We investigate the possibility that the Taylor rule should be formulated as a threshold process s... more We investigate the possibility that the Taylor rule should be formulated as a threshold process such that the Federal Reserve acts more aggressively in some circumstances than in others. It seems reasonable that the Federal Reserve would act more aggressively when inflation is high than when it is low. Similarly, it might be expected that the Federal Reserve responds more to a negative than a positive output gap. Although these specifications receive some empirical support, we find that a modified threshold model that is consistent with “opportunistic” monetary policy makes significant progress toward explaining Federal Reserve behavior.
We conduct a thorough statistical analysis of the empirical foundations for the existence of a Ta... more We conduct a thorough statistical analysis of the empirical foundations for the existence of a Taylor rule. We argue that the traditional linear specification is problematic as inflation, the output gap and the federal funds rate appear to be non-stationary or highly persistent variables that are not cointegrated. Although this lack of cointegration could be caused by missing variables or structural breaks, we are unable to 'salvage' the rule using several plausible candidate variables and break dates. As such, we investigate the possibility that the Taylor rule should be modeled as a threshold process. Although the standard types of threshold models are reasonable, we find that a modified threshold model that is consistent with "opportunistic disinflation" makes significant progress towards explaining Federal Reserve behavior.
Journal of The American Statistical Association, 2001
ABSTRACT We develop a test statistic to test hypotheses in nonlinear, weighted regression models ... more ABSTRACT We develop a test statistic to test hypotheses in nonlinear, weighted regression models with serial correlation/heteroskedasticity of unknown form. The novel aspect is that these tests are simple and do not require use of heteroskedasticity autocorrelation consistent (HAC) estimators. Furthermore, they introduce a new class of test, utilizing stochastic transformations to eliminate nuisance parameters, as a substitute for estimating the nuisance parameters. We derive the limiting null distributions of these new tests in a general nonlinear setting, and show that while the tests have nonstandard distributions, the distributions depend only upon the number of restrictions. We apply this method of testing to an empirical example and illustrate that the size of the new test is less distorted than tests utilizing HAC estimators.
Abstract We develop test statistics to test hypotheses in nonlinear, weighted regression models w... more Abstract We develop test statistics to test hypotheses in nonlinear, weighted regression models with serial correlation or conditional heteroskedasticity of unknown form. The novel aspect is that these tests are simple and do not require use of heteroskedasticity autocorrelation consistent (HAC) covariance matrix estimators. This new new class of tests utilize stochastic transformations to eliminate nuisance parameters as a substitute
ABSTRACT We develop a test statistic to test hypotheses in nonlinear, weighted regression models ... more ABSTRACT We develop a test statistic to test hypotheses in nonlinear, weighted regression models with serial correlation/heteroskedasticity of unknown form. The novel aspect is that these tests are simple and do not require use of heteroskedasticity autocorrelation consistent (HAC) estimators. Furthermore, they introduce a new class of test, utilizing stochastic transformations to eliminate nuisance parameters, as a substitute for estimating the nuisance parameters. We derive the limiting null distributions of these new tests in a general nonlinear setting, and show that while the tests have nonstandard distributions, the distributions depend only upon the number of restrictions. We apply this method of testing to an empirical example and illustrate that the size of the new test is less distorted than tests utilizing HAC estimators.
We investigate the possibility that the Taylor rule should be formulated as a threshold process s... more We investigate the possibility that the Taylor rule should be formulated as a threshold process such that the Federal Reserve acts more aggressively in some circumstances than in others. It seems reasonable that the Federal Reserve would act more aggressively when inflation is high than when it is low. Similarly, it might be expected that the Federal Reserve responds more to a negative than a positive output gap. Although these specifications receive some empirical support, we find that a modified threshold model that is consistent with “opportunistic” monetary policy makes significant progress toward explaining Federal Reserve behavior.
We conduct a thorough statistical analysis of the empirical foundations for the existence of a Ta... more We conduct a thorough statistical analysis of the empirical foundations for the existence of a Taylor rule. We argue that the traditional linear specification is problematic as inflation, the output gap and the federal funds rate appear to be non-stationary or highly persistent variables that are not cointegrated. Although this lack of cointegration could be caused by missing variables or structural breaks, we are unable to 'salvage' the rule using several plausible candidate variables and break dates. As such, we investigate the possibility that the Taylor rule should be modeled as a threshold process. Although the standard types of threshold models are reasonable, we find that a modified threshold model that is consistent with "opportunistic disinflation" makes significant progress towards explaining Federal Reserve behavior.
Journal of The American Statistical Association, 2001
ABSTRACT We develop a test statistic to test hypotheses in nonlinear, weighted regression models ... more ABSTRACT We develop a test statistic to test hypotheses in nonlinear, weighted regression models with serial correlation/heteroskedasticity of unknown form. The novel aspect is that these tests are simple and do not require use of heteroskedasticity autocorrelation consistent (HAC) estimators. Furthermore, they introduce a new class of test, utilizing stochastic transformations to eliminate nuisance parameters, as a substitute for estimating the nuisance parameters. We derive the limiting null distributions of these new tests in a general nonlinear setting, and show that while the tests have nonstandard distributions, the distributions depend only upon the number of restrictions. We apply this method of testing to an empirical example and illustrate that the size of the new test is less distorted than tests utilizing HAC estimators.
Abstract We develop test statistics to test hypotheses in nonlinear, weighted regression models w... more Abstract We develop test statistics to test hypotheses in nonlinear, weighted regression models with serial correlation or conditional heteroskedasticity of unknown form. The novel aspect is that these tests are simple and do not require use of heteroskedasticity autocorrelation consistent (HAC) covariance matrix estimators. This new new class of tests utilize stochastic transformations to eliminate nuisance parameters as a substitute
ABSTRACT We develop a test statistic to test hypotheses in nonlinear, weighted regression models ... more ABSTRACT We develop a test statistic to test hypotheses in nonlinear, weighted regression models with serial correlation/heteroskedasticity of unknown form. The novel aspect is that these tests are simple and do not require use of heteroskedasticity autocorrelation consistent (HAC) estimators. Furthermore, they introduce a new class of test, utilizing stochastic transformations to eliminate nuisance parameters, as a substitute for estimating the nuisance parameters. We derive the limiting null distributions of these new tests in a general nonlinear setting, and show that while the tests have nonstandard distributions, the distributions depend only upon the number of restrictions. We apply this method of testing to an empirical example and illustrate that the size of the new test is less distorted than tests utilizing HAC estimators.
Uploads
Papers by Helle Bunzel