and the parameters of the stochastic-cycle component. In some cases, series that look like candidates for the local-level model are best modeled as random-walk models with stochastic-cycle components. In fact, convergence can be a problem for most of the models in ucm. Convergence problems occur most often when there is insufcient variation to estimate the variances of the components in the model. When there is insufcient variation to estimate the variances of the components in the model, the optimization routine will fail to converge as it attempts to set the variance equal to 0. This usually shows up in the iteration log when the log likelihood gets stuck at a particular value
523
and the message (not concave) or (backed up) is displayed repeatedly. When this happens, use the iteration() option to limit the number of iterations, look to see which of the variances is being driven to 0, and drop that component from the model. (This technique is a method to obtain convergence to interpretable estimates, not a model-selection method.)
Example 6
We might suspect that there is some serial correlation in the idiosyncratic shock. Alternatively, we could include a cyclical component to model the stationary time-dependence in the series. In the example below, we add a stochastic-cycle model for the stationary cyclical process, but we drop the idiosyncratic term and use a random-walk model instead of the local-level model. We change the model because it is difcult to estimate the variance of the idiosyncratic term along with the parameters of a stationary cyclical component.
. ucm icsa, model(rwalk) cycle(1) searching for initial values .................... (setting technique to bhhh) Iteration 0: log likelihood = -10008.167 Iteration 1: log likelihood = -10007.272 Iteration 2: log likelihood = -10007.206 (backed up) Iteration 3: log likelihood = -10007.17 (backed up) Iteration 4: log likelihood = -10007.148 (backed up) (switching technique to nr) Iteration 5: log likelihood = -10007.137 (not concave) Iteration 6: log likelihood = -9885.1932 (not concave) Iteration 7: log likelihood = -9884.1636 Iteration 8: log likelihood = -9881.6478 Iteration 9: log likelihood = -9881.4496 Iteration 10: log likelihood = -9881.4441 Iteration 11: log likelihood = -9881.4441 Refining estimates: Iteration 0: log likelihood = -9881.4441 Iteration 1: log likelihood = -9881.4441 Unobserved-components model Components: random walk, order 1 cycle Sample: 07jan1967 - 19feb2011 Number of obs Wald chi2(2) Log likelihood = -9881.4441 Prob > chi2 OIM Std. Err. .3855657 .0349537
= = =
2303 23.04 0.0000
icsa frequency damping Variance level cycle1
Coef. 1.469633 .1644576
z 3.81 4.71
P>|z| 0.000 0.000
[95% Conf. Interval] .7139385 .0959495 2.225328 .2329656
97.90982 149.7323
8.320047 9.980798
11.77 15.00
0.000 0.000
81.60282 130.1703
114.2168 169.2943
Although the output indicates that the model ts well, the small estimate of the damping parameter indicates that the random components will be widely distributed around the central frequency. To get a better idea of the dispersion of the components, we look at the estimated spectral density of the stationary cyclical component.
524

. psdensity sdensity3 omega3 . line sdensity3 omega3
.17 .145 0 UCM cycle 1 spectral density .15 .155 .16 .165
1 Frequency
The graph shows that the random components that make up the cyclical component are diffusely distributed around a central frequency.
Comparing UCM and ARIMA, revisited Example 7

Including lags of the dependent variable is an alternative method for modeling serially correlated errors. The estimated coefcients on the lags of the dependent variable estimate the coefcients in an autoregressive model for the stationary cyclical component; see Harvey (1989, 4748) for a discussion. Including lags of the dependent variable should be viewed as an alternative to the stochastic-cycle model for the stationary cyclical component. In this example, we use the large-sample equivalence of the random-walk model with pth order autoregressive errors and an ARIMA(p, 1, 0) to illustrate this point.
525
In the output below, we include 2 lags of the dependent variable in the random-walk UCM.
. ucm icsa L(1/2).icsa, model(rwalk) searching for initial values .......... (setting technique to bhhh) Iteration 0: log likelihood = -10044.209 Iteration 1: log likelihood = -9975.8312 Iteration 2: log likelihood = -9953.5727 Iteration 3: log likelihood = -9936.7489 Iteration 4: log likelihood = -9927.2306 (switching technique to nr) Iteration 5: log likelihood = -9918.9538 Iteration 6: log likelihood = -9890.8306 Iteration 7: log likelihood = -9889.562 Iteration 8: log likelihood = -9889.5608 Iteration 9: log likelihood = -9889.5608 Refining estimates: Iteration 0: log likelihood = -9889.5608 Iteration 1: log likelihood = -9889.5608 Unobserved-components model Components: random walk Sample: 21jan1967 - 19feb2011 Log likelihood = -9889.5608 OIM Std. Err.
= = =
2301 271.88 0.0000
icsa icsa L1. L2. Variance level
Coef.
P>|z|
-.3250633 -.1794686
.0205148 .0205246
-15.85 -8.74
0.000 0.000
-.3652715 -.2196961
-.2848551 -.1392411
317.6474
9.36691
33.91
0.000
299.2886
336.0062
Now we use arima to estimate the parameters of an asymptotically equivalent ARIMA(2,1,0) model. (We specify the technique(nr) option so that arima will compute the observed information matrix standard errors that ucm computes.) We use nlcom to compute a point estimate and a standard error for the variance, which is directly comparable to the one produced by ucm.
526

. arima icsa, noconstant arima(2,1,0) technique(nr) Iteration 0: Iteration 1: Sample: log likelihood = -9896.4584 log likelihood = -9896.458 Number of obs Wald chi2(2) Prob > chi2 = = = 2302 271.95 0.0000
ARIMA regression 14jan1967 - 19feb2011
Log likelihood = -9896.458 OIM Std. Err.
D.icsa ARMA ar L1. L2. /sigma
Coef.
P>|z|
-.3249383 -.1793353 17.81606
.0205036 .0205088 .2625695
-15.85 -8.74 67.85
0.000 0.000 0.000
-.3651246 -.2195317 17.30143
-.284752 -.1391388 18.33068
Note: The test of the variance against zero is one sided, and the two-sided confidence interval is truncated at zero. . nlcom _b[sigma:_cons]^2 _nl_1: D.icsa _nl_1 _b[sigma:_cons]^2 Coef. 317.4119 Std. Err. 9.355904 z 33.93 P>|z| 0.000 [95% Conf. Interval] 299.0746 335.7491
It is no accident that the parameter estimates and the standard errors from the two estimators are so close. As the sample size grows the differences in the parameter estimates and the estimated standard errors will go to 0, because the two estimators are equivalent in large samples.
Models for the trend and idiosyncratic components

A general model that allows for xed or stochastic trends in t is given by
t = t1 + t1 + t t = t1 + t
(2) (3)
Following Harvey (1989), we dene 11 exible models for yt that specify both t and t in (1). These models place restrictions on the general model specied in (2) and (3) and on t in (1). In other words, these models jointly specify t and t . To any of these models, a cyclical component, a seasonal component, or exogenous variables may be added.
527
Table 1. Models for the trend and idiosyncratic components Model name No trend or idiosyncratic component No trend Deterministic constant Local level Random walk Deterministic trend Syntax option model(none) model(ntrend) model(dconstant) model(llevel) model(rwalk) model(dtrend) Model
yt =
t t
yt = + =
yt =t + t t =t1 + t yt =t t =t1 + t yt =t + t t =t1 + = yt =t + t t =t1 + + t = yt =t t =t1 + + t = yt =t + t t =t1 + t1 + t t =t1 + t yt =t + t t =t1 + t1 t =t1 + t yt =t t =t1 + t1 t =t1 + t
Local level with deterministic trend Random walk with drift
model(lldtrend)
model(rwdrift)
Local linear trend
model(lltrend)
Smooth trend
model(strend)
Random trend
model(rtrend)
The majority of the models available in ucm are designed for nonstationary time series. The deterministic-trend model incorporates a rst-order deterministic time-trend in the model. The locallevel, random-walk, local-level-with-deterministic-trend, and random-walk-with-drift models are for modeling series with rst-order stochastic trends. A series with a dth-order stochastic trend must be differenced d times to be stationary. The local-linear-trend, smooth-trend, and random-trend models are for modeling series with second-order stochastic trends. The no-trend-or-idiosyncratic-component model is useful for using ucm to model stationary series with cyclical components or seasonal components and perhaps exogenous variables. The no-trend and the deterministic-constant models are useful for using ucm to model stationary series with seasonal components or exogenous variables.
528
Seasonal component
A seasonal component models cyclical behavior in a time series that occurs at known seasonal periodicities. A seasonal component is modeled in the time domain; the period of the cycle is specied as the number of time periods required for the cycle to complete.
Example 8
Lets begin by considering a series that displays a seasonal effect. Below we plot a monthly series containing the number of new cases of mumps in New York City between January 1928 and December 1972. (See Hipel and McLeod [1994] for the source and further discussion of this dataset.)
. use http://www.stata-press.com/data/r12/mumps (Hipel and Mcleod (1994), http://robjhyndman.com/tsdldata/epi/mumps.dat) . tsline mumps
number of mumps cases reported in NYC 500 1000 1500 2000 0 1930m1
1940m1
1950m1 Month
1960m1
1970m1
The graph reveals recurring spikes at regular intervals, which we suspect to be seasonal effects. The series may or may not be stationary; the graph evidence is not denitive. Deterministic seasonal effects are a standard method of incorporating seasonality into a model. In a model with a constant term, the s deterministic seasonal effects are modeled as s parameters subject to the constraint that they sum to zero; formally, t + t1 + + t(s1) = 0. A stochastic-seasonal model is a more exible alternative that allows the seasonal effects at time t to sum to t , a zero-mean, nite-variance, i.i.d. random variable; formally, t + t1 + + t(s1) = t . In the output below, we model the seasonal effects by a stochastic-seasonal model, we allow for the series to follow a random walk, and we include a stationary cyclical component.

. ucm mumps, seasonal(12) cycle(1) searching for initial values ................... (setting technique to bhhh) Iteration 0: log likelihood = -3268.1838 Iteration 1: log likelihood = -3256.5253 Iteration 2: log likelihood = -3254.6124 Iteration 3: log likelihood = -3250.3465 Iteration 4: log likelihood = -3249.3536 (switching technique to nr) Iteration 5: log likelihood = -3248.9212 Iteration 6: log likelihood = -3248.7176 Iteration 7: log likelihood = -3248.7138 Iteration 8: log likelihood = -3248.7138 Refining estimates: Iteration 0: log likelihood = -3248.7138 Iteration 1: log likelihood = -3248.7138 Unobserved-components model Components: random walk, seasonal(12), order 1 cycle Sample: 1928m1 - 1972m6 Number of obs Wald chi2(2) Log likelihood = -3248.7138 Prob > chi2 OIM Std. Err. .0282037 .0197933
529
= = =
534 2141.69 0.0000
mumps frequency damping Variance level seasonal cycle1
Coef. .3863607 .8405622
z 13.70 42.47
P>|z| 0.000 0.000
[95% Conf. Interval] .3310824 .8017681 .4416389 .8793563
221.2131 4.151639 12228.17
140.5179 4.383442 813.8394
1.57 0.95 15.03
0.058 0.172 0.000
0 0 10633.08
496.6231 12.74303 13823.27
The output indicates that the trend and seasonal variances may not be necessary. When the variance of the seasonal component is zero, the seasonal component becomes deterministic. Below we estimate the parameters of a model that includes deterministic seasonal effects and a stationary cyclical component.
. ucm mumps ibn.month, model(none) cycle(1) searching for initial values ....... (setting technique to bhhh) Iteration 0: log likelihood = -4105.9711 Iteration 1: log likelihood = -3676.1708 Iteration 2: log likelihood = -3473.0335 Iteration 3: log likelihood = -3412.5661 Iteration 4: log likelihood = -3383.1948 (switching technique to nr) Iteration 5: log likelihood = -3366.493 Iteration 6: log likelihood = -3283.7964 Iteration 7: log likelihood = -3283.0325 Iteration 8: log likelihood = -3283.0284 Iteration 9: log likelihood = -3283.0284 Refining estimates: Iteration 0: log likelihood = -3283.0284 Iteration 1: log likelihood = -3283.0284
530

Unobserved-components model Components: order 1 cycle Sample: 1928m1 - 1972m6 Log likelihood = -3283.0284 OIM Std. Err.
= = =
534 3404.29 0.0000
mumps cycle1 frequency damping mumps month 1 2 3 4 5 6 7 8 9 10 11 12 Variance cycle1
Coef.
P>|z|
.3272754 .844874
.0262922 .0184994
12.45 45.67
0.000 0.000
.2757436 .8086157
.3788071 .8811322
480.5095 561.9174 832.8666 894.0747 869.6568 770.1562 433.839 218.2394 140.686 148.5876 215.0958 330.2232
32.67128 32.66999 32.67696 32.64568 32.56282 32.48587 32.50165 32.56712 32.64138 32.69067 32.70311 32.68906
14.71 17.20 25.49 27.39 26.71 23.71 13.35 6.70 4.31 4.55 6.58 10.10
0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000
416.475 497.8854 768.8209 830.0904 805.8348 706.4851 370.1369 154.409 76.7101 84.51508 150.9989 266.1538
544.544 625.9494 896.9122 958.0591 933.4787 833.8274 497.541 282.0698 204.662 212.6601 279.1927 394.2926
13031.53
798.2719
16.32
0.000
11466.95
14596.11
The output indicates that each of these components is statistically signicant.
Technical note
In a stochastic model for the seasonal component, the seasonal effects sum to the random variable 2 t i.i.d. N (0, ):
s1
t =
j=1
tj + t
Saved results
Because ucm is estimated using sspace, most of the sspace saved results appear after ucm. Not all these results are relevant for ucm; programmers wishing to treat ucm results as sspace results should see Saved results of [TS] sspace. See Methods and formulas for the state-space representation of UCMs, and see [TS] sspace for more documentation that relates to all the saved results.
531
ucm saves the following in e():

Scalars e(N) e(k) e(k aux) e(k eq) e(k dv) e(k cycles) e(df m) e(ll) e(chi2) e(p) e(tmin) e(tmax) e(stationary) e(rank) e(ic) e(rc) e(converged) Macros e(cmd) e(cmdline) e(depvar) e(covariates) e(indeps) e(tvar) e(eqnames) e(model) e(title) e(tmins) e(tmaxs) e(chi2type) e(vce) e(vcetype) e(opt) e(initial values) e(technique) e(tech steps) e(properties) e(estat cmd) e(predict) e(marginsok) e(marginsnotok) Matrices e(b) e(Cns) e(ilog) e(gradient) e(V) e(V modelbased) Functions e(sample) number of observations number of parameters number of auxiliary parameters number of equations in e(b) number of dependent variables number of stochastic cycles model degrees of freedom log likelihood
2
signicance minimum time in sample maximum time in sample 1 if the estimated parameters indicate a stationary model, 0 otherwise rank of VCE number of iterations return code 1 if converged, 0 otherwise ucm command as typed unoperated names of dependent variables in observation equations list of covariates independent variables variable denoting time within groups names of equations type of model title in estimation output formatted minimum time formatted maximum time Wald; type of model 2 test vcetype specied in vce() title used to label Std. Err. type of optimization type of initial values maximization technique iterations taken in maximization technique b V program used to implement estat program used to implement predict predictions allowed by margins predictions disallowed by margins parameter vector constraints matrix iteration log (up to 20 iterations) gradient vector variancecovariance matrix of the estimators model-based variance marks estimation sample
532

ucm is implemented as an ado-le. Methods and formulas are presented under the following headings:
Introduction State-space formulation Cyclical component extensions
Introduction
The general form of UCMs can be expressed as
yt = t + t + t + xt +
where t is the trend, t is the seasonal component, t is the cycle, is the regression coefcients for regressors xt , and t is the idiosyncratic error with variance 2 . We can decompose the trend as
t = t t = t1 + t1 + t t = t1 + t
where t is the local level, t is the local slope, and t and t are i.i.d. normal errors with mean 0 2 2 and variance and , respectively. Next consider the seasonal component, t , with a period of s time units. Ignoring a seasonal s1 disturbance term, the seasonal effects will sum to zero, j=0 tj = 0. Adding a normal error term, 2 t , with mean 0 and variance , we express the seasonal component as
s1
t =
j=1
tj + t
Finally, the cyclical component, t , is a function of the frequency , in radians, and a unit-less scaling variable , termed the damping effect, 0 < < 1. We require two equations to express the cycle: t = t1 cos + t1 sin + t
t = t1 sin + t1 cos + t
2 where the t and t disturbances are normally distributed with mean 0 and variance .
The disturbance terms
t,
t , t , t , t , and t are independent.
533
State-space formulation
ucm is an easy-to-use implementation of the state-space command sspace, with special modications, where the local linear trend components, seasonal components, and cyclical components are states of the state-space model. The state-space model can be expressed in matrix form as
yt = Dzt + Fxt + zt = Azt1 + Ct
where yt , t = 1, . . . , T , are the observations and zt are the unobserved states. The number of states, m, depends on the model specied. The k 1 vector xt contains the exogenous variables specied as indepvars, and the 1 k vector F contains the regression coefcients to be estimated. t is the observation equation disturbance, and the m0 1 vector t contains the state equation disturbances, where m0 m. Finally, C is a m m0 matrix of zeros and ones. These recursive equations are evaluated using the diffuse Kalman lter of De Jong (1991). Below we give the state-space matrix structures for a local linear trend with a stochastic seasonal component, with a period of 4 time units, and an order-2 cycle. The state vector, zt , and its transition matrix, A, have the structure
1 0 0 0 A = 0 0 0 0 0
1 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 1 1 1 0 0 0 0 0 1 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 cos sin 1 0 0 0 0 0 sin cos 0 1 0 0 0 0 0 0 cos sin 0 0 0 0 0 0 sin cos 1 0 0 0 C = 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 1
t t t t1 zt = t2 t,1 t,1 t,2 t,2
t t t = t t t
D = (1 0 1 0 0 1 0 0 0)
534
Cyclical component extensions

Recall that the stochastic cyclical model is given by
t = (t1 cos c + t1 sin c ) + t,1 t = (t1 sin c + t1 cos c ) + t,2 2 where t,j i.i.d. N (0, ) and 0 < < 1 is a damping effect. The cycle is variance-stationary 2 when < 1 because Var(t ) = /(1 ). We will express a UCM with a cyclical component added to a trend as yt = t + t + t
where t can be any of the trend parameterizations discussed earlier. Higher-order cycles, k = 2 or k = 3, are dened as
t,j = (t1,j cos c + t1,j sin c ) + t1,j+1 t,j = (t1,j sin c + t1,j cos c ) + t1,j+1
for j < k , and
t,k = (t1,k cos c + t1,k sin c ) + t,1 t,k = (t1,k sin c + t1,k cos c ) + t,2
Harvey and Trimbur (2003) discuss the properties of this model and its state-space formulation.
References
Burns, A. F., and W. C. Mitchell. 1946. Measuring Business Cycles. New York: National Bureau of Economic Research. De Jong, P. 1991. The diffuse Kalman lter. Annals of Statistics 19: 10731083. Durbin, J., and S. J. Koopman. 2001. Time Series Analysis by State Space Methods. Oxford: Oxford University Press. Fuller, W. A. 1996. Introduction to Statistical Time Series. 2nd ed. New York: Wiley. Hamilton, J. D. 1994. Time Series Analysis. Princeton: Princeton University Press. Harvey, A. C. 1989. Forecasting, Structural Time Series Models and the Kalman Filter. Cambridge: Cambridge University Press. . 1993. Time Series Models. 2nd ed. Cambridge, MA: MIT Press. Harvey, A. C., and T. M. Trimbur. 2003. General model-based lters for extracting cycles and trends in economic time series. The Review of Economics and Statistics 85: 244255. Hipel, K. W., and A. I. McLeod. 1994. Time Series Modelling of Water Resources and Environmental Systems. Amsterdam: Elsevier. Priestley, M. B. 1981. Spectral Analysis and Time Series. London: Academic Press. Trimbur, T. M. 2006. Properties of higher order stochastic cycles. Journal of Time Series Analysis 27: 117. Wei, W. W. S. 2006. Time Series Analysis: Univariate and Multivariate Methods. 2nd ed. Boston: Pearson.
535
Also see
[TS] ucm postestimation Postestimation tools for ucm [TS] tsset Declare data to be time-series data [TS] sspace State-space models [TS] var Vector autoregressive models [TS] arima ARIMA, ARMAX, and other dynamic regression models [TS] tslter Filter a time-series, keeping only selected periodicities [TS] tssmooth Smooth and forecast univariate time-series data [U] 20 Estimation and postestimation commands
Title
ucm postestimation Postestimation tools for ucm
Description
The following postestimation commands are of special interest after ucm: Command estat period psdensity Description display cycle periods in time units estimate the spectral density
For information about estat period, see below. For information about psdensity, see [TS] psdensity. The following standard postestimation commands are also available: Command estat estimates lincom lrtest nlcom predict predictnl test testnl Description
cataloging estimation results point estimates, standard errors, testing and inference for linear combinations of coefcients likelihood-ratio test point estimates, standard errors, testing and inference for nonlinear combinations of coefcients predictions, residuals, inuence statistics, and other diagnostic measures point estimates, standard errors, testing, and inference for generalized predictions Wald tests of simple and composite linear hypotheses Wald tests of nonlinear hypotheses
Special-interest postestimation commands

estat period transforms an estimated central frequency to an estimated period after ucm.
536
537
Syntax for predict

predict statistic
Main
type
stub* | newvarlist Description
if
in
, statistic options
xb trend seasonal cycle residuals rstandard
linear prediction using exogenous variables trend component seasonal component cyclical component residuals standardized residuals
options
Options
Description
put estimated root mean squared errors of predicted statistics in the new variable dynamic(time constant) begin dynamic forecast at specied time
Advanced
rmse(stub* | newvarlist)
method for predicting unobserved components Description predict using past information predict using all sample information predict using past and contemporaneous information
Menu
Statistics
>
Postestimation
>
Options for predict

Main
xb, trend, seasonal, cycle, residuals, and rstandard specify the statistic to be predicted. xb, the default, calculates the linear predictions using the exogenous variables. xb may not be used with the smethod(filter) option. trend estimates the unobserved trend component. seasonal estimates the unobserved seasonal component. cycle estimates the unobserved cyclical component.
538
residuals calculates the residuals in the equation for the dependent variable. residuals may not be specied with dynamic(). rstandard calculates the standardized residuals, which are the residuals normalized to have unit variances. rstandard may not be specied with the smethod(filter), smethod(smooth), or dynamic() option.
Options
rmse(stub* | newvarlist) puts the root mean squared errors of the predicted statistic into the specied new variable. Multiple variables are only required for predicting cycles of a model that has more than one cycle. The root mean squared errors measure the variances due to the disturbances but do not account for estimation error. The stub* syntax is for models with multiple cycles, where you provide the prex and predict will add a numeric sufx for each predicted cycle. dynamic(time constant) species when predict should start producing dynamic forecasts. The specied time constant must be in the scale of the time variable specied in tsset, and the time constant must be inside a sample for which observations on the dependent variable are available. For example, dynamic(tq(2008q4)) causes dynamic predictions to begin in the fourth quarter of 2008, assuming that your time variable is quarterly; see [D] datetime. If the model contains exogenous variables, they must be present for the whole predicted sample. dynamic() may not be specied with the rstandard, residuals, or smethod(smooth) option.
Advanced
smethod(method) species the method for predicting the unobserved components. smethod() causes different amounts of information on the dependent variable to be used in predicting the components at each time period. smethod(onestep), the default, causes predict to estimate the components at each time period using previous information on the dependent variable. The Kalman lter is performed on previous periods, but only the one-step predictions are made for the current period. smethod(smooth) causes predict to estimate the components at each time period using all the sample data by the Kalman smoother. smethod(smooth) may not be specied with the rstandard option. smethod(filter) causes predict to estimate the components at each time period using previous and contemporaneous data by the Kalman lter. The Kalman lter is performed on previous periods and the current period. smethod(filter) may not be specied with the xb option.
Syntax for estat period

estat period options
Main
, options Description set condence level; default is level(95) numeric format
level(#) cformat(% fmt)
539
Menu
Statistics
>
Postestimation
>
Reports and statistics
Options for estat period

Options
level(#) species the condence level, as a percentage, for condence intervals. The default is level(95) or as set by set level; see [U] 20.7 Specifying the width of condence intervals. cformat(% fmt) sets the display format for the table numeric values. The default is cformat(%9.0g).
Remarks
We assume that you have already read [TS] ucm. In this entry, we illustrate some features of predict after using ucm to estimate the parameters of an unobserved-components model. All predictions after ucm depend on the unobserved components, which are estimated recursively using a Kalman lter. Changing the sample can alter the state estimates, which can change all other predictions.
Example 1
We begin by modeling monthly data on the median duration of employment spells in the United States. We include a stochastic-seasonal component because the data have not been seasonally adjusted.
. use http://www.stata-press.com/data/r12/uduration2 (BLS data, not seasonally adjusted) . ucm duration, seasonal(12) cycle(1) difficult searching for initial values .................... (setting technique to bhhh) Iteration 0: log likelihood = -409.79512 Iteration 1: log likelihood = -403.383 Iteration 2: log likelihood = -403.37363 (backed up) Iteration 3: log likelihood = -403.36891 (backed up) Iteration 4: log likelihood = -403.36653 (backed up) (switching technique to nr) Iteration 5: log likelihood = -403.36649 (backed up) Iteration 6: log likelihood = -397.87764 (not concave) Iteration 7: log likelihood = -396.53184 (not concave) Iteration 8: log likelihood = -395.59253 (not concave) Iteration 9: log likelihood = -390.32744 (not concave) Iteration 10: log likelihood = -389.14967 (not concave) Iteration 11: log likelihood = -388.733 (not concave) Iteration 12: log likelihood = -388.50581 Iteration 13: log likelihood = -388.28068 Iteration 14: log likelihood = -388.25924 Iteration 15: log likelihood = -388.25676 Iteration 16: log likelihood = -388.25675 Refining estimates: Iteration 0: log likelihood = -388.25675 Iteration 1: log likelihood = -388.25675
540

Unobserved-components model Components: random walk, seasonal(12), order 1 cycle Sample: 1967m7 - 2008m12 Number of obs Wald chi2(2) Log likelihood = -388.25675 Prob > chi2 OIM Std. Err. .7250323 .1050168
= = =
498 7.17 0.0277
duration frequency damping Variance level seasonal cycle1
Coef. 1.641531 .2671232
z 2.26 2.54
P>|z| 0.024 0.011
[95% Conf. Interval] .2204938 .0612939 3.062568 .4729524
.1262922 .0017289 .0641496
.0221428 .0009647 .0211839
5.70 1.79 3.03
0.000 0.037 0.001
.0828932 0 .0226299
.1696912 .0036196 .1056693
Below we predict the trend and the seasonal components to get a look at the model t.
. . . . . predict strend, trend predict season, seasonal tsline duration strend, name(trend) nodraw legend(rows(1)) tsline season, name(season) yline(0,lwidth(vthin)) nodraw graph combine trend season, rows(2)
4 1970m1
10 12 14
1980m1
1990m1 Month
2000m1 trend, onestep
2010m1
median duration of unemployment
2 1970m1
seasonal, onestep 1 0 1
1980m1
1990m1 Month
2000m1
2010m1
The trend tracks the data well. That the seasonal component appears to change over time indicates that the stochastic-seasonal component might t better than a deterministic-seasonal component.
Example 2
In this example, we use the model to forecast the median unemployment duration. We use the root mean squared error of the prediction to compute a condence interval of our dynamic predictions. Recall that the root mean squared error accounts for variances due to the disturbances but not due to the estimation error.

. tsappend, add(12) . predict duration_f, dynamic(tm(2009m1)) rmse(rmse) . scalar z = invnormal(0.95) . generate lbound = duration_f - z*rmse if tm>=tm(2008m12) (497 missing values generated) . generate ubound = duration_f + z*rmse if tm>=tm(2008m12) (497 missing values generated) . label variable lbound "90% forecast interval" . twoway (tsline duration duration_f if tm>=tm(2006m1)) > (tsrline lbound ubound if tm>=tm(2008m12)), > ysize(2) xtitle("") legend(cols(1))
14
541
6 2006m1
10
12
2007m1
2008m1
2009m1
2010m1
median duration of unemployment xb prediction, duration, dynamic(tm(2009m1)) 90% forecast interval/ubound
The model forecasts a large temporary increase in the median duration of unemployment.

All postestimation commands listed above are implemented as ado-les. For details on the ucm postestimation methods, see [TS] sspace postestimation. See [TS] psdensity for the methods used to estimate the spectral density.
Also see
[TS] ucm Unobserved-components model [TS] psdensity Parametric spectral density estimation after arima, arma, and ucm [TS] sspace postestimation Postestimation tools for sspace
Title
var intro Introduction to vector autoregressive models
Description
Stata has a suite of commands for tting, forecasting, interpreting, and performing inference on vector autoregressive (VAR) models and structural vector autoregressive (SVAR) models. The suite includes several commands for estimating and interpreting impulseresponse functions (IRFs), dynamicmultiplier functions, and forecast-error variance decompositions (FEVDs). The table below describes the available commands.
Fitting a VAR or SVAR var [TS] var svar [TS] var svar varbasic [TS] varbasic Model diagnostics and inference varstable [TS] varstable varsoc [TS] varsoc
Fit vector autoregressive models Fit structural vector autoregressive models Fit a simple VAR and graph IRFs or FEVDs Check the stability condition of VAR or SVAR estimates Obtain lag-order selection statistics for VARs and VECMs Obtain Wald lag-exclusion statistics after var or svar Perform pairwise Granger causality tests after var or svar Perform LM test for residual autocorrelation after var or svar Test for normally distributed disturbances after var or svar
varwle vargranger varlmar varnorm
[TS] varwle [TS] vargranger [TS] varlmar [TS] varnorm
Forecasting after tting a VAR or SVAR fcast compute [TS] fcast compute Compute dynamic forecasts of dependent variables after var, svar, or vec fcast graph [TS] fcast graph Graph forecasts of variables computed by fcast compute Working with IRFs, dynamic-multiplier functions, and FEVDs irf [TS] irf Create and analyze IRFs, dynamic-multiplier functions, and FEVDs
This entry provides an overview of vector autoregressions and structural vector autoregressions. More rigorous treatments can be found in Hamilton (1994), L tkepohl (2005), and Amisano and u Giannini (1997). Stock and Watson (2001) provide an excellent nonmathematical treatment of vector autoregressions and their role in macroeconomics.
542
543
Remarks
Introduction to VARs Introduction to SVARs Short-run SVAR models Long-run restrictions IRFs and FEVDs
Introduction to VARs
A VAR is a model in which K variables are specied as linear functions of p of their own lags, p lags of the other K 1 variables, and possibly additional exogenous variables. Algebraically, a p-order VAR model, written VAR(p), with exogenous variables xt is given by
yt = v + A1 yt1 + + Ap ytp + B0 xt + B1 xt1 + + Bs xts + ut

where
t {, } (1)
yt = (y1t , . . . , yKt ) is a K 1 random vector, A1 through Ap are K K matrices of parameters, xt is an M 1 vector of exogenous variables, B0 through Bs are K M matrices of coefcients, v is a K 1 vector of parameters, and ut is assumed to be white noise; that is, E(ut ) = 0, E(ut ut ) = , and E(ut us ) = 0 for t = s
There are K 2 p + K (M (s + 1) + 1) parameters in the equation for yt , and there are {K (K + 1)}/2 parameters in the covariance matrix . One way to reduce the number of parameters is to specify an incomplete VAR, in which some of the A or B matrices are set to zero. Another way is to specify linear constraints on some of the coefcients in the VAR. A VAR can be viewed as the reduced form of a system of dynamic simultaneous equations. Consider the system
W0 yt = a + W1 yt1 + + Wp ytp + W1 xt + W2 xt2 + + Ws xts + et
(2)
where a is a K 1 vector of parameters, each Wi , i = 0, . . . , p, is a K K matrix of parameters, and et is a K 1 disturbance vector. In the traditional dynamic simultaneous equations approach, sufcient restrictions are placed on the Wi to obtain identication. Assuming that W0 is nonsingular, (2) can be rewritten as
1 1 1 yt =W0 a + W0 W1 yt1 + + W0 Wp ytp 1 1 1 1 + W0 W1 xt + W0 W2 xt2 + + W0 Ws xts + W0 et which is a VAR with 1 v = W0 a 1 Ai = W0 Wi 1 Bi = W0 Wi 1 ut = W0 et
(3)
544
The cross-equation error variancecovariance matrix contains all the information about contemporaneous correlations in a VAR and may be the VARs greatest strength and its greatest weakness. Because no questionable a priori assumptions are imposed, tting a VAR allows the dataset to speak for itself. However, without imposing some restrictions on the structure of , we cannot make a causal interpretation of the results. If we make additional technical assumptions, we can derive another representation of the VAR in (1). If the VAR is stable (see [TS] varstable), we can rewrite yt as

yt = +
i=0
Di xti +
i=0
i uti
(4)
where is the K 1 time-invariant mean of the process and Di and i are K M and K K matrices of parameters, respectively. Equation (4) states that the process by which the variables in yt uctuate about their time-invariant means, , is completely determined by the parameters in Di and i and the (innite) past history of the exogenous variables xt and the independent and identically distributed (i.i.d.) shocks or innovations, ut1 , ut2 , . . . . Equation (4) is known as the vector moving-average representation of the VAR. The Di are the dynamic-multiplier functions, or transfer functions. The moving-average coefcients i are also known as the simple IRFs at horizon i. The precise relationships between the VAR parameters and the Di and i are derived in Methods and formulas of [TS] irf create. The joint distribution of yt is determined by the distributions of xt and ut and the parameters v, Bi , and Ai . Estimating the parameters in a VAR requires that the variables in yt and xt be covariance stationary, meaning that their rst two moments exist and are time invariant. If the yt are not covariance stationary, but their rst differences are, a vector error-correction model (VECM) can be used. See [TS] vec intro and [TS] vec for more information about those models. If the ut form a zero mean, i.i.d. vector process, and yt and xt are covariance stationary and are not correlated with the ut , consistent and efcient estimates of the Bi , the Ai , and v are obtained via seemingly unrelated regression, yielding estimators that are asymptotically normally distributed. When the equations for the variables yt have the same set of regressors, equation-by-equation OLS estimates are the conditional maximum likelihood estimates. Much of the interest in VAR models is focused on the forecasts, IRFs, dynamic-multiplier functions, and the FEVDs, all of which are functions of the estimated parameters. Estimating these functions is straightforward, but their asymptotic standard errors are usually obtained by assuming that ut forms a zero mean, i.i.d. Gaussian (normal) vector process. Also, some of the specication tests for VARs have been derived using the likelihood-ratio principle and the stronger Gaussian assumption. In the absence of contemporaneous exogenous variables, the disturbance variancecovariance matrix contains all the information about contemporaneous correlations among the variables. VARs are sometimes classied into three types by how they account for this contemporaneous correlation. (See Stock and Watson [2001] for one derivation of this taxonomy.) A reduced-form VAR, aside from estimating the variancecovariance matrix of the disturbance, does not try to account for contemporaneous correlations. In a recursive VAR, the K variables are assumed to form a recursive dynamic structural equation model in which the rst variable is a function of lagged variables, the second is a function of contemporaneous values of the rst variable and lagged values, and so on. In a structural VAR, the theory you are working with places restrictions on the contemporaneous correlations that are not necessarily recursive. Stata has two commands for tting reduced-form VARs: var and varbasic. var allows for constraints to be imposed on the coefcients. varbasic allows you to t a simple VAR quickly without constraints and graph the IRFs.
545
Because tting a VAR of the correct order can be important, varsoc offers several methods for choosing the lag order p of the VAR to t. After tting a VAR, and before proceeding with inference, interpretation, or forecasting, checking that the VAR ts the data is important. varlmar can be used to check for autocorrelation in the disturbances. varwle performs Wald tests to determine whether certain lags can be excluded. varnorm tests the null hypothesis that the disturbances are normally distributed. varstable checks the eigenvalue condition for stability, which is needed to interpret the IRFs and IRFs.
Introduction to SVARs
As discussed in [TS] irf create, a problem with VAR analysis is that, because is not restricted to be a diagonal matrix, an increase in an innovation to one variable provides information about the innovations to other variables. This implies that no causal interpretation of the simple IRFs is possible: there is no way to determine whether the shock to the rst variable caused the shock in the second variable or vice versa. However, suppose that we had a matrix P such that = PP . We can then show that the variables in P1 ut have zero mean and that E{P1 ut (P1 ut ) } = IK . We could rewrite (4) as
yt = +
s=0
s PP1 uts s P1 uts

s=0
=+ =+
s=0
s wts
(5)
where s = s P and wt = P1 ut . If we had such a P, the wk would be mutually orthogonal, and the s would allow the causal interpretation that we seek.
SVAR models provide a framework for estimation of and inference about a broad class of P matrices. As described in [TS] irf create, the estimated P matrices can then be used to estimate structural IRFs and structural FEVDs. There are two types of SVAR models. Short-run SVAR models identify a P matrix by placing restrictions on the contemporaneous correlations between the variables. Long-run SVAR models, on the other hand, do so by placing restrictions on the long-term accumulated effects of the innovations.
Short-run SVAR models

A short-run SVAR model without exogenous variables can be written as
A(IK A1 L A2 L2 Ap Lp )yt = A
= Bet
(6)
where L is the lag operator; A, B, and A1 , . . . , Ap are K K matrices of parameters; t is a K 1 vector of innovations with t N (0, ) and E[ t s ] = 0K for all s = t; and et is a K 1 vector of orthogonalized disturbances; that is, et N (0, IK ) and E[et es ] = 0K for all s = t. These transformations of the innovations allow us to analyze the dynamics of the system in terms of a change to an element of et . In a short-run SVAR model, we obtain identication by placing restrictions on A and B, which are assumed to be nonsingular.
546
Equation (6) implies that Psr = A1 B, where Psr is the P matrix identied by a particular short-run SVAR model. The latter equality in (6) implies that
t tA
= Bet et B
Taking the expectation of both sides yields
= Psr Psr
Assuming that the underlying VAR is stable (see [TS] varstable for a discussion of stability), we can invert the autoregressive representation of the model in (6) to an innite-order, moving-average representation of the form
yt = +
s=0
sr ets s
(7)
whereby yt is expressed in terms of the mutually orthogonal, unit-variance structural innovations et . The sr contain the structural IRFs at horizon s. s In a short-run SVAR model, the A and B matrices model all the information about contemporaneous correlations. The B matrix also scales the innovations ut to have unit variance. This allows the structural IRFs constructed from (7) to be interpreted as the effect on variable i of a one-time unit increase in the structural innovation to variable j after s periods.
Psr identies the structural IRFs by dening a transformation of , and Psr is identied by the restrictions placed on the parameters in A and B. Because there are only K(K + 1)/2 free parameters in , only K(K + 1)/2 parameters may be estimated in an identied Psr . Because there are 2K 2 total parameters in A and B, the order condition for identication requires that at least 2K 2 K(K + 1)/2 restrictions be placed on those parameters. Just as in the simultaneous-equations framework, this order condition is necessary but not sufcient. Amisano and Giannini (1997) derive a method to check that an SVAR model is locally identied near some specied values for A and B.
Before moving on to models with long-run constraints, consider these limitations. We cannot place constraints on the elements of A in terms of the elements of B, or vice versa. This limitation is imposed by the form of the check for identication derived by Amisano and Giannini (1997). As noted in Methods and formulas of [TS] var svar, this test requires separate constraint matrices for the parameters in A and B. Also, we cannot mix short-run and long-run constraints.
Long-run restrictions
A general short-run SVAR has the form
A(IK A1 L A2 L2 Ap Lp )yt = Bet To simplify the notation, let A = (IK A1 L A2 L2 Ap Lp ). The model is assumed to be stable (see [TS] varstable), so A1 , the matrix of estimated long-run effects of the reduced-form VAR shocks, is well dened. Constraining A to be an identity matrix allows us to rewrite this equation as yt = A1 Bet which implies that = BB . Thus C = A1 B is the matrix of long-run responses to the orthogonalized shocks, and yt = Cet
547
In long-run models, the constraints are placed on the elements of C, and the free parameters are estimated. These constraints are often exclusion restrictions. For instance, constraining C[1, 2] to be zero can be interpreted as setting the long-run response of variable 1 to the structural shocks driving variable 2 to be zero. Statas svar command estimates the parameters of structural VARs. See [TS] var svar for more information and examples.
IRFs and FEVDs

IRFs describe how the K endogenous variables react over time to a one-time shock to one of the K disturbances. Because the disturbances may be contemporaneously correlated, these functions do not explain how variable i reacts to a one-time increase in the innovation to variable j after s periods, holding everything else constant. To explain this, we must start with orthogonalized innovations so that the assumption to hold everything else constant is reasonable. Recursive VARs use a Cholesky decomposition to orthogonalize the disturbances and thereby obtain structurally interpretable IRFs. Structural VARs use theory to impose sufcient restrictions, which need not be recursive, to decompose the contemporaneous correlations into orthogonal components. FEVDs are another tool for interpreting how the orthogonalized innovations affect the K variables over time. The FEVD from j to i gives the fraction of the s-step forecast-error variance of variable i that can be attributed to the j th orthogonalized innovation.
Dynamicmultiplier functions describe how the endogenous variables react over time to a unit change in an exogenous variable. This is a different experiment from that in IRFs and FEVDs because dynamic-multiplier functions consider a change in an exogenous variable instead of a shock to an endogenous variable. irf create estimates IRFs, Cholesky orthogonalized IRFs, dynamic-multiplier functions, and structural IRFs and their standard errors. It also estimates Cholesky and structural FEVDs. The irf graph, irf cgraph, irf ograph, irf table, and irf ctable commands graph and tabulate these estimates. Stata also has several other commands to manage IRF and FEVD results. See [TS] irf for a description of these commands. fcast compute computes dynamic forecasts and their standard errors from VARs. fcast graph graphs the forecasts that are generated using fcast compute.
VARs allow researchers to investigate whether one variable is useful in predicting another variable. A variable x is said to Granger-cause a variable y if, given the past values of y , past values of x are useful for predicting y . The Stata command vargranger performs Wald tests to investigate Granger causality between the variables in a VAR.
References
Amisano, G., and C. Giannini. 1997. Topics in Structural VAR Econometrics. 2nd ed. Heidelberg: Springer. Hamilton, J. D. 1994. Time Series Analysis. Princeton: Princeton University Press. L tkepohl, H. 2005. New Introduction to Multiple Time Series Analysis. New York: Springer. u Stock, J. H., and M. W. Watson. 2001. Vector autoregressions. Journal of Economic Perspectives 15: 101115. Watson, M. W. 1994. Vector autoregressions and cointegration. In Vol. IV of Handbook of Econometrics, ed. R. F. Engle and D. L. McFadden. Amsterdam: Elsevier.
548
Also see
[TS] var Vector autoregressive models [TS] var svar Structural vector autoregressive models [TS] vec intro Introduction to vector error-correction models [TS] vec Vector error-correction models [TS] irf Create and analyze IRFs, dynamic-multiplier functions, and FEVDs
Title
var Vector autoregressive models
Syntax
var depvarlist options
Model
if
in
, options Description suppress constant term use lags numlist in the VAR use exogenous variables varlist apply specied linear constraints suppress SURE iteration log set maximum number of iterations for SURE; default is iterate(1600) set convergence tolerance of SURE use one-step SURE make small-sample degrees-of-freedom adjustment report small-sample t and F statistics do not compute parameter vector for coefcients implicitly set to zero set condence level; default is level(95) report L tkepohl lag-order selection statistics u do not display constraints control column formats, row spacing, and line width display legend instead of statistics
noconstant lags(numlist) exog(varlist)

Model 2
constraints(numlist) nolog iterate(#) tolerance(#) noisure dfk small nobigf

Reporting
level(#) lutstats nocnsreport display options coeflegend
You must tsset your data before using var; see [TS] tsset. depvarlist and varlist may contain time-series operators; see [U] 11.4.4 Time-series varlists. by, rolling, statsby, and xi are allowed; see [U] 11.1.10 Prex commands. coeflegend does not appear in the dialog box. See [U] 20 Estimation and postestimation commands for more capabilities of estimation commands.
Menu
Statistics
>
>
Vector autoregression (VAR)
549
550
Description
var ts a multivariate time-series regression of each dependent variable on lags of itself and on lags of all the other dependent variables. var also ts a variant of vector autoregressive (VAR) models known as the VARX model, which also includes exogenous variables. See [TS] var intro for a list of commands that are used in conjunction with var.
Options

Model
noconstant; see [R] estimation options. lags(numlist) species the lags to be included in the model. The default is lags(1 2). This option takes a numlist and not simply an integer for the maximum lag. For example, lags(2) would include only the second lag in the model, whereas lags(1/2) would include both the rst and second lags in the model. See [U] 11.1.8 numlist and [U] 11.4.4 Time-series varlists for more discussion of numlists and lags. exog(varlist) species a list of exogenous variables to be included in the VAR.
Model 2
constraints(numlist); see [R] estimation options. nolog suppresses the log from the iterated seemingly unrelated regression algorithm. By default, the iteration log is displayed when the coefcients are estimated through iterated seemingly unrelated regression. When the constraints() option is not specied, the estimates are obtained via OLS, and nolog has no effect. For this reason, nolog can be specied only when constraints() is specied. Similarly, nolog cannot be combined with noisure. iterate(#) species an integer that sets the maximum number of iterations when the estimates are obtained through iterated seemingly unrelated regression. By default, the limit is 1,600. When constraints() is not specied, the estimates are obtained using OLS, and iterate() has no effect. For this reason, iterate() can be specied only when constraints() is specied. Similarly, iterate() cannot be combined with noisure. tolerance(#) species a number greater than zero and less than 1 for the convergence tolerance of the iterated seemingly unrelated regression algorithm. By default, the tolerance is 1e-6. When the constraints() option is not specied, the estimates are obtained using OLS, and tolerance() has no effect. For this reason, tolerance() can be specied only when constraints() is specied. Similarly, tolerance() cannot be combined with noisure. noisure species that the estimates in the presence of constraints be obtained through one-step seemingly unrelated regression. By default, var obtains estimates in the presence of constraints through iterated seemingly unrelated regression. When constraints() is not specied, the estimates are obtained using OLS, and noisure has no effect. For this reason, noisure can be specied only when constraints() is specied. dfk species that a small-sample degrees-of-freedom adjustment be used when estimating , the error variancecovariance matrix. Specically, 1/(T m) is used instead of the large-sample divisor 1/T , where m is the average number of parameters in the functional form for yt over the K equations. small causes var to report small-sample t and F statistics instead of the large-sample normal and chi-squared statistics.
551
nobigf requests that var not save the estimated parameter vector that incorporates coefcients that have been implicitly constrained to be zero, such as when some lags have been omitted from a model. e(bf) is used for computing asymptotic standard errors in the postestimation commands irf create and fcast compute; see [TS] irf create and [TS] fcast compute. Therefore, specifying nobigf implies that the asymptotic standard errors will not be available from irf create and fcast compute. See Fitting models with some lags excluded.
Reporting
level(#); see [R] estimation options. lutstats species that the L tkepohl (2005) versions of the lag-order selection statistics be reported. u See Methods and formulas in [TS] varsoc for a discussion of these statistics. nocnsreport; see [R] estimation options. display options: vsquish, cformat(% fmt), pformat(% fmt), sformat(% fmt), and nolstretch; see [R] estimation options. The following option is available with var but is not shown in the dialog box: coeflegend; see [R] estimation options.
Remarks
Introduction Fitting models with some lags excluded Fitting models with exogenous variables Fitting models with constraints on the coefcients
Introduction
A VAR is a model in which K variables are specied as linear functions of p of their own lags, p lags of the other K 1 variables, and possibly exogenous variables. A VAR with p lags is usually denoted a VAR(p). For more information, see [TS] var intro.
Example 1: VAR model

To illustrate the basic usage of var, we replicate the example in L tkepohl (2005, 7778). The u data consists of three variables: the rst difference of the natural log of investment, dln inv; the rst difference of the natural log of income, dln inc; and the rst difference of the natural log of consumption, dln consump. The dataset contains data through the fourth quarter of 1982, though L tkepohl uses only the observations through the fourth quarter of 1978. u
. use http://www.stata-press.com/data/r12/lutkepohl2 (Quarterly SA West German macro data, Bil DM, from Lutkepohl 1993 Table E.1) . tsset time variable: qtr, 1960q1 to 1982q4 delta: 1 quarter
552

. var dln_inv dln_inc dln_consump if qtr<=tq(1978q4), lutstats dfk Vector autoregression Sample: 1960q4 - 1978q4 No. of obs Log likelihood = 606.307 (lutstats) AIC FPE = 2.18e-11 HQIC Det(Sigma_ml) = 1.23e-11 SBIC Equation Parms RMSE R-sq chi2 P>chi2 dln_inv dln_inc dln_consump 7 7 7 .046148 .011719 .009445 0.1286 0.1142 0.2513 9.736909 8.508289 22.15096 0.1362 0.2032 0.0011
= 73 = -24.63163 = -24.40656 = -24.06686
Coef. dln_inv dln_inv L1. L2. dln_inc L1. L2. dln_consump L1. L2. _cons dln_inc dln_inv L1. L2. dln_inc L1. L2. dln_consump L1. L2. _cons dln_consump dln_inv L1. L2. dln_inc L1. L2. dln_consump L1. L2. _cons
Std. Err.
P>|z|
-.3196318 -.1605508
.1254564 .1249066
-2.55 -1.29
0.011 0.199
-.5655218 -.4053633
-.0737419 .0842616
.1459851 .1146009
.5456664 .5345709
0.27 0.21
0.789 0.830
-.9235013 -.9331388
1.215472 1.162341
.9612288 .9344001 -.0167221
.6643086 .6650949 .0172264
1.45 1.40 -0.97
0.148 0.160 0.332
-.3407922 -.369162 -.0504852
2.26325 2.237962 .0170409
.0439309 .0500302
.0318592 .0317196
1.38 1.58
0.168 0.115
-.018512 -.0121391
.1063739 .1121995
-.1527311 .0191634
.1385702 .1357525
-1.10 0.14
0.270 0.888
-.4243237 -.2469067
.1188615 .2852334
.2884992 -.0102 .0157672
.168699 .1688987 .0043746
1.71 -0.06 3.60
0.087 0.952 0.000
-.0421448 -.3412354 .0071932
.6191431 .3208353 .0243412
-.002423 .0338806
.0256763 .0255638
-0.09 1.33
0.925 0.185
-.0527476 -.0162235
.0479016 .0839847
.2248134 .3549135
.1116778 .1094069
2.01 3.24
0.044 0.001
.005929 .1404798
.4436978 .5693471
-.2639695 -.0222264 .0129258
.1359595 .1361204 .0035256
-1.94 -0.16 3.67
0.052 0.870 0.000
-.5304451 -.2890175 .0060157
.0025062 .2445646 .0198358
553
The output has two parts: a header and the standard Stata output table for the coefcients, standard errors, and condence intervals. The header contains summary statistics for each equation in the VAR and statistics used in selecting the lag order of the VAR. Although there are standard formulas for all the lag-order statistics, L tkepohl (2005) gives different versions of the three information criteria that u drop the constant term from the likelihood. To obtain the L tkepohl (2005) versions, we specied u the lutstats option. The formulas for the standard and L tkepohl versions of these statistics are u given in Methods and formulas of [TS] varsoc. The dfk option species that the small-sample divisor 1/(T m) be used in estimating instead of the maximum likelihood (ML) divisor 1/T , where m is the average number of parameters included in each of the K equations. All the lag-order statistics are computed using the ML estimator of . Thus, specifying dfk will not change the computed lag-order statistics, but it will change the estimated variancecovariance matrix. Also, when dfk is specied, a dfk-adjusted log likelihood is computed and saved in e(ll dfk).
The lag() option takes a numlist of lags. To specify a model that includes the rst and second lags, type
. var y1 y2 y3, lags(1/2)
not
. var y1 y2 y3, lags(2)
because the latter specication would t a model that included only the second lag.
Fitting models with some lags excluded

To t a model that has only a fourth lag, that is,
yt = v + A4 yt4 + ut
you would specify the lags(4) option. Doing so is equivalent to tting the more general model
yt = v + A1 yt1 + A2 yt2 + A3 yt3 + A4 yt4 + ut

with A1 , A2 , and A3 constrained to be 0. When you t a model with some lags excluded, var estimates the coefcients included in the specication (A4 here) and saves these estimates in e(b). To obtain the asymptotic standard errors for impulseresponse functions and other postestimation statistics, Stata needs the complete set of parameter estimates, including those that are constrained to be zero; var stores them in e(bf). Because you can specify models for which the full set of parameter estimates exceeds Statas limit on the size of matrices, the nobigf option species that var not compute and store e(bf). This means that the asymptotic standard errors of the postestimation functions cannot be obtained, although bootstrap standard errors are still available. Building e(bf) can be time consuming, so if you do not need this full matrix, and speed is an issue, use nobigf.
554
Fitting models with exogenous variables Example 2: VAR model with exogenous variables
We use the exog() option to include exogenous variables in a VAR.
. var dln_inc dln_consump if qtr<=tq(1978q4), dfk exog(dln_inv) Vector autoregression Sample: 1960q4 - 1978q4 No. of obs Log likelihood = 478.5663 AIC FPE = 9.64e-09 HQIC Det(Sigma_ml) = 6.93e-09 SBIC Equation Parms RMSE R-sq chi2 P>chi2 dln_inc dln_consump 6 6 .011917 .009197 0.0702 0.2794 5.059587 25.97262 0.4087 0.0001
= 73 = -12.78264 = -12.63259 = -12.40612
Coef. dln_inc dln_inc L1. L2. dln_consump L1. L2. dln_inv _cons dln_consump dln_inc L1. L2. dln_consump L1. L2. dln_inv _cons
Std. Err.
P>|z|
-.1343345 .0120331
.1391074 .1380346
-0.97 0.09
0.334 0.931
-.4069801 -.2585097
.1383111 .2825759
.3235342 .0754177 .0151546 .0145136
.1652769 .1648624 .0302319 .0043815
1.96 0.46 0.50 3.31
0.050 0.647 0.616 0.001
-.0004027 -.2477066 -.0440987 .0059259
.647471 .398542 .074408 .0231012
.2425719 .3487949
.1073561 .1065281
2.26 3.27
0.024 0.001
.0321578 .1400036
.452986 .5575862
-.3119629 -.0128502 .0503616 .0131013
.1275524 .1272325 .0233314 .0033814
-2.45 -0.10 2.16 3.87
0.014 0.920 0.031 0.000
-.5619611 -.2622213 .0046329 .0064738
-.0619648 .2365209 .0960904 .0197288
All the postestimation commands for analyzing VARs work when exogenous variables are included in a model, but the asymptotic standard errors for the h-step-ahead forecasts are not available.
Fitting models with constraints on the coefcients

var permits model specications that include constraints on the coefcient, though var does not allow for constraints on . See [TS] var intro and [TS] var svar for ways to constrain .
555
Example 3: VAR model with constraints

In the rst example, we t a full VAR(2) to a three-equation model. The coefcients in the equation for dln inv were jointly insignicant, as were the coefcients in the equation for dln inc; and many individual coefcients were not signicantly different from zero. In this example, we constrain the coefcient on L2.dln inc in the equation for dln inv and the coefcient on L2.dln consump in the equation for dln inc to be zero.
. constraint 1 [dln_inv]L2.dln_inc = 0 . constraint 2 [dln_inc]L2.dln_consump = 0 . var dln_inv dln_inc dln_consump if qtr<=tq(1978q4), lutstats dfk > constraints(1 2) Estimating VAR coefficients Iteration 1: tolerance = .00737681 Iteration 2: tolerance = 3.998e-06 Iteration 3: tolerance = 2.730e-09 Vector autoregression Sample: 1960q4 - 1978q4 Log likelihood = 606.2804 FPE = 1.77e-14 Det(Sigma_ml) = 1.05e-14 Equation Parms dln_inv dln_inc dln_consump ( 1) ( 2) 6 6 7
(lutstats)
RMSE .043895 .011143 .008981
R-sq 0.1280 0.1141 0.2512
No. of obs AIC HQIC SBIC chi2 P>chi2 0.0798 0.1268 0.0008
= 73 = -31.69254 = -31.46747 = -31.12777
9.842338 8.584446 22.86958
[dln_inv]L2.dln_inc = 0 [dln_inc]L2.dln_consump = 0 Coef. Std. Err. z P>|z| [95% Conf. Interval]
dln_inv dln_inv L1. L2. dln_inc L1. L2. dln_consump L1. L2. _cons
-.320713 -.1607084
.1247512 .124261
-2.57 -1.29
0.010 0.196
-.5652208 -.4042555
-.0762051 .0828386
.1195448 5.66e-19
.5295669 9.33e-18
0.23 0.06
0.821 0.952
-.9183873 -1.77e-17
1.157477 1.89e-17
1.009281 1.008079 -.0162102
.623501 .5713486 .016893
1.62 1.76 -0.96
0.106 0.078 0.337
-.2127586 -.1117438 -.0493199
2.231321 2.127902 .0168995
556

dln_inc dln_inv L1. L2. dln_inc L1. L2. dln_consump L1. L2. _cons dln_consump dln_inv L1. L2. dln_inc L1. L2. dln_consump L1. L2. _cons
.0435712 .0496788
.0309078 .0306455
1.41 1.62
0.159 0.105
-.017007 -.0103852
.1041495 .1097428
-.1555119 .0122353
.1315854 .1165811
-1.18 0.10
0.237 0.916
-.4134146 -.2162595
.1023908 .2407301
.29286 -1.53e-18 .015689
.1568345 1.89e-17 .003819
1.87 -0.08 4.11
0.062 0.935 0.000
-.01453 -3.85e-17 .0082039
.6002501 3.55e-17 .0231741
-.0026229 .0337245
.0253538 .0252113
-0.10 1.34
0.918 0.181
-.0523154 -.0156888
.0470696 .0831378
.2224798 .3469758
.1094349 .1006026
2.03 3.45
0.042 0.001
.0079912 .1497984
.4369683 .5441532
-.2600227 -.0146825 .0129149
.1321622 .1117618 .003376
-1.97 -0.13 3.83
0.049 0.895 0.000
-.519056 -.2337315 .0062981
-.0009895 .2043666 .0195317
None of the free parameter estimates changed by much. Whereas the coefcients in the equation dln inv are now signicant at the 10% level, the coefcients in the equation for dln inc remain jointly insignicant.
557
Saved results
var saves the following in e():
Scalars e(N) e(N gaps) e(k) e(k eq) e(k dv) e(df eq) e(df m) e(df r) e(ll) e(ll dfk) e(obs #) e(k #) e(df m#) e(df r#) e(r2 #) e(ll #) e(chi2 #) e(F #) e(rmse #) e(aic) e(hqic) e(sbic) e(fpe) e(mlag) e(tmin) e(tmax) e(detsig) e(detsig ml) e(rank) number of observations number of gaps in sample number of parameters number of equations in e(b) number of dependent variables average number of parameters in an equation model degrees of freedom residual degrees of freedom (small only) log likelihood dfk adjusted log likelihood (dfk only) number of observations on equation # number of parameters in equation # model degrees of freedom for equation # residual degrees of freedom for equation # (small only) R-squared for equation # log likelihood for equation # x2 for equation # F statistic for equation # (small only) root mean squared error for equation # Akaike information criterion HannanQuinn information criterion SchwarzBayesian information criterion nal prediction error highest lag in VAR rst time period in sample maximum time determinant of e(Sigma) determinant of ml rank of e(V)
558
Macros e(cmd) e(cmdline) e(depvar) e(endog) e(exog) e(exogvars) e(eqnames) e(lags) e(exlags) e(title) e(nocons) e(constraints) e(cnslist var) e(small) e(lutstats) e(timevar) e(tsfmt) e(dfk) e(properties) e(predict) e(marginsok) e(marginsnotok) Matrices e(b) e(Cns) e(Sigma) e(V) e(bf) e(exlagsm) e(G) Functions e(sample)
var command as typed names of dependent variables names of endogenous variables, if specied names of exogenous variables, and their lags, if specied names of exogenous variables, if specied names of equations lags in model lags of exogenous variables in model, if specied title in estimation output nocons, if noconstant is specied constraints, if specied list of specied constraints small, if specied lutstats, if specied time variable specied in tsset format for the current time variable dfk, if specied b V program used to implement predict predictions allowed by margins predictions disallowed by margins coefcient vector constraints matrix matrix variancecovariance matrix of the estimators constrained coefcient vector matrix mapping lags to exogenous variables Gamma matrix; see Methods and formulas marks estimation sample

var is implemented as an ado-le. When there are no constraints placed on the coefcients, the VAR(p) is a seemingly unrelated regression model with the same explanatory variables in each equation. As discussed in L tkepohl (2005) and u Greene (2008, 696), performing linear regression on each equation produces the maximum likelihood estimates of the coefcients. The estimated coefcients can then be used to calculate the residuals, which in turn are used to estimate the cross-equation error variancecovariance matrix . Per L tkepohl (2005), we write the VAR(p) with exogenous variables as u
yt = AYt1 + B0 xt + ut
where
(5)
yt is the K 1 vector of endogenous variables, A is a K Kp matrix of coefcients, B0 is a K M matrix of coefcients,
559
xt is the M 1 vector of exogenous variables, ut is the K 1 vector of white noise innovations, and yt . . Yt is the Kp 1 matrix given by Yt = . ytp+1 Although (5) is easier to read, the formulas are much easier to manipulate if it is instead written
as
Y = BZ + U
where
Y= (y1 , . . . , yT ) B= (A, B0 ) Z= Y0 . . . , YT 1 x1 . . . , xT
Y is K T B is K (Kp + M ) Z is (Kp + M ) T U is K T
U= (u1 , . . . , uT )
Intercept terms in the model are included in xt . If there are no exogenous variables and no intercept terms in the model, xt is empty. The coefcients are estimated by iterated seemingly unrelated regression. Because the estimation is actually performed by reg3, the methods are documented in [R] reg3. See [P] makecns for more on estimation with constraints. Let U be the matrix of residuals that are obtained via Y BZ, where B is the matrix of estimated coefcients. Then the estimator of is
1 T
UU
By default, the maximum likelihood divisor of T = T is used. When dfk is specied, a small-sample degrees-of-freedom adjustment is used; then, T = T m where m is the average number of parameters per equation in the functional form for yt over the K equations. small species that Wald tests after var be assumed to have F or t distributions instead of chi-squared or standard normal distributions. The standard errors from each equation are computed using the degrees of freedom for the equation. The gamma matrix saved in e(G) referred to in Saved results is the (Kp + 1) (Kp + 1) matrix given by
1 T
(1, Yt )(1, Yt )
t=1
The formulas for the lag-order selection criteria and the log likelihood are discussed in [TS] varsoc.
560
Acknowledgment
We thank Christopher F. Baum of Boston College for his helpful comments.
References
Greene, W. H. 2008. Econometric Analysis. 6th ed. Upper Saddle River, NJ: Prentice Hall. Hamilton, J. D. 1994. Time Series Analysis. Princeton: Princeton University Press. L tkepohl, H. 1993. Introduction to Multiple Time Series Analysis. 2nd ed. New York: Springer. u . 2005. New Introduction to Multiple Time Series Analysis. New York: Springer. Stock, J. H., and M. W. Watson. 2001. Vector autoregressions. Journal of Economic Perspectives 15: 101115. Watson, M. W. 1994. Vector autoregressions and cointegration. In Vol. IV of Handbook of Econometrics, ed. R. F. Engle and D. L. McFadden. Amsterdam: Elsevier.
Also see
[TS] var postestimation Postestimation tools for var [TS] tsset Declare data to be time-series data [TS] dfactor Dynamic-factor models [TS] mgarch Multivariate GARCH models [TS] sspace State-space models [TS] var svar Structural vector autoregressive models [TS] varbasic Fit a simple VAR and graph IRFs or FEVDs [TS] vec Vector error-correction models [U] 20 Estimation and postestimation commands [TS] var intro Introduction to vector autoregressive models
Title
var postestimation Postestimation tools for var
Description
The following postestimation commands are of special interest after var: Command fcast compute fcast graph irf vargranger varlmar varnorm varsoc varstable varwle Description obtain dynamic forecasts graph dynamic forecasts obtained from fcast compute create and analyze IRFs and FEVDs Granger causality tests LM test for autocorrelation in residuals test for normally distributed residuals lag-order selection criteria check stability condition of estimates Wald lag-exclusion statistics
For information about these commands, see the corresponding entries in this manual. The following standard postestimation commands are also available: Command estat estimates lincom Description
561
562
Syntax for predict

predict statistic
Main
type
newvar
if
in
, statistic equation(eqno | eqname)
xb stdp residuals
Menu
Statistics
>
Postestimation
>
Options for predict

Main
xb, the default, calculates the linear prediction for the specied equation. stdp calculates the standard error of the linear prediction for the specied equation. residuals calculates the residuals. equation(eqno | eqname) species the equation to which you are referring. equation() is lled in with one eqno or eqname for options xb, stdp, and residuals. For example, equation(#1) would mean that the calculation is to be made for the rst equation, equation(#2) would mean the second, and so on. You could also refer to the equation by its name; thus, equation(income) would refer to the equation named income and equation(hours), to the equation named hours. If you do not specify equation(), the results are the same as if you specied equation(#1). For more information on using predict after multiple-equation estimation commands, see [R] predict.
Remarks
Model selection and inference Forecasting
Model selection and inference

See the following sections for information on model selection and inference after var. [TS] [TS] [TS] [TS] [TS] [TS] [TS]
irf Create and analyze IRFs, dynamic-multiplier functions, and FEVDs vargranger Perform pairwise Granger causality tests after var or svar varlmar Perform LM test for residual autocorrelation after var or svar varnorm Test for normally distributed disturbances after var or svar varsoc Obtain lag-order selection statistics for VARs and VECMs varstable Check the stability condition of VAR or SVAR estimates varwle Obtain Wald lag-exclusion statistics after var or svar
563
Forecasting
Two types of forecasts are available after you t a VAR(p): a one-step-ahead forecast and a dynamic h-step-ahead forecast. The one-step-ahead forecast produces a prediction of the value of an endogenous variable in the current period by using the estimated coefcients, the past values of the endogenous variables, and any exogenous variables. If you include contemporaneous values of exogenous variables in your model, you must have observations on the exogenous variables that are contemporaneous with the period in which the prediction is being made to compute the prediction. In Stata terms, these one-stepahead predictions are just the standard linear predictions available after any estimation command. Thus predict, xb eq(eqno | eqname) produces one-step-ahead forecasts for the specied equation. predict, stdp eq(eqno | eqname) produces the standard error of the linear prediction for the specied equation. The standard error of the forecast includes an estimate of the variability due to innovations, whereas the standard error of the linear prediction does not. The dynamic h-step-ahead forecast begins by using the estimated coefcients, the lagged values of the endogenous variables, and any exogenous variables to predict one step ahead for each endogenous variable. Then the one-step-ahead forecast produces two-step-ahead forecasts for each endogenous variable. The process continues for h periods. Because each step uses the predictions of the previous steps, these forecasts are known as dynamic forecasts. See the following sections for information on obtaining forecasts after svar:
[TS] fcast compute Compute dynamic forecasts of dependent variables after var, svar, or vec [TS] fcast graph Graph forecasts of variables computed by fcast compute

All postestimation commands listed above are implemented as ado-les. Formulas for predict predict with the xb option provides the one-step-ahead forecast. If exogenous variables are specied, the forecast is conditional on the exogenous xt variables. Specifying the residuals option causes predict to calculate the errors of the one-step-ahead forecasts. Specifying the stdp option causes predict to calculate the standard errors of the one-step-ahead forecasts.
Also see
[TS] var Vector autoregressive models [U] 20 Estimation and postestimation commands
Title
var svar Structural vector autoregressive models
Syntax
Short-run constraints svar depvarlist if in , aconstraints(constraintsa ) aeq(matrixaeq )
acns(matrixacns ) bconstraints(constraintsb ) beq(matrixbeq ) bcns(matrixbcns ) short run options Long-run constraints svar depvarlist if in , lrconstraints(constraintslr ) lreq(matrixlreq )
lrcns(matrixlrcns )
long run options
564
565
short run options

Model
Description suppress constant term apply previously dened constraintsa to A dene and apply to A equality constraint matrix matrixaeq dene and apply to A cross-parameter constraint matrix matrixacns apply previously dened constraintsb to B dene and apply to B equality constraint matrix matrixbeq dene and apply to B cross-parameter constraint matrixbcns use lags numlist in the underlying VAR use exogenous variables varlist apply constraintsv to underlying VAR suppress SURE iteration log set maximum number of iterations for SURE; default is isiterate(1600) set convergence tolerance of SURE use one-step SURE make small-sample degrees-of-freedom adjustment report small-sample t and F statistics do not check for local identication do not compute parameter vector for coefcients implicitly set to zero set condence level; default is level(95) show constrained parameters in table display underlying var output report L tkepohl lag-order selection statistics u do not display constraints control column formats control the maximization process; seldom used display legend instead of statistics
noconstant aconstraints(constraintsa ) aeq(matrixaeq ) acns(matrixacns )
bconstraints(constraintsb ) beq(matrixbeq ) bcns(matrixbcns ) lags(numlist)

Model 2
exog(varlistexog ) varconstraints(constraintsv ) noislog isiterate(#) istolerance(#) noisure dfk small noidencheck nobigf
Reporting
level(#) full var lutstats nocnsreport display options

Maximization
aconstraints(constraintsa ), aeq(matrixaeq ), acns(matrixacns ), bconstraints(constraintsb ), beq(matrixbeq ), bcns(matrixbcns ): at least one of these options must be specied. coeflegend does not appear in the dialog box.
566
long run options

Model
Description suppress constant term apply previously dened constraintslr to C dene and apply to C equality constraint matrix matrixlreq dene and apply to C cross-parameter constraint matrix matrixlrcns use lags numlist in the underlying VAR use exogenous variables varlist apply constraintsv to underlying VAR suppress SURE iteration log set maximum number of iterations for SURE; default is isiterate(1600) set convergence tolerance of SURE use one-step SURE make small-sample degrees-of-freedom adjustment report small-sample t and F statistics do not check for local identication do not compute parameter vector for coefcients implicitly set to zero set condence level; default is level(95) show constrained parameters in table display underlying var output report L tkepohl lag-order selection statistics u do not display constraints control column formats control the maximization process; seldom used display legend instead of statistics
noconstant lrconstraints(constraintslr ) lreq(matrixlreq ) lrcns(matrixlrcns )
lags(numlist)
Model 2
exog(varlistexog ) varconstraints(constraintsv ) noislog isiterate(#) istolerance(#) noisure dfk small noidencheck nobigf
Reporting
level(#) full var lutstats nocnsreport display options

Maximization
lrconstraints(constraintslr ), lreq(matrixlreq ), lrcns(matrixlrcns ): at least one of these options must be specied. coeflegend does not appear in the dialog box. You must tsset your data before using svar; see [TS] tsset. depvarlist and varlistexog may contain time-series operators; see [U] 11.4.4 Time-series varlists. by, rolling, statsby, and xi are allowed; see [U] 11.1.10 Prex commands. See [U] 20 Estimation and postestimation commands for more capabilities of estimation commands.
567
Menu
Statistics
>
>
Structural vector autoregression (SVAR)
Description
svar ts a vector autoregressive model subject to short- or long-run constraints you place on the resulting impulseresponse functions (IRFs). Economic theory typically motivates the constraints, allowing a causal interpretation of the IRFs to be made. See [TS] var intro for a list of commands that are used in conjunction with svar.
Options

Model
noconstant; see [R] estimation options. aconstraints(constraintsa ), aeq(matrixaeq ), acns(matrixacns ) bconstraints(constraintsb ), beq(matrixbeq ), bcns(matrixbcns ) These options specify the short-run constraints in an SVAR. To specify a short-run SVAR model, you must specify at least one of these options. The rst list of options species constraints on the parameters of the A matrix; the second list species constraints on the parameters of the B matrix (see Short-run SVAR models). If at least one option is selected from the rst list and none are selected from the second list, svar sets B to the identity matrix. Similarly, if at least one option is selected from the second list and none are selected from the rst list, svar sets A to the identity matrix. None of these options may be specied with any of the options that dene long-run constraints. aconstraints(constraintsa ) species a numlist of previously dened Stata constraints that are to be applied to A during estimation. aeq(matrixaeq ) species a matrix that denes a set of equality constraints. This matrix must be square with dimension equal to the number of equations in the underlying VAR. The elements of this matrix must be missing or real numbers. A missing value in the (i, j ) element of this matrix species that the (i, j ) element of A is a free parameter. A real number in the (i, j ) element of this matrix constrains the (i, j ) element of A to this real number. For example,
A=
1 0 . 1.5
species that A[1, 1] = 1, A[1, 2] = 0, A[2, 2] = 1.5, and A[2, 1] is a free parameter.
568
acns(matrixacns ) species a matrix that denes a set of exclusion or cross-parameter equality constraints on A. This matrix must be square with dimension equal to the number of equations in the underlying VAR. Each element of this matrix must be missing, 0, or a positive integer. A missing value in the (i, j ) element of this matrix species that no constraint be placed on this element of A. A zero in the (i, j ) element of this matrix constrains the (i, j ) element of A to be zero. Any strictly positive integers must be in two or more elements of this matrix. A strictly positive integer in the (i, j ) element of this matrix constrains the (i, j ) element of A to be equal to all the other elements of A that correspond to elements in this matrix that contain the same integer. For example, consider the matrix
A=
. 1 1 0
Specifying acns(A) in a two-equation SVAR constrains A[2, 1] = A[1, 2] and A[2, 2] = 0 while leaving A[1, 1] free. bconstraints(constraintsb ) species a numlist of previously dened Stata constraints to be applied to B during estimation. beq(matrixbeq ) species a matrix that denes a set of equality constraints. This matrix must be square with dimension equal to the number of equations in the underlying VAR. The elements of this matrix must be either missing or real numbers. The syntax of implied constraints is analogous to the one described in aeq(), except that it applies to B rather than to A. bcns(matrixbcns ) species a matrix that denes a set of exclusion or cross-parameter equality constraints on B. This matrix must be square with dimension equal to the number of equations in the underlying VAR. Each element of this matrix must be missing, 0, or a positive integer. The format of the implied constraints is the same as the one described in the acns() option above. lrconstraints(constraintslr ), lreq(matrixlreq ), lrcns(matrixlrcns ) These options specify the long-run constraints in an SVAR. To specify a long-run SVAR model, you must specify at least one of these options. The list of options species constraints on the parameters of the long-run C matrix (see Long-run SVAR models for the denition of C). None of these options may be specied with any of the options that dene short-run constraints. lrconstraints(constraintslr ) species a numlist of previously dened Stata constraints to be applied to C during estimation. lreq(matrixlreq ) species a matrix that denes a set of equality constraints on the elements of C. This matrix must be square with dimension equal to the number of equations in the underlying VAR. The elements of this matrix must be either missing or real numbers. The syntax of implied constraints is analogous to the one described in option aeq(), except that it applies to C. lrcns(matrixlrcns ) species a matrix that denes a set of exclusion or cross-parameter equality constraints on C. This matrix must be square with dimension equal to the number of equations in the underlying VAR. Each element of this matrix must be missing, 0, or a positive integer. The syntax of the implied constraints is the same as the one described for the acns() option above. lags(numlist) species the lags to be included in the underlying VAR model. The default is lags(1 2). This option takes a numlist and not simply an integer for the maximum lag. For instance, lags(2) would include only the second lag in the model, whereas lags(1/2) would include both the rst and second lags in the model. See [U] 11.1.8 numlist and [U] 11.4.4 Time-series varlists for further discussion of numlists and lags.
569
Model 2
exog(varlistexog ) species a list of exogenous variables to be included in the underlying VAR. varconstraints(constraintsv ) species a list of constraints to be applied to coefcients in the underlying VAR. Because svar estimates multiple equations, the constraints must specify the equation name for all but the rst equation. noislog prevents svar from displaying the iteration log from the iterated seemingly unrelated regression algorithm. When the varconstraints() option is not specied, the VAR coefcients are estimated via OLS, a noniterative procedure. As a result, noislog may be specied only with varconstraints(). Similarly, noislog may not be combined with noisure. isiterate(#) sets the maximum number of iterations for the iterated, seemingly unrelated regression algorithm. The default limit is 1,600. When the varconstraints() option is not specied, the VAR coefcients are estimated via OLS, a noniterative procedure. As a result, isiterate() may be specied only with varconstraints(). Similarly, isiterate() may not be combined with noisure. istolerance(#) species the convergence tolerance of the iterated seemingly unrelated regression algorithm. The default tolerance is 1e-6. When the varconstraints() option is not specied, the VAR coefcients are estimated via OLS, a noniterative procedure. As a result, istolerance() may be specied only with varconstraints(). Similarly, istolerance() may not be combined with noisure. noisure species that the VAR coefcients be estimated via one-step seemingly unrelated regression when varconstraints() is specied. By default, svar estimates the coefcients in the VAR via iterated, seemingly unrelated regression when varconstraints() is specied. When the varconstraints() option is not specied, the VAR coefcient estimates are obtained via OLS, a noniterative procedure. As a result, noisure may be specied only with varconstraints(). dfk species that a small-sample degrees-of-freedom adjustment be used when estimating , the covariance matrix of the VAR disturbances. Specically, 1/(T m) is used instead of the largesample divisor 1/T , where m is the average number of parameters in the functional form for yt over the K equations. small causes svar to calculate and report small-sample t and F statistics instead of the large-sample normal and chi-squared statistics. noidencheck requests that the Amisano and Giannini (1997) check for local identication not be performed. This check is local to the starting values used. Because of this dependence on the starting values, you may wish to suppress this check by specifying the noidencheck option. However, be careful in specifying this option. Models that are not structurally identied can still converge, thereby producing meaningless results that only appear to have meaning. nobigf requests that svar not save the estimated parameter vector that incorporates coefcients that have been implicitly constrained to be zero, such as when some lags have been omitted from a model. e(bf) is used for computing asymptotic standard errors in the postestimation commands irf create and fcast compute. Therefore, specifying nobigf implies that the asymptotic standard errors will not be available from irf create and fcast compute. See Fitting models with some lags excluded in [TS] var.
Reporting
level(#); see [R] estimation options. full shows constrained parameters in table.
570
var species that the output from var also be displayed. By default, the underlying VAR is t quietly. lutstats species that the L tkepohl versions of the lag-order selection statistics be reported. See u Methods and formulas in [TS] varsoc for a discussion of these statistics. nocnsreport; see [R] estimation options. display options: cformat(% fmt), pformat(% fmt), and sformat(% fmt); see [R] estimation options.
Maximization
maximize options: difficult, technique(algorithm spec), iterate(#), no log, trace, gradient, showstep, hessian, showtolerance, tolerance(#), ltolerance(#), nrtolerance(#), nonrtolerance, and from(init specs); see [R] maximize. These options are seldom used. The following option is available with svar but is not shown in the dialog box: coeflegend; see [R] estimation options.
Remarks
Introduction Short-run SVAR models Long-run SVAR models
Introduction
This entry assumes that you have already read [TS] var intro and [TS] var; if not, please do. Here we illustrate how to t SVARs in Stata subject to short-run and long-run restrictions. For more detailed information on SVARs, see Amisano and Giannini (1997) and Hamilton (1994). For good introductions to VARs, see L tkepohl (2005), Hamilton (1994), and Stock and Watson (2001). u
Short-run SVAR models

A short-run SVAR model without exogenous variables can be written as
A(IK A1 L A2 L2 Ap Lp )yt = A
= Bet
where L is the lag operator, A, B, and A1 , . . . , Ap are K K matrices of parameters, t is a K 1 vector of innovations with t N (0, ) and E[ t s ] = 0K for all s = t, and et is a K 1 vector of orthogonalized disturbances; that is, et N (0, IK ) and E[et es ] = 0K for all s = t. These transformations of the innovations allow us to analyze the dynamics of the system in terms of a change to an element of et . In a short-run SVAR model, we obtain identication by placing restrictions on A and B, which are assumed to be nonsingular.
Example 1: Short-run just-identied SVAR model

Following Sims (1980), the Cholesky decomposition is one method of identifying the impulse response functions in a VAR; thus, this method corresponds to an SVAR. There are several sets of constraints on A and B that are easily manipulated back to the Cholesky decomposition, and the following example illustrates this point.
571
One way to impose the Cholesky restrictions is to assume an SVAR model of the form
A(IK A1 A2 L2 Ap Lp )yt = Bet

where A is a lower triangular matrix with ones on the diagonal and B is a diagonal matrix. Because the P matrix for this model is Psr = A1 B, its estimate, Psr , obtained by plugging in estimates of A and B, should equal the Cholesky decomposition of . To illustrate, we use the German macroeconomic data discussed in L tkepohl (2005) and used u in [TS] var. In this example, yt = (dln inv, dln inc, dln consump), where dln inv is the rst difference of the log of investment, dln inc is the rst difference of the log of income, and dln consump is the rst difference of the log of consumption. Because the rst difference of the natural log of a variable can be treated as an approximation of the percent change in that variable, we will refer to these variables as percent changes in inv, inc, and consump, respectively. We will impose the Cholesky restrictions on this system by applying constraint matrices . 0 1 0 0 and B = 0 . A = . 1 0 0 0 . . 1 equality constraints with the 0 0 .
With these structural restrictions, we assume that the percent change in inv is not contemporaneously affected by the percent changes in either inc or consump. We also assume that the percent change of inc is affected by contemporaneous changes in inv but not consump. Finally, we assume that percent changes in consump are affected by contemporaneous changes in both inv and inc. The following commands t an SVAR model with these constraints.
. use http://www.stata-press.com/data/r12/lutkepohl2 (Quarterly SA West German macro data, Bil DM, from Lutkepohl 1993 Table E.1) . matrix A = (1,0,0\.,1,0\.,.,1) . matrix B = (.,0,0\0,.,0\0,0,.)
572

. svar dln_inv dln_inc dln_consump if qtr<=tq(1978q4), aeq(A) beq(B) Estimating short-run parameters (output omitted ) Structural vector autoregression ( 1) [a_1_1]_cons = 1 ( 2) [a_1_2]_cons = 0 ( 3) [a_1_3]_cons = 0 ( 4) [a_2_2]_cons = 1 ( 5) [a_2_3]_cons = 0 ( 6) [a_3_3]_cons = 1 ( 7) [b_1_2]_cons = 0 ( 8) [b_1_3]_cons = 0 ( 9) [b_2_1]_cons = 0 (10) [b_2_3]_cons = 0 (11) [b_3_1]_cons = 0 (12) [b_3_2]_cons = 0 Sample: 1960q4 - 1978q4 Exactly identified model Coef. /a_1_1 /a_2_1 /a_3_1 /a_1_2 /a_2_2 /a_3_2 /a_1_3 /a_2_3 /a_3_3 /b_1_1 /b_2_1 /b_3_1 /b_1_2 /b_2_2 /b_3_2 /b_1_3 /b_2_3 /b_3_3 1 -.0336288 -.0435846 0 1 -.424774 0 0 1 .0438796 0 0 0 .0110449 0 0 0 .0072243 Std. Err. z
No. of obs Log likelihood P>|z|
= =
73 606.307
(constrained) .0294605 -1.14 .0194408 -2.24 (omitted) (constrained) .0765548 -5.55 (omitted) (omitted) (constrained) .0036315 (omitted) (omitted) (omitted) .0009141 (omitted) (omitted) (omitted) .0005979 12.08
0.254 0.025
-.0913702 -.0816879
.0241126 -.0054812
0.000
-.5748187
-.2747293
0.000
.036762
.0509972
12.08
0.000
.0092534
.0128365
12.08
0.000
.0060525
.0083962
The SVAR output has four parts: an iteration log, a display of the constraints imposed, a header with sample and SVAR log-likelihood information, and a table displaying the estimates of the parameters from the A and B matrices. From the output above, we can see that the equality constraint matrices supplied to svar imposed the intended constraints and that the SVAR header informs us that the model we t is just identied. The estimates of a 2 1, a 3 1, and a 3 2 are all negative. Because the off-diagonal elements of the A matrix contain the negative of the actual contemporaneous effects, the estimated effects are positive, as expected. The estimates A and B are stored in e(A) and e(B), respectively, allowing us to compute the estimated Cholesky decomposition.
. matrix Aest = e(A) . matrix Best = e(B) . matrix chol_est = inv(Aest)*Best

. matrix list chol_est chol_est[3,3] dln_inv dln_inv .04387957 dln_inc .00147562 dln_consump .00253928
573
dln_inc 0 .01104494 .0046916
dln_consump 0 0 .00722432
svar saves the estimated from the underlying var in e(Sigma). The output below illustrates the computation of the Cholesky decomposition of e(Sigma). It is the same as the output computed from the SVAR estimates.
. matrix sig_var = e(Sigma) . matrix chol_var = cholesky(sig_var) . matrix list chol_var chol_var[3,3] dln_inv dln_inv .04387957 dln_inc .00147562 dln_consump .00253928
dln_inc 0 .01104494 .0046916
dln_consump 0 0 .00722432
We might now wonder why we bother obtaining parameter estimates via nonlinear estimation if we can obtain them simply by a transform of the estimates produced by var. When the model is just identied, as in the previous example, the SVAR parameter estimates can be computed via a transform of the VAR estimates. However, when the model is overidentied, such is not the case.
Example 2: Short-run overidentied SVAR model

The Cholesky decomposition example above t a just-identied model. This example considers an overidentied model. In the previous example, the a 2 1 parameter was not signicant, which is consistent with a theory in which changes in our measure of investment affect only changes in income with a lag. We can impose the restriction that a 2 1 is zero and then test this overidentifying restriction. Our A and B matrices are now 1 0 0 . 0 0 A = 0 1 0 and B = 0 . 0 . . 1 0 0 . The output below contains the commands and results we obtained by tting this model on the L tkepohl data. u
. matrix B = (.,0,0\0,.,0\0,0,.) . matrix A = (1,0,0\0,1,0\.,.,1)
574

. svar dln_inv dln_inc dln_consump if qtr<=tq(1978q4), aeq(A) beq(B) Estimating short-run parameters (output omitted ) Structural vector autoregression ( 1) [a_1_1]_cons = 1 ( 2) [a_1_2]_cons = 0 ( 3) [a_1_3]_cons = 0 ( 4) [a_2_1]_cons = 0 ( 5) [a_2_2]_cons = 1 ( 6) [a_2_3]_cons = 0 ( 7) [a_3_3]_cons = 1 ( 8) [b_1_2]_cons = 0 ( 9) [b_1_3]_cons = 0 (10) [b_2_1]_cons = 0 (11) [b_2_3]_cons = 0 (12) [b_3_1]_cons = 0 (13) [b_3_2]_cons = 0 Sample: 1960q4 - 1978q4 Overidentified model Coef. /a_1_1 /a_2_1 /a_3_1 /a_1_2 /a_2_2 /a_3_2 /a_1_3 /a_2_3 /a_3_3 /b_1_1 /b_2_1 /b_3_1 /b_1_2 /b_2_2 /b_3_2 /b_1_3 /b_2_3 /b_3_3 1 0 -.0435911 0 1 -.4247741 0 0 1 .0438796 0 0 0 .0111431 0 0 0 .0072243 Std. Err. z
No. of obs Log likelihood P>|z|
= =
73 605.6613
(constrained) (omitted) .0192696 -2.26 (omitted) (constrained) .0758806 -5.60 (omitted) (omitted) (constrained) .0036315 (omitted) (omitted) (omitted) .0009222 (omitted) (omitted) (omitted) .0005979 12.08
0.024
-.0813589
-.0058233
0.000
-.5734973
-.2760508
0.000
.036762
.0509972
12.08
0.000
.0093356
.0129506
12.08 chi2( 1)=
0.000 1.292
.0060525
.0083962
LR test of identifying restrictions:
Prob > chi2 = 0.256
The footer in this example reports a test of the overidentifying restriction. The null hypothesis of this test is that any overidentifying restrictions are valid. In the case at hand, we cannot reject this null hypothesis at any of the conventional levels.
Example 3: Short-run SVAR model with constraints

svar also allows us to place constraints on the parameters of the underlying VAR. We begin by looking at the underlying VAR for the SVARs that we have used in the previous examples.

. var dln_inv dln_inc dln_consump if qtr<=tq(1978q4) Vector autoregression Sample: 1960q4 - 1978q4 No. of obs Log likelihood = 606.307 AIC FPE = 2.18e-11 HQIC Det(Sigma_ml) = 1.23e-11 SBIC Equation Parms RMSE R-sq chi2 P>chi2 dln_inv dln_inc dln_consump 7 7 7 .046148 .011719 .009445 0.1286 0.1142 0.2513 10.76961 9.410683 24.50031 0.0958 0.1518 0.0004
575
= 73 = -16.03581 = -15.77323 = -15.37691
Std. Err.
P>|z|
-.3196318 -.1605508
.1192898 .118767
-2.68 -1.35
0.007 0.176
-.5534355 -.39333
-.0858282 .0722283
.1459851 .1146009
.5188451 .508295
0.28 0.23
0.778 0.822
-.8709326 -.881639
1.162903 1.110841
.9612288 .9344001 -.0167221
.6316557 .6324034 .0163796
1.52 1.48 -1.02
0.128 0.140 0.307
-.2767936 -.3050877 -.0488257
2.199251 2.173888 .0153814
.0439309 .0500302
.0302933 .0301605
1.45 1.66
0.147 0.097
-.0154427 -.0090833
.1033046 .1091437
-.1527311 .0191634
.131759 .1290799
-1.16 0.15
0.246 0.882
-.4109741 -.2338285
.1055118 .2721552
.2884992 -.0102 .0157672
.1604069 .1605968 .0041596
1.80 -0.06 3.79
0.072 0.949 0.000
-.0258926 -.3249639 .0076146
.6028909 .3045639 .0239198
-.002423 .0338806
.0244142 .0243072
-0.10 1.39
0.921 0.163
-.050274 -.0137607
.045428 .0815219
.2248134 .3549135
.1061884 .1040292
2.12 3.41
0.034 0.001
.0166879 .1510199
.4329389 .558807
-.2639695 -.0222264 .0129258
.1292766 .1294296 .0033523
-2.04 -0.17 3.86
0.041 0.864 0.000
-.517347 -.2759039 .0063554
-.010592 .231451 .0194962
576
The equation-level model tests reported in the header indicate that we cannot reject the null hypotheses that all the coefcients in the rst equation are zero, nor can we reject the null that all the coefcients in the second equation are zero at the 5% signicance level. We use a combination of theory and the p-values from the output above to place some exclusion restrictions on the underlying VAR(2). Specically, in the equation for the percent change of inv, we constrain the coefcients on L2.dln inv, L.dln inc, L2.dln inc, and L2.dln consump to be zero. In the equation for dln inc, we constrain the coefcients on L2.dln inv, L2.dln inc, and L2.dln consump to be zero. Finally, in the equation for dln consump, we constrain L.dln inv and L2.dln consump to be zero. We then ret the SVAR from the previous example.
. . . . constraint constraint constraint constraint 1 2 3 4 [dln_inv]L2.dln_inv = 0 [dln_inv]L.dln_inc = 0 [dln_inv]L2.dln_inc = 0 [dln_inv]L2.dln_consump = 0
. constraint 5 [dln_inc]L2.dln_inv = 0 . constraint 6 [dln_inc]L2.dln_inc = 0 . constraint 7 [dln_inc]L2.dln_consump = 0 . constraint 8 [dln_consump]L.dln_inv = 0 . constraint 9 [dln_consump]L2.dln_consump = 0 . svar dln_inv dln_inc dln_consump if qtr<=tq(1978q4), aeq(A) beq(B) varconst(1/9) > noislog Estimating short-run parameters (output omitted ) Structural vector autoregression ( 1) ( 2) ( 3) ( 4) ( 5) ( 6) ( 7) ( 8) ( 9) (10) (11) (12) (13) [a_1_1]_cons [a_1_2]_cons [a_1_3]_cons [a_2_1]_cons [a_2_2]_cons [a_2_3]_cons [a_3_3]_cons [b_1_2]_cons [b_1_3]_cons [b_2_1]_cons [b_2_3]_cons [b_3_1]_cons [b_3_2]_cons = = = = = = = = = = = = = 1 0 0 0 1 0 1 0 0 0 0 0 0

Sample: 1960q4 - 1978q4 Overidentified model Coef. /a_1_1 /a_2_1 /a_3_1 /a_1_2 /a_2_2 /a_3_2 /a_1_3 /a_2_3 /a_3_3 /b_1_1 /b_2_1 /b_3_1 /b_1_2 /b_2_2 /b_3_2 /b_1_3 /b_2_3 /b_3_3 1 0 -.0418708 0 1 -.4255808 0 0 1 .0451851 0 0 0 .0113723 0 0 0 .0072417 Std. Err. z No. of obs Log likelihood P>|z| = = 73 601.8591
577
(constrained) (omitted) .0187579 -2.23 (omitted) (constrained) .0745298 -5.71 (omitted) (omitted) (constrained) .0037395 (omitted) (omitted) (omitted) .0009412 (omitted) (omitted) (omitted) .0005993 12.08
0.026
-.0786356
-.0051061
0.000
-.5716565
-.2795051
0.000
.0378557
.0525145
12.08
0.000
.0095276
.013217
12.08 chi2( 1)=
0.000 .8448
.006067
.0084164
Prob > chi2 = 0.358
If we displayed the underlying VAR(2) results by using the var option, we would see that most of the unconstrained coefcients are now signicant at the 10% level and that none of the equation-level model statistics fail to reject the null hypothesis at the 10% level. The svar output reveals that the p-value of the overidentication test rose and that the coefcient on a 3 1 is still insignicant at the 1% level but not at the 5% level.
Before moving on to models with long-run constraints, consider these limitations. We cannot place constraints on the elements of A in terms of the elements of B, or vice versa. This limitation is imposed by the form of the check for identication derived by Amisano and Giannini (1997). As noted in Methods and formulas, this test requires separate constraint matrices for the parameters in A and B. Another limitation is that we cannot mix short-run and long-run constraints.
Long-run SVAR models

As discussed in [TS] var intro, a long-run SVAR has the form
yt = Cet
In long-run models, the constraints are placed on the elements of C, and the free parameters are estimated. These constraints are often exclusion restrictions. For instance, constraining C[1, 2] to be zero can be interpreted as setting the long-run response of variable 1 to the structural shocks driving variable 2 to be zero. Similar to the short-run model, the Plr matrix such that Plr Plr = identies the structural impulseresponse functions. Plr = C is identied by the restrictions placed on the parameters in C. There are K 2 parameters in C, and the order condition for identication requires that there be at least K 2 K(K + 1)/2 restrictions placed on those parameters. As in the short-run model, this order condition is necessary but not sufcient, so the Amisano and Giannini (1997) check for local identication is performed by default.
578
Example 4: Long-run SVAR model

Suppose that we have a theory in which unexpected changes to the money supply have no long-run effects on changes in output and, similarly, that unexpected changes in output have no long-run effects on changes in the money supply. The C matrix implied by this theory is
C=
. 0 0 .
. use http://www.stata-press.com/data/r12/m1gdp . matrix lr = (.,0\0,.) . svar d.ln_m1 d.ln_gdp, lreq(lr) Estimating long-run parameters (output omitted ) Structural vector autoregression ( 1) [c_1_2]_cons = 0 ( 2) [c_2_1]_cons = 0 Sample: 1959q4 - 2002q2 Overidentified model Coef. /c_1_1 /c_2_1 /c_1_2 /c_2_2 .0301007 0 0 .0129691 Std. Err. .0016277 (omitted) (omitted) .0007013 z 18.49
No. of obs Log likelihood P>|z| 0.000
= =
171 1151.614
[95% Conf. Interval] .0269106 .0332909
18.49 chi2( 1)=
0.000 .1368
.0115946
.0143436
Prob > chi2 = 0.712
We have assumed that the underlying VAR has 2 lags; four of the ve selection-order criteria computed by varsoc (see [TS] varsoc) recommended this choice. The test of the overidentifying restrictions provides no indication that it is not valid.
579
Saved results
svar saves the following in e():
Scalars e(N) e(N cns) e(k eq) e(k dv) e(k aux) e(ll) e(ll #) e(N gaps var) e(k var) e(k eq var) e(k dv var) e(df eq var) e(df m var) e(df r var) e(obs # var) e(k # var) e(df m# var) e(df r# var) e(r2 # var) e(ll # var) e(chi2 # var) e(F # var) e(rmse # var) e(mlag var) e(tparms var) e(aic var) e(hqic var) e(sbic var) e(fpe var) e(ll var) e(detsig var) e(detsig ml var) e(tmin) e(tmax) e(chi2 oid) e(oid df) e(rank) e(ic ml) e(rc ml) number of observations number of constraints number of equations in e(b) number of dependent variables number of auxiliary parameters log likelihood from svar log likelihood for equation # number of gaps in the sample number of coefcients in VAR number of equations in underlying VAR number of dependent variables in underlying VAR average number of parameters in an equation model degrees of freedom if small, residual degrees of freedom number of observations on equation # number of coefcients in equation # model degrees of freedom for equation # residual degrees of freedom for equation # (small only) R-squared for equation # log likelihood for equation # VAR 2 statistic for equation # F statistic for equation # (small only) root mean squared error for equation # highest lag in VAR number of parameters in all equations Akaike information criterion HannanQuinn information criterion SchwarzBayesian information criterion nal prediction error log likelihood from var determinant of e(Sigma) determinant of ml rst time period in the sample maximum time overidentication test number of overidentifying restrictions rank of e(V) number of iterations return code from ml
580
Macros e(cmd) e(cmdline) e(lrmodel) e(lags var) e(depvar var) e(endog var) e(exog var) e(nocons var) e(cns lr) e(cns a) e(cns b) e(dfk var) e(eqnames var) e(lutstats var) e(constraints var) e(small) e(tsfmt) e(timevar) e(title) e(properties) e(predict) Matrices e(b) e(Cns) e(Sigma) e(V) e(b var) e(V var) e(bf var) e(G var) e(aeq) e(acns) e(beq) e(bcns) e(lreq) e(lrcns) e(Cns var) e(A) e(B) e(C) e(A1) Functions e(sample)
svar command as typed long-run model, if specied lags in model names of dependent variables names of endogenous variables names of exogenous variables, if specied noconstant, if noconstant specied long-run constraints cross-parameter equality constraints on A cross-parameter equality constraints on B alternate divisor (dfk), if specied names of equations lutstats, if specied constraints var, if there are constraints on VAR small, if specied format of timevar name of timevar title in estimation output b V program used to implement predict coefcient vector constraints matrix matrix variancecovariance matrix of the estimators coefcient vector of underlying VAR model VCE of underlying VAR model full coefcient vector with zeros in dropped lags G matrix saved by var; see [TS] var Methods and formulas aeq(matrix), if specied acns(matrix), if specied beq(matrix), if specied bcns(matrix), if specied lreq(matrix), if specied lrcns(matrix), if specied constraint matrix from var, if varconstraints() is specied estimated A matrix, if a short-run model estimated B matrix estimated C matrix, if a long-run model estimated A matrix, if a long-run model marks estimation sample

svar is implemented as an ado-le. The log-likelihood function for models with short-run constraints is
L(A, B) =
where W = B1 A.
NK N N ln(2) + ln(|W|2 ) tr(W W) 2 2 2
581
When there are long-run constraints, because C = A1 B and A = IK , W = B1 = C1 A1 = 1 . Substituting the last term for W in the short-run log likelihood produces the long-run log (AC) likelihood NK N N L(C) = ln(2) + ln(|W|2 ) tr(W W) 2 2 2 where W = (AC)1 .
For both the short-run and the long-run models, the maximization is performed by the scoring method. See Harvey (1990) for a discussion of this method. Based on results from Amisano and Giannini (1997), the score vector for the short-run model is
L(A, B) = N {vec(W 1 )} {vec(W)} ( IK ) [vec(A), vec(B)] (IK B1 ), (A B 1 B1 )

and the expected information matrix is
I [vec(A), vec(B)] = N
(W1 B 1 ) (IK 2 + ) (W 1 B1 ), (IK B1 ) (IK B 1 )
where is the commutation matrix dened in Magnus and Neudecker (1999, 4648). Using results from Amisano and Giannini (1997), we can derive the score vector and the expected information matrix for the case with long-run restrictions. The score vector is
L(C) = N {vec(W 1 )} {vec(W)} ( IK ) vec(C)

and the expected information matrix is
(A 1 C 1 C1 )
I [vec(C)] = N (IK C 1 )(IK 2 + )(IK C 1 )
Checking for identication This section describes the methods used to check for identication of models with short-run or long-run constraints. Both methods depend on the starting values. By default, svar uses starting values constructed by taking a vector of appropriate dimension and applying the constraints. If there are m parameters in the model, the j th element of the 1 m vector is 1 + m/100. svar also allows the user to provide starting values. For the short-run case, the model is identied if the matrix
NK NK NK NK Vsr = Ra (W B) 0K 2 0K 2 Ra (IK B)
has full column rank of 2K 2 , where NK = (1/2)(IK 2 + ), Ra is the constraint matrix for the parameters in A (that is, Ra vec(A) = ra ), and Rb is the constraint matrix for the parameters in B (that is, Rb vec(B) = rb ).
582
For the long-run case, based on results from the C model in Amisano and Giannini (1997), the model is identied if the matrix
Vlr =
(I C 1 )(2NK )(I C1 ) Rc
has full column rank of K 2 , where Rc is the constraint matrix for the parameters in C; that is, Rc vec(C) = rc . The test of the overidentifying restrictions is computed as
LR = 2(LLvar LLsvar )
where LR is the value of the test statistic against the null hypothesis that the overidentifying restrictions are valid, LLvar is the log likelihood from the underlying VAR(p) model, and LLsvar is the log likelihood from the SVAR model. The test statistic is asymptotically distributed as 2 (q), where q is the number of overidentifying restrictions. Amisano and Giannini (1997, 3839) emphasize that, because this test of the validity of the overidentifying restrictions is an omnibus test, it can be interpreted as a test of the null hypothesis that all the restrictions are valid. Because constraints might not be independent either by construction or because of the data, the number of restrictions is not necessarily equal to the number of constraints. The rank of e(V) gives the number of parameters that were independently estimated after applying the constraints. The maximum number of parameters that can be estimated in an identied short-run or long-run SVAR is K(K + 1)/2. This implies that the number of overidentifying restrictions, q , is equal to K(K + 1)/2 minus the rank of e(V). The number of overidentifying restrictions is also linked to the order condition for each model. In a short-run SVAR model, there are 2K 2 parameters. Because no more than K(K + 1)/2 parameters may be estimated, the order condition for a short-run SVAR model is that at least 2K 2 K(K + 1)/2 restrictions be placed on the model. Similarly, there are K 2 parameters in long-run SVAR model. Because no more than K(K + 1)/2 parameters may be estimated, the order condition for a long-run SVAR model is that at least K 2 K(K + 1)/2 restrictions be placed on the model.
Acknowledgment
We thank Gianni Amisano, Universit` degli Studi di Brescia, for his helpful comments. a
References
Amisano, G., and C. Giannini. 1997. Topics in Structural VAR Econometrics. 2nd ed. Heidelberg: Springer. Christiano, L. J., M. Eichenbaum, and C. L. Evans. 1999. Monetary policy shocks: What have we learned and to what end? In Handbook of Macroeconomics: Volume 1A, ed. J. B. Taylor and M. Woodford. New York: Elsevier. Hamilton, J. D. 1994. Time Series Analysis. Princeton: Princeton University Press. Harvey, A. C. 1990. The Econometric Analysis of Time Series. 2nd ed. Cambridge, MA: MIT Press. L tkepohl, H. 1993. Introduction to Multiple Time Series Analysis. 2nd ed. New York: Springer. u . 2005. New Introduction to Multiple Time Series Analysis. New York: Springer. Magnus, J. R., and H. Neudecker. 1999. Matrix Differential Calculus with Applications in Statistics and Econometrics. Rev. ed. New York: Wiley. Rothenberg, T. J. 1971. Identication in parametric models. Econometrica 39: 577591.

Sims, C. A. 1980. Macroeconomics and reality. Econometrica 48: 148. Stock, J. H., and M. W. Watson. 2001. Vector autoregressions. Journal of Economic Perspectives 15: 101115.
583
Watson, M. W. 1994. Vector autoregressions and cointegration. In Vol. IV of Handbook of Econometrics, ed. R. F. Engle and D. L. McFadden. Amsterdam: Elsevier.
Also see
[TS] var svar postestimation Postestimation tools for svar [TS] tsset Declare data to be time-series data [TS] var Vector autoregressive models [TS] varbasic Fit a simple VAR and graph IRFs or FEVDs [TS] vec Vector error-correction models [U] 20 Estimation and postestimation commands [TS] var intro Introduction to vector autoregressive models
Title
var svar postestimation Postestimation tools for svar
Description
The following postestimation commands are of special interest after svar: Command fcast compute fcast graph irf vargranger varlmar varnorm varsoc varstable varwle Description obtain dynamic forecasts graph dynamic forecasts obtained from fcast compute create and analyze IRFs and FEVDs Granger causality tests LM test for autocorrelation in residuals test for normally distributed residuals lag-order selection criteria check stability condition of estimates Wald lag-exclusion statistics
For information about these commands, see the corresponding entries in this manual. The following standard postestimation commands are also available: Command estat estimates lincom lrtest nlcom predict predictnl test testnl Description
cataloging estimation results point estimates, standard errors, testing, and inference for linear combinations of coefcients likelihood-ratio test point estimates, standard errors, testing, and inference for nonlinear combinations of coefcients predictions, residuals, inuence statistics, and other diagnostic measures point estimates, standard errors, testing, and inference for generalized predictions Wald tests of simple and composite linear hypotheses Wald tests of nonlinear hypotheses
Syntax for predict

predict statistic
Main
type
newvar
if
in
xb stdp residuals
These statistics are available both in and out of sample; type predict the estimation sample. 584
585
Menu
Statistics
>
Postestimation
>
Options for predict

Main
xb, the default, calculates the linear prediction for the specied equation. stdp calculates the standard error of the linear prediction for the specied equation. residuals calculates the residuals. equation(eqno | eqname) species the equation to which you are referring. equation() is lled in with one eqno or eqname for options xb, stdp, and residuals. For example, equation(#1) would mean that the calculation is to be made for the rst equation, equation(#2) would mean the second, and so on. You could also refer to the equation by its name; thus, equation(income) would refer to the equation named income and equation(hours), to the equation named hours. If you do not specify equation(), the results are the same as if you specied equation(#1). For more information on using predict after multiple-equation estimation commands, see [R] predict.
Remarks

See the following sections for information on model selection and inference after var. [TS] [TS] [TS] [TS] [TS] [TS] [TS]
irf Create and analyze IRFs, dynamic-multiplier functions, and FEVDs vargranger Perform pairwise Granger causality tests after var or svar varlmar Perform LM test for residual autocorrelation after var or svar varnorm Test for normally distributed disturbances after var or svar varsoc Obtain lag-order selection statistics for VARs and VECMs varstable Check the stability condition of VAR or SVAR estimates varwle Obtain Wald lag-exclusion statistics after var or svar
Forecasting
See the following sections for information on obtaining forecasts after svar: [TS] fcast compute Compute dynamic forecasts of dependent variables after var, svar, or vec [TS] fcast graph Graph forecasts of variables computed by fcast compute
586

Also see
[TS] var svar Structural vector autoregressive models [U] 20 Estimation and postestimation commands
Title
varbasic Fit a simple VAR and graph IRFs or FEVDs
Syntax
varbasic depvarlist options
Main
if
in
, options
Description use lags numlist in the model; default is lags(1 2) produce matrix graph of IRFs produce matrix graph of FEVDs do not produce a graph set forecast horizon # for estimating the OIRFs, IRFs, and FEVDs; default is step(8)
lags(numlist) irf fevd nograph step(#)
You must tsset your data before using varbasic; see [TS] tsset. depvarlist may contain time-series operators; see [U] 11.4.4 Time-series varlists. rolling, statsby, and xi are allowed; see [U] 11.1.10 Prex commands. See [U] 20 Estimation and postestimation commands for more capabilities of estimation commands.
Menu
Statistics
>
>
Basic VAR
Description
varbasic ts a basic vector autoregressive (VAR) model and graphs the impulseresponse functions (IRFs), the orthogonalized impulseresponse functions (OIRFs), or the forecast-error variance decompositions (FEVDs).
Options

Main
lags(numlist) species the lags to be included in the model. The default is lags(1 2). This option takes a numlist and not simply an integer for the maximum lag. For instance, lags(2) would include only the second lag in the model, whereas lags(1/2) would include both the rst and second lags in the model. See [U] 11.1.8 numlist and [U] 11.4.4 Time-series varlists for more discussion of numlists and lags. irf causes varbasic to produce a matrix graph of the IRFs instead of a matrix graph of the OIRFs, which is produced by default. fevd causes varbasic to produce a matrix graph of the FEVDs instead of a matrix graph of the OIRFs, which is produced by default. nograph species that no graph be produced. The IRFs, OIRFs, and FEVDs are still estimated and saved in the IRF le varbasic.irf.
587
588
step(#) species the forecast horizon for estimating the IRFs, OIRFs, and FEVDs. The default is eight periods.
Remarks
varbasic simplies tting simple VARs and graphing the IRFs, the OIRFs, or the FEVDs. See [TS] var and [TS] var svar for tting more advanced VAR models and structural vector autoregressive (SVAR) models. All the postestimation commands discussed in [TS] var postestimation work after varbasic. This entry does not discuss the methods for tting a VAR or the methods surrounding the IRFs, OIRFs, and FEVDs. See [TS] var and [TS] irf create for more on these methods. This entry illustrates how to use varbasic to easily obtain results. It also illustrates how varbasic serves as an entry point to further analysis.
Example 1
We t a three-variable VAR with two lags to the German macro data used by L tkepohl (2005). u The three variables are the rst difference of natural log of investment, dln inv; the rst difference of the natural log of income, dln inc; and the rst difference of the natural log of consumption, dln consump. In addition to tting the VAR, we want to see the OIRFs. Below we use varbasic to t a VAR(2) model on the data from the second quarter of 1961 through the fourth quarter of 1978. By default, varbasic produces graphs of the OIRFs.
. use http://www.stata-press.com/data/r12/lutkepohl2 (Quarterly SA West German macro data, Bil DM, from Lutkepohl 1993 Table E.1) . varbasic dln_inv dln_inc dln_consump if qtr<=tq(1978q4) Vector autoregression Sample: 1960q4 - 1978q4 No. of obs = 73 Log likelihood = 606.307 AIC = -16.03581 FPE = 2.18e-11 HQIC = -15.77323 Det(Sigma_ml) = 1.23e-11 SBIC = -15.37691 Equation Parms RMSE R-sq chi2 P>chi2 dln_inv dln_inc dln_consump 7 7 7 .046148 .011719 .009445 0.1286 0.1142 0.2513 10.76961 9.410683 24.50031 0.0958 0.1518 0.0004
589
Std. Err.
P>|z|
-.3196318 -.1605508
.1192898 .118767
-2.68 -1.35
0.007 0.176
-.5534355 -.39333
-.0858282 .0722283
.1459851 .1146009
.5188451 .508295
0.28 0.23
0.778 0.822
-.8709326 -.881639
1.162903 1.110841
.9612288 .9344001 -.0167221
.6316557 .6324034 .0163796
1.52 1.48 -1.02
0.128 0.140 0.307
-.2767936 -.3050877 -.0488257
2.199251 2.173888 .0153814
.0439309 .0500302
.0302933 .0301605
1.45 1.66
0.147 0.097
-.0154427 -.0090833
.1033046 .1091437
-.1527311 .0191634
.131759 .1290799
-1.16 0.15
0.246 0.882
-.4109741 -.2338285
.1055118 .2721552
.2884992 -.0102 .0157672
.1604069 .1605968 .0041596
1.80 -0.06 3.79
0.072 0.949 0.000
-.0258926 -.3249639 .0076146
.6028909 .3045639 .0239198
-.002423 .0338806
.0244142 .0243072
-0.10 1.39
0.921 0.163
-.050274 -.0137607
.045428 .0815219
.2248134 .3549135
.1061884 .1040292
2.12 3.41
0.034 0.001
.0166879 .1510199
.4329389 .558807
-.2639695 -.0222264 .0129258
.1292766 .1294296 .0033523
-2.04 -0.17 3.86
0.041 0.864 0.000
-.517347 -.2759039 .0063554
-.010592 .231451 .0194962
590
varbasic, dln_consump, dln_consump

.06 .04 .02 0 .02
varbasic, dln_consump, dln_inc
varbasic, dln_consump, dln_inv
varbasic, dln_inc, dln_consump

.06 .04 .02 0 .02
varbasic, dln_inc, dln_inc
varbasic, dln_inc, dln_inv
varbasic, dln_inv, dln_consump

.06 .04 .02 0 .02 0 2 4 6 8 0
varbasic, dln_inv, dln_inc
varbasic, dln_inv, dln_inv
Because we are also interested in looking at the FEVDs, we can use irf graph to obtain the graphs. Although the details are available in [TS] irf and [TS] irf graph, the command below produces what we want after the call to varbasic.
. irf graph fevd, lstep(1)
varbasic, dln_consump, dln_consump
1
varbasic, dln_consump, dln_inc
varbasic, dln_consump, dln_inv
.5
varbasic, dln_inc, dln_consump

1
varbasic, dln_inc, dln_inc
varbasic, dln_inc, dln_inv
.5
varbasic, dln_inv, dln_consump

1
varbasic, dln_inv, dln_inc
varbasic, dln_inv, dln_inv
.5
0 0 2 4 6 8 0 2 4 6 8 0 2 4 6 8
step 95% CI fraction of mse due to impulse
Technical note
Stata stores the estimated IRFs, OIRFs, and FEVDs in a IRF le called varbasic.irf in the current working directory. varbasic replaces any varbasic.irf that already exists. Finally, varbasic makes varbasic.irf the active IRF le. This means that the graph and table commands irf graph,
591
irf cgraph, irf ograph, irf table, and irf ctable will all display results that correspond to the VAR t by varbasic.
Saved results
See Saved results in [TS] var.

varbasic is implemented as an ado-le. varbasic uses var and irf graph to obtain its results. See [TS] var and [TS] irf graph for a discussion of how those commands obtain their results.
References
L tkepohl, H. 1993. Introduction to Multiple Time Series Analysis. 2nd ed. New York: Springer. u . 2005. New Introduction to Multiple Time Series Analysis. New York: Springer.
Also see
[TS] varbasic postestimation Postestimation tools for varbasic [TS] tsset Declare data to be time-series data [TS] var Vector autoregressive models [TS] var svar Structural vector autoregressive models [U] 20 Estimation and postestimation commands [TS] var intro Introduction to vector autoregressive models
Title
varbasic postestimation Postestimation tools for varbasic
Description
The following postestimation commands are of special interest after varbasic: Command fcast compute fcast graph irf vargranger varlmar varnorm varsoc varstable varwle Description obtain dynamic forecasts graph dynamic forecasts obtained from fcast compute create and analyze IRFs and FEVDs Granger causality tests LM test for autocorrelation in residuals test for normally distributed residuals lag-order selection criteria check stability condition of estimates Wald lag-exclusion statistics
592
593
Syntax for predict

predict statistic
Main
type
newvar
if
in
xb stdp residuals
Menu
Statistics
>
Postestimation
>
Options for predict

Main
xb, the default, calculates the linear prediction for the specied equation. stdp calculates the standard error of the linear prediction for the specied equation. residuals calculates the residuals. equation(eqno | eqname) species the equation to which you are referring. equation() is lled in with one eqno or eqname for the xb, stdp, and residuals options. For example, equation(#1) would mean that the calculation is to be made for the rst equation, equation(#2) would mean the second, and so on. You could also refer to the equation by its name; thus, equation(income) would refer to the equation named income and equation(hours), to the equation named hours. If you do not specify equation(), the results are the same as if you specied equation(#1). For more information on using predict after multiple-equation estimation commands, see [R] predict.
Remarks
Example 1
All the postestimation commands discussed in [TS] var postestimation work after varbasic. Suppose that we are interested in testing the hypothesis that there is no autocorrelation in the VAR disturbances. Continuing example 1 from [TS] varbasic, we now use varlmar to test this hypothesis.
594

. use http://www.stata-press.com/data/r12/lutkepohl2 (Quarterly SA West German macro data, Bil DM, from Lutkepohl 1993 Table E.1) . varbasic dln_inv dln_inc dln_consump if qtr<=tq(1978q4) (output omitted ) . varlmar Lagrange-multiplier test lag 1 2 chi2 5.5871 6.3189 df 9 9 Prob > chi2 0.78043 0.70763
H0: no autocorrelation at lag order
Because we cannot reject the null hypothesis of no autocorrelation in the residuals, this test does not indicate any model misspecication.

Also see
[TS] varbasic Fit a simple VAR and graph IRFs or FEVDs [U] 20 Estimation and postestimation commands
Title
vargranger Perform pairwise Granger causality tests after var or svar
Syntax
vargranger , estimates(estname) separator(#)
vargranger can be used only after var or svar; see [TS] var and [TS] var svar.
Menu
Statistics
>
>
VAR diagnostics and tests
>
Granger causality tests
Description
vargranger performs a set of Granger causality tests for each equation in a VAR, providing a convenient alternative to test; see [R] test.
Options
estimates(estname) requests that vargranger use the previously obtained set of var or svar estimates saved as estname. By default, vargranger uses the active results. See [R] estimates for information about saving and restoring estimation results. separator(#) species how often separator lines should be drawn between rows. By default, separator lines appear every K lines, where K is the number of equations in the VAR under analysis. For example, separator(1) would draw a line between each row, separator(2) between every other row, and so on. separator(0) species that lines not appear in the table.
Remarks
After tting a VAR, we may want to know whether one variable Granger-causes another (Granger 1969). A variable x is said to Granger-cause a variable y if, given the past values of y , past values of x are useful for predicting y . A common method for testing Granger causality is to regress y on its own lagged values and on lagged values of x and test the null hypothesis that the estimated coefcients on the lagged values of x are jointly zero. Failure to reject the null hypothesis is equivalent to failing to reject the hypothesis that x does not Granger-cause y . For each equation and each endogenous variable that is not the dependent variable in that equation, vargranger computes and reports Wald tests that the coefcients on all the lags of an endogenous variable are jointly zero. For each equation in a VAR, vargranger tests the hypotheses that each of the other endogenous variables does not Granger-cause the dependent variable in that equation. Because it may be interesting to investigate these types of hypotheses by using the VAR that underlies an SVAR, vargranger can also produce these tests by using the e() results from an svar. When vargranger uses svar e() results, the hypotheses concern the underlying var estimates. See [TS] var and [TS] var svar for information about tting VARs and SVARs in Stata. See L tkepohl (2005), Hamilton (1994), and Amisano and Giannini (1997) for information about Granger u causality and on VARs and SVARs in general.
595
596
Example 1: After var

Here we ret the model with German data described in [TS] var and then perform Granger causality tests with vargranger.
. use http://www.stata-press.com/data/r12/lutkepohl2 (Quarterly SA West German macro data, Bil DM, from Lutkepohl 1993 Table E.1) . var dln_inv dln_inc dln_consump if qtr<=tq(1978q4), dfk small (output omitted ) . vargranger Granger causality Wald tests Equation dln_inv dln_inv dln_inv dln_inc dln_inc dln_inc dln_consump dln_consump dln_consump Excluded dln_inc dln_consump ALL dln_inv dln_consump ALL dln_inv dln_inc ALL F .04847 1.5004 1.5917 1.7683 1.7184 1.9466 .97147 6.1465 3.7746 df 2 2 4 2 2 4 2 2 4 df_r 66 66 66 66 66 66 66 66 66 Prob > F 0.9527 0.2306 0.1869 0.1786 0.1873 0.1130 0.3839 0.0036 0.0080
Because the estimates() option was not specied, vargranger used the active e() results. Consider the results of the three tests for the rst equation. The rst is a Wald test that the coefcients on the two lags of dln inc that appear in the equation for dln inv are jointly zero. The null hypothesis that dln inc does not Granger-cause dln inv cannot be rejected. Similarly, we cannot reject the null hypothesis that the coefcients on the two lags of dln consump in the equation for dln inv are jointly zero, so we cannot reject the hypothesis that dln consump does not Grangercause dln inv. The third test is with respect to the null hypothesis that the coefcients on the two lags of all the other endogenous variables are jointly zero. Because this cannot be rejected, we cannot reject the null hypothesis that dln inc and dln consump, jointly, do not Granger-cause dln inv. Because we failed to reject most of these null hypotheses, we might be interested in imposing some constraints on the coefcients. See [TS] var for more on tting VAR models with constraints on the coefcients.
Example 2: Using test instead of vargranger

We could have used test to compute these Wald tests, but vargranger saves a great deal of typing. Still, seeing how to use test to obtain the results reported by vargranger is useful.
. test [dln_inv]L.dln_inc [dln_inv]L2.dln_inc ( 1) [dln_inv]L.dln_inc = 0 ( 2) [dln_inv]L2.dln_inc = 0 F( 2, 66) = 0.05 Prob > F = 0.9527

. test ( 1) ( 2) ( 3) ( 4) [dln_inv]L.dln_consump [dln_inv]L2.dln_consump, accumulate [dln_inv]L.dln_inc = 0 [dln_inv]L2.dln_inc = 0 [dln_inv]L.dln_consump = 0 [dln_inv]L2.dln_consump = 0 F( 4, 66) = 1.59 Prob > F = 0.1869 . test [dln_inv]L.dln_inv [dln_inv]L2.dln_inv, accumulate ( 1) [dln_inv]L.dln_inc = 0 ( 2) [dln_inv]L2.dln_inc = 0 ( 3) [dln_inv]L.dln_consump = 0 ( 4) [dln_inv]L2.dln_consump = 0 ( 5) [dln_inv]L.dln_inv = 0 ( 6) [dln_inv]L2.dln_inv = 0 F( 6, 66) = 1.62 Prob > F = 0.1547
597
The rst two calls to test show how vargranger obtains its results. The rst test reproduces the rst test reported for the dln inv equation. The second test reproduces the ALL entry for the rst equation. The third test reproduces the standard F statistic for the dln inv equation, reported in the header of the var output in the previous example. The standard F statistic also includes the lags of the dependent variable, as well as any exogenous variables in the equation. This illustrates that the test performed by vargranger of the null hypothesis that the coefcients on all the lags of all the other endogenous variables are jointly zero for a particular equation; that is, the All test is not the same as the standard F statistic for that equation.
Example 3: After svar

When vargranger is run on svar estimates, the null hypotheses are with respect to the underlying var estimates. We run vargranger after using svar to t an SVAR that has the same underlying VAR as our model in the previous example.
. matrix A = (., 0,0 \ ., ., 0\ .,.,.) . matrix B = I(3) . svar dln_inv dln_inc dln_consump if qtr<=tq(1978q4), dfk small aeq(A) beq(B) (output omitted ) . vargranger Granger causality Wald tests Equation dln_inv dln_inv dln_inv dln_inc dln_inc dln_inc dln_consump dln_consump dln_consump Excluded dln_inc dln_consump ALL dln_inv dln_consump ALL dln_inv dln_inc ALL F .04847 1.5004 1.5917 1.7683 1.7184 1.9466 .97147 6.1465 3.7746 df 2 2 4 2 2 4 2 2 4 df_r 66 66 66 66 66 66 66 66 66 Prob > F 0.9527 0.2306 0.1869 0.1786 0.1873 0.1130 0.3839 0.0036 0.0080
As we expected, the vargranger results are identical to those in the rst example.
598
Saved results
vargranger saves the following in r():
Matrices r(gstats) r(gstats)
2 , df, and p-values (if e(small)=="") F , df, df r, and p-values (if e(small)!="")

vargranger is implemented as an ado-le. vargranger uses test to obtain Wald statistics of the hypothesis that all coefcients on the lags of variable x are jointly zero in the equation for variable y . vargranger uses the e() results saved by var or svar to determine whether to calculate and report small-sample F statistics or large-sample 2 statistics.
Clive William John Granger (19342009) was born in Swansea, Wales, and earned degrees at the University of Nottingham in mathematics and statistics. Joining the staff there, he also worked at Princeton on the spectral analysis of economic time series, before moving in 1973 to the University of California, San Diego. He was awarded the 2003 Nobel Prize in Economics for methods of analyzing economic time series with common trends (cointegration). He was knighted in 2005, thus becoming Sir Clive Granger.
References
Amisano, G., and C. Giannini. 1997. Topics in Structural VAR Econometrics. 2nd ed. Heidelberg: Springer. Granger, C. W. J. 1969. Investigating causal relations by econometric models and cross-spectral methods. Econometrica 37: 424438. Hamilton, J. D. 1994. Time Series Analysis. Princeton: Princeton University Press. L tkepohl, H. 1993. Introduction to Multiple Time Series Analysis. 2nd ed. New York: Springer. u . 2005. New Introduction to Multiple Time Series Analysis. New York: Springer. Phillips, P. C. B. 1997. The ET Interview: Professor Clive Granger. Econometric Theory 13: 253303.
Also see
[TS] var Vector autoregressive models [TS] var svar Structural vector autoregressive models [TS] varbasic Fit a simple VAR and graph IRFs or FEVDs [TS] var intro Introduction to vector autoregressive models
Title
varlmar Perform LM test for residual autocorrelation after var or svar
Syntax
varlmar options mlag(#) estimates(estname) separator(#) , options Description use # for the maximum order of autocorrelation; default is mlag(2) use previously saved results estname; default is to use active results draw separator line after every # rows
varlmar can be used only after var or svar; see [TS] var and [TS] var svar. You must tsset your data before using varlmar; see [TS] tsset.
Menu
Statistics
>
>
>
LM test for residual autocorrelation
Description
varlmar implements a Lagrange-multiplier (LM) test for autocorrelation in the residuals of VAR models, which was presented in Johansen (1995).
Options
mlag(#) species the maximum order of autocorrelation to be tested. The integer specied in mlag() must be greater than 0; the default is 2. estimates(estname) requests that varlmar use the previously obtained set of var or svar estimates saved as estname. By default, varlmar uses the active results. See [R] estimates for information on saving and restoring estimation results. separator(#) species how often separator lines should be drawn between rows. By default, separator lines do not appear. For example, separator(1) would draw a line between each row, separator(2) between every other row, and so on.
Remarks
Most postestimation analyses of VAR models and SVAR models assume that the disturbances are not autocorrelated. varlmar implements the LM test for autocorrelation in the residuals of a VAR model discussed in Johansen (1995, 2122). The test is performed at lags j = 1, . . . , mlag(). For each j , the null hypothesis of the test is that there is no autocorrelation at lag j . varlmar uses the estimation results saved by var or svar. By default, varlmar uses the active estimation results. However, varlmar can use any previously saved var or svar estimation results specied in the estimates() option.
599
600

Here we ret the model with German data described in [TS] var and then call varlmar.
. use http://www.stata-press.com/data/r12/lutkepohl2 (Quarterly SA West German macro data, Bil DM, from Lutkepohl 1993 Table E.1) . var dln_inv dln_inc dln_consump if qtr<=tq(1978q4), dfk (output omitted ) . varlmar, mlag(5) Lagrange-multiplier test lag 1 2 3 4 5 chi2 5.5871 6.3189 8.4022 11.8742 5.2914 df 9 9 9 9 9 Prob > chi2 0.78043 0.70763 0.49418 0.22049 0.80821
Because we cannot reject the null hypothesis that there is no autocorrelation in the residuals for any of the ve orders tested, this test gives no hint of model misspecication. Although we t the VAR with the dfk option to be consistent with the example in [TS] var, varlmar always uses the ML estimator of . The results obtained from varlmar are the same whether or not dfk is specied.

When varlmar is applied to estimation results produced by svar, the sequence of LM tests is applied to the underlying VAR. See [TS] var svar for a description of how an SVAR model builds on a VAR. In this example, we t an SVAR that has an underlying VAR with two lags that is identical to the one t in the previous example.
. matrix A = (.,.,0\0,.,0\.,.,.) . matrix B = I(3) . svar dln_inv dln_inc dln_consump if qtr<=tq(1978q4), dfk aeq(A) beq(B) (output omitted ) . varlmar, mlag(5) Lagrange-multiplier test lag 1 2 3 4 5 chi2 5.5871 6.3189 8.4022 11.8742 5.2914 df 9 9 9 9 9 Prob > chi2 0.78043 0.70763 0.49418 0.22049 0.80821
Because the underlying VAR(2) is the same as the previous example (we assure you that this is true), the output from varlmar is also the same.
601
Saved results
varlmar saves the following in r():
Matrices r(lm)
2 , df, and p-values

varlmar is implemented as an ado-le. The formula for the LM test statistic at lag j is
LMs = (T d .5) ln
|| |s |
where T is the number of observations in the VAR; d is explained below; is the maximum likelihood estimate of , the variancecovariance matrix of the disturbances from the VAR; and s is the maximum likelihood estimate of from the following augmented VAR. If there are K equations in the VAR, we can dene et to be a K 1 vector of residuals. After we create the K new variables e1, e2, . . . , eK containing the residuals from the K equations, we can augment the original VAR with lags of these K new variables. For each lag s, we form an augmented regression in which the new residual variables are lagged s times. Per the method of Davidson and MacKinnon (1993, 358), the missing values from these s lags are replaced with zeros. s is the maximum likelihood estimate of from this augmented VAR, and d is the number of coefcients estimated in the augmented VAR. See [TS] var for a discussion of the maximum likelihood estimate of in a VAR. The asymptotic distribution of LMs is 2 with K 2 degrees of freedom.
References
Davidson, R., and J. G. MacKinnon. 1993. Estimation and Inference in Econometrics. New York: Oxford University Press. Johansen, S. 1995. Likelihood-Based Inference in Cointegrated Vector Autoregressive Models. Oxford: Oxford University Press.
Also see
Title
varnorm Test for normally distributed disturbances after var or svar
Syntax
varnorm options jbera skewness kurtosis estimates(estname) cholesky separator(#) , options Description report JarqueBera statistic; default is to report all three statistics report skewness statistic; default is to report all three statistics report kurtosis statistic; default is to report all three statistics use previously saved results estname; default is to use active results use Cholesky decomposition draw separator line after every # rows
varnorm can be used only after var or svar; see [TS] var and [TS] var svar. You must tsset your data before using varnorm; see [TS] tsset.
Menu
Statistics
>
>
>
Test for normally distributed disturbances
Description
varnorm computes and reports a series of statistics against the null hypothesis that the disturbances in a VAR are normally distributed. For each equation, and for all equations jointly, up to three statistics may be computed: a skewness statistic, a kurtosis statistic, and the JarqueBera statistic. By default, all three statistics are reported.
Options
jbera requests that the JarqueBera statistic and any other explicitly requested statistic be reported. By default, the JarqueBera, skewness, and kurtosis statistics are reported. skewness requests that the skewness statistic and any other explicitly requested statistic be reported. By default, the JarqueBera, skewness, and kurtosis statistics are reported. kurtosis requests that the kurtosis statistic and any other explicitly requested statistic be reported. By default, the JarqueBera, skewness, and kurtosis statistics are reported. estimates(estname) species that varnorm use the previously obtained set of var or svar estimates saved as estname. By default, varnorm uses the active results. See [R] estimates for information on saving and restoring estimation results. cholesky species that varnorm use the Cholesky decomposition of the estimated variancecovariance matrix of the disturbances, , to orthogonalize the residuals when varnorm is applied to svar results. By default, when varnorm is applied to svar results, it uses the estimated structural decomposition A1 B on C to orthogonalize the residuals. When applied to var e() results, varnorm always uses the Cholesky decomposition of . For this reason, the cholesky option may not be specied when using var results.
602
603
separator(#) species how often separator lines should be drawn between rows. By default, separator lines do not appear. For example, separator(1) would draw a line between each row, separator(2) between every other row, and so on.
Remarks
Some of the postestimation statistics for VAR and SVAR assume that the K disturbances have a K -dimensional multivariate normal distribution. varnorm uses the estimation results produced by var or svar to produce a series of statistics against the null hypothesis that the K disturbances in the VAR are normally distributed. Per the notation in L tkepohl (2005), call the skewness statistic 1 , the kurtosis statistic 2 , and u the JarqueBera statistic 3 . The JarqueBera statistic is a combination of the other two statistics. The single-equation results are from tests against the null hypothesis that the disturbance for that particular equation is normally distributed. The results for all the equations are from tests against the null hypothesis that the K disturbances follow a K -dimensional multivariate normal distribution. Failure to reject the null hypothesis indicates a lack of model misspecication.

We ret the model with German data described in [TS] var and then call varnorm.
. use http://www.stata-press.com/data/r12/lutkepohl2 (Quarterly SA West German macro data, Bil DM, from Lutkepohl 1993 Table E.1) . var dln_inv dln_inc dln_consump if qtr<=tq(1978q4), dfk (output omitted ) . varnorm Jarque-Bera test Equation dln_inv dln_inc dln_consump ALL Skewness test Equation dln_inv dln_inc dln_consump ALL Kurtosis test Equation dln_inv dln_inc dln_consump ALL Kurtosis 3.9331 3.7396 2.6484 chi2 2.648 1.664 0.376 4.688 df 1 1 1 3 Prob > chi2 0.10367 0.19710 0.53973 0.19613 Skewness .11935 -.38316 -.31275 chi2 0.173 1.786 1.190 3.150 df 1 1 1 3 Prob > chi2 0.67718 0.18139 0.27532 0.36913 chi2 2.821 3.450 1.566 7.838 df 2 2 2 6 Prob > chi2 0.24397 0.17817 0.45702 0.25025
dfk estimator used in computations
604
In this example, neither the single-equation JarqueBera statistics nor the joint JarqueBera statistic come close to rejecting the null hypothesis. The skewness and kurtosis results have similar structures. The JarqueBera results use the sum of the skewness and kurtosis statistics. The skewness and kurtosis results are based on the skewness and kurtosis coefcients, respectively. See Methods and formulas.

The test statistics are computed on the orthogonalized VAR residuals; see Methods and formulas. When varnorm is applied to var results, varnorm uses a Cholesky decomposition of the estimated variancecovariance matrix of the disturbances, , to orthogonalize the residuals. By default, when varnorm is applied to svar estimation results, it uses the estimated structural decomposition A1 B on C to orthogonalize the residuals of the underlying VAR. Alternatively, when varnorm is applied to svar results and the cholesky option is specied, varnorm uses the Cholesky decomposition of to orthogonalize the residuals of the underlying VAR. We t an SVAR that is based on an underlying VAR with two lags that is the same as the one t in the previous example. We impose a structural decomposition that is the same as the Cholesky decomposition, as illustrated in [TS] var svar.
. matrix a = (.,0,0\.,.,0\.,.,.) . matrix b = I(3) . svar dln_inv dln_inc dln_consump if qtr<=tq(1978q4), dfk aeq(a) beq(b) (output omitted ) . varnorm Jarque-Bera test Equation dln_inv dln_inc dln_consump ALL Skewness test Equation dln_inv dln_inc dln_consump ALL Kurtosis test Equation dln_inv dln_inc dln_consump ALL Kurtosis 3.9331 3.7396 2.6484 chi2 2.648 1.664 0.376 4.688 df 1 1 1 3 Prob > chi2 0.10367 0.19710 0.53973 0.19613 Skewness .11935 -.38316 -.31275 chi2 0.173 1.786 1.190 3.150 df 1 1 1 3 Prob > chi2 0.67718 0.18139 0.27532 0.36913 chi2 2.821 3.450 1.566 7.838 df 2 2 2 6 Prob > chi2 0.24397 0.17817 0.45702 0.25025
dfk estimator used in computations
605
Because the estimated structural decomposition is the same as the Cholesky decomposition, the varnorm results are the same as those from the previous example.
Technical note
The statistics computed by varnorm depend on , the estimated variancecovariance matrix of the disturbances. var uses the maximum likelihood estimator of this matrix by default, but the dfk option produces an estimator that uses a small-sample correction. Thus specifying dfk in the call to var or svar will affect the test results produced by varnorm.
Saved results
varnorm saves the following in r():
Macros r(dfk) Matrices r(kurtosis) r(skewness) r(jb) dfk, if specied kurtosis test, df, and p-values skewness test, df, and p-values JarqueBera test, df, and p-values

varnorm is implemented as an ado-le. varnorm is based on the derivations found in L tkepohl (2005, 174181). Let ut be the K 1 u vector of residuals from the K equations in a previously tted VAR or the residuals from the K equations of the VAR underlying a previously tted SVAR. Similarly, let be the estimated covariance matrix of the disturbances. (Note that depends on whether the dfk option was specied.) The skewness, kurtosis, and JarqueBera statistics must be computed using the orthogonalized residuals. Because
= PP
implies that
P1 P1 = IK
premultiplying ut by P is one way of performing the orthogonalization. When varnorm is applied to var results, P is dened to be the Cholesky decomposition of . When varnorm is applied to svar results, P is set, by default, to the estimated structural decomposition; that is, P = A1 B, where A and B are the svar estimates of the A and B matrices, or C, where C is the long-run SVAR estimation of C . (See [TS] var svar for more on the origin and estimation of the A and B matrices.) When varnorm is applied to svar results and the cholesky option is specied, P is set to the Cholesky decomposition of . Dene wt to be the orthogonalized VAR residuals given by
wt = (w1t , . . . , wKt ) = P1 ut
606
The K 1 vectors of skewness and kurtosis coefcients are then computed using the orthogonalized residuals by
b1 = (b11 , . . . , bK1 ) ;
bk1 =
1 T 1 T
T 3 wkt i=1 T 4 wkt i=1
b2 = (b12 , . . . , bK2 ) ;
bk2 =
Under the null hypothesis of multivariate Gaussian disturbances,
1 =
T b1 b1 6
2 (K)
2 =
and
T (b2 3) (b2 3) 24 3 = 1 + 2
d
2 (K)
2 (2K)
1 is the skewness statistic, 2 is the kurtosis statistic, and 3 is the JarqueBera statistic. 1 , 2 , and 3 are for tests of the null hypothesis that the K 1 vector of disturbances follows a multivariate normal distribution. The corresponding statistics against the null hypothesis that the disturbances from the k th equation come from a univariate normal distribution are 1k = 2k =
and
T b2 k1 6
2 (1) 2 (1)
d d
T ( b 2 3)2 k2 24
3k = 1 + 2
2 (2)
References
Hamilton, J. D. 1994. Time Series Analysis. Princeton: Princeton University Press. Jarque, C. M., and A. K. Bera. 1987. A test for normality of observations and regression residuals. International Statistical Review 2: 163172. L tkepohl, H. 1993. Introduction to Multiple Time Series Analysis. 2nd ed. New York: Springer. u . 2005. New Introduction to Multiple Time Series Analysis. New York: Springer.
Also see
Title
varsoc Obtain lag-order selection statistics for VARs and VECMs
Syntax
Preestimation syntax varsoc depvarlist Postestimation syntax varsoc , estimates(estname) Description set maximum lag order to #; default is maxlag(4) use varlist as exogenous variables apply constraints to exogenous variables suppress constant term use L tkepohls version of information criteria u set condence level; default is level(95) draw separator line after every # rows if in , preestimation options
preestimation options
Main
maxlag(#) exog(varlist) constraints(constraints) noconstant lutstats level(#) separator(#)
You must tsset your data before using varsoc; see [TS] tsset. by is allowed with the preestimation version of varsoc; see [U] 11.1.10 Prex commands.
Menu
Preestimation for VARs
Statistics
>
>
>
Lag-order selection statistics (preestimation)
Postestimation for VARs

Statistics
>
>
>
Lag-order selection statistics (postestimation)
Preestimation for VECMs

Statistics
>
>
VEC diagnostics and tests
>
Lag-order selection statistics (preestimation)
Postestimation for VECMs

Statistics
>
>
>
Lag-order selection statistics (postestimation)
Description
varsoc reports the nal prediction error (FPE), Akaikes information criterion (AIC), Schwarzs Bayesian information criterion (SBIC), and the Hannan and Quinn information criterion (HQIC) lagorder selection statistics for a series of vector autoregressions of order 1, . . . , maxlag(). A sequence of likelihood-ratio test statistics for all the full VARs of order less than or equal to the highest lag order is also reported. In the postestimation version, the maximum lag and estimation options are based on the model just t or the model specied in estimates(estname).
607
608
The preestimation version of varsoc can also be used to select the lag order for a vector errorcorrection model (VECM). As shown by Nielsen (2001), the lag-order selection statistics discussed here can be used in the presence of I(1) variables.
Preestimation options

Main
maxlag(#) species the maximum lag order for which the statistics are to be obtained. exog(varlist) species exogenous variables to include in the VARs t by varsoc. constraints(constraints) species a list of constraints on the exogenous variables to be applied. Do not specify constraints on the lags of the endogenous variables because specifying one would mean that at least one of the VAR models considered by varsoc will not contain the lag specied in the constraint. Use var directly to obtain selection-order criteria with constraints on lags of the endogenous variables. noconstant suppresses the constant terms from the model. By default, constant terms are included. lutstats species that the L tkepohl (2005) versions of the lag-order selection statistics be reported. u See Methods and formulas for a discussion of these statistics. level(#) species the condence level, as a percentage, that is used to identify the rst likelihoodratio test that rejects the null hypothesis that the additional parameters from adding a lag are jointly zero. The default is level(95) or as set by set level; see [U] 20.7 Specifying the width of condence intervals. separator(#) species how often separator lines should be drawn between rows. By default, separator lines do not appear. For example, separator(1) would draw a line between each row, separator(2) between every other row, and so on.
Postestimation option
estimates(estname) species the name of a previously stored set of var or svar estimates. When no depvarlist is specied, varsoc uses the postestimation syntax and uses the currently active estimation results or the results specied in estimates(estname). See [R] estimates for information about saving and restoring estimation results.
Remarks
Many selection-order statistics have been developed to assist researchers in tting a VAR of the correct order. Several of these selection-order statistics appear in the [TS] var output. The varsoc command computes these statistics over a range of lags p while maintaining a common sample and option specication. varsoc can be used as a preestimation or a postestimation command. When it is used as a preestimation command, a depvarlist is required, and the default maximum lag is 4. When it is used as a postestimation command, varsoc uses the model specication stored in estname or the previously tted model. varsoc computes four information criteria as well as a sequence of likelihood ratio (LR) tests. The information criteria include the FPE, AIC, the HQIC, and SBIC.
609
For a given lag p, the LR test compares a VAR with p lags with one with p 1 lags. The null hypothesis is that all the coefcients on the pth lags of the endogenous variables are zero. To use this sequence of LR tests to select a lag order, we start by looking at the results of the test for the model with the most lags, which is at the bottom of the table. Proceeding up the table, the rst test that rejects the null hypothesis is the lag order selected by this process. See L tkepohl (2005, 143144) u for more information on this procedure. An * appears next to the LR statistic indicating the optimal lag. For the remaining statistics, the lag with the smallest value is the order selected by that criterion. An * indicates the optimal lag. Strictly speaking, the FPE is not an information criterion, though we include it in this discussion because, as with an information criterion, we select the lag length corresponding to the lowest value; and, naturally, we want to minimize the prediction error. The AIC measures the discrepancy between the given model and the true model, which, of course, we want to minimize. Amemiya (1985) provides an intuitive discussion of the arguments in Akaike (1973). The SBIC and the HQIC can be interpreted similarly to the AIC, though the SBIC and the HQIC have a theoretical advantage over the AIC and the FPE. As L tkepohl (2005, 148152) demonstrates, choosing u p to minimize the SBIC or the HQIC provides consistent estimates of the true lag order, p. In contrast, minimizing the AIC or the FPE will overestimate the true lag order with positive probability, even with an innite sample size.
Example 1: Preestimation
Here we use varsoc as a preestimation command.
. use http://www.stata-press.com/data/r12/lutkepohl2 (Quarterly SA West German macro data, Bil DM, from Lutkepohl 1993 Table E.1) . varsoc dln_inv dln_inc dln_consump if qtr<=tq(1978q4), lutstats Selection-order criteria (lutstats) Sample: 1961q2 - 1978q4 Number of obs = 71 lag 0 1 2 3 4 LL 564.784 576.409 588.859 591.237 598.457 LR df p FPE AIC HQIC SBIC -24.423* -24.2102 -24.0205 -23.5472 -23.2102
23.249 24.901* 4.7566 14.438
9 9 9 9
0.006 0.003 0.855 0.108
2.7e-11 -24.423 -24.423* 2.5e-11 -24.497 -24.3829 2.3e-11* -24.5942* -24.3661 2.7e-11 -24.4076 -24.0655 2.9e-11 -24.3575 -23.9012
Endogenous: Exogenous:
dln_inv dln_inc dln_consump _cons
The sample used begins in 1961q2 because all the VARs are t to the sample dened by any if or in conditions and the available data for the maximum lag specied. The default maximum number of lags is four. Because we specied the lutstats option, the table contains the L tkepohl (2005) u versions of the information criteria, which differ from the standard denitions in that they drop the constant term from the log likelihood. In this example, the likelihood-ratio tests selected a model with two lags. AIC and FPE have also both chosen a model with two lags, whereas SBIC and HQIC have both selected a model with zero lags.
610
Example 2: Postestimation
varsoc works as a postestimation command when no dependent variables are specied.
. var dln_inc dln_consump if qtr<=tq(1978q4), lutstats exog(l.dln_inv) (output omitted ) . varsoc Selection-order criteria (lutstats) Sample: 1960q4 - 1978q4 Number of obs = lag 0 1 2 LL 460.646 467.606 477.087 LR df p FPE AIC HQIC SBIC
73
13.919 18.962*
4 4
0.008 0.001
1.3e-08 -18.2962 -18.2962 -18.2962* 1.2e-08 -18.3773 -18.3273 -18.2518 1.0e-08* -18.5275* -18.4274* -18.2764
dln_inc dln_consump L.dln_inv _cons
Because we included one lag of dln inv in our original model, varsoc did likewise with each model it t.
Based on the work of Tsay (1984), Paulsen (1984), and Nielsen (2001), these lag-order selection criteria can be used to determine the lag length of the VAR underlying a VECM. See [TS] vec intro for an example in which we use varsoc to choose the lag order for a VECM.
Saved results
varsoc saves the following in r():
Scalars r(N) r(tmax) r(tmin) Macros r(endog) r(lutstats) r(cns#) Matrices r(stats) number of observations last time period in sample rst time period in sample names of endogenous variables lutstats, if specied the #th constraint LL, LR, FPE, AIC, HQIC, SBIC, and p-values r(mlag) r(N gaps) maximum lag order the number of gaps in the sample names of exogenous variables rmlutstats, if specied
r(exog) r(rmlutstats)

varsoc is implemented as an ado-le. As shown by Hamilton (1994, 295296), the log likelihood for a VAR(p) is
LL
T 2
ln |1 | K ln(2) K
611
where T is the number of observations, K is the number of equations, and is the maximum likelihood estimate of E[ut ut ], where ut is the K 1 vector of disturbances. Because ln |1 | = ln || the log likelihood can be rewritten as
LL
T 2
ln || + K ln(2) + K
Letting LL(j ) be the value of the log likelihood with j lags yields the LR statistic for lag order j as
LR(j)
= 2 LL(j) LL(j 1)
Model-order statistics The formula for the FPE given in L tkepohl (2005, 147) is u
FPE
= |u |
T + Kp + 1 T Kp 1
This formula, however, assumes that there is a constant in the model and that none of the variables are dropped because of collinearity. To deal with these problems, the FPE is implemented as
FPE
= |u |
T +m T m
where m is the average number of parameters over the K equations. This implementation accounts for variables dropped because of collinearity. By default, the AIC, SBIC, and HQIC are computed according to their standard denitions, which include the constant term from the log likelihood. That is,
AIC
=2 =2 =2
LL
T
LL
SBIC
T
LL
2tp T ln(T ) + tp T + + 2ln ln(T ) tp T
HQIC
where tp is the total number of parameters in the model and LL is the log likelihood.
612
Lutstats L tkepohl (2005) advocates dropping the constant term from the log likelihood because it does u not affect inference. The L tkepohl versions of the information criteria are u
2pK 2 T ln(T ) SBIC = ln |u | + pK 2 T 2ln ln(T ) HQIC = ln |u | + pK 2 T

AIC
= ln |u | +
References
Akaike, H. 1973. Information theory and an extension of the maximum likelihood principle. In Second International Symposium on Information Theory, ed. B. N. Petrov and F. Csaki, 267281. Budapest: AkailseoniaiKiudo. Amemiya, T. 1985. Advanced Econometrics. Cambridge, MA: Harvard University Press. Hamilton, J. D. 1994. Time Series Analysis. Princeton: Princeton University Press. L tkepohl, H. 1993. Introduction to Multiple Time Series Analysis. 2nd ed. New York: Springer. u . 2005. New Introduction to Multiple Time Series Analysis. New York: Springer. Nielsen, B. 2001. Order determination in general vector autoregressions. Working paper, Department of Economics, University of Oxford and Nufeld College. http://ideas.repec.org/p/nuf/econwp/0110.html. Paulsen, J. 1984. Order determination of multivariate autoregressive time series with unit roots. Journal of Time Series Analysis 5: 115127. Tsay, R. S. 1984. Order selection in nonstationary autoregressive models. Annals of Statistics 12: 14251433.
Also see
[TS] var Vector autoregressive models [TS] var svar Structural vector autoregressive models [TS] varbasic Fit a simple VAR and graph IRFs or FEVDs [TS] vec Vector error-correction models [TS] var intro Introduction to vector autoregressive models [TS] vec intro Introduction to vector error-correction models
Title
varstable Check the stability condition of VAR or SVAR estimates
Syntax
varstable options
Main
, options Description use previously saved results estname; default is to use active results save the companion matrix as matrix name graph eigenvalues of the companion matrix label eigenvalues with the distance from the unit circle label eigenvalues with the modulus change look of markers (color, size, etc.) affect rendition of reference unit circle suppress polar grid circles specify radii and appearance of polar grid circles; see Options for details add other plots to the generated graph any options other than by() documented in [G-3] twoway options
estimates(estname) amat(matrix name) graph dlabel modlabel marker options rlopts(cline options) nogrid pgrid( . . . )
Add plots
varstable can be used only after var or svar; see [TS] var and [TS] var svar.
Menu
Statistics
>
>
>
Check stability condition of VAR estimates
Description
varstable checks the eigenvalue stability condition after estimating the parameters of a vector autoregression using var or svar.
Options

Main
estimates(estname) requests that varstable use the previously obtained set of var estimates saved as estname. By default, varstable uses the active estimation results. See [R] estimates for information about saving and restoring estimation results. amat(matrix name) species a valid Stata matrix name by which the companion matrix A can be saved (see Methods and formulas for the denition of the matrix A). The default is not to save the A matrix.
613
614
graph causes varstable to draw a graph of the eigenvalues of the companion matrix. dlabel labels each eigenvalue with its distance from the unit circle. dlabel cannot be specied with modlabel. modlabel labels the eigenvalues with their moduli. modlabel cannot be specied with dlabel. marker options specify the look of markers. This look includes the marker symbol, the marker size, and its color and outline; see [G-3] marker options. rlopts(cline options) affect the rendition of the reference unit circle; see [G-3] cline options. nogrid suppresses the polar grid circles. pgrid( numlist , line options ) determines the radii and appearance of the polar grid circles. By default, the graph includes nine polar grid circles with radii .1, .2, . . . , .9 that have the grid line style. The numlist species the radii for the polar grid circles. The line options determine the appearance of the polar grid circles; see [G-3] line options. Because the pgrid() option can be repeated, circles with different radii can have distinct appearances.
Add plots
addplot(plot) adds specied plots to the generated graph. See [G-3] addplot option.
twoway options are any of the options documented in [G-3] twoway options, except by(). These include options for titling the graph (see [G-3] title options) and for saving the graph to disk (see [G-3] saving option).
Remarks
Inference after var and svar requires that variables be covariance stationary. The variables in yt are covariance stationary if their rst two moments exist and are independent of time. More explicitly, a variable yt is covariance stationary if 1. E[yt ] is nite and independent of t. 2. Var[yt ] is nite and independent of t 3. Cov[yt , ys ] is a nite function of |t s| but not of t or s alone. Interpretation of VAR models, however, requires that an even stricter stability condition be met. If a VAR is stable, it is invertible and has an innite-order vector moving-average representation. If the VAR is stable, impulseresponse functions and forecast-error variance decompositions have known interpretations. L tkepohl (2005) and Hamilton (1994) both show that if the modulus of each eigenvalue of the u matrix A is strictly less than one, the estimated VAR is stable (see Methods and formulas for the denition of the matrix A).
Example 1
After tting a VAR with var, we can use varstable to check the stability condition. Using the same VAR model that was used in [TS] var, we demonstrate the use of varstable.
. use http://www.stata-press.com/data/r12/lutkepohl2 (Quarterly SA West German macro data, Bil DM, from Lutkepohl 1993 Table E.1) . var dln_inv dln_inc dln_consump if qtr>=tq(1961q2) & qtr<=tq(1978q4) (output omitted )

. varstable, graph Eigenvalue stability condition Eigenvalue .5456253 -.3785754 -.3785754 -.0643276 -.0643276 -.3698058 Modulus .545625 .540232 .540232 .464074 .464074 .369806
615
+ + -
.3853982i .3853982i .4595944i .4595944i
All the eigenvalues lie inside the unit circle. VAR satisfies stability condition.
Because the modulus of each eigenvalue is strictly less than 1, the estimates satisfy the eigenvalue stability condition. Specifying the graph option produced a graph of the eigenvalues with the real components on the x axis and the complex components on the y axis. The graph below indicates visually that these eigenvalues are well inside the unit circle.
Roots of the companion matrix
1 1 1 .5 Imaginary 0 .5
.5
0 Real
.5
Example 2
This example illustrates two other features of the varstable command. First, varstable can check the stability of the estimates of the VAR underlying an SVAR t by var svar. Second, varstable can check the stability of any previously stored var or var svar estimates. We begin by retting the previous VAR and storing the results as var1. Because this is the same VAR that was t in the previous example, the stability results should be identical.
. var dln_inv dln_inc dln_consump if qtr>=tq(1961q2) & qtr<=tq(1978q4) (output omitted ) . estimates store var1
Now we use svar to t an SVAR with a different underlying VAR and check the estimates of that underlying VAR for stability.
616

. matrix A = (.,0\.,.) . matrix B = I(2) . svar d.ln_inc d.ln_consump, aeq(A) beq(B) (output omitted ) . varstable Eigenvalue stability condition Eigenvalue .548711 -.2979493 + -.2979493 -.3570825 Modulus .548711 .525443 .525443 .357082
.4328013i .4328013i
The estimates() option allows us to check the stability of the var results stored as var1.
. varstable, est(var1) Eigenvalue stability condition Eigenvalue .5456253 -.3785754 -.3785754 -.0643276 -.0643276 -.3698058 Modulus .545625 .540232 .540232 .464074 .464074 .369806
+ + -
.3853982i .3853982i .4595944i .4595944i
The results are identical to those obtained in the previous example, conrming that we were checking the results in var1.
Saved results
varstable saves the following in r():
Matrices r(Re) r(Im) r(Modulus) real part of the eigenvalues of A imaginary part of the eigenvalues of A modulus of the eigenvalues of A

varstable is implemented as an ado-le.
617
varstable forms the companion matrix
A1 I A= 0 . . . 0
A2 0 I . . . 0
. . . Ap1 ... 0 ... 0 . .. . . . ... I
Ap 0 0 . . . 0
and obtains its eigenvalues by using matrix eigenvalues. The modulus of the complex eigenvalue r + ci is r2 + c2 . As shown by L tkepohl (2005) and Hamilton (1994), the VAR is stable if the u modulus of each eigenvalue of A is strictly less than 1.
References
Hamilton, J. D. 1994. Time Series Analysis. Princeton: Princeton University Press. L tkepohl, H. 1993. Introduction to Multiple Time Series Analysis. 2nd ed. New York: Springer. u . 2005. New Introduction to Multiple Time Series Analysis. New York: Springer.
Also see
Title
varwle Obtain Wald lag-exclusion statistics after var or svar
Syntax
varwle , estimates(estname) separator(#)
varwle can be used only after var or svar; see [TS] var and [TS] var svar.
Menu
Statistics
>
>
>
Wald lag-exclusion statistics
Description
varwle reports Wald tests the hypothesis that the endogenous variables at a given lag are jointly zero for each equation and for all equations jointly.
Options
estimates(estname) requests that varwle use the previously obtained set of var or svar estimates saved as estname. By default, varwle uses the active estimation results. See [R] estimates for information about saving and restoring estimation results. separator(#) species how often separator lines should be drawn between rows. By default, separator lines do not appear. For example, separator(1) would draw a line between each row, separator(2) between every other row, and so on.
Remarks
After tting a VAR, one hypothesis of interest is that all the endogenous variables at a given lag are jointly zero. varwle reports Wald tests of this hypothesis for each equation and for all equations jointly. varwle uses the estimation results from a previously tted var or svar. By default, varwle uses the active estimation results, but you may also use a stored set of estimates by specifying the estimates() option. If the VAR was t with the small option, varwle also presents small-sample F statistics; otherwise, varwle presents large-sample chi-squared statistics.

We analyze the model with the German data described in [TS] var using varwle.
618

. use http://www.stata-press.com/data/r12/lutkepohl2 (Quarterly SA West German macro data, Bil DM, from Lutkepohl 1993 Table E.1) . var dln_inv dln_inc dln_consump if qtr<=tq(1978q4), dfk small (output omitted ) . varwle Equation: dln_inv lag 1 2 F 2.64902 1.25799 df 3 3 df_r 66 66 Prob > F 0.0560 0.2960
619
Equation: dln_inc lag 1 2 F 2.19276 .907499 df 3 3 df_r 66 66 Prob > F 0.0971 0.4423
Equation: dln_consump lag 1 2 F 1.80804 5.57645 df 3 3 df_r 66 66 Prob > F 0.1543 0.0018
Equation: All lag 1 2 F 3.78884 2.96811 df 9 9 df_r 66 66 Prob > F 0.0007 0.0050
Because the VAR was t with the dfk and small options, varwle used the small-sample estimator of in constructing the VCE, producing an F statistic. The rst two equations appear to have a different lag structure from that of the third. In the rst two equations, we cannot reject the null hypothesis that all three endogenous variables have zero coefcients at the second lag. The hypothesis that all three endogenous variables have zero coefcients at the rst lag can be rejected at the 10% level for both of the rst two equations. In contrast, in the third equation, the coefcients on the second lag of the endogenous variables are jointly signicant, but not those on the rst lag. However, we strongly reject the hypothesis that the coefcients on the rst lag of the endogenous variables are zero in all three equations jointly. Similarly, we can also strongly reject the hypothesis that the coefcients on the second lag of the endogenous variables are zero in all three equations jointly. If we believe these results strongly enough, we might want to ret the original VAR, placing some constraints on the coefcients. See [TS] var for details on how to t VAR models with constraints.

Here we t a simple SVAR and then run varwle:
. matrix a = (.,0\.,.) . matrix b = I(2)
620

. svar dln_inc dln_consump, aeq(a) beq(b) Estimating short-run parameters Iteration Iteration Iteration Iteration Iteration Iteration Iteration Iteration ( ( ( ( ( 1) 2) 3) 4) 5) 0: 1: 2: 3: 4: 5: 6: 7: log log log log log log log log likelihood likelihood likelihood likelihood likelihood likelihood likelihood likelihood = = = = = 0 1 0 0 1 No. of obs Log likelihood Std. Err. 6.725107 10.67698 . 9.466318 . . . . z 13.34 -6.06 . 13.34 . . . . P>|z| 0.000 0.000 . 0.000 . . . . = = 89 578.277 = -159.21683 = 490.92264 = 528.66126 = 573.96363 = 578.05136 = 578.27633 = 578.27699 = 578.27699
Structural vector autoregression [a_1_2]_cons [b_1_1]_cons [b_1_2]_cons [b_2_1]_cons [b_2_2]_cons
Sample: 1960q4 - 1982q4 Exactly identified model Coef. /a_1_1 /a_2_1 /a_1_2 /a_2_2 /b_1_1 /b_2_1 /b_1_2 /b_2_2 89.72411 -64.73622 0 126.2964 1 0 0 1
[95% Conf. Interval] 76.54315 -85.66271 . 107.7428 . . . . 102.9051 -43.80973 . 144.8501 . . . .
The output table from var svar gives information about the estimates of the parameters in the A and B matrices in the structural VAR. But, as discussed in [TS] var svar, an SVAR model builds on an underlying VAR. When varwle uses the estimation results produced by svar, it performs Wald lag-exclusion tests on the underlying VAR model. Next we run varwle on these svar results.
. varwle Equation: dln_inc lag 1 2 chi2 6.88775 1.873546 df 2 2 Prob > chi2 0.032 0.392
Equation: dln_consump lag 1 2 chi2 9.938547 13.89996 df 2 2 Prob > chi2 0.007 0.001
Equation: All lag 1 2 chi2 34.54276 19.44093 df 4 4 Prob > chi2 0.000 0.001
621
Now we t the underlying VAR with two lags and apply varwle to these results.
. var dln_inc dln_consump (output omitted ) . varwle Equation: dln_inc lag 1 2 chi2 6.88775 1.873546 df 2 2 Prob > chi2 0.032 0.392
Equation: dln_consump lag 1 2 chi2 9.938547 13.89996 df 2 2 Prob > chi2 0.007 0.001
Equation: All lag 1 2 chi2 34.54276 19.44093 df 4 4 Prob > chi2 0.000 0.001
Because varwle produces the same results in these two cases, we can conclude that when varwle is applied to svar results, it performs Wald lag-exclusion tests on the underlying VAR.
Saved results
varwle saves the following in r():
Matrices if e(small)=="" r(chi2) r(df) r(p) if e(small)!="" r(F) r(df r) r(df) r(p)
2 test statistics
degrees of freedom p-values

F test statistics numerator degrees of freedom denominator degree of freedom p-values

varwle is implemented as an ado-le. varwle uses test to obtain Wald statistics of the hypotheses that all the endogenous variables at a given lag are jointly zero for each equation and for all equations jointly. Like the test command, varwle uses estimation results saved by var or var svar to determine whether to calculate and report small-sample F statistics or large-sample chi-squared statistics.
622
Abraham Wald (19021950) was born in Cluj, in what is now Romania. He studied mathematics at the University of Vienna, publishing at rst on geometry, but then became interested in economics and econometrics. He moved to the United States in 1938 and later joined the faculty at Columbia. His major contributions to statistics include work in decision theory, optimal sequential sampling, large-sample distributions of likelihood-ratio tests, and nonparametric inference. Wald died in a plane crash in India.
References
Amisano, G., and C. Giannini. 1997. Topics in Structural VAR Econometrics. 2nd ed. Heidelberg: Springer. Hamilton, J. D. 1994. Time Series Analysis. Princeton: Princeton University Press. L tkepohl, H. 1993. Introduction to Multiple Time Series Analysis. 2nd ed. New York: Springer. u Mangel, M., and F. J. Samaniego. 1984. Abraham Walds work on aircraft survivability. Journal of the American Statistical Association 79: 259267. Wolfowitz, J. 1952. Abraham Wald, 19021950. Annals of Mathematical Statistics 23: 113 (and other reports in same issue).
Also see
Title
vec intro Introduction to vector error-correction models
Description
Stata has a suite of commands for tting, forecasting, interpreting, and performing inference on vector error-correction models (VECMs) with cointegrating variables. After tting a VECM, the irf commands can be used to obtain impulseresponse functions (IRFs) and forecast-error variance decompositions (FEVDs). The table below describes the available commands.
Fitting a VECM vec
[TS] vec
Fit vector error-correction models
Model diagnostics and inference vecrank [TS] vecrank veclmar [TS] veclmar
vecnorm vecstable varsoc
[TS] vecnorm [TS] vecstable [TS] varsoc
Estimate the cointegrating rank of a VECM Perform LM test for residual autocorrelation after vec Test for normally distributed disturbances after vec Check the stability condition of VECM estimates Obtain lag-order selection statistics for VARs and VECMs
Forecasting from a VECM fcast compute [TS] fcast compute
fcast graph
[TS] fcast graph
Compute dynamic forecasts of dependent variables after var, svar, or vec Graph forecasts of variables computed by fcast compute
Working with IRFs and FEVDs irf [TS] irf
Create and analyze IRFs and FEVDs
This manual entry provides an overview of the commands for VECMs; provides an introduction to integration, cointegration, estimation, inference, and interpretation of VECM models; and gives an example of how to use Statas vec commands.
Remarks
vec estimates the parameters of cointegrating VECMs. You may specify any of the ve trend specications in Johansen (1995, sec. 5.7). By default, identication is obtained via the Johansen normalization, but vec allows you to obtain identication by placing your own constraints on the parameters of the cointegrating vectors. You may also put more restrictions on the adjustment coefcients. vecrank is the command for determining the number of cointegrating equations. vecrank implements Johansens multiple trace test procedure, the maximum eigenvalue test, and a method based on minimizing either of two different information criteria.
623
624
Because Nielsen (2001) has shown that the methods implemented in varsoc can be used to choose the order of the autoregressive process, no separate vec command is needed; you can simply use varsoc. veclmar tests that the residuals have no serial correlation, and vecnorm tests that they are normally distributed. All the irf routines described in [TS] irf are available for estimating, interpreting, and managing estimated IRFs and FEVDs for VECMs. Remarks are presented under the following headings:
Introduction to cointegrating VECMs What is cointegration? The multivariate VECM specication Trends in the Johansen VECM framework VECM estimation in Stata Selecting the number of lags Testing for cointegration Fitting a VECM Fitting VECMs with Johansens normalization Postestimation specication testing Impulseresponse functions for VECMs Forecasting with VECMs
Introduction to cointegrating VECMs

This section provides a brief introduction to integration, cointegration, and cointegrated vector error-correction models. For more details about these topics, see Hamilton (1994), Johansen (1995), L tkepohl (2005), and Watson (1994). u What is cointegration? Standard regression techniques, such as ordinary least squares (OLS), require that the variables be covariance stationary. A variable is covariance stationary if its mean and all its autocovariances are nite and do not change over time. Cointegration analysis provides a framework for estimation, inference, and interpretation when the variables are not covariance stationary. Instead of being covariance stationary, many economic time series appear to be rst-difference stationary. This means that the level of a time series is not stationary but its rst difference is. Firstdifference stationary processes are also known as integrated processes of order 1, or I(1) processes. Covariance-stationary processes are I(0). In general, a process whose dth difference is stationary is an integrated process of order d, or I(d). The canonical example of a rst-difference stationary process is the random walk. This is a variable xt that can be written as xt = xt1 + t (1) where the t are independently and identically distributed (i.i.d.) with mean zero and a nite variance 2 . Although E[xt ] = 0 for all t, Var[xt ] = T 2 is not time invariant, so xt is not covariance stationary. Because xt = xt xt1 = t and t is covariance stationary, xt is rst-difference stationary. These concepts are important because, although conventional estimators are well behaved when applied to covariance-stationary data, they have nonstandard asymptotic distributions and different rates of convergence when applied to I(1) processes. To illustrate, consider several variants of the model yt = a + bxt + et (2) Throughout the discussion, we maintain the assumption that E[et ] = 0.
625
If both yt and xt are covariance-stationary processes, et must also be covariance stationary. As long as E[xt et ] = 0, we can consistently estimate the parameters a and b by using OLS. Furthermore, the distribution of the OLS estimator converges to a normal distribution centered at the true value as the sample size grows. If yt and xt are independent random walks and b = 0, there is no relationship between yt and xt , and (2) is called a spurious regression. Granger and Newbold (1974) performed Monte Carlo experiments and showed that the usual t statistics from OLS regression provide spurious results: given a large enough dataset, we can almost always reject the null hypothesis of the test that b = 0 even though b is in fact zero. Here the OLS estimator does not converge to any well-dened population parameter. Phillips (1986) later provided the asymptotic theory that explained the Granger and Newbold (1974) results. He showed that the random walks yt and xt are rst-difference stationary processes and that the OLS estimator does not have its usual asymptotic properties when the variables are rst-difference stationary. Because yt and xt are covariance stationary, a simple regression of yt on xt appears to be a viable alternative. However, if yt and xt cointegrate, as dened below, the simple regression of yt on xt is misspecied. If yt and xt are I(1) and b = 0, et could be either I(0) or I(1). Phillips and Durlauf (1986) have derived the asymptotic theory for the OLS estimator when et is I(1), though it has not been widely used in applied work. More interesting is the case in which et = yt a bxt is I(0). yt and xt are then said to be cointegrated. Two variables are cointegrated if each is an I(1) process but a linear combination of them is an I(0) process. It is not possible for yt to be a random walk and xt and et to be covariance stationary. As Granger (1981) pointed out, because a random walk cannot be equal to a covariance-stationary process, the equation does not balance. An equation balances when the processes on each side of the equal sign are of the same order of integration. Before attacking any applied problem with integrated variables, make sure that the equation balances before proceeding. An example from Engle and Granger (1987) provides more intuition. Redene yt and xt to be
yt + xt = t , yt + xt = t ,
= t1 + t t = t1 + t ,
t
|| < 1
(3) (4)
where t and t are i.i.d. disturbances over time that are correlated with each other. Because t is I(1), (3) and (4) imply that both xt and yt are I(1). The condition that || < 1 implies that t and yt + xt are I(0). Thus yt and xt cointegrate, and (1, ) is the cointegrating vector. Using a bit of algebra, we can rewrite (3) and (4) as
yt =zt1 + 1t xt = zt1 + 2t
(5) (6)
where = (1)/(), zt = yt +xt , and 1t and 2t are distinct, stationary, linear combinations of t and t . This representation is known as the vector error-correction model (VECM). One can think of zt = 0 as being the point at which yt and xt are in equilibrium. The coefcients on zt1 describe how yt and xt adjust to zt1 being nonzero, or out of equilibrium. zt is the error in the system, and (5) and (6) describe how system adjusts or corrects back to the equilibrium. As goes to 1, the system degenerates into a pair of correlated random walks. The VECM parameterization highlights this point, because 0 as 1.
626
If we knew , we would know zt , and we could work with the stationary system of (5) and . Although knowing seems silly, we can conduct much of the analysis as if we knew because there is an estimator for the cointegrating parameter that converges to its true value at a faster rate than the estimator for the adjustment parameters and . The denition of a bivariate cointegrating relation requires simply that there exist a linear combination of the I(1) variables that is I(0). If yt and xt are I(1) and there are two nite real numbers a = 0 and b = 0, such that ayt + bxt is I(0), then yt and xt are cointegrated. Although there are two parameters, a and b, only one will be identiable because if ayt + bxt is I(0), so is cayt + cbxt for any nite, nonzero, real number c. Obtaining identication in the bivariate case is relatively simple. The coefcient on yt in (4) is unity. This natural construction of the model placed the necessary identication restriction on the cointegrating vector. As we discuss below, identication in the multivariate case is more involved. If yt is a K 1 vector of I(1) variables and there exists a vector , such that yt is a vector of I(0) variables, then yt is said to be cointegrating of order (1,0) with cointegrating vector . We say that the parameters in are the parameters in the cointegrating equation. For a vector of length K , there may be at most K 1 distinct cointegrating vectors. Engle and Granger (1987) provide a more general denition of cointegration, but this one is sufcient for our purposes. The multivariate VECM specication In practice, most empirical applications analyze multivariate systems, so the rest of our discussion focuses on that case. Consider a VAR with p lags
yt = v + A1 yt1 + A2 yt2 + + Ap ytp +
(7)
where yt is a K 1 vector of variables, v is a K 1 vector of parameters, A1 Ap are K K matrices of parameters, and t is a K 1 vector of disturbances. t has mean 0, has covariance matrix , and is i.i.d. normal over time. Any VAR(p) can be rewritten as a VECM. Using some algebra, we can rewrite (7) in VECM form as
p1
yt = v + yt1 +
i=1
i yti +
(8)
where =
j=p j=1
Aj Ik and i =
j=p j=i+1
Aj . The v and
in (7) and (8) are identical.
Engle and Granger (1987) show that if the variables yt are I(1) the matrix in (8) has rank 0 r < K , where r is the number of linearly independent cointegrating vectors. If the variables cointegrate, 0 < r < K and (8) shows that a VAR in rst differences is misspecied because it omits the lagged level term yt1 . Assume that has reduced rank 0 < r < K so that it can be expressed as = , where and are both r K matrices of rank r. Without further restrictions, the cointegrating vectors are not identied: the parameters (, ) are indistinguishable from the parameters (Q, Q1 ) for any r r nonsingular matrix Q. Because only the rank of is identied, the VECM is said to identify the rank of the cointegrating space, or equivalently, the number of cointegrating vectors. In practice, the estimation of the parameters of a VECM requires at least r2 identication restrictions. Statas vec command can apply the conventional Johansen restrictions discussed below or use constraints that the user supplies. The VECM in (8) also nests two important special cases. If the variables in yt are I(1) but not cointegrated, is a matrix of zeros and thus has rank 0. If all the variables are I(0), has full rank K.
627
There are several different frameworks for estimation and inference in cointegrating systems. Although the methods in Stata are based on the maximum likelihood (ML) methods developed by Johansen (1988, 1991, 1995), other useful frameworks have been developed by Park and Phillips (1988, 1989); Sims, Stock, and Watson (1990); Stock (1987); and Stock and Watson (1988); among others. The ML framework developed by Johansen was independently developed by Ahn and Reinsel (1990). Maddala and Kim (1998) and Watson (1994) survey all these methods. The cointegration methods in Stata are based on Johansens maximum likelihood framework because it has been found to be particularly useful in several comparative studies, including Gonzalo (1994) and Hubrich, L tkepohl, u and Saikkonen (2001). Trends in the Johansen VECM framework Deterministic trends in a cointegrating VECM can stem from two distinct sources; the mean of the cointegrating relationship and the mean of the differenced series. Allowing for a constant and a linear trend and assuming that there are r cointegrating relations, we can rewrite the VECM in (8) as
p1
yt = yt1 +
i=1
i yti + v + t +
(9)
where is a K 1 vector of parameters. Because (9) models the differences of the data, the constant implies a linear time trend in the levels, and the time trend t implies a quadratic time trend in the levels of the data. Often we may want to include a constant or a linear time trend for the differences without allowing for the higher-order trend that is implied for the levels of the data. VECMs exploit the properties of the matrix to achieve this exibility. Because is a K r rank matrix, we can rewrite the deterministic components in (9) as
v = + t = t + t
(10a) (10b)
where and are r 1 vectors of parameters and and are K 1 vectors of parameters. is orthogonal to , and is orthogonal to ; that is, = 0 and = 0, allowing us to rewrite (9) as
p1
yt = ( yt1 + + t) +
i=1
i yti + + t +
(11)
Placing restrictions on the trend terms in (11) yields ve cases.

CASE 1: Unrestricted trend
If no restrictions are placed on the trend parameters, (11) implies that there are quadratic trends in the levels of the variables and that the cointegrating equations are stationary around time trends (trend stationary).
CASE 2: Restricted trend,
=0
By setting = 0, we assume that the trends in the levels of the data are linear but not quadratic. This specication allows the cointegrating equations to be trend stationary.
CASE 3: Unrestricted constant,
= 0 and = 0
By setting = 0 and = 0, we exclude the possibility that the levels of the data have quadratic trends, and we restrict the cointegrating equations to be stationary around constant means. Because is not restricted to zero, this specication still puts a linear time trend in the levels of the data.
628
CASE 4: Restricted constant,
= 0, = 0, and = 0
By adding the restriction that = 0, we assume there are no linear time trends in the levels of the data. This specication allows the cointegrating equations to be stationary around a constant mean, but it allows no other trends or constant terms.
CASE 5: No trend,
= 0, = 0, = 0, and = 0
This specication assumes that there are no nonzero means or trends. It also assumes that the cointegrating equations are stationary with means of zero and that the differences and the levels of the data have means of zero. This exibility does come at a price. Below we discuss testing procedures for determining the number of cointegrating equations. The asymptotic distribution of the LR for hypotheses about r changes with the trend specication, so we must rst specify a trend specication. A combination of theory and graphical analysis will aid in specifying the trend before proceeding with the analysis.
VECM estimation in Stata

We provide an overview of the vec commands in Stata through an extended example. We have monthly data on the average selling prices of houses in four cities in Texas: Austin, Dallas, Houston, and San Antonio. In the dataset, these average housing prices are contained in the variables austin, dallas, houston, and sa. The series begin in January of 1990 and go through December 2003, for a total of 168 observations. The following graph depicts our data.
11.2 1990m1
11.4
11.6
11.8
12
12.2
1995m1 t ln of house prices in austin ln of house prices in houston
2000m1
2005m1
ln of house prices in dallas ln of house prices in san antonio
The plots on the graph indicate that all the series are trending and potential I(1) processes. In a competitive market, the current and past prices contain all the information available, so tomorrows price will be a random walk from todays price. Some researchers may opt to use [TS] dfgls to investigate the presence of a unit root in each series, but the test for cointegration we use includes the case in which all the variables are stationary, so we defer formal testing until we test for cointegration. The time trends in the data appear to be approximately linear, so we will specify trend(constant) when modeling these series, which is the default with vec. The next graph shows just Dallas and Houstons data, so we can more carefully examine their relationship.

12.2
629
11.2 1990m1 1991m11
11.4
11.6
11.8
12
1994m1
1996m1 t
1998m1
2000m1
2002m1
2004m1
ln of house prices in dallas
ln of house prices in houston
Except for the crash at the end of 1991, housing prices in Dallas and Houston appear closely related. Although average prices in the two cities will differ because of resource variations and other factors, if the housing markets become too dissimilar, people and businesses will migrate, bringing the average housing prices back toward each other. We therefore expect the series of average housing prices in Houston to be cointegrated with the series of average housing prices in Dallas. Selecting the number of lags To test for cointegration or t cointegrating VECMs, we must specify how many lags to include. Building on the work of Tsay (1984) and Paulsen (1984), Nielsen (2001) has shown that the methods implemented in varsoc can be used to determine the lag order for a VAR model with I(1) variables. As can be seen from (9), the order of the corresponding VECM is always one less than the VAR. vec makes this adjustment automatically, so we will always refer to the order of the underlying VAR. The output below uses varsoc to determine the lag order of the VAR of the average housing prices in Dallas and Houston.
. use http://www.stata-press.com/data/r12/txhprice . varsoc dallas houston Selection-order criteria Sample: 1990m5 - 2003m12 Number of obs lag 0 1 2 3 4 LL 299.525 577.483 590.978 593.437 596.364 LR df p FPE AIC HQIC -3.61301 -6.92326 -7.00837* -6.95888 -6.9151
= SBIC
164
555.92 26.991* 4.918 5.8532
4 4 4 4
0.000 0.000 0.296 0.210
.000091 -3.62835 3.2e-06 -6.9693 2.9e-06* -7.0851* 2.9e-06 -7.06631 3.0e-06 -7.05322
-3.59055 -6.85589 -6.89608* -6.80168 -6.71299
dallas houston _cons
We will use two lags for this bivariate model because the HannanQuinn information criterion (HQIC) method, Schwarz Bayesian information criterion (SBIC) method, and sequential likelihood-ratio (LR) test all chose two lags, as indicated by the * in the output. The reader can verify that when all four cities data are used, the LR test selects three lags, the HQIC method selects two lags, and the SBIC method selects one lag. We will use three lags in our four-variable model.
630
Testing for cointegration The tests for cointegration implemented in vecrank are based on Johansens method. If the log likelihood of the unconstrained model that includes the cointegrating equations is signicantly different from the log likelihood of the constrained model that does not include the cointegrating equations, we reject the null hypothesis of no cointegration. Here we use vecrank to determine the number of cointegrating equations:
. vecrank dallas houston Johansen tests for cointegration Trend: constant Number of obs = Sample: 1990m3 - 2003m12 Lags = 5% trace critical statistic value 46.8252 15.41 0.1785* 3.76 166 2
maximum rank 0 1 2
parms 6 9 10
LL 576.26444 599.58781 599.67706
eigenvalue . 0.24498 0.00107
Besides presenting information about the sample size and time span, the header indicates that test statistics are based on a model with two lags and a constant trend. The body of the table presents test statistics and their critical values of the null hypotheses of no cointegration (line 1) and one or fewer cointegrating equations (line 2). The eigenvalue shown on the last line is used to compute the trace statistic in the line above it. Johansens testing procedure starts with the test for zero cointegrating equations (a maximum rank of zero) and then accepts the rst null hypothesis that is not rejected. In the output above, we strongly reject the null hypothesis of no cointegration and fail to reject the null hypothesis of at most one cointegrating equation. Thus we accept the null hypothesis that there is one cointegrating equation in the bivariate model. Using all four series and a model with three lags, we nd that there are two cointegrating relationships.
. vecrank austin dallas houston sa, lag(3) Johansen tests for cointegration Trend: constant Number of obs = Sample: 1990m4 - 2003m12 Lags = 5% trace critical statistic value 101.6070 47.21 41.6768 29.68 9.8865* 15.41 0.3354 3.76 165 3
maximum rank 0 1 2 3 4
parms 36 43 48 51 52
LL 1107.7833 1137.7484 1153.6435 1158.4191 1158.5868
eigenvalue . 0.30456 0.17524 0.05624 0.00203
Fitting a VECM vec estimates the parameters of cointegrating VECMs. There are four types of parameters of interest: 1. The parameters in the cointegrating equations 2. The adjustment coefcients 3. The short-run coefcients 4. Some standard functions of and that have useful interpretations
631
Although all four types are discussed in [TS] vec, here we discuss only types 13 and how they appear in the output of vec. Having determined that there is a cointegrating equation between the Dallas and Houston series, we now want to estimate the parameters of a bivariate cointegrating VECM for these two series by using vec.
. vec dallas houston Vector error-correction model Sample: 1990m3 - 2003m12 Log likelihood = Det(Sigma_ml) = Equation D_dallas D_houston 599.5878 2.50e-06 Parms 4 4
RMSE .038546 .045348
R-sq 0.1692 0.3737
No. of obs AIC HQIC SBIC chi2 P>chi2 32.98959 96.66399 0.0000 0.0000
= 166 = -7.115516 = -7.04703 = -6.946794
Coef. D_dallas _ce1 L1. dallas LD. houston LD. _cons D_houston _ce1 L1. dallas LD. houston LD. _cons
Std. Err.
P>|z|
-.3038799
.0908504
-3.34
0.001
-.4819434
-.1258165
-.1647304
.0879356
-1.87
0.061
-.337081
.0076202
-.0998368 .0056128
.0650838 .0030341
-1.53 1.85
0.125 0.064
-.2273988 -.0003339
.0277251 .0115595
.5027143
.1068838
4.70
0.000
.2932258
.7122028
-.0619653
.1034547
-0.60
0.549
-.2647327
.1408022
-.3328437 .0033928
.07657 .0035695
-4.35 0.95
0.000 0.342
-.4829181 -.0036034
-.1827693 .010389
Cointegrating equations Equation Parms _ce1 Identification: 1
chi2 1640.088
P>chi2 0.0000
beta is exactly identified Johansen normalization restriction imposed Coef. Std. Err. z P>|z| [95% Conf. Interval]
beta _ce1 dallas houston _cons
1 -.8675936 -1.688897
. .0214231 .
. -40.50 .
. 0.000 .
. -.9095821 .
. -.825605 .
632
The header contains information about the sample, the t of each equation, and overall model t statistics. The rst estimation table contains the estimates of the short-run parameters, along with their standard errors, z statistics, and condence intervals. The two coefcients on L. ce1 are the parameters in the adjustment matrix for this model. The second estimation table contains the estimated parameters of the cointegrating vector for this model, along with their standard errors, z statistics, and condence intervals. Using our previous notation, we have estimated = (0.304, 0.503) and = (1, 0.868)
v = (0.0056, 0.0034)
0.165 0.0998 0.062 0.333
Overall, the output indicates that the model ts well. The coefcient on houston in the cointegrating equation is statistically signicant, as are the adjustment parameters. The adjustment parameters in this bivariate example are easy to interpret, and we can see that the estimates have the correct signs and imply rapid adjustment toward equilibrium. When the predictions from the cointegrating equation are positive, dallas is above its equilibrium value because the coefcient on dallas in the cointegrating equation is positive. The estimate of the coefcient [D dallas]L. ce1 is .3. Thus when the average housing price in Dallas is too high, it quickly falls back toward the Houston level. The estimated coefcient [D houston]L. ce1 of .5 implies that when the average housing price in Dallas is too high, the average price in Houston quickly adjusts toward the Dallas level at the same time that the Dallas prices are adjusting. Fitting VECMs with Johansens normalization As discussed by Johansen (1995), if there are r cointegrating equations, then at least r2 restrictions are required to identify the free parameters in . Johansen proposed a default identication scheme that has become the conventional method of identifying models in the absence of theoretically justied restrictions. Johansens identication scheme is = (Ir , ) where Ir is the r r identity matrix and is an (K r) r matrix of identied parameters. vec applies Johansens normalization by default. To illustrate, we t a VECM with two cointegrating equations and three lags on all four series. We are interested only in the estimates of the parameters in the cointegrating equations, so we can specify the noetable option to suppress the estimation table for the adjustment and short-run parameters.

. vec austin dallas houston sa, lags(3) rank(2) noetable Vector error-correction model Sample: 1990m4 - 2003m12 No. of obs AIC Log likelihood = 1153.644 HQIC Det(Sigma_ml) = 9.93e-12 SBIC Cointegrating equations Equation Parms chi2 P>chi2 _ce1 _ce2 Identification: 2 2 586.3044 2169.826 0.0000 0.0000
633
= 165 = -13.40174 = -13.03496 = -12.49819
beta is exactly identified Johansen normalization restrictions imposed Coef. Std. Err. z P>|z| [95% Conf. Interval]
beta _ce1 austin dallas houston sa _cons _ce2 austin dallas houston sa _cons
1 -2.37e-17 -.2623782 -1.241805 5.577099
. . .1893625 .229643 .
. . -1.39 -5.41 .
. . 0.166 0.000 .
. . -.6335219 -1.691897 .
. . .1087655 -.7917128 .
-1.58e-17 1 -1.095652 .2883986 -2.351372
. . .0669898 .0812396 .
. . -16.36 3.55 .
. . 0.000 0.000 .
. . -1.22695 .1291718 .
. . -.9643545 .4476253 .
The Johansen identication scheme has placed four constraints on the parameters in : [ ce1]austin=1, [ ce1]dallas=0, [ ce2]austin=0, and [ ce2]dallas=1. (The computational method used imposes zero restrictions that are numerical rather than exact. The values 3.48e 17 and 1.26e17 are indistinguishable from zero.) We interpret the results of the rst equation as indicating the existence of an equilibrium relationship between the average housing price in Austin and the average prices of houses in Houston and San Antonio. The Johansen normalization restricted the coefcient on dallas to be unity in the second cointegrating equation, but we could instead constrain the coefcient on houston. Both sets of restrictions dene just-identied models, so tting the model with the latter set of restrictions will yield the same maximized log likelihood. To impose the alternative set of constraints, we use the constraint command.
. constraint 1 [_ce1]austin = 1 . constraint 2 [_ce1]dallas = 0 . constraint 3 [_ce2]austin = 0 . constraint 4 [_ce2]houston = 1
634

. vec austin dallas houston sa, lags(3) rank(2) noetable bconstraints(1/4) Iteration 1: log likelihood = 1148.8745 (output omitted ) Iteration 25: log likelihood = 1153.6435 Vector error-correction model Sample: 1990m4 - 2003m12 No. of obs = 165 AIC = -13.40174 Log likelihood = 1153.644 HQIC = -13.03496 Det(Sigma_ml) = 9.93e-12 SBIC = -12.49819 Cointegrating equations Equation Parms chi2 P>chi2 _ce1 _ce2 2 2 586.3392 3455.469 0.0000 0.0000
Identification: beta is exactly identified ( 1) [_ce1]austin = 1 ( 2) [_ce1]dallas = 0 ( 3) [_ce2]austin = 0 ( 4) [_ce2]houston = 1 beta _ce1 austin dallas houston sa _cons _ce2 austin dallas houston sa _cons 0 -.9126985 1 -.2632209 2.146094 (omitted) .0595804 . .0628791 . -15.32 . -4.19 . 0.000 . 0.000 . -1.029474 . -.3864617 . -.7959231 . -.1399802 . 1 0 -.2623784 -1.241805 5.577099 . (omitted) .1876727 .2277537 . . -1.40 -5.45 . . 0.162 0.000 . . -.6302102 -1.688194 . . .1054534 -.7954157 . Coef. Std. Err. z P>|z| [95% Conf. Interval]
Only the estimates of the parameters in the second cointegrating equation have changed, and the new estimates are simply the old estimates divided by 1.095652 because the new constraints are just an alternative normalization of the same just-identied model. With the new normalization, we can interpret the estimates of the parameters in the second cointegrating equation as indicating an equilibrium relationship between the average house price in Houston and the average prices of houses in Dallas and San Antonio. Postestimation specication testing Inference on the parameters in depends crucially on the stationarity of the cointegrating equations, so we should check the specication of the model. As a rst check, we can predict the cointegrating equations and graph them over time.
. predict ce1, ce equ(#1) . predict ce2, ce equ(#2)

. twoway line ce1 t
.4
635
.4 1990m1
Predicted cointegrated equation .2 0 .2
1995m1 t
2000m1
2005m1
. twoway line ce2 t

.2 .3 1990m1 Predicted cointegrated equation .2 .1 0 .1
1995m1 t
2000m1
2005m1
Although the large shocks apparent in the graph of the levels have clear effects on the predictions from the cointegrating equations, our only concern is the negative trend in the rst cointegrating equation since the end of 2000. The graph of the levels shows that something put a signicant brake on the growth of housing prices after 2000 and that the growth of housing prices in San Antonio slowed during 2000 but then recuperated while Austin maintained slower growth. We suspect that this indicates that the end of the high-tech boom affected Austin more severely than San Antonio. This difference is what causes the trend in the rst cointegrating equation. Although we could try to account for this effect with a more formal analysis, we will proceed as if the cointegrating equations are stationary. We can use vecstable to check whether we have correctly specied the number of cointegrating equations. As discussed in [TS] vecstable, the companion matrix of a VECM with K endogenous variables and r cointegrating equations has K r unit eigenvalues. If the process is stable, the moduli of the remaining r eigenvalues are strictly less than one. Because there is no general distribution
636
theory for the moduli of the eigenvalues, ascertaining whether the moduli are too close to one can be difcult.
. vecstable, graph Eigenvalue stability condition Eigenvalue 1 1 -.6698661 .3740191 .3740191 -.386377 -.386377 .540117 -.0749239 -.0749239 -.2023955 .09923966 Modulus 1 1 .669866 .583297 .583297 .553246 .553246 .540117 .532715 .532715 .202395 .09924
+ + + -
.4475996i .4475996i .395972i .395972i .5274203i .5274203i
The VECM specification imposes 2 unit moduli.

1 1 1 .5 Imaginary 0 .5
.5
0 Real
.5
The VECM specification imposes 2 unit moduli
Because we specied the graph option, vecstable plotted the eigenvalues of the companion matrix. The graph of the eigenvalues shows that none of the remaining eigenvalues appears close to the unit circle. The stability check does not indicate that our model is misspecied. Here we use veclmar to test for serial correlation in the residuals.
. veclmar, mlag(4) Lagrange-multiplier test lag 1 2 3 4 chi2 56.8757 31.1970 30.6818 14.6493 df 16 16 16 16 Prob > chi2 0.00000 0.01270 0.01477 0.55046
637
The results clearly indicate serial correlation in the residuals. The results in Gonzalo (1994) indicate that underspecifying the number of lags in a VECM can signicantly increase the nite-sample bias in the parameter estimates and lead to serial correlation. For this reason, we ret the model with ve lags instead of three.
. vec austin dallas houston sa, lags(5) rank(2) noetable bconstraints(1/4) Iteration 1: log likelihood = 1200.5402 (output omitted ) Iteration 20: log likelihood = 1203.9465 Vector error-correction model Sample: 1990m6 - 2003m12 No. of obs AIC HQIC SBIC chi2 P>chi2 0.0000 0.0000 = 163 = -13.79075 = -13.1743 = -12.27235
Log likelihood = 1203.946 Det(Sigma_ml) = 4.51e-12 Cointegrating equations Equation Parms _ce1 _ce2 2 2
498.4682 4125.926
Identification: beta is exactly identified ( 1) [_ce1]austin = 1 ( 2) [_ce1]dallas = 0 ( 3) [_ce2]austin = 0 ( 4) [_ce2]houston = 1 beta _ce1 austin dallas houston sa _cons _ce2 austin dallas houston sa _cons 0 -.932048 1 -.2363915 2.065719 (omitted) .0564332 . .0599348 . -16.52 . -3.94 . 0.000 . 0.000 . -1.042655 . -.3538615 . -.8214409 . -.1189215 . 1 0 -.6525574 -.6960166 3.846275 . (omitted) .2047061 .2494167 . . -3.19 -2.79 . . 0.001 0.005 . . -1.053774 -1.184864 . . -.2513407 -.2071688 . Coef. Std. Err. z P>|z| [95% Conf. Interval]
Comparing these results with those from the previous model reveals that 1. there is now evidence that the coefcient [ ce1]houston is not equal to zero, 2. the two sets of estimated coefcients for the rst cointegrating equation are different, and 3. the two sets of estimated coefcients for the second cointegrating equation are similar. The assumption that the errors are independently, identically, and normally distributed with zero mean and nite variance allows us to derive the likelihood function. If the errors do not come from a normal distribution but are just independently and identically distributed with zero mean and nite variance, the parameter estimates are still consistent, but they are not efcient.
638
We use vecnorm to test the null hypothesis that the errors are normally distributed.
. qui vec austin dallas houston sa, lags(5) rank(2) bconstraints(1/4) . vecnorm Jarque-Bera test Equation D_austin D_dallas D_houston D_sa ALL Skewness test Equation D_austin D_dallas D_houston D_sa ALL Kurtosis test Equation D_austin D_dallas D_houston D_sa ALL Kurtosis 6.0807 3.6896 8.6316 3.8139 chi2 64.458 3.229 215.397 4.499 287.583 df 1 1 1 1 4 Prob > chi2 0.00000 0.07232 0.00000 0.03392 0.00000 Skewness .60265 .09996 -1.0444 .38019 chi2 9.867 0.271 29.635 3.927 43.699 df 1 1 1 1 4 Prob > chi2 0.00168 0.60236 0.00000 0.04752 0.00000 chi2 74.324 3.501 245.032 8.426 331.283 df 2 2 2 2 8 Prob > chi2 0.00000 0.17370 0.00000 0.01481 0.00000
The results indicate that we can strongly reject the null hypothesis of normally distributed errors. Most of the errors are both skewed and kurtotic. Impulseresponse functions for VECMs With a model that we now consider acceptably well specied, we can use the irf commands to estimate and interpret the IRFs. Whereas IRFs from a stationary VAR die out over time, IRFs from a cointegrating VECM do not always die out. Because each variable in a stationary VAR has a timeinvariant mean and nite, time-invariant variance, the effect of a shock to any one of these variables must die out so that the variable can revert to its mean. In contrast, the I(1) variables modeled in a cointegrating VECM are not mean reverting, and the unit moduli in the companion matrix imply that the effects of some shocks will not die out over time. These two possibilities gave rise to new terms. When the effect of a shock dies out over time, the shock is said to be transitory. When the effect of a shock does not die out over time, the shock is said to be permanent. Below we use irf create to estimate the IRFs and irf graph to graph two of the orthogonalized
IRFs.

. irf (file (file (file . irf create vec1, set(vecintro, replace) step(24) vecintro.irf created) vecintro.irf now active) vecintro.irf updated) graph oirf, impulse(austin dallas) response(sa) yline(0)
639
vec1, austin, sa
vec1, dallas, sa
.015
.01
.005
10
20
30
10
20
30
step
The graphs indicate that an orthogonalized shock to the average housing price in Austin has a permanent effect on the average housing price in San Antonio but that an orthogonalized shock to the average price of housing in Dallas has a transitory effect. According to this model, unexpected shocks that are local to the Austin housing market will have a permanent effect on the housing market in San Antonio, but unexpected shocks that are local to the Dallas housing market will have only a transitory effect on the housing market in San Antonio. Forecasting with VECMs Cointegrating VECMs are also used to produce forecasts of both the rst-differenced variables and the levels of the variables. Comparing the variances of the forecast errors of stationary VARs with those from a cointegrating VECM reveals a fundamental difference between the two models. Whereas the variances of the forecast errors for a stationary VAR converge to a constant as the prediction horizon grows, the variances of the forecast errors for the levels of a cointegrating VECM diverge with the forecast horizon. (See sec. 6.5 of L tkepohl [2005] for more about this result.) Because all u the variables in the model for the rst differences are stationary, the forecast errors for the dynamic forecasts of the rst differences remain nite. In contrast, the forecast errors for the dynamic forecasts of the levels diverge to innity. We use fcast compute to obtain dynamic forecasts of the levels and fcast graph to graph these dynamic forecasts, along with their asymptotic condence intervals.
640

. tsset time variable: t, 1990m1 to 2003m12 delta: 1 month . fcast compute m1_, step(24) . fcast graph m1_austin m1_dallas m1_houston m1_sa
Forecast for austin

12.1 12.2 12.3 12.4 12.5 12 12.1 12.2 12.3 12.4
Forecast for dallas
Forecast for houston

11.9 12 12.1 12.2 12.3 11.7 11.8 11.9 12 12.1
Forecast for sa
2004m1 2004m7 2005m1 2005m7 2006m1 2004m1 2004m7 2005m1 2005m7 2006m1
95% CI
forecast
As expected, the widths of the condence intervals grow with the forecast horizon.
References
Ahn, S. K., and G. C. Reinsel. 1990. Estimation for partially nonstationary multivariate autoregressive models. Journal of the American Statistical Association 85: 813823. Engle, R. F., and C. W. J. Granger. 1987. Co-integration and error correction: Representation, estimation, and testing. Econometrica 55: 251276. Gonzalo, J. 1994. Five alternative methods of estimating long-run equilibrium relationships. Journal of Econometrics 60: 203233. Granger, C. W. J. 1981. Some properties of time series data and their use in econometric model specication. Journal of Econometrics 16: 121130. Granger, C. W. J., and P. Newbold. 1974. Spurious regressions in econometrics. Journal of Econometrics 2: 111120. Hamilton, J. D. 1994. Time Series Analysis. Princeton: Princeton University Press. Hubrich, K., H. L tkepohl, and P. Saikkonen. 2001. A review of systems cointegration tests. Econometric Reviews u 20: 247318. Johansen, S. 1988. Statistical analysis of cointegration vectors. Journal of Economic Dynamics and Control 12: 231254. . 1991. Estimation and hypothesis testing of cointegration vectors in Gaussian vector autoregressive models. Econometrica 59: 15511580. . 1995. Likelihood-Based Inference in Cointegrated Vector Autoregressive Models. Oxford: Oxford University Press. L tkepohl, H. 2005. New Introduction to Multiple Time Series Analysis. New York: Springer. u Maddala, G. S., and I.-M. Kim. 1998. Unit Roots, Cointegration, and Structural Change. Cambridge: Cambridge University Press.
641
Nielsen, B. 2001. Order determination in general vector autoregressions. Working paper, Department of Economics, University of Oxford and Nufeld College. http://ideas.repec.org/p/nuf/econwp/0110.html. Park, J. Y., and P. C. B. Phillips. 1988. Statistical inference in regressions with integrated processes: Part I. Econometric Theory 4: 468497. . 1989. Statistical inference in regressions with integrated processes: Part II. Econometric Theory 5: 95131. Paulsen, J. 1984. Order determination of multivariate autoregressive time series with unit roots. Journal of Time Series Analysis 5: 115127. Phillips, P. C. B. 1986. Understanding spurious regressions in econometrics. Journal of Econometrics 33: 311340. Phillips, P. C. B., and S. N. Durlauf. 1986. Multiple time series regressions with integrated processes. Review of Economic Studies 53: 473495. Sims, C. A., J. H. Stock, and M. W. Watson. 1990. Inference in linear time series models with some unit roots. Econometrica 58: 113144. Stock, J. H. 1987. Asymptotic properties of least squares estimators of cointegrating vectors. Econometrica 55: 10351056. Stock, J. H., and M. W. Watson. 1988. Testing for common trends. Journal of the American Statistical Association 83: 10971107. Tsay, R. S. 1984. Order selection in nonstationary autoregressive models. Annals of Statistics 12: 14251433. Watson, M. W. 1994. Vector autoregressions and cointegration. In Vol. IV of Handbook of Econometrics, ed. R. F. Engle and D. L. McFadden. Amsterdam: Elsevier.
Also see
[TS] vec Vector error-correction models [TS] irf Create and analyze IRFs, dynamic-multiplier functions, and FEVDs
Title
vec Vector error-correction models
Syntax
vec varlist options
Model
if
in
, options Description use # cointegrating equations; default is rank(1) use # for the maximum lag in underlying VAR model include an unrestricted constant in model; the default include a restricted constant in model include a linear trend in the cointegrating equations and a quadratic trend in the undifferenced data include a restricted trend in model do not include a trend or a constant place constraintsbc on cointegrating vectors place constraintsac on adjustment parameters include normalized seasonal indicator variables varlistsi do not perform checks and corrections for collinearity among lags of dependent variables set condence level; default is level(95) do not report parameters in the cointegrating equations do not report the likelihood-ratio test of overidentifying restrictions report adjustment parameters in separate table report parameters in = do not report elements of matrix report parameters in the moving-average impact matrix do not report adjustment and short-run parameters force reporting of short-run, beta, and alpha parameters when the parameters in beta are not identied; advanced option do not display constraints control column formats, row spacing, and line width control the maximization process; seldom used display legend instead of statistics
rank(#) lags(#) trend(constant) trend(rconstant) trend(trend) trend(rtrend) trend(none) bconstraints(constraintsbc ) aconstraints(constraintsac )

Adv. model
sindicators(varlistsi ) noreduce
Reporting
level(#) nobtable noidtest alpha pi noptable mai noetable dforce nocnsreport display options
Maximization
642

vec does not allow gaps in the data. You must tsset your data before using vec; see [TS] tsset.
643
varlist must contain at least two variables and may contain time-series operators; see [U] 11.4.4 Time-series varlists. by, rolling, statsby, and xi are allowed; see [U] 11.1.10 Prex commands. coeflegend does not appear in the dialog box. See [U] 20 Estimation and postestimation commands for more capabilities of estimation commands.
Menu
Statistics
>
>
Vector error-correction model (VECM)
Description
vec ts a type of vector autoregression in which some of the variables are cointegrated by using Johansens (1995) maximum likelihood method. Constraints may be placed on the parameters in the cointegrating equations or on the adjustment terms. See [TS] vec intro for a list of commands that are used in conjunction with vec.
Options

Model
rank(#) species the number of cointegrating equations; rank(1) is the default. lags(#) species the maximum lag to be included in the underlying VAR model. The maximum lag in a VECM is one smaller than the maximum lag in the corresponding VAR in levels; the number of lags must be greater than zero but small enough so that the degrees of freedom used up by the model are fewer than the number of observations. The default is lags(2). trend(trend spec) species which of Johansens ve trend specications to include in the model. These specications are discussed in Specication of constants and trends below. The default is trend(constant). bconstraints(constraintsbc ) species the constraints to be placed on the parameters of the cointegrating equations. When no constraints are placed on the adjustment parametersthat is, when the aconstraints() option is not speciedthe default is to place the constraints dened by Johansens normalization on the parameters of the cointegrating equations. When constraints are placed on the adjustment parameters, the default is not to place constraints on the parameters in the cointegrating equations. aconstraints(constraintsac ) species the constraints to be placed on the adjustment parameters. By default, no constraints are placed on the adjustment parameters.
Adv. model
sindicators(varlistsi ) species the normalized seasonal indicator variables to include in the model. The indicator variables specied in this option must be normalized as discussed in Johansen (1995). If the indicators are not properly normalized, the estimator of the cointegrating vector does not converge to the asymptotic distribution derived by Johansen (1995). More details about how these variables are handled are provided in Methods and formulas. sindicators() cannot be specied with trend(none) or with trend(rconstant).
644
noreduce causes vec to skip the checks and corrections for collinearity among the lags of the dependent variables. By default, vec checks to see whether the current lag specication causes some of the regressions performed by vec to contain perfectly collinear variables; if so, it reduces the maximum lag until the perfect collinearity is removed.
Reporting
level(#); see [R] estimation options. nobtable suppresses the estimation table for the parameters in the cointegrating equations. By default, vec displays the estimation table for the parameters in the cointegrating equations. noidtest suppresses the likelihood-ratio test of the overidentifying restrictions, which is reported by default when the model is overidentied. alpha displays a separate estimation table for the adjustment parameters, which is not displayed by default. pi displays a separate estimation table for the parameters in = , which is not displayed by default. noptable suppresses the estimation table for the elements of the matrix, which is displayed by default when the parameters in the cointegrating equations are not identied. mai displays a separate estimation table for the parameters in the moving-average impact matrix, which is not displayed by default. noetable suppresses the main estimation table that contains information about the estimated adjustment parameters and the short-run parameters, which is displayed by default. dforce displays the estimation tables for the short-run parameters and and if the last two are requestedwhen the parameters in are not identied. By default, when the specied constraints do not identify the parameters in the cointegrating equations, estimation tables are displayed only for and the MAI. nocnsreport; see [R] estimation options. display options: vsquish, cformat(% fmt), pformat(% fmt), sformat(% fmt), and nolstretch; see [R] estimation options.
Maximization
maximize options: iterate(#), nolog, trace, toltrace, tolerance(#), ltolerance(#), afrom(matrixa ), and bfrom(matrixb ); see [R] maximize. toltrace displays the relative differences for the log likelihood and the coefcient vector at every iteration. This option cannot be specied if no constraints are dened or if nolog is specied. afrom(matrixa ) species a 1 (K r) row vector with starting values for the adjustment parameters, where K is the number of endogenous variables and r is the number of cointegrating equations specied in the rank() option. The starting values should be ordered as they are reported in e(alpha). This option cannot be specied if no constraints are dened. bfrom(matrixb ) species a 1 (m1 r) row vector with starting values for the parameters of the cointegrating equations, where m1 is the number of variables in the trend-augmented system and r is the number of cointegrating equations specied in the rank() option. (See Methods and formulas for more details about m1 .) The starting values should be ordered as they are reported in e(betavec). As discussed in Methods and formulas, for some trend specications, e(beta) contains parameter estimates that are not obtained directly from the optimization algorithm. bfrom() should specify only starting values for the parameters reported in e(betavec). This option cannot be specied if no constraints are dened.
645
The following option is available with vec but is not shown in the dialog box: coeflegend; see [R] estimation options.
Remarks
Introduction Specication of constants and trends Collinearity
Introduction
VECMs are used to model the stationary relationships between multiple time series that contain unit roots. vec implements Johansens approach for estimating the parameters of a VECM.
[TS] vec intro reviews the basics of integration and cointegration and highlights why we need special methods for modeling the relationships between processes that contain unit roots. This manual entry assumes familiarity with the material in [TS] vec intro and provides examples illustrating how to use the vec command. See Johansen (1995) and Hamilton (1994) for more in-depth introductions to cointegration analysis.
Example 1
This example uses annual data on the average per-capita disposable personal income in the eight U.S. Bureau of Economic Analysis (BEA) regions of the United States. We use data from 19482002 in logarithms. Unit-root tests on these series fail to reject the null hypothesis that per-capita disposable income in each region contains a unit root. Because capital and labor can move easily between the different regions of the United States, we would expect that no one series will diverge from all the remaining series and that cointegrating relationships exist. Below we graph the natural logs of average disposal income in the New England and the Southeast regions.
. use http://www.stata-press.com/data/r12/rdinc . line ln_ne ln_se year
7 1950
10
11
1960
1970 year ln(new_england)
1980
1990 ln(southeast)
2000
646
The graph indicates a differential between the two series that shrinks between 1960 and about 1980 and then grows until it stabilizes around 1990. We next estimate the parameters of a bivariate VECM with one cointegrating relationship.
. vec ln_ne ln_se Vector error-correction model Sample: 1950 - 2002 300.6224 4.06e-08 Parms 4 4 RMSE .017896 .018723 R-sq 0.9313 0.9292 No. of obs AIC HQIC SBIC chi2 664.4668 642.7179 P>chi2 0.0000 0.0000 = 53 = -11.00462 = -10.87595 = -10.67004
Log likelihood = Det(Sigma_ml) = Equation D_ln_ne D_ln_se
Coef. D_ln_ne _ce1 L1. ln_ne LD. ln_se LD. _cons D_ln_se _ce1 L1. ln_ne LD. ln_se LD. _cons -.3543935 -.4337524
Std. Err.
P>|z|
.0721365
-6.01
0.000
-.5751373
-.2923675
.7168658
.1889085
3.79
0.000
.3466119
1.08712
-.6748754 -.0019846
.2117975 .0080291
-3.19 -0.25
0.001 0.805
-1.089991 -.0177214
-.2597599 .0137521
.0754725
-4.70
0.000
-.5023168
-.2064701
.3366786
.1976448
1.70
0.088
-.050698
.7240553
-.1605811 .002429
.2215922 .0084004
-0.72 0.29
0.469 0.772
-.5948939 -.0140355
.2737317 .0188936
Cointegrating equations Equation _ce1 Identification: Parms 1 chi2 29805.02 P>chi2 0.0000
beta is exactly identified Johansen normalization restriction imposed Coef. Std. Err. z P>|z| [95% Conf. Interval]
beta _ce1 ln_ne ln_se _cons
1 -.9433708 -.8964065
. .0054643 .
. -172.64 .
. 0.000 .
. -.9540807 .
. -.9326609 .
The default output has three parts. The header provides information about the sample, the model t, and the identication of the parameters in the cointegrating equation. The main estimation table
647
contains the estimates of the short-run parameters, along with their standard errors and condence intervals. The second estimation table reports the estimates of the parameters in the cointegrating equation, along with their standard errors and condence intervals. The results indicate strong support for a cointegrating equation such that ln ne .943 ln se .896 should be a stationary series. Identication of the parameters in the cointegrating equation is achieved by constraining some of them to be xed, and xed parameters do not have standard errors. In this example, the coefcient on ln ne has been normalized to 1, so its standard error is missing. As discussed in Methods and formulas, the constant term in the cointegrating equation is not directly estimated in this trend specication but rather is backed out from other estimates. Not all the elements of the VCE that correspond to this parameter are readily available, so the standard error for the cons parameter is missing. To get a better idea of how our model ts, we predict the cointegrating equation and graph it over time:
. predict ce, ce . line ce year
.05 .25 1950 Predicted cointegrated equation .2 .15 .1
1960
1970 year
1980
1990
2000
Although the predicted cointegrating equation has the right appearance for the time before the mid-1960s, afterward the predicted cointegrating equation does not look like a stationary series. A better model would account for the trends in the size of the differential.
As discussed in [TS] vec intro, simply normalizing one of the coefcients to be one is sufcient to identify the parameters of the single cointegrating vector. When there is more than one cointegrating equation, more restrictions are required.
Example 2
We have data on monthly unemployment rates in Indiana, Illinois, Kentucky, and Missouri from January 1978 through December 2003. We suspect that factor mobility will keep the unemployment rates in equilibrium. The following graph plots the data.
648

. use http://www.stata-press.com/data/r12/urates, clear . line missouri indiana kentucky illinois t
2 1980m1
10
12
1985m1
1990m1 t missouri kentucky
1995m1
2000m1
2005m1
indiana illinois
The graph shows that although the series do appear to move together, the relationship is not as clear as in the previous example. There are periods when Indiana has the highest rate and others when Indiana has the lowest rate. Although the Kentucky rate moves closely with the other series for most of the sample, there is a period in the mid-1980s when the unemployment rate in Kentucky does not fall at the same rate as the other series. We will model the series with two cointegrating equations and no linear or quadratic time trends in the original series. Because we are focusing on the cointegrating vectors, we use the noetable option to suppress displaying the short-run estimation table.

. vec missouri indiana kentucky illinois, trend(rconstant) rank(2) > noetable Vector error-correction model Sample: 1978m5 - 2003m12 No. of obs AIC Log likelihood = 417.1314 HQIC Det(Sigma_ml) = 7.83e-07 SBIC Cointegrating equations Equation Parms chi2 P>chi2 _ce1 _ce2 Identification: 2 2 133.3885 195.6324 0.0000 0.0000 lags(4)
649
= 308 = -2.306048 = -2.005818 = -1.555184
beta is exactly identified Johansen normalization restrictions imposed Coef. Std. Err. z P>|z| [95% Conf. Interval]
beta _ce1 missouri indiana kentucky illinois _cons _ce2 missouri indiana kentucky illinois _cons
1 -2.19e-18 .3493902 -1.135152 -.3880707
. . .2005537 .2069063 .4974323
. . 1.74 -5.49 -0.78
. . 0.081 0.000 0.435
. . -.0436879 -1.540681 -1.36302
. . .7424683 -.7296235 .5868787
1.09e-16 1 .2059473 -1.51962 2.92857
. . .2718678 .2804792 .6743122
. . 0.76 -5.42 4.34
. . 0.449 0.000 0.000
. . -.3269038 -2.069349 1.606942
. . .7387985 -.9698907 4.250197
Except for the coefcients on kentucky in the two cointegrating equations and the constant term in the rst, all the parameters are signicant at the 5% level. We can ret the model with the Johansen normalization and the overidentifying constraint that the coefcient on kentucky in the second cointegrating equation is zero.
. . . . constraint constraint constraint constraint 1 2 3 4 [_ce1]missouri [_ce1]indiana [_ce2]missouri [_ce2]indiana = = = = 1 0 0 1
. constraint 5 [_ce2]kentucky = 0
650

. vec missouri indiana kentucky illinois, trend(rconstant) rank(2) > noetable bconstraints(1/5) Iteration 1: log likelihood = 416.97177 (output omitted ) Iteration 20: log likelihood = 416.9744 Vector error-correction model Sample: 1978m5 - 2003m12 No. of obs AIC Log likelihood = 416.9744 HQIC Det(Sigma_ml) = 7.84e-07 SBIC Cointegrating equations Equation Parms _ce1 _ce2 Identification: ( ( ( ( ( 1) 2) 3) 4) 5) 2 1 chi2 145.233 209.9344 P>chi2 0.0000 0.0000 lags(4)
= 308 = -2.311522 = -2.016134 = -1.572769
beta is overidentified
[_ce1]missouri = 1 [_ce1]indiana = 0 [_ce2]missouri = 0 [_ce2]indiana = 1 [_ce2]kentucky = 0 beta Coef. Std. Err. z P>|z| [95% Conf. Interval]
_ce1 missouri indiana kentucky illinois _cons _ce2 missouri indiana kentucky illinois _cons
1 0 .2521685 -1.037453 -.3891102
. (omitted) .1649653 .1734165 .4726968
. 1.53 -5.98 -0.82
. 0.126 0.000 0.410
. -.0711576 -1.377343 -1.315579
. .5754946 -.6975626 .5373586
0 1 0 -1.314265 2.937016
(omitted) . (omitted) .0907071 .6448924
. -14.49 4.55 chi2(
. 0.000 0.000 .3139
. -1.492048 1.67305
. -1.136483 4.200982
1) =
Prob > chi2 = 0.575
The test of the overidentifying restriction does not reject the null hypothesis that the restriction is valid, and the p-value on the coefcient on kentucky in the rst cointegrating equation indicates that it is not signicant. We will leave the variable in the model and attribute the lack of signicance to whatever caused the kentucky series to temporarily rise above the others from 1985 until 1990, though we could instead consider removing kentucky from the model. Next, we look at the estimates of the adjustment parameters. In the output below, we replay the previous results. We specify the alpha option so that vec will display an estimation table for the estimates of the adjustment parameters, and we specify nobtable to suppress the table for the parameters of the cointegrating equations because we have already looked at those.

. vec, alpha nobtable noetable Vector error-correction model Sample: 1978m5 - 2003m12 Log likelihood = Det(Sigma_ml) = 416.9744 7.84e-07 chi2 19.39607 6.426086 8.524901 22.32893 P>chi2 0.0001 0.0402 0.0141 0.0000
651
No. of obs AIC HQIC SBIC
= 308 = -2.311522 = -2.016134 = -1.572769
Adjustment parameters Equation Parms D_missouri D_indiana D_kentucky D_illinois 2 2 2 2
alpha D_missouri _ce1 L1. _ce2 L1. D_indiana _ce1 L1. _ce2 L1. D_kentucky _ce1 L1. _ce2 L1. D_illinois _ce1 L1. _ce2 L1.
Coef.
Std. Err.
P>|z|
-.0683152
.0185763
-3.68
0.000
-.1047242
-.0319063
.0405613
.0112417
3.61
0.000
.018528
.0625946
-.0342096
.0220955
-1.55
0.122
-.0775159
.0090967
.0325804
.0133713
2.44
0.015
.0063732
.0587877
-.0482012
.0231633
-2.08
0.037
-.0936004
-.0028021
.0374395
.0140175
2.67
0.008
.0099657
.0649133
.0138224
.0227041
0.61
0.543
-.0306768
.0583215
.0567664
.0137396
4.13 chi2(
0.000 .3139
.0298373
.0836955
1) =
Prob > chi2 = 0.575
All the coefcients are signicant at the 5% level, except those on Indiana and Illinois in the rst cointegrating equation. From an economic perspective, the issue is whether the unemployment rates in Indiana and Illinois adjust when the rst cointegrating equation is out of equilibrium. We could impose restrictions on one or both of those parameters and ret the model, or we could just decide to use the current results.
652
Technical note
vec can be used to t models in which the parameters in are not identied, in which case only the parameters in and the moving-average impact matrix C are identied. When the parameters in are not identied, the values of and can vary depending on the starting values. However, the estimates of and C are identied and have known asymptotic distributions. This method is valid because these additional normalization restrictions impose no restriction on or C.
Specication of constants and trends

As discussed in [TS] vec intro, allowing for a constant term and linear time trend allow us to write the VECM as
p1
yt = (yt1 + + t) +
i=1
i yti + + t +
Five different trend specications are available: Option in trend() trend rtrend constant rconstant none Parameter restrictions none =0 = 0, and = 0 = 0, = 0 and = 0 = 0, = 0, = 0, and = 0 Johansen (1995) notation
H(r) H (r) H1 (r) H1 (r) H2 (r)
trend(trend) allows for a linear trend in the cointegrating equations and a quadratic trend in the undifferenced data. A linear trend in the cointegrating equations implies that the cointegrating equations are assumed to be trend stationary. trend(rtrend) denes a restricted trend model that excludes linear trends in the differenced data but allows for linear trends in the cointegrating equations. As in the previous case, a linear trend in a cointegrating equation implies that the cointegrating equation is trend stationary. trend(constant) denes a model with an unrestricted constant. This allows for a linear trend in the undifferenced data and cointegrating equations that are stationary around a nonzero mean. This is the default. trend(rconstant) denes a model with a restricted constant in which there is no linear or quadratic trend in the undifferenced data. A nonzero allows for the cointegrating equations to be stationary around nonzero means, which provide the only intercepts for differenced data. Seasonal indicators are not allowed with this specication. trend(none) denes a model that does not include a trend or a constant. When there is no trend or constant, the cointegrating equations are restricted to being stationary with zero means. Also, after adjusting for the effects of lagged endogenous variables, the differenced data are modeled as having mean zero. Seasonal indicators are not allowed with this specication.
653
Technical note
vec uses a switching algorithm developed by Boswijk (1995) to maximize the log-likelihood function when constraints are placed on the parameters. The starting values affect both the ability of the algorithm to nd a maximum and its speed in nding that maximum. By default, vec uses the parameter estimates that correspond to Johansens normalization. Sometimes, other starting values will cause the algorithm to nd a maximum faster. To specify starting values for the parameters in , we specify a 1 (K r) matrix in the afrom() option. Specifying starting values for the parameters in is slightly more complicated. As explained in Methods and formulas, specifying trend(constant), trend(rtrend), or trend(trend) causes some of the estimates of the trend parameters appearing in to be backed out. The switching algorithm estimates only the parameters of the cointegrating equations whose estimates are saved in e(betavec). For this reason, only the parameters saved in e(betavec) can have their initial values set via bfrom(). The table below describes which trend parameters in the cointegrating equations are estimated by the switching algorithm for each of the ve specications. Trend specication none rconstant constant rtrend trend Trend parameters in cointegrating equations none cons cons cons, trend cons, trend Trend parameter estimated via switching algorithm none cons none trend none
Collinearity
As expected, collinearity among variables causes some parameters to be unidentied numerically. If vec encounters perfect collinearity among the dependent variables, it exits with an error. In contrast, if vec encounters perfect collinearity that appears to be due to too many lags in the model, vec displays a warning message and reduces the maximum lag included in the model in an effort to nd a model with fewer lags in which all the parameters are identied by the data. Specifying the noreduce option causes vec to skip over these additional checks and corrections for collinearity. Thus the noreduce option can be used to force the estimation to proceed when not all the parameters are identied by the data. When some parameters are not identied because of collinearity, the results cannot be interpreted but can be used to nd the source of the collinearity.
654
Saved results
vec saves the following in e():
Scalars e(N) e(k rank) e(k eq) e(k dv) e(k ce) e(n lags) e(df m) e(ll) e(chi2 res) e(df lr) e(beta iden) e(beta icnt) e(k #) e(df m#) e(r2 #) e(chi2 #) e(rmse #) e(aic) e(hqic) e(sbic) e(tmin) e(tmax) e(detsig ml) e(rank) e(converge) Macros e(cmd) e(cmdline) e(trend) e(tsfmt) e(tvar) e(endog) e(covariates) e(eqnames) e(cenames) e(reduce opt) e(reduce lags) e(title) e(aconstraints) e(bconstraints) e(sindicators) e(properties) e(predict) e(marginsok) e(marginsnotok) number of observations number of unconstrained parameters number of equations in e(b) number of dependent variables number of cointegrating equations number of lags model degrees of freedom log likelihood value of test of overidentifying restrictions degrees of freedom of the test of overidentifying restrictions 1 if the parameters in are identied and 0 otherwise number of independent restrictions placed on number of variables in equation # model degrees of freedom in equation # R2 of equation # 2 statistic for equation # RMSE of equation # value of AIC value of HQIC value of SBIC minimum time maximum time determinant of the estimated covariance matrix rank of e(V) 1 if the switching algorithm converged, 0 if it did not converge vec command as typed trend specied format of the time variable variable denoting time within groups endogenous variables list of covariates equation names names of cointegrating equations noreduce, if noreduce is specied list of maximum lags to which the model has been reduced title in estimation output constraints placed on constraints placed on sindicators, if specied b V program used to implement predict predictions allowed by margins predictions disallowed by margins

Matrices e(b) e(V) e(beta) e(V beta) e(betavec) e(pi) e(V pi) e(alpha) e(V alpha) e(omega) e(mai) e(V mai) Functions e(sample)
655
estimates of short-run parameters VCE of short-run parameter estimates estimates of VCE of directly obtained estimates of estimates of VCE of estimates of VCE of estimates of estimates of C VCE of C marks estimation sample

vec is implemented as an ado-le. Methods and formulas are presented under the following headings:
General specication of the VECM The log-likelihood function Unrestricted trend Restricted trend Unrestricted constant Restricted constant No trend Estimation with Johansen identication Estimation with constraints: identied Estimation with constraints: not identied Formulas for the information criteria Formulas for predict
General specication of the VECM

vec estimates the parameters of a VECM that can be written as
p1
yt = yt1 +
i=1
i yti + v + t + w1 s1 + + wm sm +
(1)
where
yt is a K 1 vector of endogenous variables,

is a K r matrix of parameters, is a K r matrix of parameters,
1 , . . . , p1 are K K matrices of parameters,

v is a K 1 vector of parameters,
is a K 1 vector of trend coefcients,
t is a linear time trend,
656
s1 , . . . , sm are orthogonalized seasonal indicators specied in the sindicators() option, and w1 , . . . , wm are K 1 vectors of coefcients on the orthogonalized seasonal indicators.
There are two types of deterministic elements in (1): the trend, v + t, and the orthogonalized seasonal terms, w1 s1 + + wm sm . Johansen (1995, chap. 11) shows that inference about the number of cointegrating equations is based on nonstandard distributions and that the addition of any term that generalizes the deterministic specication in (1) changes the asymptotic distributions of the statistics used for inference on the number of cointegrating equations and the asymptotic distribution of the ML estimator of the cointegrating equations. In fact, Johansen (1995, 84) notes that including event indicators causes the statistics used for inference on the number of cointegrating equations to have asymptotic distributions that must be computed case by case. For this reason, event indicators may not be specied in the present version of vec. If seasonal indicators are included in the model, they cannot be collinear with a constant term. If they are collinear with a constant term, one of the indicator variables is omitted. As discussed in Specication of constants and trends, we can reparameterize the model as
p1
yt = (yt1 + + t) +
i=1
i yti + + t +
(2)
The log-likelihood function

We can maximize the log-likelihood function much more easily by writing it in concentrated form. In fact, as discussed below, in the simple case with the Johansen normalization on and no constraints on , concentrating the log-likelihood function produces an analytical solution for the parameter estimates. To concentrate the log likelihood, rewrite (2) as
Z0t = Z1t + Z2t +
(3)
where Z0t is a K 1 vector of variables yt , is the K r matrix of adjustment coefcients, and t is a K 1 vector of independently and identically distributed normal vectors with mean 0 and contemporaneous covariance matrix . Z1t , Z2t , , and depend on the trend specication and are dened below. The log-likelihood function for the model in (3) is
L=
1 T K ln(2) + T ln(||) 2
T
+
t=1
(Z0t Z1t Z2t ) 1 (Z0t Z1t Z2t )
(4)
with the constraints that and have rank r. Johansen (1995, chap. 6), building on Anderson (1951), shows how the parameters can be expressed as analytic functions of , , and the data, yielding the concentrated log-likelihood function
657
Lc =
1 T K ln(2) + T ln(||) 2
T
+
t=1
(R0t R1t ) 1 (R0t R1t )
(5)
where
Mij = T 1
T t=1
Zit Zjt ,
i, j {0, 1, 2};
R0t = Z0t M02 M1 Z2t ; and 22 R1t = Z1t M12 M1 Z2t . 22

The denitions of Z1t , Z2t , , and change with the trend specications, although some of their components stay the same.
Unrestricted trend
When the trend in the VECM is unrestricted, we can dene the variables in (3) directly in terms of the variables in (1):
Z1t = yt1 is K 1 Z2t = (yt1 , . . . , ytp+1 , 1, t, s1 , . . . , sm ) is {K(p 1) + 2 + m} 1; = (1 , . . . , p1 , v, , w1 , . . . , wm ) is K {K(p 1) + 2 + m}

= is the K r matrix composed of the r cointegrating vectors. In the unrestricted trend specication, m1 = K, m2 = K(p 1) + 2 + m, and there are nparms = Kr + Kr + K{K(p 1) + 2 + m} parameters in (3).
Restricted trend
When there is a restricted trend in the VECM in (2), = 0, but the intercept v = + is unrestricted. The VECM with the restricted trend can be written as
yt = ( , )
yt1 t
p1
+
i=1
i yti + v + w1 s1 + + wm sm +
This equation can be written in the form of (3) by dening
Z1t = yt1 , t
is (K + 1) 1
Z2t = (yt1 , . . . , ytp+1 , 1, s1 , . . . , sm ) is {K(p 1) + 1 + m} 1 = (1 , . . . , p1 , v, w1 , . . . , wm ) is K {K(p 1) + 1 + m}

= , is the (K + 1) r matrix composed of the r cointegrating vectors and the r trend coefcients In the restricted trend specication, m1 = K + 1, m2 = {K(p 1) + 1 + m}, and there are nparms = Kr + (K + 1)r + K{K(p 1) + 1 + m} parameters in (3).
Unrestricted constant
An unrestricted constant in the VECM in (2) is equivalent to setting = 0 in (1), which can be written in the form of (3) by dening
658
Z1t = yt1 is (K 1) Z2t = (yt1 , . . . , ytp+1 , 1, s1 , . . . , sm ) is {K(p 1) + 1 + m} 1; = (1 , . . . , p1 , v, w1 , . . . , wm ) is K {K(p 1) + 1 + m}

= is the K r matrix composed of the r cointegrating vectors In the unrestricted constant specication, m1 = K, m2 = {K(p 1) + 1 + m}, and there are nparms = Kr + Kr + K{K(p 1) + 1 + m} parameters in (3).
Restricted constant
When there is a restricted constant in the VECM in (2), it can be written in the form of (3) by dening
Z1t = yt1 , 1
is (K + 1) 1
Z2t = (yt1 , . . . , ytp+1 ) is K(p 1) 1 = (1 , . . . , p1 ) is K K(p 1)

= , is the (K + 1) r matrix composed of the r cointegrating vectors and the r constants in the cointegrating relations. In the restricted trend specication, m1 = K + 1, m2 = K(p 1), and there are nparms = Kr + (K + 1)r + K{K(p 1)} parameters in (3).
No trend
When there is no trend in the VECM in (2), it can be written in the form of (3) by dening
Z1t = yt1 is K 1 Z2t = (yt1 , . . . , ytp+1 ) is K(p 1) + m 1 = (1 , . . . , p1 ) is K K(p 1)

= is K r matrix of r cointegrating vectors In the no-trend specication, m1 = K , m2 = K(p 1), and there are nparms = Kr + Kr + K{K(p 1)} parameters in (3).
Estimation with Johansen identication

Not all the parameters in and are identied. Consider the simple case in which is K r and let Q be a nonsingular r r matrix. Then = QQ1 = Q(Q 1 ) = Substituting into the log likelihood in (5) for would not change the value of the log likelihood, so some a priori identication restrictions must be found to identify and . As discussed in Johansen (1995, chap. 5 and 6) and Boswijk (1995), if the restrictions exactly identify or overidentify , the estimates of the unconstrained parameters in will be superconsistent, meaning that the estimates of the free parameters in will converge at a faster rate than estimates of the short-run parameters in and i . This allows the distribution of the estimator of the short-run parameters to be derived conditional on the estimated .
659
Johansen (1995, chap. 6) has proposed a normalization method for use when theory does not provide sufcient a priori restrictions to identify the cointegrating vector. This method has become widely adopted by researchers. Johansens identication scheme is
= (Ir , ) where Ir is the r r identity matrix and is a (m1 r) r matrix of identied parameters.
(6)
Johansens identication method places r2 linearly independent constraints on the parameters in , thereby dening an exactly identied model. The total number of freely estimated parameters is nparms r2 = {K + m2 + (K + m1 r)r}, and the degrees of freedom d is calculated as the integer part of (nparms r2 )/K. When only the rank and the Johansen identication restrictions are placed on the model, we can further manipulate the log likelihood in (5) to obtain analytic formulas for the parameters in , , and . For a given value of , and can be found by regressing R0t on R1t . This allows a further simplication of the problem in which () = S01 ( S11 )1
() = S00 S01 ( S11 )1 S10 Sij = (1/T )

T t=1
Rit Rjt
i, j {0, 1}
Johansen (1995) shows that by inserting these solutions into equation (5), is given by the r eigenvectors v1 , . . . , vr corresponding to the r largest eigenvalues 1 , . . . , r that solve the generalized eigenvalue problem |i S11 S10 S1 S01 | = 0 (7) 00 The eigenvectors corresponding to 1 , . . . , r that solve (7) are the unidentied parameter estimates. To impose the identication restrictions in (6), we normalize the eigenvectors such that
i S11 vi = S01 S1 S01 vi 00

and
(8)
vi S11 vj =
1 if i = j 0 otherwise
(9)
At the optimum the log-likelihood function with the Johansen identication restrictions can be expressed in terms of T, K, S00 , and the r largest eigenvalues
1 Lc = T K ln(2) + K + ln(|S00 |) + ln(1 i ) 2 i=1

where the i are the eigenvalues that solve (7), (8), and (9). Using the normalized , we can then obtain the estimates = S01 ( S11 )1 and
(10)
= S00 S10
660
Let y be a K r matrix that contains the estimates of the parameters in in (1). y differs from in that any trend parameter estimates are omitted from . We can then use y to obtain predicted values for the r nondemeaned cointegrating equations
Et = y yt
The r series in E t are called the predicted, nondemeaned cointegrating equations because they still contain the terms and . We want to work with the predicted, demeaned cointegrating equations. Thus we need estimates of and . In the trend(rconstant) specication, the algorithm directly produces the estimator . Similarly, in the trend(rtrend) specication, the algorithm directly produces the estimator . In the remaining cases, to back out estimates of and , we need estimates of v and , which we can obtain by estimating the parameters of the following VAR:
p1
yt = Et1 +
i=1
i yti + v + t + w1 s1 + + wm sm +
(11)
Depending on the trend specication, we use to back out the estimates of = ( )1 v = ( )1 if they are not already in and are included in the trend specication. We then augment y to f = (y , , ) where the estimates of and are either obtained from or backed out using (12) and (13). We next use f to obtain the r predicted, demeaned cointegrating equations, Et , via
(12) (13)
Et = f (yt , 1, t)
We last obtain estimates of all the short-run parameters from the VAR:
p1
yt = Et1 +
i=1
i yti + + t + w1 s1 + + wm sm +
(14)
Because the estimator f converges in probability to its true value at a rate faster than T 2 , we can take our estimated Et1 as given data in (14). This allows us to estimate the variancecovariance (VCE) matrix of the estimates of the parameters in (14) by using the standard VAR VCE estimator. Equation (11) can be used to obtain consistent estimates of all the parameters and of the VCE of all the parameters, except v and . The standard VAR VCE of v and is incorrect because these estimates converge at a faster rate. This is why it is important to use the predicted, demeaned cointegrating equations, Et1 , when estimating the short-run parameters and trend terms. In keeping with the cointegration literature, vec makes a small-sample adjustment to the VCE estimator so that the divisor is (T d) instead of T , where d represents the degrees of freedom of the model. d is calculated as the integer part of nparms /K, where nparms is the total number of freely estimated parameters in the model.
661
In the trend(rconstant) specication, the estimation procedure directly estimates . For trend(constant), trend(rtrend), and trend(trend), the estimates of are backed out using (12). In the trend(rtrend) specication, the estimation procedure directly estimates . In the trend(trend) specication, the estimates of are backed out using (13). Because the elements of the estimated VCE are readily available only when the estimates are obtained directly, when the trend parameter estimates are backed out, their elements in the VCE for f are missing. Under the Johansen identication restrictions, vec obtains , the estimates of the parameters in the r m1 matrix in (5). The VCE of vec() is rm1 rm1 . Per Johansen (1995), the asymptotic distribution of is mixed Gaussian, and its VCE is consistently estimated by
1 T d
(Ir HJ ) ( 1 ) (HJ S11 HJ )
(Ir HJ )
(15)
where HJ is the m1 (m1 r) matrix given by HJ = (0r(m1 r) , Im1 r ) . The VCE reported in e(V beta) is the estimated VCE in (15) augmented with missing values to account for any backed-out estimates of or . The parameter estimates can be found either as a function of , using (10) or from the VAR in (14). The estimated VCE of reported in e(V alpha) is given by
1 B (T d)
where B = ( S11 )1 . As we would expect, the estimator of = is
=
and its estimated VCE is given by
1 (B ) (T d)
The moving-average impact matrix C is estimated by
C = ( )1
where is the orthogonal complement of y , is the orthogonal complement of , and
= IK i=1 i . The orthogonal complement of a K r matrix Q that has rank r is a matrix Q of rank K r, such that Q Q = 0. Although this operation is not uniquely dened, the results used by vec do not depend on the method of obtaining the orthogonal complement. vec uses the following method: the orthogonal complement of Q is given by the r eigenvectors with the highest eigenvalues from the matrix Q (Q Q)1 Q .
Per Johansen (1995, chap. 13) and Drukker (2004), the VCE of C is estimated by
p1
T d Sq V Sq T
(16)
662
where
Sq = C
=
(1 , 2 ) if p > 1 1 if p = 1
1 = (C IK ) = ( )1 2 = p1 C p1 is a (p 1) 1 vector of ones
V is the estimated VCE of = (, 1 , . . . p1 )
Estimation with constraints: identied

vec can also t models in which the adjustment parameters are subject to homogeneous linear constraints and the cointegrating vectors are subject to general linear restrictions. Mathematically, vec allows for constraints of the form
R vec() = 0
where R is a known Kr n constraint matrix, and
(17)
R vec() = b
where R
(18)
is a known m1 r n constraint matrix and b is a known n 1 vector of constants. Although (17) and (18) are intuitive, they can be rewritten in a form to facilitate computation. Specically, (17) can be written as vec( ) = Ga (19) where G is Kr n and a is n 1. Equation (18) can be rewritten as
vec() = Hb + h0
(20)
where H is a known n1 r n matrix, b is an n 1 matrix of parameters, and h0 is a known n1 r 1 matrix. See [P] makecns for a discussion of the different ways of specifying the constraints. When constraints are specied via the aconstraints() and bconstraints() options, the Boswijk (1995) rank method determines whether the parameters in are underidentied, exactly identied, or overidentied. Boswijk (1995) uses the Rothenberg (1971) method to determine whether the parameters in are identied. Thus the parameters in are exactly identied if = r2 , and the parameters in are overidentied if > r2 , where
= rank R (Ir ) and is a full-rank matrix with the same dimensions as . The computed is saved in e(beta icnt).
663
Similarly, the number of freely estimated parameters in and is given by jacob , where
jacob = rank ( Im1 )H, (IK )G

Using jacob , we can calculate several other parameter counts of interest. In particular, the degrees of freedom of the overidentifying test are given by (K + m1 r)r jacob , and the number of freely estimated parameters in the model is nparms = Km2 + jacob . Although the problem of maximizing the log-likelihood function in (4), subject to the constraints in (17) and (18), could be handled by the algorithms in [R] ml, the switching algorithm of Boswijk (1995) has proven to be more convergent. For this reason, vec uses the Boswijk (1995) switching algorithm to perform the optimization. Given starting values (b0 , a0 , 0 ), the algorithm iteratively updates the estimates until convergence is achieved, as follows: j is constructed from (19) and aj j is constructed from (20) and bj
bj+1 = {H (j 1 j S11 )H}1 H (j 1 S11 ){vec(P) (j InZ1 )h0 } j j aj+1 = {G(1 j S11 j )G}1 G (1 j S11 )vec(P) j j j+1 = S00 S01 j j j j S10 + j j S11 j j
The estimated VCE of is given by
1 H{H (W S11 )H}1 H (T d)

where W is 1 . As in the case without constraints, the estimated VCE of can be obtained either from the VCE of the short-run parameters, as described below, or via the formula
1 V = G G 1 ( S11 )G (T d)
Boswijk (1995) notes that, as long as the parameters of the cointegrating equations are exactly identied or overidentied, the constrained ML estimator produces superconsistent estimates of . This implies that the method of estimating the short-run parameters described above applies in the presence of constraints, as well, albeit with a caveat: when there are constraints placed on , the VARs must be estimated subject to these constraints. With these estimates and the estimated VCE of the short-run parameter matrix V , Drukker (2004) shows that the estimated VCE for is given by
( IK )V ( IK )
Drukker (2004) also shows that the estimated VCE of C can be obtained from (16) with the extension that V is the estimated VCE of that takes into account any constraints on .
664
Estimation with constraints: not identied

When the parameters in are not identied, only the parameters in = and C are identied. The estimates of and C would not change if more identication restrictions were imposed to achieve exact identication. Thus the VCE matrices for and C can be derived as if the model exactly identied .
Formulas for the information criteria

The AIC, SBIC, and HQIC are calculated according to their standard denitions, which include the constant term from the log likelihood; that is,
L T L SBIC = 2 T
AIC
=2
2nparms T ln(T ) + nparms T + + 2ln ln(T ) nparms T
HQIC
=2
L T
where nparms is the total number of parameters in the model and L is the value of the log likelihood at the optimum.
Formulas for predict

xb, residuals and stdp are standard and are documented in [R] predict. ce causes predict to compute Et = f yt for the requested cointegrating equation.
d levels causes predict to compute the predictions for the levels of the data. Let yt be the predicted value of yt . Because the computations are performed for a given equation, yt is a scalar. d d Using yt , we can predict the level by yt = yt + yt1 .
Because the residuals from the VECM for the differences and the residuals from the corresponding VAR in levels are identical, there is no need for an option for predicting the residuals in levels.
References
Anderson, T. W. 1951. Estimating linear restrictions on regression coefcients for multivariate normal distributions. Annals of Mathematical Statistics 22: 327351. Boswijk, H. P. 1995. Identiability of cointegrated systems. Discussion Paper #95-78, Tinbergen Institute. http://www1.fee.uva.nl/pp/bin/258fulltext.pdf. Boswijk, H. P., and J. A. Doornik. 2004. Identifying, estimating and testing restricted cointegrating systems: An overview. Statistica Neerlandica 58: 440465. Drukker, D. M. 2004. Some further results on estimation and inference in the presence of constraints on alpha in a cointegrating VECM. Working paper, StataCorp. Engle, R. F., and C. W. J. Granger. 1987. Co-integration and error correction: Representation, estimation, and testing. Econometrica 55: 251276. Hamilton, J. D. 1994. Time Series Analysis. Princeton: Princeton University Press. Johansen, S. 1988. Statistical analysis of cointegration vectors. Journal of Economic Dynamics and Control 12: 231254.
665
. 1991. Estimation and hypothesis testing of cointegration vectors in Gaussian vector autoregressive models. Econometrica 59: 15511580. . 1995. Likelihood-Based Inference in Cointegrated Vector Autoregressive Models. Oxford: Oxford University Press. Maddala, G. S., and I.-M. Kim. 1998. Unit Roots, Cointegration, and Structural Change. Cambridge: Cambridge University Press. Park, J. Y., and P. C. B. Phillips. 1988. Statistical inference in regressions with integrated processes: Part I. Econometric Theory 4: 468497. . 1989. Statistical inference in regressions with integrated processes: Part II. Econometric Theory 5: 95131. Phillips, P. C. B. 1986. Understanding spurious regressions in econometrics. Journal of Econometrics 33: 311340. Phillips, P. C. B., and S. N. Durlauf. 1986. Multiple time series regressions with integrated processes. Review of Economic Studies 53: 473495. Rothenberg, T. J. 1971. Identication in parametric models. Econometrica 39: 577591. Sims, C. A., J. H. Stock, and M. W. Watson. 1990. Inference in linear time series models with some unit roots. Econometrica 58: 113144. Stock, J. H. 1987. Asymptotic properties of least squares estimators of cointegrating vectors. Econometrica 55: 10351056. Stock, J. H., and M. W. Watson. 1988. Testing for common trends. Journal of the American Statistical Association 83: 10971107. Watson, M. W. 1994. Vector autoregressions and cointegration. In Vol. IV of Handbook of Econometrics, ed. R. F. Engle and D. L. McFadden. Amsterdam: Elsevier.
Also see
[TS] vec postestimation Postestimation tools for vec [TS] tsset Declare data to be time-series data [TS] var Vector autoregressive models [TS] var svar Structural vector autoregressive models [U] 20 Estimation and postestimation commands [TS] vec intro Introduction to vector error-correction models
Title
vec postestimation Postestimation tools for vec
Description
The following postestimation commands are of special interest after vec: Command fcast compute fcast graph irf veclmar vecnorm vecstable Description obtain dynamic forecasts graph dynamic forecasts obtained from fcast compute create and analyze IRFs and FEVDs LM test for autocorrelation in residuals test for normally distributed residuals check stability condition of estimates
666
667
Syntax for predict

predict statistic
Main
type
newvar
if
in
Description linear prediction; the default standard error of the linear prediction residuals the predicted value of specied cointegrating equation one-step prediction of the level of the endogenous variable compute the predictions using previously predicted cointegrating equations
xb stdp residuals ce levels usece(varlistce )
Menu
Statistics
>
Postestimation
>
Options for predict

Main
xb, the default, calculates the linear prediction for the specied equation. The form of the VECM implies that these tted values are the one-step predictions for the rst-differenced variables. stdp calculates the standard error of the linear prediction for the specied equation. residuals calculates the residuals from the specied equation of the VECM. ce calculates the predicted value of the specied cointegrating equation. levels calculates the one-step prediction of the level of the endogenous variable in the requested equation. usece(varlistce ) species that previously predicted cointegrating equations saved under the names in varlistce be used to compute the predictions. The number of variables in the varlistce must equal the number of cointegrating equations specied in the model. equation(eqno | eqname) species to which equation you are referring. equation() is lled in with one eqno or eqname for xb, residuals, stdp, ce, and levels options. equation(#1) would mean that the calculation is to be made for the rst equation, equation(#2) would mean the second, and so on. You could also refer to the equation by its name. equation(D income) would refer to the equation named D income and equation( ce1), to the rst cointegrating equation, which is named ce1 by vec. If you do not specify equation(), the results are as if you specied equation(#1). For more information on using predict after multiple-equation estimation commands, see [R] predict.
668
Remarks

See the following sections for information on model selection and inference after vec. [TS] [TS] [TS] [TS] [TS] [TS]
irf Create and analyze IRFs, dynamic-multiplier functions, and FEVDs varsoc Obtain lag-order selection statistics for VARs and VECMs veclmar Perform LM test for residual autocorrelation after vec vecnorm Test for normally distributed disturbances after vec vecrank Estimate the cointegrating rank of a VECM vecstable Check the stability condition of VECM estimates
Forecasting
See the following sections for information on obtaining forecasts after vec:
[TS] fcast compute Compute dynamic forecasts of dependent variables after var, svar, or vec [TS] fcast graph Graph forecasts of variables computed by fcast compute

Also see
[TS] vec Vector error-correction models [U] 20 Estimation and postestimation commands [TS] vec intro Introduction to vector error-correction models
Title
veclmar Perform LM test for residual autocorrelation after vec
Syntax
veclmar options mlag(#) estimates(estname) separator(#) , options Description use # for the maximum order of autocorrelation; default is mlag(2) use previously saved results estname; default is to use active results draw separator line after every # rows
veclmar can be used only after vec; see [TS] vec. You must tsset your data before using veclmar; see [TS] tsset.
Menu
Statistics
>
>
>
LM test for residual autocorrelation
Description
veclmar implements a Lagrange-multiplier (LM) test for autocorrelation in the residuals of vector error-correction models (VECMs).
Options
mlag(#) species the maximum order of autocorrelation to be tested. The integer specied in mlag() must be greater than 0; the default is 2. estimates(estname) requests that veclmar use the previously obtained set of vec estimates saved as estname. By default, veclmar uses the active results. See [R] estimates for information on saving and restoring estimation results. separator(#) species how many rows should appear in the table between separator lines. By default, separator lines do not appear. For example, separator(1) would draw a line between each row, separator(2) between every other row, and so on.
Remarks
Estimation, inference, and postestimation analysis of VECMs is predicated on the errors not being autocorrelated. veclmar implements the LM test for autocorrelation in the residuals of a VECM discussed in Johansen (1995, 2122). The test is performed at lags j = 1, . . . , mlag(). For each j , the null hypothesis of the test is that there is no autocorrelation at lag j .
669
670
Example 1
We t a VECM using the regional income data described in [TS] vec and then call veclmar to test for autocorrelation.
. use http://www.stata-press.com/data/r12/rdinc . vec ln_ne ln_se (output omitted ) . veclmar, mlag(4) Lagrange-multiplier test lag 1 2 3 4 chi2 8.9586 4.9809 4.8519 0.3270 df 4 4 4 4 Prob > chi2 0.06214 0.28926 0.30284 0.98801
At the 5% level, we cannot reject the null hypothesis that there is no autocorrelation in the residuals for any of the orders tested. Thus this test nds no evidence of model misspecication.
Saved results
veclmar saves the following in r():
Matrices r(lm)
2 , df, and p-values

veclmar is implemented as an ado-le. Consider a VECM without any trend:
p1
yt = yt1 +
i=1
i yti +
As discussed in [TS] vec, as long as the parameters in the cointegrating vectors, , are exactly identied or overidentied, the estimates of these parameters are superconsistent. This implies that the r 1 vector of estimated cointegrating relations
Et = yt
(1)
can be used as data with standard estimation and inference methods. When the parameters of the cointegrating equations are not identied, (1) does not provide consistent estimates of Et ; in these cases, veclmar exits with an error message.
671
The VECM above can be rewritten as

p1
yt = Et +
i=1
i yti +
which is just a VAR with p 1 lags where the endogenous variables have been rst-differenced and is augmented with the exogenous variables E. veclmar ts this VAR and then calls varlmar to compute the LM test for autocorrelation. The above discussion assumes no trend and implicitly ignores constraints on the parameters in . As discussed in vec, the other four trend specications considered by Johansen (1995, sec. 5.7) complicate the estimation of the free parameters in but do not alter the basic result that the Et can be used as data in the subsequent VAR. Similarly, constraints on the parameters in imply that the subsequent VAR must be estimated with these constraints applied, but Et can still be used as data in the VAR. See [TS] varlmar for more information on the Johansen LM test.
Reference
Johansen, S. 1995. Likelihood-Based Inference in Cointegrated Vector Autoregressive Models. Oxford: Oxford University Press.
Also see
[TS] vec Vector error-correction models [TS] varlmar Perform LM test for residual autocorrelation after var or svar [TS] vec intro Introduction to vector error-correction models
Title
vecnorm Test for normally distributed disturbances after vec
Syntax
vecnorm options jbera skewness kurtosis estimates(estname) dfk separator(#) , options Description report JarqueBera statistic; default is to report all three statistics report skewness statistic; default is to report all three statistics report kurtosis statistic; default is to report all three statistics use previously saved results estname; default is to use active results make small-sample adjustment when computing the estimated variancecovariance matrix of the disturbances draw separator line after every # rows
vecnorm can be used only after vec; see [TS] vec.
Menu
Statistics
>
>
>
Test for normally distributed disturbances
Description
vecnorm computes and reports a series of statistics against the null hypothesis that the disturbances in a VECM are normally distributed.
Options
jbera requests that the JarqueBera statistic and any other explicitly requested statistic be reported. By default, the JarqueBera, skewness, and kurtosis statistics are reported. skewness requests that the skewness statistic and any other explicitly requested statistic be reported. By default, the JarqueBera, skewness, and kurtosis statistics are reported. kurtosis requests that the kurtosis statistic and any other explicitly requested statistic be reported. By default, the JarqueBera, skewness, and kurtosis statistics are reported. estimates(estname) requests that vecnorm use the previously obtained set of vec estimates saved as estname. By default, vecnorm uses the active results. See [R] estimates for information about saving and restoring estimation results. dfk requests that a small-sample adjustment be made when computing the estimated variance covariance matrix of the disturbances. separator(#) species how many rows should appear in the table between separator lines. By default, separator lines do not appear. For example, separator(1) would draw a line between each row, separator(2) between every other row, and so on.
672
673
Remarks
vecnorm computes a series of test statistics of the null hypothesis that the disturbances in a VECM are normally distributed. For each equation and all equations jointly, up to three statistics may be computed: a skewness statistic, a kurtosis statistic, and the JarqueBera statistic. By default, all three statistics are reported; if you specify only one statistic, the others are not reported. The JarqueBera statistic tests skewness and kurtosis jointly. The single-equation results are against the null hypothesis that the disturbance for that particular equation is normally distributed. The results for all the equations are against the null that all K disturbances have a K -dimensional multivariate normal distribution. Failure to reject the null hypothesis indicates lack of model misspecication. As noted by Johansen (1995, 141), the log likelihood for the VECM is derived assuming the errors are independently and identically distributed (i.i.d.) normal, though many of the asymptotic properties can be derived under the weaker assumption that the errors are merely i.i.d. Many researchers still prefer to test for normality. vecnorm uses the results from vec to produce a series of statistics against the null hypothesis that the K disturbances in the VECM are normally distributed.
Example 1
This example uses vecnorm to test for normality after estimating the parameters of a VECM using the regional income data.
. use http://www.stata-press.com/data/r12/rdinc . vec ln_ne ln_se (output omitted ) . vecnorm Jarque-Bera test Equation D_ln_ne D_ln_se ALL Skewness test Equation D_ln_ne D_ln_se ALL Kurtosis test Equation D_ln_ne D_ln_se ALL Kurtosis 3.1679 2.8294 chi2 0.062 0.064 0.126 df 1 1 2 Prob > chi2 0.80302 0.79992 0.93873 Skewness .05982 .243 chi2 0.032 0.522 0.553 df 1 1 2 Prob > chi2 0.85890 0.47016 0.75835 chi2 0.094 0.586 0.680 df 2 2 4 Prob > chi2 0.95417 0.74608 0.95381
The JarqueBera results present test statistics for each equation and for all equations jointly against the null hypothesis of normality. For the individual equations, the null hypothesis is that the disturbance term in that equation has a univariate normal distribution. For all equations jointly, the null hypothesis is that the K disturbances come from a K -dimensional normal distribution. In this example, the single-equation and overall JarqueBera statistics do not reject the null of normality.
674
The single-equation skewness test statistics are of the null hypotheses that the disturbance term in each equation has zero skewness, which is the skewness of a normally distributed variable. The row marked ALL shows the results for a test that the disturbances in all equations jointly have zero skewness. The skewness results shown above do not suggest nonnormality. The kurtosis of a normally distributed variable is three, and the kurtosis statistics presented in the table test the null hypothesis that the disturbance terms have kurtosis consistent with normality. The results in this example do not reject the null hypothesis.
The statistics computed by vecnorm are based on the estimated variancecovariance matrix of the disturbances. vec saves the ML estimate of this matrix, which vecnorm uses by default. Specifying the dfk option instructs vecnorm to make a small-sample adjustment to the variancecovariance matrix before computing the test statistics.
Saved results
vecnorm saves the following in r():
Macros r(dfk) Matrices r(jb) r(skewness) r(kurtosis) dfk, if specied JarqueBera 2 , df, and p-values skewness 2 , df, and p-values kurtosis 2 , df, and p-values

vecnorm is implemented as an ado-le. As discussed in Methods and formulas of [TS] vec, a cointegrating VECM can be rewritten as a VAR in rst differences that includes the predicted cointegrating equations as exogenous variables. vecnorm computes the tests discussed in [TS] varnorm for the corresponding augmented VAR in rst differences. See Methods and formulas of [TS] veclmar for more information on this approach. When the parameters of the cointegrating equations are not identied, the consistent estimates of the cointegrating equations are not available, and, in these cases, vecnorm exits with an error message.
References
Hamilton, J. D. 1994. Time Series Analysis. Princeton: Princeton University Press. Jarque, C. M., and A. K. Bera. 1987. A test for normality of observations and regression residuals. International Statistical Review 2: 163172. Johansen, S. 1995. Likelihood-Based Inference in Cointegrated Vector Autoregressive Models. Oxford: Oxford University Press. L tkepohl, H. 2005. New Introduction to Multiple Time Series Analysis. New York: Springer. u
675
Also see
[TS] vec Vector error-correction models [TS] varnorm Test for normally distributed disturbances after var or svar [TS] vec intro Introduction to vector error-correction models
Title
vecrank Estimate the cointegrating rank of a VECM
Syntax
vecrank depvar options
Model
if
in
, options Description use # for the maximum lag in underlying VAR model include an unrestricted constant in model; the default include a restricted constant in model include a linear trend in the cointegrating equations and a quadratic trend in the undifferenced data include a restricted trend in model do not include a trend or a constant include normalized seasonal indicator variables varlistsi do not perform checks and corrections for collinearity among lags of dependent variables do not report the trace statistic report maximum-eigenvalue statistic report information criteria report 1% critical values instead of 5% critical values report both 1% and 5% critical values
lags(#) trend(constant) trend(rconstant) trend(trend) trend(rtrend) trend(none)

Adv. model
sindicators(varlistsi ) noreduce
Reporting
notrace max ic level99 levela
You must tsset your data before using vecrank; see [TS] tsset. depvar may contain time-series operators; see [U] 11.4.4 Time-series varlists. by, rolling, and statsby are allowed; see [U] 11.1.10 Prex commands. vecrank does not allow gaps in the data.
Menu
Statistics
>
>
Cointegrating rank of a VECM
Description
vecrank produces statistics used to determine the number of cointegrating equations in a vector error-correction model (VECM).
676
677
Options

Model
lags(#) species the number of lags in the VAR representation of the model. The VECM will include one fewer lag of the rst differences. The number of lags must be greater than zero but small enough so that the degrees of freedom used by the model are less than the number of observations. trend(trend spec) species one of ve trend specications to include in the model. See [TS] vec intro and [TS] vec for descriptions. The default is trend(constant).
Adv. model
sindicators(varlistsi ) species normalized seasonal indicator variables to be included in the model. The indicator variables specied in this option must be normalized as discussed in Johansen (1995, 84). If the indicators are not properly normalized, the likelihood-ratiobased tests for the number of cointegrating equations do not converge to the asymptotic distributions derived by Johansen. For details, see Methods and formulas of [TS] vec. sindicators() cannot be specied with trend(none) or trend(rconstant) noreduce causes vecrank to skip the checks and corrections for collinearity among the lags of the dependent variables. By default, vecrank checks whether the current lag specication causes some of the regressions performed by vecrank to contain perfectly collinear variables and reduces the maximum lag until the perfect collinearity is removed. See Collinearity in [TS] vec for more information.
Reporting
notrace requests that the output for the trace statistic not be displayed. The default is to display the trace statistic. max requests that the output for the maximum-eigenvalue statistic be displayed. The default is to not display this output. ic causes the output for the information criteria to be displayed. The default is to not display this output. level99 causes the 1% critical values to be displayed instead of the default 5% critical values. levela causes both the 1% and the 5% critical values to be displayed.
Remarks
Introduction The trace statistic The maximum-eigenvalue statistic Minimizing an information criterion
Introduction
Before estimating the parameters of a VECM models, you must choose the number of lags in the underlying VAR, the trend specication, and the number of cointegrating equations. vecrank offers several ways of determining the number of cointegrating vectors conditional on a trend specication and lag order.
678
vecrank implements three types of methods for determining r, the number of cointegrating equations in a VECM. The rst is Johansens trace statistic method. The second is his maximum eigenvalue statistic method. The third method chooses r to minimize an information criterion. All three methods are based on Johansens maximum likelihood (ML) estimator of the parameters of a cointegrating VECM. The basic VECM is
p1
yt = yt1 +
t=1
i yti +
where y is a (K 1) vector of I(1) variables, and are (K r) parameter matrices with rank r < K , 1 , . . . , p1 are (K K) matrices of parameters, and t is a (K 1) vector of normally distributed errors that is serially uncorrelated but has contemporaneous covariance matrix . Building on the work of Anderson (1951), Johansen (1995) derives an ML estimator for the parameters and two likelihood-ratio (LR) tests for inference on r. These LR tests are known as the trace statistic and the maximum-eigenvalue statistic because the log likelihood can be written as the log of the determinant of a matrix plus a simple function of the eigenvalues of another matrix. Let 1 , . . . , K be the K eigenvalues used in computing the log likelihood at the optimum. Furthermore, assume that these eigenvalues are sorted from the largest 1 to the smallest K . If there are r < K cointegrating equations, and have rank r and the eigenvalues r+1 , . . . , K are zero.
The trace statistic

The null hypothesis of the trace statistic is that there are no more than r cointegrating relations. Restricting the number of cointegrating equations to be r or less implies that the remaining K r eigenvalues are zero. Johansen (1995, chap. 11 and 12) derives the distribution of the trace statistic
K
T
i=r+1
ln(1 i )
where T is the number of observations and the i are the estimated eigenvalues. For any given value of r, large values of the trace statistic are evidence against the null hypothesis that there are r or fewer cointegrating relations in the VECM. One of the problems in determining the number of cointegrating equations is that the process involves more than one statistical test. Johansen (1995, chap. 6, 11, and 12) derives a method based on the trace statistic that has nominal coverage despite evaluating multiple tests. This method can be interpreted as being an estimator r of the true number of cointegrating equations r0 . The method starts testing at r = 0 and accepts as r the rst value of r for which the trace statistic fails to reject the null.
Example 1
We have quarterly data on the natural logs of aggregate consumption, investment, and GDP in the United States from the rst quarter of 1959 through the fourth quarter of 1982. As discussed in King et al. (1991), the balanced-growth hypothesis in economics implies that we would expect to nd two cointegrating equations among these three variables. In the output below, we use vecrank to determine the number of cointegrating equations using Johansens multiple-trace test method.

. use http://www.stata-press.com/data/r12/balance2 (macro data for VECM/balance study) . vecrank y i c, lags(5) Johansen tests for cointegration Trend: constant Number of obs = Sample: 1960q2 - 1982q4 Lags = 5% trace critical statistic value 46.1492 29.68 17.5810 15.41 3.3465* 3.76
679
91 5
maximum rank 0 1 2 3
parms 39 44 47 48
LL 1231.1041 1245.3882 1252.5055 1254.1787
eigenvalue . 0.26943 0.14480 0.03611
The header produces information about the sample, the trend specication, and the number of lags included in the model. The main table contains a separate row for each possible value of r, the number of cointegrating equations. When r = 3, all three variables in this model are stationary. In this example, because the trace statistic at r = 0 of 46.1492 exceeds its critical value of 29.68, we reject the null hypothesis of no cointegrating equations. Similarly, because the trace statistic at r = 1 of 17.581 exceeds its critical value of 15.41, we reject the null hypothesis that there is one or fewer cointegrating equation. In contrast, because the trace statistic at r = 2 of 3.3465 is less than its critical value of 3.76, we cannot reject the null hypothesis that there are two or fewer cointegrating equations. Because Johansens method for estimating r is to accept as r the rst r for which the null hypothesis is not rejected, we accept r = 2 as our estimate of the number of cointegrating equations between these three variables. The * by the trace statistic at r = 2 indicates that this is the value of r selected by Johansens multiple-trace test procedure. The eigenvalue shown in the last line of output computes the trace statistic in the preceding line.
Example 2
In the previous example, we used the default 5% critical values. We can estimate r with 1% critical values instead by specifying the level99 option.
. vecrank y i c, lags(5) level99 Johansen tests for cointegration Trend: constant Number of obs = Sample: 1960q2 - 1982q4 Lags = 1% trace critical statistic value 46.1492 35.65 17.5810* 20.04 3.3465 6.65
91 5
maximum rank 0 1 2 3
parms 39 44 47 48
LL 1231.1041 1245.3882 1252.5055 1254.1787
eigenvalue . 0.26943 0.14480 0.03611
The output indicates that switching from the 5% to the 1% level changes the resulting estimate from r = 2 to r = 1.
680
The maximum-eigenvalue statistic

The alternative hypothesis of the trace statistic is that the number of cointegrating equations is strictly larger than the number r assumed under the null hypothesis. Instead, we could assume a given r under the null hypothesis and test this against the alternative that there are r + 1 cointegrating equations. Johansen (1995, chap. 6, 11, and 12) derives an LR test of the null of r cointegrating relations against the alternative of r + 1 cointegrating relations. Because the part of the log likelihood that changes with r is a simple function of the eigenvalues of a (K K) matrix, this test is known as the maximum-eigenvalue statistic. This method is used less often than the trace statistic method because no solution to the multiple-testing problem has yet been found.
Example 3
In the output below, we reexamine the balanced-growth hypothesis. We use the levela option to obtain both the 5% and 1% critical values, and we use the notrace option to suppress the table of trace statistics.
. vecrank y i c, lags(5) max levela notrace Johansen tests for cointegration Trend: constant Number of obs = Sample: 1960q2 - 1982q4 Lags = maximum rank 0 1 2 3 max statistic 28.5682 14.2346 3.3465 5% critical value 20.97 14.07 3.76
91 5
parms 39 44 47 48
LL 1231.1041 1245.3882 1252.5055 1254.1787
eigenvalue 0.26943 0.14480 0.03611
1% critical value 25.52 18.63 6.65
We can reject r = 1 in favor of r = 2 at the 5% level but not at the 1% level. As with the trace statistic method, whether we choose to specify one or two cointegrating equations in our VECM will depend on the signicance level we use here.
Minimizing an information criterion

Many multiple-testing problems in the time-series literature have been solved by dening an estimator that minimizes an information criterion with known asymptotic properties. Selecting the lag length in an autoregressive model is probably the best-known example. Gonzalo and Pitarakis (1998) and Aznar and Salvador (2002) have shown that this approach can be applied to determining the number of cointegrating equations in a VECM. As in the lag-length selection problem, choosing the number of cointegrating equations that minimizes either the Schwarz Bayesian information criterion (SBIC) or the Hannan and Quinn information criterion (HQIC) provides a consistent estimator of the number of cointegrating equations.
Example 4
We use these information-criteria methods to estimate the number of cointegrating equations in our balanced-growth data.

. vecrank y i c, lags(5) ic notrace Johansen tests for cointegration Trend: constant Number of obs = Sample: 1960q2 - 1982q4 Lags = maximum rank 0 1 2 3 91 5
681
parms 39 44 47 48
LL 1231.1041 1245.3882 1252.5055 1254.1787
eigenvalue 0.26943 0.14480 0.03611
SBIC -25.12401 -25.19009 -25.19781* -25.18501
HQIC -25.76596 -25.91435 -25.97144* -25.97511
AIC -26.20009 -26.40414 -26.49463 -26.50942
Both the SBIC and the HQIC estimators suggest that there are two cointegrating equations in the balanced-growth data.
Saved results
vecrank saves the following in e():
Scalars e(N) e(k eq) e(k dv) e(tmin) e(tmax) e(n lags) e(k ce95) e(k ce99) e(k cesbic) e(k cehqic) Macros e(cmd) e(cmdline) e(trend) e(reduced lags) e(reduce opt) e(tsfmt) Matrices e(max) e(trace) e(lambda) e(k rank) e(hqic) e(sbic) e(aic) e(ll) number of observations number of equations in e(b) number of dependent variables minimum time maximum time number of lags number of cointegrating equations number of cointegrating equations number of cointegrating equations number of cointegrating equations
chosen chosen chosen chosen
by by by by
multiple trace tests with level(95) multiple trace tests with level(99) minimizing SBIC minimizing HQIC
vecrank command as typed trend specied list of maximum lags to which the model has been reduced noreduce, if noreduce is specied format for current time variable vector vector vector vector vector vector vector vector of of of of of of of of maximum-eigenvalue statistics trace statistics eigenvalues numbers of unconstrained parameters HQIC values SBIC values AIC values log-likelihood values

vecrank is implemented as an ado-le. As shown in Methods and formulas of [TS] vec, given a lag, trend, and seasonal specication when there are 0 r K cointegrating equations, the log likelihood with the Johansen identication restrictions can be written as
682
1 L = T K {ln (2) + 1} + ln (|S00 |) + ln 1 i 2 i=1
(1)
where the (K K) matrix S00 and the eigenvalues i are dened in Methods and formulas of [TS] vec. The trace statistic compares the null hypothesis that there are r or fewer cointegrating relations with the alternative hypothesis that there are more than r cointegrating equations. Under the alternative hypothesis, the log likelihood is
1 ln 1 i LA = T K {ln (2) + 1} + ln (|S00 |) + 2 i=1
(2)
Thus the LR test that compares the unrestricted model in (2) with the restricted model in (1) is given by
K
LRtrace = T
i=r+1
ln 1 i
As discussed by Johansen (1995), the trace statistic has a nonstandard distribution under the null hypothesis because the null hypothesis places restrictions on the coefcients on yt1 , which is assumed to have K r random-walk components. vecrank reports the Osterwald-Lenum (1992) critical values. The maximum-eigenvalue statistic compares the null model containing r cointegrating relations with the alternative model that has r + 1 cointegrating relations. Thus using these two values for r in (1) and a few lines of algebra implies that the LR test of this hypothesis is
LRmax = T ln 1 r+1
As for the trace statistic, because this test involves restrictions on the coefcients on a vector of I(1) variables, the test statistics distribution will be nonstandard. vecrank reports the OsterwaldLenum (1992) critical values. The formulas for the AIC, SBIC, and HQIC are given in Methods and formulas of [TS] vec.
Sren Johansen (1939 ) earned degrees in mathematical statistics at the University of Copenhagen, where he is now based. In addition to making contributions to mathematical statistics, probability theory, and medical statistics, he has worked mostly in econometricsin particular, on the theory of cointegration.
References
Anderson, T. W. 1951. Estimating linear restrictions on regression coefcients for multivariate normal distributions. Annals of Mathematical Statistics 22: 327351. Aznar, A., and M. Salvador. 2002. Selecting the rank of the cointegration space and the form of the intercept using an information criterion. Econometric Theory 18: 926947. Engle, R. F., and C. W. J. Granger. 1987. Co-integration and error correction: Representation, estimation, and testing. Econometrica 55: 251276.
683
Gonzalo, J., and J.-Y. Pitarakis. 1998. Specication via model selection in vector error correction models. Economics Letters 60: 321328. Hamilton, J. D. 1994. Time Series Analysis. Princeton: Princeton University Press. Hubrich, K., H. L tkepohl, and P. Saikkonen. 2001. A review of systems cointegration tests. Econometric Reviews u 20: 247318. Johansen, S. 1988. Statistical analysis of cointegration vectors. Journal of Economic Dynamics and Control 12: 231254. . 1991. Estimation and hypothesis testing of cointegration vectors in Gaussian vector autoregressive models. Econometrica 59: 15511580. . 1995. Likelihood-Based Inference in Cointegrated Vector Autoregressive Models. Oxford: Oxford University Press. King, R. G., C. I. Plosser, J. H. Stock, and M. W. Watson. 1991. Stochastic trends and economic uctuations. American Economic Review 81: 819840. L tkepohl, H. 2005. New Introduction to Multiple Time Series Analysis. New York: Springer. u Maddala, G. S., and I.-M. Kim. 1998. Unit Roots, Cointegration, and Structural Change. Cambridge: Cambridge University Press. Osterwald-Lenum, M. 1992. A note with quantiles of the asymptotic distribution of the maximum likelihood cointegration rank test statistics. Oxford Bulletin of Economics and Statistics 54: 461472. Park, J. Y., and P. C. B. Phillips. 1988. Statistical inference in regressions with integrated processes: Part I. Econometric Theory 4: 468497. . 1989. Statistical inference in regressions with integrated processes: Part II. Econometric Theory 5: 95131. Phillips, P. C. B. 1986. Understanding spurious regressions in econometrics. Journal of Econometrics 33: 311340. Phillips, P. C. B., and S. N. Durlauf. 1986. Multiple time series regressions with integrated processes. Review of Economic Studies 53: 473495. Sims, C. A., J. H. Stock, and M. W. Watson. 1990. Inference in linear time series models with some unit roots. Econometrica 58: 113144. Stock, J. H. 1987. Asymptotic properties of least squares estimators of cointegrating vectors. Econometrica 55: 10351056. Stock, J. H., and M. W. Watson. 1988. Testing for common trends. Journal of the American Statistical Association 83: 10971107. Watson, M. W. 1994. Vector autoregressions and cointegration. In Vol. IV of Handbook of Econometrics, ed. R. F. Engle and D. L. McFadden. Amsterdam: Elsevier.
Also see
[TS] tsset Declare data to be time-series data [TS] vec Vector error-correction models [TS] vec intro Introduction to vector error-correction models
Title
vecstable Check the stability condition of VECM estimates
Syntax
vecstable options
Main
, options Description use previously saved results estname; default is to use active results save the companion matrix as matrix name graph eigenvalues of the companion matrix label eigenvalues with the distance from the unit circle label eigenvalues with the modulus change look of markers (color, size, etc.) affect rendition of reference unit circle suppress polar grid circles specify radii and appearance of polar grid circles; see Options for details add other plots to the generated graph any options other than by() documented in [G-3] twoway options
estimates(estname) amat(matrix name) graph dlabel modlabel marker options rlopts(cline options) nogrid pgrid( . . . )
Add plots
vecstable can be used only after vec; see [TS] vec.
Menu
Statistics
>
>
>
Check stability condition of VEC estimates
Description
vecstable checks the eigenvalue stability condition in a vector error-correction model (VECM) t using vec.
Options

Main
estimates(estname) requests that vecstable use the previously obtained set of vec estimates saved as estname By default, vecstable uses the active results. See [R] estimates for information about saving and restoring estimation results. amat(matrix name) species a valid Stata matrix name by which the companion matrix can be saved. The companion matrix is referred to as the A matrix in L tkepohl (2005) and [TS] varstable. The u default is not to save the companion matrix.
684
685
graph causes vecstable to draw a graph of the eigenvalues of the companion matrix. dlabel labels the eigenvalues with their distances from the unit circle. dlabel cannot be specied with modlabel. modlabel labels the eigenvalues with their moduli. modlabel cannot be specied with dlabel. marker options specify the look of markers. This look includes the marker symbol, the marker size, and its color and outline; see [G-3] marker options. rlopts(cline options) affects the rendition of the reference unit circle; see [G-3] cline options. nogrid suppresses the polar grid circles. pgrid( numlist , line options ) pgrid( numlist , line options ) . . . pgrid( numlist , line options ) determines the radii and appearance of the polar grid circles. By default, the graph includes nine polar grid circles with radii .1, .2, . . . , .9 that have the grid linestyle. The numlist species the radii for the polar grid circles. The line options determine the appearance of the polar grid circles; see [G-3] line options. Because the pgrid() option can be repeated, circles with different radii can have distinct appearances.
Add plots
Remarks
Inference after vec requires that the cointegrating equations be stationary and that the number of cointegrating equations be correctly specied. Although the methods implemented in vecrank identify the number of stationary cointegrating equations, they assume that the individual variables are I(1). vecstable provides indicators of whether the number of cointegrating equations is misspecied or whether the cointegrating equations, which are assumed to be stationary, are not stationary. vecstable is analogous to varstable. vecstable uses the coefcient estimates from the previously tted VECM to back out estimates of the coefcients of the corresponding VAR and then compute the eigenvalues of the companion matrix. See [TS] varstable for details about how the companion matrix is formed and about how to interpret the resulting eigenvalues for covariancestationary VAR models. If a VECM has K endogenous variables and r cointegrating vectors, there will be K r unit moduli in the companion matrix. If any of the remaining moduli computed by vecrank are too close to one, either the cointegrating equations are not stationary or there is another common trend and the rank() specied in the vec command is too high. Unfortunately, there is no general distribution theory that allows you to determine whether an estimated root is too close to one for all the cases that commonly arise in practice.
686
Example 1
In example 1 of [TS] vec, we estimated the parameters of a bivariate VECM of the natural logs of the average disposable incomes in two of the economic regions created by the U.S. Bureau of Economic Analysis. In that example, we concluded that the predicted cointegrating equation was probably not stationary. Here we continue that example by retting that model and using vecstable to analyze the eigenvalues of the companion matrix of the corresponding VAR.
. use http://www.stata-press.com/data/r12/rdinc . vec ln_ne ln_se (output omitted ) . vecstable Eigenvalue stability condition Eigenvalue 1 .9477854 .2545357 + .2545357 Modulus 1 .947785 .343914 .343914
.2312756i .2312756i
The VECM specification imposes a unit modulus.
The output contains a table showing the eigenvalues of the companion matrix and their associated moduli. The table shows that one of the roots is 1. The table footer reminds us that the specied VECM imposes one unit modulus on the companion matrix. The output indicates that there is a real root at about 0.95. Although there is no distribution theory to measure how close this root is to one, per other discussions in the literature (for example, Johansen [1995, 137138]), we conclude that the root of 0.95 supports our earlier analysis, in which we concluded that the predicted cointegrating equation is probably not stationary. If we had included the graph option with vecstable, the following graph would have been displayed:
1 1 1 .5 Imaginary 0 .5
.5
0 Real
.5
The VECM specification imposes 1 unit modulus
The graph plots the eigenvalues of the companion matrix with the real component on the x axis and the imaginary component on the y axis. Although the information is the same as in the table, the graph shows visually how close the root with modulus 0.95 is to the unit circle.
687
Saved results
vecstable saves the following in r():
Scalars r(unitmod) Matrices r(Re) r(Im) r(Modulus) number of unit moduli imposed on the companion matrix real part of the eigenvalues of A imaginary part of the eigenvalues of A moduli of the eigenvalues of A
where A is the companion matrix of the VAR that corresponds to the VECM.

vecstable is implemented as an ado-le. vecstable uses the formulas given in Methods and formulas of [TS] irf create to obtain estimates of the parameters in the corresponding VAR from the vec estimates. With these estimates, the calculations are identical to those discussed in [TS] varstable. In particular, the derivation of the companion matrix, A, from the VAR point estimates is given in [TS] varstable.
References
Hamilton, J. D. 1994. Time Series Analysis. Princeton: Princeton University Press. Johansen, S. 1995. Likelihood-Based Inference in Cointegrated Vector Autoregressive Models. Oxford: Oxford University Press. L tkepohl, H. 2005. New Introduction to Multiple Time Series Analysis. New York: Springer. u
Also see
[TS] vec Vector error-correction models [TS] vec intro Introduction to vector error-correction models
Title
wntestb Bartletts periodogram-based test for white noise
Syntax
wntestb varname options
Main
if
in
, options
Description display a table instead of graphical output set condence level; default is level(95) change look of markers (color, size, etc.) add marker labels; change look or position add connecting lines; change look add other plots to the generated graph any options other than by() documented in [G-3] twoway options
table level(#)
Plot
marker options marker label options cline options

Add plots
You must tsset your data before using wntestb; see [TS] tsset. In addition, the time series must be dense (nonmissing with no gaps in the time variable) in the specied sample. varname may contain time-series operators; see [U] 11.4.4 Time-series varlists.
Menu
Statistics
>
Time series
>
Tests
>
Bartletts periodogram-based white-noise test
Description
wntestb performs Bartletts periodogram-based test for white noise. The result is presented graphically by default but optionally may be presented as text in a table.
Options

Main
table displays the test results as a table instead of as the default graph. level(#) species the condence level, as a percentage, for the condence bands included on the graph. The default is level(95) or as set by set level; see [U] 20.7 Specifying the width of condence intervals.
688
689
Plot
marker options specify the look of markers. This look includes the marker symbol, the marker size, and its color and outline; see [G-3] marker options. marker label options specify if and how the markers are to be labeled; see [G-3] marker label options. cline options specify if the points are to be connected with lines and the rendition of those lines; see [G-3] cline options.
Add plots
Remarks
Bartletts test is a test of the null hypothesis that the data come from a white-noise process of uncorrelated random variables having a constant mean and a constant variance. For a discussion of this test, see Bartlett (1955, 9294), Newton (1988, 172), or Newton (1996).
Example 1
In this example, we generate two time series and show the graphical and statistical tests that can be obtained from this command. The rst time series is a white-noise process, and the second is a white-noise process with an embedded deterministic cosine curve.
. drop _all . set seed 12393 . set obs 100 obs was 0, now 100 . generate x1 = rnormal() . generate x2 = rnormal() + cos(2*_pi*(_n-1)/10) . generate time = _n . tsset time time variable: time, 1 to 100 delta: 1 unit
We can then submit the white-noise data to the wntestb command by typing
690

. wntestb x1
0.00
Cumulative periodogram for x1 0.20 0.40 0.60 0.80
1.00
Cumulative Periodogram WhiteNoise Test
0.00
0.10
0.20 0.30 Frequency

0.71 Prob > B = 0.6957
0.40
0.50
Bartletts (B) statistic =
We can see in the graph that the values never appear outside the condence bands. The test statistic has a p-value of 0.91, so we conclude that the process is not different from white noise. If we had wanted only the statistic without the plot, we could have used the table option. Turning our attention to the other series (x2), we type
. wntestb x2
0.00
Cumulative periodogram for x2 0.20 0.40 0.60 0.80
1.00
Cumulative Periodogram WhiteNoise Test
0.00
0.10
0.20 0.30 Frequency

1.83 Prob > B = 0.0024
0.40
0.50
Bartletts (B) statistic =
Here the process does appear outside of the bands. In fact, it steps out of the bands at a frequency of 0.1 (exactly as we synthesized this process). We also have conrmation from the test statistic, at a p-value of 0.001, that the process is signicantly different from white noise.
691
Saved results
wntestb saves the following in r():
Scalars r(stat) Bartletts statistic r(p) probability value

wntestb is implemented as an ado-le. If x(1), . . . , x(T ) is a realization from a white-noise process with variance 2 , the spectral distribution would be given by F () = 2 for [ 0, 1 ], and we would expect the cumulative periodogram (see [TS] cumsp) of the data to be close to the points Sk = k/q for q = n/2 + 1, k = 1, . . . , q . n/2 is the greatest integer less than or equal to n/2. Except for = 0 and = .5, the random variables 2f (k )/ 2 are asymptotically independently and identically distributed as 2 . Because 2 is the same as twice a random variable distributed 2 2 exponentially with mean 1, the cumulative periodogram has approximately the same distribution as the ordered values from a uniform (on the unit interval) distribution. Feller (1948) shows that this results in 2 2 k lim Pr max q Uk a = (1)j e2a j = G(a) q 1kq q
j=
where Uk is the ordered uniform quantile. The Bartlett statistic is computed as
B = max
1kq
n k Fk 2 q
where Fk is the cumulative periodogram dened in terms of the sample spectral density f (see [TS] pergram) as k j=1 f (j ) Fk = q j=1 f (j ) The associated p-value for the Bartlett statistic and the condence bands on the graph are computed as 1 G(B) using Fellers result.
Maurice Stevenson Bartlett (19102002) was a British statistician. Apart from a short period in industry, he spent his career teaching and researching at the universities of Cambridge, Manchester, London (University College), and Oxford. His many contributions include work on the statistical analysis of multivariate data (especially factor analysis) and time series and on stochastic models of population growth, epidemics, and spatial processes.
Acknowledgment
wntestb is based on the wntestf command by H. Joseph Newton (1996), Department of Statistics, Texas A&M University.
692
References
Bartlett, M. S. 1955. An Introduction to Stochastic Processes with Special Reference to Methods and Applications. Cambridge: Cambridge University Press. Feller, W. 1948. On the KolmogorovSmirnov limit theorems for empirical distributions. Annals of Mathematical Statistics 19: 177189. Gani, J. 2002. Professor M. S. Bartlett FRS, 19102002. Statistician 51: 399402. Newton, H. J. 1988. TIMESLAB: A Time Series Analysis Laboratory. Belmont, CA: Wadsworth. . 1996. sts12: A periodogram-based test for white noise. Stata Technical Bulletin 34: 3639. Reprinted in Stata Technical Bulletin Reprints, vol. 6, pp. 203207. College Station, TX: Stata Press. Olkin, I. 1989. A conversation with Maurice Bartlett. Statistical Science 4: 151163.
Also see
[TS] tsset Declare data to be time-series data [TS] corrgram Tabulate and graph autocorrelations [TS] cumsp Cumulative spectral distribution [TS] pergram Periodogram [TS] wntestq Portmanteau (Q) test for white noise
Title
wntestq Portmanteau (Q) test for white noise
Syntax
wntestq varname if in , lags(#)
You must tsset your data before using wntestq; see [TS] tsset. Also the time series must be dense (nonmissing with no gaps in the time variable) in the specied sample. varname may contain time-series operators; see [U] 11.4.4 Time-series varlists.
Menu
Statistics
>
Time series
>
Tests
>
Portmanteau white-noise test
Description
wntestq performs the portmanteau (or Q) test for white noise.
Option
lags(#) species the number of autocorrelations to calculate. The default is to use min( n/2 2, 40), where n/2 is the greatest integer less than or equal to n/2.
Remarks
Box and Pierce (1970) developed a portmanteau test of white noise that was rened by Ljung and Box (1978). See also Diggle (1990, sec. 2.5).
Example 1
In the example shown in [TS] wntestb, we generated two time series. One (x1) was a white-noise process, and the other (x2) was a white-noise process with an embedded cosine curve. Here we compare the output of the two tests.
. drop _all . set seed 12393 . set obs 100 obs was 0, now 100 . generate x1 = rnormal() . generate x2 = rnormal() + cos(2*_pi*(_n-1)/10) . generate time = _n . tsset time time variable: time, 1 to 100 delta: 1 unit
693
694

. wntestb x1, table Cumulative periodogram white-noise test Bartletts (B) statistic = 0.7093 Prob > B = 0.6957 . wntestq x1 Portmanteau test for white noise Portmanteau (Q) statistic = Prob > chi2(40) = 32.6863 0.7875
. wntestb x2, table Cumulative periodogram white-noise test Bartletts (B) statistic = 1.8323 Prob > B = 0.0024 . wntestq x2 Portmanteau test for white noise Portmanteau (Q) statistic = Prob > chi2(40) = 129.4436 0.0000
This example shows that both tests agree. For the rst process, the Bartlett and portmanteau tests result in nonsignicant test statistics: a p-value of 0.9053 for wntestb and one of 0.9407 for wntestq. For the second process, each test has a signicant result to 0.0010.
Saved results
wntestq saves the following in r():
Scalars r(stat) r(df)
Q statistic degrees of freedom
r(p)
probability value

wntestq is implemented as an ado-le. The portmanteau test relies on the fact that if x(1), . . . , x(n) is a realization from a white-noise process. Then m 1 2 (j) 2 Q = n(n + 2) m nj
j=1
where m is the number of autocorrelations calculated (equal to the number of lags specied) and indicates convergence in distribution to a 2 distribution with m degrees of freedom. j is the estimated autocorrelation for lag j ; see [TS] corrgram for details.
695
References
Box, G. E. P., and D. A. Pierce. 1970. Distribution of residual autocorrelations in autoregressive-integrated moving average time series models. Journal of the American Statistical Association 65: 15091526. Diggle, P. J. 1990. Time Series: A Biostatistical Introduction. Oxford: Oxford University Press. Ljung, G. M., and G. E. P. Box. 1978. On a measure of lack of t in time series models. Biometrika 65: 297303. Sperling, R., and C. F. Baum. 2001. sts19: Multivariate portmanteau (Q) test for white noise. Stata Technical Bulletin 60: 3941. Reprinted in Stata Technical Bulletin Reprints, vol. 10, pp. 373375. College Station, TX: Stata Press.
Also see
[TS] tsset Declare data to be time-series data [TS] corrgram Tabulate and graph autocorrelations [TS] cumsp Cumulative spectral distribution [TS] pergram Periodogram [TS] wntestb Bartletts periodogram-based test for white noise
Title
xcorr Cross-correlogram for bivariate time series
Syntax
xcorr varname1 varname2 options
Main
if
in
, options
Description create newvar containing cross-correlation values display a table instead of graphical output do not include the character-based plot in tabular output include # lags and leads in graph value to drop to; default is 0 change look of markers (color, size, etc.) add marker labels; change look or position change look of dropped lines add other plots to the generated graph any options other than by() documented in [G-3] twoway options
generate(newvar) table noplot lags(#)

Plot
base(#) marker options marker label options line options

Add plots
You must tsset your data before using xcorr; see [TS] tsset. varname1 and varname2 may contain time-series operators; see [U] 11.4.4 Time-series varlists.
Menu
Statistics
>
Time series
>
Graphs
>
Cross-correlogram for bivariate time series
Description
xcorr plots the sample cross-correlation function.
Options

Main
generate(newvar) species a new variable to contain the cross-correlation values. table requests that the results be presented as a table rather than the default graph. noplot requests that the table not include the character-based plot of the cross-correlations. lags(#) indicates the number of lags and leads to include in the graph. The default is to use min( n/2 2, 20).
696
697
Plot
base(#) species the value from which the lines should extend. The default is base(0). marker options, marker label options, and line options affect the rendition of the plotted crosscorrelations. marker options specify the look of markers. This look includes the marker symbol, the marker size, and its color and outline; see [G-3] marker options. marker label options specify if and how the markers are to be labeled; see [G-3] marker label options. line options specify the look of the dropped lines, including pattern, width, and color; see [G-3] line options.
Add plots
addplot(plot) provides a way to add other plots to the generated graph. See [G-3] addplot option.
Remarks
Example 1
We have a bivariate time series (Box, Jenkins, and Reinsel 2008, Series J) on the input and output of a gas furnace, where 296 paired observations on the input (gas rate) and output (% CO2 ) were recorded every 9 seconds. The cross-correlation function is given by
. use http://www.stata-press.com/data/r12/furnace (TIMESLAB: Gas furnace) . xcorr input output, xline(5) lags(40)
Crosscorrelations of input and output 0.50 0.00 0.50 1.00
1.00
40
20
0 Lag
20
40
1.00
0.50
0.00
0.50
1.00
Crosscorrelogram
698
We included a vertical line at lag 5, because there is a well-dened peak at this value. This peak indicates that the output lags the input by ve periods. Further, the fact that the correlations are negative indicates that as input (coded gas rate) is increased, output (% CO2 ) decreases. We may obtain the table of autocorrelations and the character-based plot of the cross-correlations (analogous to the univariate time-series command corrgram) by specifying the table option.
. xcorr input output, table -1 0 1 LAG CORR [Cross-correlation] -20 -19 -18 -17 -16 -15 -14 -13 -12 -11 -10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 -0.1033 -0.1027 -0.0998 -0.0932 -0.0832 -0.0727 -0.0660 -0.0662 -0.0751 -0.0927 -0.1180 -0.1484 -0.1793 -0.2059 -0.2266 -0.2429 -0.2604 -0.2865 -0.3287 -0.3936 -0.4845 -0.5985 -0.7251 -0.8429 -0.9246 -0.9503 -0.9146 -0.8294 -0.7166 -0.5998 -0.4952 -0.4107 -0.3479 -0.3049 -0.2779 -0.2632 -0.2548 -0.2463 -0.2332 -0.2135 -0.1869
Once again, the well-dened peak is apparent in the plot.
699

xcorr is implemented as an ado-le. The cross-covariance function of lag k for time series x1 and x2 is given by
Cov x1 (t), x2 (t + k) = R12 (k)

This function is not symmetric about lag zero; that is,
R12 (k) = R12 (k)

We dene the cross-correlation function as
ij (k) = Corr xi (t), xj (t + k) =
Rij (k) Rii (0)Rjj (0)
where 11 and 22 are the autocorrelation functions for x1 and x2 , respectively. The sequence 12 (k) is the cross-correlation function and is drawn for lags k (Q, Q + 1, . . . , 1, 0, 1, . . . , Q 1, Q). If 12 (k) = 0 for all lags, x1 and x2 are not cross-correlated.
References
Box, G. E. P., G. M. Jenkins, and G. C. Reinsel. 2008. Time Series Analysis: Forecasting and Control. 4th ed. Hoboken, NJ: Wiley. Hamilton, J. D. 1994. Time Series Analysis. Princeton: Princeton University Press. Newton, H. J. 1988. TIMESLAB: A Time Series Analysis Laboratory. Belmont, CA: Wadsworth.
Also see
[TS] tsset Declare data to be time-series data [TS] corrgram Tabulate and graph autocorrelations [TS] pergram Periodogram
700
Glossary
Glossary
ARCH model. An autoregressive conditional heteroskedasticity (ARCH) model is a regression model in which the conditional variance is modeled as an autoregressive (AR) process. The ARCH(m) model is yt = x t + t
E( 2 | t
2 2 t1 , t2 , . . .)
= 0 + 1
2 t1
+ + m
2 tm
where t is a white-noise error term. The equation for yt represents the conditional mean of the process, and the equation for E( 2 | 2 , 2 , . . .) species the conditional variance as an t t1 t2 autoregressive function of its past realizations. Although the conditional variance changes over time, the unconditional variance is time invariant because yt is a stationary process. Modeling the conditional variance as an AR process raises the implied unconditional variance, making this model particularly appealing to researchers modeling fat-tailed data, such as nancial data.
ARFIMA model. An autoregressive fractionally integrated moving-average (ARFIMA) model is a timeseries model suitable for use with long-memory processes. ARFIMA models generalize autoregressive integrated moving-average (ARIMA) models by allowing the differencing parameter to be a real number in (0.5, 0.5) instead of requiring it to be an integer. ARIMA model. An autoregressive integrated moving-average (ARIMA) model is a time-series model suitable for use with integrated processes. In an ARIMA(p, d, q) model, the data is differenced d times to obtain a stationary series, and then an ARMA(p, q) model is t to this differenced data. ARIMA models that include exogenous explanatory variables are known as ARMAX models. ARMA model. An autoregressive moving-average (ARMA) model is a time-series model in which the current periods realization is the sum of an autoregressive (AR) process and a moving-average (MA) process. An ARMA(p, q) model includes p AR terms and q MA terms. ARMA models with just a few lags are often able to t data, as well as pure AR or MA models with many more lags. ARMAX model. An ARMAX model is a time-series model in which the current periods realization is an ARMA process plus a linear function of a set of a exogenous variables. Equivalently, an ARMAX model is a linear regression model in which the error term is specied to follow an ARMA process. autocorrelation function. The autocorrelation function (ACF) expresses the correlation between periods t and t k of a time series as function of the time t and the lag k . For a stationary time series, the ACF does not depend on t and is symmetric about k = 0, meaning that the correlation between periods t and t k is equal to the correlation between periods t and t + k . autoregressive process. An autoregressive process is a time-series model in which the current value of a variable is a linear function of its own past values and a white-noise error term. A rst-order autoregressive process, denoted as an AR(1) process, is yt = yt1 + t . An AR(p) model contains p lagged values of the dependent variable. band-pass lter. Time-series lters are designed to pass or block stochastic cycles at specied frequencies. Band-pass lters, such as those implemented in tsfilter bk and tsfilter cf, pass through stochastic cycles in the specied range of frequencies and block all other stochastic cycles. Cholesky ordering. Cholesky ordering is a method used to orthogonalize the error term in a VAR or VECM to impose a recursive structure on the dynamic model, so that the resulting impulseresponse functions can be given a causal interpretation. The method is so named because it uses the Cholesky decomposition of the error covariance matrix.
Glossary
701
CochraneOrcutt estimator. This estimation is a linear regression estimator that can be used when the error term exhibits rst-order autocorrelation. An initial estimate of the autocorrelation parameter is obtained from OLS residuals, and then OLS is performed on the transformed data yt = yt yt1 and xt = xt xt1 . cointegrating vector. A cointegrating vector species a stationary linear combination of nonstationary variables. Specically, if each of the variables x1 , x2 , . . . , xk is integrated of order one and there exists a set of parameters 1 , 2 , . . . , k such that zt = 1 x1 + 2 x2 + + k xk is a stationary process, the variables x1 , x2 , . . . , xk are said to be cointegrated, and the vector is known as a cointegrating vector. conditional variance. Although the conditional variance is simply the variance of a conditional distribution, in time-series analysis the conditional variance is often modeled as an autoregressive process, giving rise to ARCH models. correlogram. A correlogram is a table or graph showing the sample autocorrelations or partial autocorrelations of a time series. covariance stationarity. A process is covariance stationary if the mean of the process is nite and independent of t, the unconditional variance of the process is nite and independent of t, and the covariance between periods t and t s is nite and depends on t s but not on t or s themselves. Covariance-stationary processes are also known as weakly stationary processes. cross-correlation function. The cross-correlation function expresses the correlation between one series at time t and another series at time t k as a function of the time t and lag k . If both series are stationary, the function does not depend on t. The function is not symmetric about k = 0: 12 (k) = 12 (k). cyclical component. A cyclical component is a part of a time series that is a periodic function of time. Deterministic functions of time are deterministic cyclical components, and random functions of time are stochastic cyclical components. For example, xed seasonal effects are deterministic cyclical components and random seasonal effects are stochastic seasonal components.
Random coefcients on time inside of periodic functions form an especially useful class of stochastic cyclical components; see [TS] ucm.
deterministic trend. A deterministic trend is a deterministic function of time that species the long-run tendency of a time series. difference operator. The difference operator denotes the change in the value of a variable from period t 1 to period t. Formally, yt = yt yt1 , and 2 yt = (yt yt1 ) = (yt yt1 ) (yt1 yt2 ) = yt 2yt1 + yt2 . drift. Drift is the constant term in a unit-root process. In
yt = + yt1 + is the drift when

t
is a stationary, zero-mean process.
dynamic forecast. A dynamic forecast is one in which the current periods forecast is calculated using forecasted values for prior periods. dynamic-multiplier function. A dynamic-multiplier function measures the effect of a shock to an exogenous variable on an endogenous variable. The k th dynamic-multiplier function of variable i on variable j measures the effect on variable j in period t + k in response to a one-unit shock to variable i in period t, holding everything else constant. endogenous variable. An endogenous variable is a regressor that is correlated with the unobservable error term. Equivalently, an endogenous variable is one whose values are determined by the equilibrium or outcome of a structural model.
702
Glossary
exogenous variable. An exogenous variable is one that is correlated with none of the unobservable error terms in the model. Equivalently, an exogenous variable is one whose values change independently of the other variables in a structural model. exponential smoothing. Exponential smoothing is a method of smoothing a time series in which the smoothed value at period t is equal to a fraction of the series value at time t plus a fraction 1 of the previous periods smoothed value. The fraction is known as the smoothing parameter. forecast-error variance decomposition. Forecast-error variance decompositions measure the fraction of the error in forecasting variable i after h periods that is attributable to the orthogonalized shocks to variable j . forward operator. The forward operator F denotes the value of a variable at time t + 1. Formally, F yt = yt+1 , and F 2 yt = F yt+1 = yt+2 . frequency-domain analysis. Frequency-domain analysis is analysis of time-series data by considering its frequency properties. The spectral density and distribution functions are key components of frequency-domain analysis, so it is often called spectral analysis. In Stata, the cumsp and pergram commands are used to analyze the sample spectral distribution and density functions, respectively. psdensity estimates the spectral density or the spectral distribution function after estimating the parameters of a parametric model using arfima, arima, or ucm. gain (of a linear lter). The gain of a linear lter scales the spectral density of the unltered series into the spectral density of the ltered series for each frequency. Specically, at each frequency, multiplying the spectral density of the unltered series by the square of the gain of a linear lter yields the spectral density of the ltered series. If the gain at a particular frequency is 1, the ltered and unltered spectral densities are the same at that frequency and the corresponding stochastic cycles are passed through perfectly. If the gain at a particular frequency is 0, the lter removes all the corresponding stochastic cycles from the unltered series. GARCH model. A generalized autoregressive conditional heteroskedasticity (GARCH) model is a regression model in which the conditional variance is modeled as an ARMA process. The GARCH(m, k) model is yt = xt + t
2 t = 0 + 1 2 t1
+ + m
2 tm
2 2 + 1 t1 + + k tk
where the equation for yt represents the conditional mean of the process and t represents the conditional variance. See [TS] arch or Hamilton (1994, chap. 21) for details on how the conditional variance equation can be viewed as an ARMA process. GARCH models are often used because the ARMA specication often allows the conditional variance to be modeled with fewer parameters than are required by a pure ARCH model. Many extensions to the basic GARCH model exist; see [TS] arch for those that are implemented in Stata. See also ARCH model.
generalized least-squares estimator. A generalized least-squares (GLS) estimator is used to estimate the parameters of a regression function when the error term is heteroskedastic or autocorrelated. In the linear case, GLS is sometimes described as OLS on transformed data because the GLS estimator can be implemented by applying an appropriate transformation to the dataset and then using OLS. Granger causality. The variable x is said to Granger-cause variable y if, given the past values of y, past values of x are useful for predicting y. high-pass lter. Time-series lters are designed to pass or block stochastic cycles at specied frequencies. High-pass lters, such as those implemented in tsfilter bw and tsfilter hp, pass through stochastic cycles above the cutoff frequency and block all other stochastic cycles.
Glossary
703
HoltWinters smoothing. A set of methods for smoothing time-series data that assume that the value of a time series at time t can be approximated as the sum of a mean term that drifts over time, as well as a time trend whose strength also drifts over time. Variations of the basic method allow for seasonal patterns in data, as well. impulseresponse function. An impulseresponse function (IRF) measures the effect of a shock to an endogenous variable on itself or another endogenous variable. The k th impulseresponse function of variable i on variable j measures the effect on variable j in period t + k in response to a one-unit shock to variable i in period t, holding everything else constant. independent and identically distributed. A series of observations is independently and identically distributed (i.i.d.) if each observation is an independent realization from the same underlying distribution. In some contexts, the denition is relaxed to mean only that the observations are independent and have identical means and variances; see Davidson and MacKinnon (1993, 42). integrated process. A nonstationary process is integrated of order d, written I(d), if the process must be differenced d times to produce a stationary series. An I(1) process yt is one in which yt is stationary. Kalman lter. The Kalman lter is a recursive procedure for predicting the state vector in a state-space model. lag operator. The lag operator L denotes the value of a variable at time t 1. Formally, Lyt = yt1 , and L2 yt = Lyt1 = yt2 . linear lter. A linear lter is a sequence of weights used to compute a weighted average of a time series at each time period. More formally, a linear lter (L) is
(L) = 0 + 1 L + 2 L2 + =
=0
where L is the lag operator. Applying the linear lter (L) to the time series xt yields a sequence of weighted averages of xt :
(L)xt =
=0
L xt
long-memory process. A long-memory process is a stationary process whose autocorrelations decay at a slower rate than a short-memory process. ARFIMA models are typically used to represent long-memory processes, and ARMA models are typically used to represent short-memory processes. moving-average process. A moving-average process is a time-series process in which the current value of a variable is modeled as a weighted average of current and past realizations of a whitenoise process and, optionally, a time-invariant constant. By convention, the weight on the current realization of the white-noise process is equal to one, and the weights on the past realizations are known as the moving-average (MA) coefcients. A rst-order moving-average process, denoted as an MA(1) process, is yt = t1 + t . multivariate GARCH models. Multivariate GARCH models are multivariate time-series models in which the conditional covariance matrix of the errors depends on its own past and its past shocks. The acute trade-off between parsimony and exibility has given rise to a plethora of models; see [TS] mgarch. NeweyWest covariance matrix. The NeweyWest covariance matrix is a member of the class of heteroskedasticity- and autocorrelation-consistent (HAC) covariance matrix estimators used with time-series data that produces covariance estimates that are robust to both arbitrary heteroskedasticity and autocorrelation up to a prespecied lag.
704
Glossary
orthogonalized impulseresponse function. An orthogonalized impulseresponse function (OIRF) measures the effect of an orthogonalized shock to an endogenous variable on itself or another endogenous variable. An orthogonalized shock is one that affects one variable at time t but no other variables. See [TS] irf create for a discussion of the difference between IRFs and OIRFs. partial autocorrelation function. The partial autocorrelation function (PACF) expresses the correlation between periods t and t k of a time series as a function of the time t and lag k , after controlling for the effects of intervening lags. For a stationary time series, the PACF does not depend on t. The PACF is not symmetric about k = 0: the partial autocorrelation between yt and ytk is not equal to the partial autocorrelation between yt and yt+k . periodogram. A periodogram is a graph of the spectral density function of a time series as a function of frequency. The pergram command rst standardizes the amplitude of the density by the sample variance of the time series, and then plots the logarithm of that standardized density. Peaks in the periodogram represent cyclical behavior in the data. phase function. The phase function of a linear lter species how the lter changes the relative importance of the random components at different frequencies in the frequency domain. portmanteau statistic. The portmanteau, or Q, statistic is used to test for white noise and is calculated using the rst m autocorrelations of the series, where m is chosen by the user. Under the null hypothesis that the series is a white-noise process, the portmanteau statistic has a 2 distribution with m degrees of freedom. PraisWinsten estimator. A PraisWinsten estimator is a linear regression estimator that is used when the error term exhibits rst-order autocorrelation; see also CochraneOrcutt estimator. Here the rst observation in the dataset is transformed as y1 = 1 2 y1 and x1 = 1 2 x1 , so that the rst observation is not lost. The PraisWinsten estimator is a generalized least-squares estimator. priming values. Priming values are the initial, preestimation values used to begin a recursive process. random walk. A random walk is a time-series process in which the current periods realization is equal to the previous periods realization plus a white-noise error term: yt = yt1 + t . A random walk with drift also contains a nonzero time-invariant constant: yt = + yt1 + t . The constant term is known as the drift parameter. An important property of random-walk processes is that the best predictor of the value at time t + 1 is the value at time t plus the value of the drift parameter. recursive regression analysis. A recursive regression analysis involves performing a regression at time t by using all available observations from some starting time t0 through time t, performing another regression at time t + 1 by using all observations from time t0 through time t + 1, and so on. Unlike a rolling regression analysis, the rst period used for all regressions is held xed. rolling regression analysis. A rolling, or moving window, regression analysis involves performing regressions for each period by using the most recent m periods data, where m is known as the window size. At time t the regression is t using observations for times t 19 through time t; at time t + 1 the regression is t using the observations for time t 18 through t + 1; and so on. seasonal difference operator. The period-s seasonal difference operator s denotes the difference in the value of a variable at time t and time t s. Formally, s yt = yt yts , and 2 yt = s s (yt yts ) = (yt yts ) (yts yt2s ) = yt 2yts + yt2s . serial correlation. Serial correlation refers to regression errors that are correlated over time. If a regression model does not contained lagged dependent variables as regressors, the OLS estimates are consistent in the presence of mild serial correlation, but the covariance matrix is incorrect. When the model includes lagged dependent variables and the residuals are serially correlated, the
Glossary
705
OLS estimates are biased and inconsistent. See, for example, Davidson and MacKinnon (1993, chap. 10) for more information.
serial correlation tests. Because OLS estimates are at least inefcient and potentially biased in the presence of serial correlation, econometricians have developed many tests to detect it. Popular ones include the DurbinWatson (1950, 1951, 1971) test, the BreuschPagan (1980) test, and Durbins (1970) alternative test. See [R] regress postestimation time series. smoothing. Smoothing a time series refers to the process of extracting an overall trend in the data. The motivation behind smoothing is the belief that a time series exhibits a trend component as well as an irregular component and that the analyst is interested only in the trend component. Some smoothers also account for seasonal or other cyclical patterns. spectral analysis. See frequency-domain analysis. spectral density function. The spectral density function is the derivative of the spectral distribution function. Intuitively, the spectral density function f () indicates the amount of variance in a time series that is attributable to sinusoidal components with frequency . See also spectral distribution function. The spectral density function is sometimes called the spectrum. spectral distribution function. The (normalized) spectral distribution function F () of a process describes the proportion of variance that can be explained by sinusoids with frequencies in the range (0, ), where 0 . The spectral distribution and density functions used in frequencydomain analysis are closely related to the autocorrelation function used in time-domain analysis; see Chateld (2004, chap. 6) and Wei (2006, chap. 12). spectrum. See spectral density function. state-space model. A state-space model describes the relationship between an observed time series and an unobservable state vector that represents the state of the world. The measurement equation expresses the observed series as a function of the state vector, and the transition equation describes how the unobserved state vector evolves over time. By dening the parameters of the measurement and transition equations appropriately, one can write a wide variety of time-series models in the state-space form. steady-state equilibrium. The steady-state equilibrium is the predicted value of a variable in a dynamic model, ignoring the effects of past shocks, or, equivalently, the value of a variable, assuming that the effects of past shocks have fully died out and no longer affect the variable of interest. stochastic trend. A stochastic trend is a nonstationary random process. Unit-root process and random coefcients on time are two common stochastic trends. See [TS] ucm for examples and discussions of more commonly applied stochastic trends. strict stationarity. A process is strictly stationary if the joint distribution of y1 , . . . , yk is the same as the joint distribution of y1+ , . . . , yk+ for all k and . Intuitively, shifting the origin of the series by units has no effect on the joint distributions. structural model. In time-series analysis, a structural model is one that describes the relationship among a set of variables, based on underlying theoretical considerations. Structural models may contain both endogenous and exogenous variables. SVAR. A structural vector autoregressive (SVAR) model is a type of VAR in which short- or long-run constraints are placed on the resulting impulseresponse functions. The constraints are usually motivated by economic theory and therefore allow causal interpretations of the IRFs to be made. time-domain analysis. Time-domain analysis is analysis of data viewed as a sequence of observations observed over time. The autocorrelation function, linear regression, ARCH models, and ARIMA models are common tools used in time-domain analysis.
706
Glossary
trend. The trend species the long-run behavior in a time series. The trend can be deterministic or stochastic. Many economic, biological, health, and social time series have long-run tendencies to increase or decrease. Before the 1980s, most time-series analysis specied the long-run tendencies as deterministic functions of time. Since the 1980s, the stochastic trends implied by unit-root processes have become a standard part of the toolkit. unit-root process. A unit-root process is one that is integrated of order one, meaning that the process is nonstationary but that rst-differencing the process produces a stationary series. The simplest example of a unit-root process is the random walk. See Hamilton (1994, chap. 15) for a discussion of when general ARMA processes may contain a unit root. unit-root tests. Whether a process has a unit root has both important statistical and economic ramications, so a variety of tests have been developed to test for them. Among the earliest tests proposed is the one by Dickey and Fuller (1979), though most researchers now use an improved variant called the augmented DickeyFuller test instead of the original version. Other common unit-root tests implemented in Stata include the DFGLS test of Elliott, Rothenberg, and Stock (1996) and the PhillipsPerron (1988) test. See [TS] dfuller, [TS] dfgls, and [TS] pperron.
Variants of unit-root tests suitable for panel data have also been developed; see [XT] xtunitroot.
VAR. A vector autoregressive (VAR) model is a multivariate regression technique in which each dependent variable is regressed on lags of itself and on lags of all the other dependent variables in the model. Occasionally, exogenous variables are also included in the model. VECM. A vector error-correction model (VECM) is a type of VAR that is used with variables that are cointegrated. Although rst-differencing variables that are integrated of order one makes them stationary, tting a VAR to such rst-differenced variables results in misspecication error if the variables are cointegrated. See The multivariate VECM specication in [TS] vec intro for more on this point. white noise. A variable ut represents a white-noise process if the mean of ut is zero, the variance of ut is 2 , and the covariance between ut and us is zero for all s = t. Gaussian white noise refers to white noise in which ut is normally distributed. YuleWalker equations. The YuleWalker equations are a set of difference equations that describe the relationship among the autocovariances and autocorrelations of an autoregressive moving-average (ARMA) process.
References
Breusch, T. S., and A. R. Pagan. 1980. The Lagrange multiplier test and its applications to model specication in econometrics. Review of Economic Studies 47: 239253. Chateld, C. 2004. The Analysis of Time Series: An Introduction. 6th ed. Boca Raton, FL: Chapman & Hall/CRC. Davidson, R., and J. G. MacKinnon. 1993. Estimation and Inference in Econometrics. New York: Oxford University Press. Dickey, D. A., and W. A. Fuller. 1979. Distribution of the estimators for autoregressive time series with a unit root. Journal of the American Statistical Association 74: 427431. Durbin, J. 1970. Testing for serial correlation in least-squares regressions when some of the regressors are lagged dependent variables. Econometrica 38: 410421. Durbin, J., and G. S. Watson. 1950. Testing for serial correlation in least squares regression. I. Biometrika 37: 409428. . 1951. Testing for serial correlation in least squares regression. II. Biometrika 38: 159177. . 1971. Testing for serial correlation in least squares regression. III. Biometrika 58: 119. Elliott, G., T. J. Rothenberg, and J. H. Stock. 1996. Efcient tests for an autoregressive unit root. Econometrica 64: 813836.
Glossary
Hamilton, J. D. 1994. Time Series Analysis. Princeton: Princeton University Press.
707
Phillips, P. C. B., and P. Perron. 1988. Testing for a unit root in time series regression. Biometrika 75: 335346. Wei, W. W. S. 2006. Time Series Analysis: Univariate and Multivariate Methods. 2nd ed. Boston: Pearson.
Subject and author index 709
Subject and author index

This is the subject and author index for the Time-Series Reference Manual. Readers interested in topics other than time series should see the combined subject index (and the combined author index) in the Quick Reference and Index. Semicolons set off the most important entries from the rest. Sometimes no entry will be set off with semicolons, meaning that all entries are equally important.
autoregressive conditional heteroskedasticity, see ARCH fractionally integrated moving-average model, see ARFIMA integrated moving average, see ARIMA model, [TS] dfactor, [TS] psdensity, [TS] sspace, [TS] ucm moving average, see ARMA process, [TS] Glossary Aznar, A., [TS] vecrank
A
Abraham, B., [TS] tssmooth, [TS] tssmooth dexponential, [TS] tssmooth exponential, [TS] tssmooth hwinters, [TS] tssmooth shwinters ac command, [TS] corrgram add, irf subcommand, [TS] irf add Adkins, L. C., [TS] arch Ahn, S. K., [TS] vec intro Aielli, G. P., [TS] mgarch, [TS] mgarch dcc Akaike, H., [TS] varsoc Amemiya, T., [TS] varsoc Amisano, G., [TS] irf create, [TS] var intro, [TS] var svar, [TS] vargranger, [TS] varwle An, S., [TS] arma Anderson, B. D. O., [TS] sspace Anderson, T. W., [TS] vec, [TS] vecrank Ansley, C. F., [TS] arima A-PARCH, [TS] arch AR, [TS] arch, [TS] arma, [TS] arima, [TS] dfactor, [TS] sspace, [TS] ucm ARCH, also see multivariate GARCH effects, [TS] arch model, [TS] arch, [TS] Glossary postestimation, [TS] arch postestimation arch command, [TS] arch, [TS] arch postestimation ARFIMA model, [TS] arma, [TS] Glossary postestimation, [TS] arma postestimation, [TS] psdensity arfima command, [TS] arma, [TS] arma postestimation ARIMA model, [TS] arima, [TS] Glossary postestimation, [TS] arima postestimation, [TS] psdensity arima command, [TS] arima, [TS] arima postestimation ARMA, [TS] arch, [TS] arma, [TS] arima, [TS] sspace, [TS] ucm, [TS] Glossary ARMAX, [TS] arma, [TS] arima, [TS] dfactor, [TS] sspace, [TS] ucm, [TS] Glossary autocorrelation, [TS] arch, [TS] arma, [TS] arima, [TS] corrgram, [TS] dfactor, [TS] newey, [TS] prais, [TS] psdensity, [TS] sspace, [TS] ucm, [TS] var, [TS] varlmar, [TS] Glossary
B
Baillie, R. T., [TS] arma band-pass lters, [TS] tslter bk, [TS] tslter cf, [TS] Glossary Bartlett, M. S., [TS] wntestb Bartletts bands, [TS] corrgram periodogram test, [TS] wntestb Baum, C. F., [TS] time series, [TS] arch, [TS] arima, [TS] dfgls, [TS] rolling, [TS] tsset, [TS] var, [TS] wntestq Bauwens, L., [TS] mgarch BaxterKing lter, [TS] tslter, [TS] tslter bk Baxter, M., [TS] tslter, [TS] tslter bk, [TS] tslter cf Becketti, S., [TS] corrgram Bera, A. K., [TS] arch, [TS] varnorm, [TS] vecnorm Beran, J., [TS] arma, [TS] arma postestimation Berkes, I., [TS] mgarch Berndt, E. K., [TS] arch, [TS] arima Bianchi, G., [TS] tslter, [TS] tslter bw bk, tsfilter subcommand, [TS] tslter bk Black, F., [TS] arch block exogeneity, [TS] vargranger Bloomeld, P., [TS] arma Bollerslev, T., [TS] arch, [TS] arima, [TS] mgarch, [TS] mgarch ccc, [TS] mgarch dvech Boswijk, H. P., [TS] vec Bowerman, B. L., [TS] tssmooth, [TS] tssmooth dexponential, [TS] tssmooth exponential, [TS] tssmooth hwinters, [TS] tssmooth shwinters Box, G. E. P., [TS] arma, [TS] arima, [TS] corrgram, [TS] cumsp, [TS] dfuller, [TS] pergram, [TS] pperron, [TS] psdensity, [TS] wntestq, [TS] xcorr Breusch, T. S., [TS] Glossary Brockwell, P. J., [TS] corrgram, [TS] sspace Burns, A. F., [TS] tslter, [TS] tslter bk, [TS] tslter bw, [TS] tslter cf, [TS] tslter hp, [TS] ucm business days, [TS] intro Butterworth lter, [TS] tslter, [TS] tslter bw Butterworth, S., [TS] tslter, [TS] tslter bw bw, tsfilter subcommand, [TS] tslter bw
710 Subject and author index
C
Caines, P. E., [TS] sspace calendars, [TS] intro Casals, J., [TS] sspace ccc, mgarch subcommand, [TS] mgarch ccc cf, tsfilter subcommand, [TS] tslter cf cgraph, irf subcommand, [TS] irf cgraph Chang, Y., [TS] sspace Chateld, C., [TS] arima, [TS] corrgram, [TS] pergram, [TS] tssmooth, [TS] tssmooth dexponential, [TS] tssmooth exponential, [TS] tssmooth hwinters, [TS] tssmooth ma, [TS] tssmooth shwinters, [TS] Glossary Chatterjee, S., [TS] prais Cheung, Y.-W., [TS] dfgls Cholesky ordering, [TS] Glossary Chou, R. Y., [TS] arch ChristianoFitzgerald lter, [TS] tslter, [TS] tslter cf Christiano, L. J., [TS] irf create, [TS] tslter, [TS] tslter cf, [TS] var svar Chu-Chun-Lin, S., [TS] sspace clock time, [TS] tsset cluster estimator of variance, PraisWinsten and CochraneOrcutt regression, [TS] prais Cochrane, D., [TS] prais CochraneOrcutt regression, [TS] prais, [TS] Glossary cointegration, [TS] fcast compute, [TS] fcast graph, [TS] vec intro, [TS] vec, [TS] veclmar, [TS] vecnorm, [TS] vecrank, [TS] vecstable, [TS] Glossary compute, fcast subcommand, [TS] fcast compute Comte, F., [TS] mgarch conditional variance, [TS] arch, [TS] Glossary constant conditional-correlation model, [TS] mgarch, [TS] mgarch ccc constrained estimation ARCH, [TS] arch ARFIMA, [TS] arma ARIMA and ARMAX, [TS] arima dynamic factor model, [TS] dfactor GARCH model, [TS] mgarch ccc, [TS] mgarch dcc, [TS] mgarch dvech, [TS] mgarch vcc state-space model, [TS] sspace structural vector autoregressive models, [TS] var svar unobserved-components model, [TS] ucm vector autoregressive models, [TS] var vector error-correction models, [TS] vec correlogram, [TS] corrgram, [TS] Glossary corrgram command, [TS] corrgram covariance stationarity, [TS] Glossary Cox, N. J., [TS] tsline, [TS] tssmooth hwinters, [TS] tssmooth shwinters create, irf subcommand, [TS] irf create cross-correlation function, [TS] xcorr, [TS] Glossary cross-correlogram, [TS] xcorr ctable, irf subcommand, [TS] irf ctable
cumsp command, [TS] cumsp, [TS] psdensity cumulative spectral distribution, empirical, [TS] cumsp, [TS] psdensity cyclical component, [TS] Glossary
D
data manipulation, [TS] tsappend, [TS] tsll, [TS] tsreport, [TS] tsrevar, [TS] tsset David, J. S., [TS] arima Davidson, R., [TS] arch, [TS] arima, [TS] prais, [TS] sspace, [TS] varlmar, [TS] Glossary Davis, R. A., [TS] corrgram, [TS] sspace dcc, mgarch subcommand, [TS] mgarch dcc De Jong, P., [TS] dfactor, [TS] sspace, [TS] sspace postestimation, [TS] ucm DeGroot, M. H., [TS] arima Deistler, M., [TS] sspace del Rio, A., [TS] tslter hp describe, irf subcommand, [TS] irf describe deterministic trend, [TS] Glossary dexponential, tssmooth subcommand, [TS] tssmooth dexponential dfactor command, [TS] dfactor dfgls command, [TS] dfgls dfuller command, [TS] dfuller diagonal vech model, [TS] mgarch, [TS] mgarch dvech Dickens, R., [TS] prais Dickey, D. A., [TS] dfgls, [TS] dfuller, [TS] pperron, [TS] Glossary DickeyFuller test, [TS] dfgls, [TS] dfuller Diebold, F. X., [TS] arch difference operator, [TS] Glossary Diggle, P. J., [TS] arima, [TS] wntestq Ding, Z., [TS] arch Doornik, J. A., [TS] arma, [TS] vec double-exponential smoothing, [TS] tssmooth dexponential drift, [TS] Glossary drop, irf subcommand, [TS] irf drop Drukker, D. M., [TS] arma postestimation, [TS] sspace, [TS] vec Durbin, J., [TS] prais, [TS] ucm, [TS] Glossary DurbinWatson statistic, [TS] prais Durlauf, S. N., [TS] vec intro, [TS] vec, [TS] vecrank dvech, mgarch subcommand, [TS] mgarch dvech dynamic conditional-correlation model, [TS] mgarch, [TS] mgarch dcc dynamic factor model, [TS] dfactor, [TS] dfactor postestimation, also see state-space model dynamic forecast, [TS] arch, [TS] arma, [TS] fcast compute, [TS] fcast graph, [TS] mgarch, [TS] Glossary dynamic-multiplier function, [TS] irf, [TS] irf cgraph, [TS] irf create, [TS] irf ctable, [TS] irf ograph, [TS] irf table, [TS] var intro, [TS] Glossary dynamic regression model, [TS] arma, [TS] arima, [TS] var

dynamic structural simultaneous equations, [TS] var svar forecast, continued vector error-correction model, [TS] vec postestimation forecasting, [TS] arch, [TS] arma, [TS] arima, [TS] fcast compute, [TS] fcast graph, [TS] irf create, [TS] mgarch, [TS] tsappend, [TS] tssmooth, [TS] tssmooth dexponential, [TS] tssmooth exponential, [TS] tssmooth hwinters, [TS] tssmooth ma, [TS] tssmooth shwinters, [TS] ucm, [TS] var intro, [TS] var, [TS] vec intro, [TS] vec forward operator, [TS] Glossary fractionally integrated autoregressive moving average model, [TS] psdensity frequency-domain analysis, [TS] cumsp, [TS] pergram, [TS] psdensity, [TS] Glossary Friedman, M., [TS] arima Fuller, W. A., [TS] dfgls, [TS] dfuller, [TS] pperron, [TS] psdensity, [TS] tslter, [TS] tslter bk, [TS] ucm, [TS] Glossary
E
EGARCH, [TS] arch Eichenbaum, M., [TS] irf create, [TS] var svar eigenvalue stability condition, [TS] varstable, [TS] vecstable Elliott, G., [TS] dfgls, [TS] Glossary Enders, W., [TS] arch, [TS] arima, [TS] corrgram endogenous variable, [TS] Glossary Engle, R. F., [TS] arch, [TS] arima, [TS] dfactor, [TS] mgarch, [TS] mgarch dcc, [TS] mgarch dvech, [TS] mgarch vcc, [TS] vec intro, [TS] vec, [TS] vecrank Evans, C. L., [TS] irf create, [TS] var svar exogenous variable, [TS] Glossary exp list, [TS] rolling exponential smoothing, [TS] tssmooth, [TS] tssmooth exponential, [TS] Glossary exponential, tssmooth subcommand, [TS] tssmooth exponential
G
gain, [TS] Glossary Gani, J., [TS] wntestb GARCH, [TS] arch, [TS] Glossary, also see multivariate GARCH Gardiner, J. S., [TS] tssmooth, [TS] tssmooth dexponential, [TS] tssmooth exponential, [TS] tssmooth hwinters, [TS] tssmooth shwinters Gardner, E. S., Jr., [TS] tssmooth dexponential, [TS] tssmooth hwinters generalized autoregressive conditional heteroskedasticity, see ARCH model, see GARCH generalized least-squares estimator, [TS] Glossary Geweke, J., [TS] dfactor Giannini, C., [TS] irf create, [TS] var intro, [TS] var svar, [TS] vargranger, [TS] varwle Giles, D. E. A., [TS] prais GJR, [TS] arch Glosten, L. R., [TS] arch Golub, G. H., [TS] arma, [TS] arma postestimation G mez, V., [TS] tslter, [TS] tslter hp o Gonzalo, J., [TS] vec intro, [TS] vecrank Gourieroux, C., [TS] arima, [TS] mgarch ccc, [TS] mgarch dcc, [TS] mgarch vcc Gradshteyn, I. S., [TS] arma Granger, C. W. J., [TS] arch, [TS] arma, [TS] vargranger, [TS] vec intro, [TS] vec, [TS] vecrank Granger causality, [TS] vargranger, [TS] Glossary graph, fcast subcommand, [TS] fcast graph graph, irf subcommand, [TS] irf graph graphs, autocorrelations, [TS] corrgram correlogram, [TS] corrgram cross-correlogram, [TS] xcorr cumulative spectral density, [TS] cumsp
F
factor model, [TS] dfactor fcast compute command, [TS] fcast compute fcast graph command, [TS] fcast graph feasible generalized least squares, see FGLS Feller, W., [TS] wntestb FEVD, [TS] irf, [TS] irf create, [TS] irf ograph, [TS] irf table, [TS] var intro, [TS] varbasic, [TS] vec intro, [TS] Glossary FGLS, [TS] dfgls, [TS] prais, [TS] var lters, [TS] tslter, also see smoothers BaxterKing, [TS] tslter bk Butterworth, [TS] tslter bw ChristianoFitzgerald, [TS] tslter cf HodrickPrescott, [TS] tslter hp Fiorentini, G., [TS] mgarch Fitzgerald, T. J., [TS] tslter, [TS] tslter cf Flannery, B. P., [TS] arch, [TS] arima forecast-error variance decomposition, see FEVD forecast, ARCH model, [TS] arch postestimation ARFIMA model, [TS] arma postestimation ARIMA model, [TS] arima postestimation dynamic-factor model, [TS] dfactor postestimation MGARCH model, see multivariate GARCH postestimation state-space model, [TS] sspace postestimation structural vector autoregressive model, [TS] var svar postestimation unobserved-components model, [TS] ucm postestimation vector autoregressive model, [TS] var postestimation

graphs, continued forecasts, [TS] fcast graph impulseresponse functions, [TS] irf, [TS] irf cgraph, [TS] irf graph, [TS] irf ograph partial correlogram, [TS] corrgram periodogram, [TS] pergram white-noise test, [TS] wntestb Greene, W. H., [TS] arch, [TS] arima, [TS] corrgram, [TS] var Grifths, W. E., [TS] arch, [TS] prais Horv th, L., [TS] mgarch a Hosking, J. R. M., [TS] arma hp, tsfilter subcommand, [TS] tslter hp Hubrich, K., [TS] vec intro, [TS] vecrank Hurst, H. E., [TS] arma hwinters, tssmooth subcommand, [TS] tssmooth hwinters
I
impulseresponse functions, see IRF independent and identically distributed, [TS] Glossary information criterion, [TS] varsoc innovation accounting, [TS] irf integrated autoregressive moving average model, [TS] psdensity integrated process, [TS] Glossary IRF, [TS] irf, [TS] irf add, [TS] irf cgraph, [TS] irf create, [TS] irf ctable, [TS] irf describe, [TS] irf drop, [TS] irf graph, [TS] irf ograph, [TS] irf rename, [TS] irf set, [TS] irf table, [TS] var intro, [TS] varbasic, [TS] vec intro, [TS] Glossary irf, [TS] irf add command, [TS] irf add cgraph command, [TS] irf cgraph create command, [TS] irf create ctable command, [TS] irf ctable describe command, [TS] irf describe drop command, [TS] irf drop graph command, [TS] irf graph ograph command, [TS] irf ograph rename command, [TS] irf rename set command, [TS] irf set table command, [TS] irf table
H
Hadi, A. S., [TS] prais Hall, B. H., [TS] arch, [TS] arima Hall, R. E., [TS] arch, [TS] arima Hamilton, J. D., [TS] time series, [TS] arch, [TS] arma, [TS] arima, [TS] corrgram, [TS] dfuller, [TS] fcast compute, [TS] irf, [TS] irf create, [TS] pergram, [TS] pperron, [TS] psdensity, [TS] sspace, [TS] sspace postestimation, [TS] tslter, [TS] ucm, [TS] var intro, [TS] var, [TS] var svar, [TS] vargranger, [TS] varnorm, [TS] varsoc, [TS] varstable, [TS] varwle, [TS] vec intro, [TS] vec, [TS] vecnorm, [TS] vecrank, [TS] vecstable, [TS] xcorr, [TS] Glossary Hannan, E. J., [TS] sspace Hardin, J. W., [TS] newey, [TS] prais Harvey, A. C., [TS] intro, [TS] arch, [TS] arima, [TS] prais, [TS] psdensity, [TS] sspace, [TS] sspace postestimation, [TS] tslter, [TS] tslter hp, [TS] tssmooth hwinters, [TS] ucm, [TS] var svar Hauser, M. A., [TS] arma Hausman, J. A., [TS] arch, [TS] arima Haver Analytics, [TS] haver haver command, [TS] haver heteroskedasticity, ARCH model, see ARCH model GARCH model, see GARCH NeweyWest estimator, see NeweyWest regression Higgins, M. L., [TS] arch high-pass lter, [TS] tslter bw, [TS] tslter hp, [TS] Glossary Hildreth, C., [TS] prais HildrethLu regression, [TS] prais Hill, R. C., [TS] arch, [TS] prais Hipel, K. W., [TS] arima, [TS] ucm HodrickPrescott lter, [TS] tslter, [TS] tslter hp Hodrick, R. J., [TS] tslter, [TS] tslter hp Holt, C. C., [TS] tssmooth, [TS] tssmooth dexponential, [TS] tssmooth exponential, [TS] tssmooth hwinters, [TS] tssmooth shwinters HoltWinters smoothing, [TS] tssmooth, [TS] tssmooth dexponential, [TS] tssmooth exponential, [TS] tssmooth hwinters, [TS] tssmooth shwinters, [TS] Glossary
J
Jaeger, A., [TS] tslter, [TS] tslter hp Jagannathan, R., [TS] arch Jarque, C. M., [TS] varnorm, [TS] vecnorm JarqueBera statistic, [TS] varnorm, [TS] vecnorm Jeantheau, T., [TS] mgarch Jenkins, G. M., [TS] arma, [TS] arima, [TS] corrgram, [TS] cumsp, [TS] dfuller, [TS] pergram, [TS] pperron, [TS] psdensity, [TS] xcorr Jerez, M., [TS] sspace Johansen, S., [TS] irf create, [TS] varlmar, [TS] vec intro, [TS] vec, [TS] veclmar, [TS] vecnorm, [TS] vecrank, [TS] vecstable Johnson, L. A., [TS] tssmooth, [TS] tssmooth dexponential, [TS] tssmooth exponential, [TS] tssmooth hwinters, [TS] tssmooth shwinters Joyeux, R., [TS] arma Judge, G. G., [TS] arch, [TS] prais
K
Kalman lter, [TS] arima, [TS] dfactor, [TS] dfactor postestimation, [TS] sspace, [TS] sspace postestimation, [TS] ucm, [TS] ucm postestimation, [TS] Glossary forecast, [TS] dfactor postestimation, [TS] sspace postestimation, [TS] ucm postestimation smoothing, [TS] dfactor postestimation, [TS] sspace postestimation, [TS] ucm postestimation Kalman, R. E., [TS] arima Kim, I.-M., [TS] vec intro, [TS] vec, [TS] vecrank King, M. L., [TS] prais King, R. G., [TS] tslter, [TS] tslter bk, [TS] tslter cf, [TS] tslter hp, [TS] vecrank Kmenta, J., [TS] arch, [TS] prais, [TS] rolling Koehler, A. B., [TS] tssmooth, [TS] tssmooth dexponential, [TS] tssmooth exponential, [TS] tssmooth hwinters, [TS] tssmooth shwinters Kohn, R., [TS] arima Koopman, S. J., [TS] ucm Kroner, K. F., [TS] arch kurtosis, [TS] varnorm, [TS] vecnorm
L tkepohl, H., continued u [TS] varbasic, [TS] vargranger, [TS] varnorm, [TS] varsoc, [TS] varstable, [TS] varwle, [TS] vec intro, [TS] vecnorm, [TS] vecrank, [TS] vecstable
M
MA, [TS] arch, [TS] arma, [TS] arima, [TS] sspace, [TS] ucm ma, tssmooth subcommand, [TS] tssmooth ma MacKinnon, J. G., [TS] arch, [TS] arima, [TS] dfuller, [TS] pperron, [TS] prais, [TS] sspace, [TS] varlmar, [TS] Glossary Maddala, G. S., [TS] vec intro, [TS] vec, [TS] vecrank Magnus, J. R., [TS] var svar Mandelbrot, B., [TS] arch Mangel, M., [TS] varwle Maravall, A., [TS] tslter hp McAleer, M., [TS] mgarch McCullough, B. D., [TS] corrgram McDowell, A. W., [TS] arima McLeod, A. I., [TS] arima, [TS] ucm Meiselman, D., [TS] arima MGARCH, see multivariate GARCH mgarch ccc command, [TS] mgarch ccc dcc command, [TS] mgarch dcc dvech command, [TS] mgarch dvech vcc command, [TS] mgarch vcc Miller, J. I., [TS] sspace Mitchell, W. C., [TS] tslter, [TS] tslter bk, [TS] tslter bw, [TS] tslter cf, [TS] tslter hp, [TS] ucm Monfort, A., [TS] arima, [TS] mgarch ccc, [TS] mgarch dcc, [TS] mgarch vcc Montgomery, D. C., [TS] tssmooth, [TS] tssmooth dexponential, [TS] tssmooth exponential, [TS] tssmooth hwinters, [TS] tssmooth shwinters Moore, J. B., [TS] sspace moving average model, see MA process, [TS] Glossary smoother, [TS] tssmooth, [TS] tssmooth ma multiplicative heteroskedasticity, [TS] arch multivariate GARCH, [TS] mgarch, [TS] Glossary model, constant conditional correlation, [TS] mgarch ccc diagonal vech, [TS] mgarch dvech dynamic conditional correlation, [TS] mgarch dcc varying conditional correlation, [TS] mgarch vcc postestimation, after ccc model, [TS] mgarch ccc postestimation after dcc model, [TS] mgarch dcc postestimation after dvech model, [TS] mgarch dvech postestimation
L
lag-exclusion statistics, [TS] varwle lag operator, [TS] Glossary lag-order selection statistics, [TS] var, [TS] var intro, [TS] var svar, [TS] vec intro, [TS] varsoc Lagrange-multiplier test, [TS] varlmar, [TS] veclmar Lai, K. S., [TS] dfgls Laurent, S., [TS] mgarch leap seconds, [TS] tsset Ledolter, J., [TS] tssmooth, [TS] tssmooth dexponential, [TS] tssmooth exponential, [TS] tssmooth hwinters, [TS] tssmooth shwinters Lee, T.-C., [TS] arch, [TS] prais Leser, C. E. V., [TS] tslter, [TS] tslter hp Lieberman, O., [TS] mgarch Lilien, D. M., [TS] arch Lim, G. C., [TS] arch linear lter, [TS] Glossary regression, [TS] newey, [TS] prais, [TS] var intro, [TS] var, [TS] var svar, [TS] varbasic Ling, S., [TS] mgarch Ljung, G. M., [TS] wntestq long-memory process, [TS] arma, [TS] Glossary Lu, J. Y., [TS] prais L tkepohl, H., [TS] time series, [TS] arch, u [TS] dfactor, [TS] fcast compute, [TS] irf, [TS] irf create, [TS] mgarch dvech, [TS] prais, [TS] sspace, [TS] sspace postestimation, [TS] var intro, [TS] var, [TS] var svar,

multivariate GARCH, postestimation, continued after vcc model, [TS] mgarch vcc postestimation multivariate time-series estimators, dynamic-factor models, [TS] dfactor MGARCH models, see multivariate GARCH state-space models, [TS] sspace structural vector autoregressive models, [TS] var svar vector autoregressive models, [TS] var, [TS] varbasic vector error-correction models, [TS] vec
P
pac command, [TS] corrgram Pagan, A. R., [TS] Glossary Palma, W., [TS] arma, [TS] arma postestimation unobserved components model, [TS] psdensity PARCH, [TS] arch Park, J. Y., [TS] sspace, [TS] vec intro, [TS] vec, [TS] vecrank partial autocorrelation function, [TS] corrgram, [TS] Glossary Paulsen, J., [TS] varsoc, [TS] vec intro pergram command, [TS] pergram periodogram, [TS] pergram, [TS] psdensity, [TS] Glossary Perron, P., [TS] dfgls, [TS] pperron, [TS] Glossary phase function, [TS] Glossary Phillips, P. C. B., [TS] pperron, [TS] vargranger, [TS] vec intro, [TS] vec, [TS] vecrank, [TS] Glossary PhillipsPerron test, [TS] pperron Pierce, D. A., [TS] wntestq Pisati, M., [TS] time series Pitarakis, J.-Y., [TS] vecrank Plosser, C. I., [TS] vecrank Pollock, D. S. G., [TS] tslter, [TS] tslter bk, [TS] tslter bw, [TS] tslter cf, [TS] tslter hp portmanteau statistic, [TS] Glossary portmanteau test, [TS] corrgram, [TS] wntestq postestimation command, [TS] fcast compute, [TS] fcast graph, [TS] irf, [TS] psdensity, [TS] vargranger, [TS] varlmar, [TS] varnorm, [TS] varsoc, [TS] varstable, [TS] varwle, [TS] veclmar, [TS] vecnorm, [TS] vecstable pperron command, [TS] pperron prais command, [TS] prais, [TS] prais postestimation, [TS] Glossary Prais, S. J., [TS] prais PraisWinsten regression, see prais command Prescott, E. C., [TS] tslter, [TS] tslter hp Press, W. H., [TS] arch, [TS] arima Priestley, M. B., [TS] psdensity, [TS] tslter, [TS] ucm priming values, [TS] Glossary psdensity command, [TS] psdensity
N
NARCH, [TS] arch NARCHK, [TS] arch Nelson, D. B., [TS] arch, [TS] arima, [TS] mgarch Neudecker, H., [TS] var svar Newbold, P., [TS] arima, [TS] vec intro newey command, [TS] newey, [TS] newey postestimation Newey, W. K., [TS] newey, [TS] pperron NeweyWest covariance matrix, [TS] Glossary postestimation, [TS] newey postestimation regression, [TS] newey Newton, H. J., [TS] arima, [TS] corrgram, [TS] cumsp, [TS] dfuller, [TS] pergram, [TS] wntestb, [TS] xcorr Ng, S., [TS] dfgls Nielsen, B., [TS] varsoc, [TS] vec intro nl, tssmooth subcommand, [TS] tssmooth nl nonlinear estimation, [TS] arch smoothing, [TS] tssmooth nl nonstationary time series, [TS] dfgls, [TS] dfuller, [TS] pperron, [TS] vec intro, [TS] vec normality test after VAR or SVAR, [TS] varnorm after VEC, [TS] vecnorm NPARCH, [TS] arch
O
OConnell, R. T., [TS] tssmooth, [TS] tssmooth dexponential, [TS] tssmooth exponential, [TS] tssmooth hwinters, [TS] tssmooth shwinters ograph, irf subcommand, [TS] irf ograph Olkin, I., [TS] wntestb Ooms, M., [TS] arma Orcutt, G. H., [TS] prais orthogonalized impulseresponse function, [TS] irf, [TS] var intro, [TS] vec intro, [TS] vec, [TS] Glossary Osterwald-Lenum, M., [TS] vecrank
Q
Q statistic, [TS] wntestq
R
random walk, [TS] Glossary Ravn, M. O., [TS] tslter, [TS] tslter hp Rebelo, S. T., [TS] tslter, [TS] tslter hp recursive estimation, [TS] rolling recursive regression analysis, [TS] Glossary

Reinsel, G. C., [TS] arma, [TS] arima, [TS] corrgram, [TS] cumsp, [TS] dfuller, [TS] pergram, [TS] pperron, [TS] psdensity, [TS] vec intro, [TS] xcorr rename, irf subcommand, [TS] irf rename Robins, R. P., [TS] arch robust, Huber/White/sandwich estimator of variance ARCH, [TS] arch ARFIMA, [TS] arma ARIMA and ARMAX, [TS] arima dynamic-factor model, [TS] dfactor GARCH model, see GARCH NeweyWest regression, [TS] newey PraisWinsten and CochraneOrcutt regression, [TS] prais state-space model, [TS] sspace unobserved-components model, [TS] ucm rolling command, [TS] rolling rolling regression, [TS] rolling, [TS] Glossary Rombouts, J. V. K., [TS] mgarch Room, T., [TS] arima Rothenberg, T. J., [TS] dfgls, [TS] sspace, [TS] var svar, [TS] vec, [TS] Glossary Runkle, D. E., [TS] arch Ryzhik, I. M., [TS] arma smoothers, continued HoltWinters, nonseasonal, [TS] tssmooth hwinters seasonal, [TS] tssmooth shwinters moving average, [TS] tssmooth ma nonlinear, [TS] tssmooth nl Sorrentino, R., [TS] tslter, [TS] tslter bw Sotoca, S., [TS] sspace Sowell, F., [TS] arma spectral analysis, [TS] Glossary density, [TS] psdensity, [TS] Glossary distribution, [TS] cumsp, [TS] pergram, [TS] psdensity, [TS] Glossary spectrum, [TS] psdensity, [TS] Glossary Sperling, R., [TS] arch, [TS] arima, [TS] dfgls, [TS] wntestq sspace command, [TS] sspace stability, [TS] var intro, [TS] var, [TS] var svar, [TS] vecstable after VAR or SVAR, [TS] varstable after VEC, [TS] vec intro, [TS] vec state-space model, [TS] sspace, [TS] sspace postestimation, [TS] Glossary, also see ARIMA model, also see dynamic factor model stationary time series, [TS] dfgls, [TS] dfuller, [TS] pperron, [TS] var intro, [TS] var, [TS] vec intro, [TS] vec steady-state equilibrium, [TS] Glossary stochastic trend, [TS] Glossary Stock, J. H., [TS] time series, [TS] arch, [TS] dfactor, [TS] dfgls, [TS] irf create, [TS] rolling, [TS] sspace, [TS] var intro, [TS] var, [TS] var svar, [TS] vec intro, [TS] vec, [TS] vecrank, [TS] Glossary strict stationarity, [TS] Glossary structural time-series model, [TS] psdensity, [TS] sspace, [TS] ucm, [TS] Glossary structural vector autoregression, see SVAR SUR, [TS] dfactor SVAR, [TS] var intro, [TS] var svar, [TS] Glossary postestimation, [TS] fcast compute, [TS] fcast graph, [TS] irf, [TS] irf create, [TS] vargranger, [TS] varlmar, [TS] varnorm, [TS] varsoc, [TS] varstable, [TS] varwle, [TS] var svar postestimation svar command, [TS] var svar, [TS] var svar postestimation
S
SAARCH, [TS] arch Saikkonen, P., [TS] vec intro, [TS] vecrank Salvador, M., [TS] vecrank Samaniego, F. J., [TS] varwle Sargan, J. D., [TS] prais Sargent, T. J., [TS] dfactor Schmidt, T. J., [TS] tslter Schneider, W., [TS] sspace Schwert, G. W., [TS] dfgls seasonal ARIMA, [TS] arima difference operator, [TS] Glossary smoothing, [TS] tssmooth, [TS] tssmooth shwinters selection-order statistics, [TS] varsoc Sentana, E., [TS] mgarch Sering, R. J., [TS] irf create serial correlation, see autocorrelation set, irf subcommand, [TS] irf set Shumway, R. H., [TS] arima shwinters, tssmooth subcommand, [TS] tssmooth shwinters Silvennoinen, A., [TS] mgarch, [TS] mgarch ccc Sims, C. A., [TS] dfactor, [TS] irf create, [TS] var svar, [TS] vec intro, [TS] vec, [TS] vecrank skewness, [TS] varnorm smoothers, [TS] tssmooth, [TS] Glossary double exponential, [TS] tssmooth dexponential exponential, [TS] tssmooth exponential
T
tables, [TS] irf ctable, [TS] irf table TARCH, [TS] arch Ter svirta, T., [TS] mgarch, [TS] mgarch ccc a test, DickeyFuller, [TS] dfgls, [TS] dfuller granger causality, [TS] vargranger Lagrange-multiplier, [TS] varlmar, [TS] veclmar

test, continued normality, [TS] varnorm, [TS] vecnorm Wald, [TS] vargranger, [TS] varwle Teukolsky, S. A., [TS] arch, [TS] arima Theil, H., [TS] prais time-domain analysis, [TS] arch, [TS] arma, [TS] arima, [TS] Glossary time-series lter, [TS] psdensity, [TS] ucm operators, [TS] tsset time-varying variance, [TS] arch trend, [TS] Glossary Trimbur, T. M., [TS] psdensity, [TS] tslter, [TS] tslter hp, [TS] ucm tsappend command, [TS] tsappend Tsay, R. S., [TS] varsoc, [TS] vec intro Tse, Y. K., [TS] mgarch, [TS] mgarch vcc tsfill command, [TS] tsll tsfilter, [TS] tslter bk command, [TS] tslter bk bw command, [TS] tslter bw cf command, [TS] tslter cf hp command, [TS] tslter hp tsline command, [TS] tsline tsreport command, [TS] tsreport tsrevar command, [TS] tsrevar tsrline command, [TS] tsline tsset command, [TS] tsset tssmooth, [TS] tssmooth shwinters command, [TS] tssmooth shwinters dexponential command, [TS] tssmooth dexponential exponential command, [TS] tssmooth exponential hwinters command, [TS] tssmooth hwinters ma command, [TS] tssmooth ma nl command, [TS] tssmooth nl Tsui, A. K. C., [TS] mgarch, [TS] mgarch vcc
V
Van Loan, C. F., [TS] arma, [TS] arma postestimation VAR, [TS] dfactor, [TS] sspace, [TS] ucm, [TS] var intro, [TS] var, [TS] var svar, [TS] varbasic, [TS] Glossary postestimation, [TS] fcast compute, [TS] fcast graph, [TS] irf, [TS] irf create, [TS] var postestimation, [TS] vargranger, [TS] varlmar, [TS] varnorm, [TS] varsoc, [TS] varstable, [TS] varwle var command, [TS] var, [TS] var postestimation varbasic command, [TS] varbasic, [TS] varbasic postestimation vargranger command, [TS] vargranger variance decompositions, see FEVD varlmar command, [TS] varlmar varnorm command, [TS] varnorm varsoc command, [TS] varsoc varstable command, [TS] varstable varwle command, [TS] varwle varying conditional-correlation model, [TS] mgarch, [TS] mgarch vcc vcc, mgarch subcommand, [TS] mgarch vcc vec command, [TS] vec, [TS] vec postestimation veclmar command, [TS] veclmar VECM, [TS] vec intro, [TS] vec, [TS] Glossary, also see multivariate GARCH postestimation, [TS] fcast compute, [TS] fcast graph, [TS] irf, [TS] irf create, [TS] varsoc, [TS] vec postestimation, [TS] veclmar, [TS] vecnorm, [TS] vecrank, [TS] vecstable vecnorm command, [TS] vecnorm vecrank command, [TS] vecrank vecstable command, [TS] vecstable vector autoregression, see VAR vector autoregressive forecast, [TS] fcast compute, [TS] fcast graph vector autoregressive moving-average model, [TS] dfactor, [TS] sspace, [TS] ucm vector error-correction model, see VECM Vetterling, W. T., [TS] arch, [TS] arima
U
UCM model, [TS] ucm postestimation, [TS] psdensity, [TS] ucm postestimation ucm command, [TS] ucm Uhlig, H., [TS] tslter, [TS] tslter hp unit-root models, [TS] vec intro, [TS] vec process, [TS] Glossary test, [TS] dfgls, [TS] dfuller, [TS] pperron, [TS] Glossary univariate time series, [TS] arch, [TS] arma, [TS] arima, [TS] newey, [TS] prais, [TS] ucm unobserved-components model, see UCM
W
Wald, A., [TS] varwle Wald test, [TS] vargranger, [TS] varwle Watson, G. S., [TS] prais, [TS] Glossary Watson, M. W., [TS] time series, [TS] arch, [TS] dfactor, [TS] dfgls, [TS] irf create, [TS] rolling, [TS] sspace, [TS] var intro, [TS] var, [TS] var svar, [TS] vec intro, [TS] vec, [TS] vecrank Wei, W. W. S., [TS] psdensity, [TS] tslter, [TS] ucm, [TS] Glossary weighted moving average, [TS] tssmooth, [TS] tssmooth ma West, K. D., [TS] newey, [TS] pperron

White, H., [TS] newey, [TS] prais white noise, [TS] Glossary white-noise test, [TS] wntestb, [TS] wntestq Wiggins, V. L., [TS] arch, [TS] arima, [TS] sspace Winsten, C. B., [TS] prais Winters, P. R., [TS] tssmooth, [TS] tssmooth dexponential, [TS] tssmooth exponential, [TS] tssmooth hwinters, [TS] tssmooth shwinters wntestb command, [TS] wntestb wntestq command, [TS] wntestq Wolfowitz, J., [TS] varwle Wooldridge, J. M., [TS] arch, [TS] mgarch, [TS] mgarch dvech, [TS] prais
X
xcorr command, [TS] xcorr
Y
Yar, M., [TS] tssmooth, [TS] tssmooth dexponential, [TS] tssmooth exponential, [TS] tssmooth hwinters, [TS] tssmooth shwinters YuleWalker equations, [TS] corrgram, [TS] Glossary
Z
Zakoian, J. M., [TS] arch Zellner, A., [TS] prais

Ts

Uploaded by

Copyright:

Available Formats

Ts

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Ts

Uploaded by

Copyright:

Available Formats

STATA TIME-SERIES REFERENCE MANUAL

A Stata Press Publication StataCorp LP College Station, Texas

Copyright c 19852011 StataCorp LP All rights reserved Version 12

mgarch mgarch mgarch mgarch mgarch mgarch mgarch mgarch mgarch

Glossary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Subject and author index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Cross-referencing the documentation

intro Introduction to time-series manual

time series Introduction to time-series commands

Univariate time series Estimators [TS] arma

time series Introduction to time-series commands

time series Introduction to time-series commands

We also offer a NetCourse on Statas time-series capabilities; see http://www.stata.com/netcourse/nc461.html.

Data-management tools and time-series operators

time series Introduction to time-series commands

Univariate time series

time series Introduction to time-series commands

time series Introduction to time-series commands

Multivariate time series

time series Introduction to time-series commands

Description suppress constant term

archm archmlags(numlist) archmexp(exp) arima(# p ,# d ,# q ) ar(numlist) ma(numlist)

distribution(dist # ) het(varlist) savespace

arch Autoregressive conditional heteroskedasticity (ARCH) family of estimators

level(#) detail nocnsreport display options

maximize options coeflegend

To t an ARCH(# m ) model with Gaussian errors, type

. . . , arch(1/#m ) garch(1/#k ) distribution(t 7)

You can also t many other models.

arch Autoregressive conditional heteroskedasticity (ARCH) family of estimators

2 Var( t ) = t = 0 + A(, ) + B(, )2

The yt equation may optionally include ARCH-in-mean and ARMA terms:

A() = A()+ 1,1 A() = A() = A() = A()+ 3,1

2 2 t1 + 1,2 t2 + 2 2 A()+ 2,1 t1 + 2,2 t2 + t1

A() = A()+ 5,1 (| A() = A()+ 6,1 ( A() = A()+ 7,1 (

6,1 ) + 6,2 ( 7 ) + 7,2 (

B() = B()+ 8,1 | B() = B()+ 9,1 |

B() = B()+ 10,1 t1 + 10,2 t2 +

arch Autoregressive conditional heteroskedasticity (ARCH) family of estimators

Option earch() egarch()

Terms added to C()

2 2 C() = C() +12,1 lnt1 + 12,2 lnt2 +

{Var( t )}/2 = t = 0 + D(, ) + A(, ) + B(, )2

D() = D()+ 13,1 D() = D() = D() = D() = D() =

arch Autoregressive conditional heteroskedasticity (ARCH) family of estimators

arch Autoregressive conditional heteroskedasticity (ARCH) family of estimators

Reading arch output

... ... ... ... ...

... ... ... ... ...

... ... ... ... ...

arch Autoregressive conditional heteroskedasticity (ARCH) family of estimators

arch Autoregressive conditional heteroskedasticity (ARCH) family of estimators

= [ARCH] = [ARCH] = [ARCH] = [ARCH] = [ARCH] = [ARCH] = [ARCH]

b[arch] b[garch] b[saarch] b[tarch] b[aarch] b[narch] b[narch]

5 = [ARCH] b[aarch e] 6 = [ARCH] b[narch k] 7 = [ARCH] b[narch k]

ARCH and GARCH models

Nelsons EGARCH model

Threshold ARCH model

GJR form of threshold ARCH model

Simple asymmetric ARCH model