Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                
Scribd
Upload
3 views

Assignment 1-418 Answers - Copy - Copy (2)

The document presents calculations and analyses related to a regression model, including coefficient estimates, significance tests, and potential issues such as heteroskedasticity and multicollinearity. It discusses the implications of these issues on the model's credibility and the relationship between variables X and Y, concluding that changes in scale do not affect the slope. Additionally, it mentions the use of the Newey-West technique for autocorrelation corrections and highlights the need for further testing of multicollinearity.

Uploaded by

Mostafa Allam
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
3 views

Assignment 1-418 Answers - Copy - Copy (2)

The document presents calculations and analyses related to a regression model, including coefficient estimates, significance tests, and potential issues such as heteroskedasticity and multicollinearity. It discusses the implications of these issues on the model's credibility and the relationship between variables X and Y, concluding that changes in scale do not affect the slope. Additionally, it mentions the use of the Newey-West technique for autocorrelation corrections and highlights the need for further testing of multicollinearity.

Uploaded by

Mostafa Allam
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 4

Key Answers Assignment 1- 418

Question (2)
(1)
44 106
𝑋′𝑋 =
106 270
0.41925466 −0.16459627
(𝑋 ′ 𝑋) − 1 =
−0.16459627 0.06832298
4.9
𝑋′𝑌 =
13.8

̂ = −𝟎. 𝟐𝟏𝟕𝟎𝟖𝟎𝟕𝟓
𝜷
𝟎. 𝟏𝟑𝟔𝟑𝟑𝟓𝟒
(2) Signifiance :

−0.2142857
𝜀 = 𝑌 − 𝑋𝛽̂ = 0.4968944
−0.0009317
−0.139441
′
𝜺 𝜺 𝟎. 𝟑𝟏𝟐𝟐𝟔
𝝈𝟐 = = = 𝟎. 𝟏𝟓𝟔𝟏
𝑵−𝑲 𝟐

Please Note I divide by (N-K) not (N-K-1) as the model has no constant!

0.41925466
̂ ) = 𝝈𝟐 (𝑿′ 𝑿)−𝟏 = 𝟎. 𝟏𝟓𝟔𝟏 × −0.16459627
𝑽𝒂𝒓 (𝜷
−0.16459627 0.06832298
𝟎. 𝟎𝟔𝟓𝟒𝟒𝟓𝟔 −0.02569
=
−0.02569 𝟎. 𝟎𝟏𝟎𝟔𝟔𝟓𝟐𝟐

−𝟎. 𝟐𝟏𝟕𝟎𝟖𝟎𝟕𝟓
𝒕𝜷𝟏 = = 𝟎. 𝟖𝟒𝟖
√𝟎. 𝟎𝟔𝟓𝟒𝟒𝟓𝟔
𝟎. 𝟏𝟑𝟔𝟑𝟑𝟓𝟒
𝒕𝜷𝟐 = = 𝟏. 𝟑𝟐
√𝟎. 𝟎𝟏𝟎𝟔𝟔𝟓𝟐𝟐

The Two t are far from (3) , no significant coefficients

(3) Potential problems that affect variance of the matrix:


a. Heteroskedasticity.
b. Multicollinearity (This also affect the slope itself).
c. Autocorrelation.
Other problems can be noted such as Cov(X, residual).

(4) White Test


𝒏. 𝑹𝟐 = 𝟒 × 𝟏 = 𝟒. 𝟎𝟎
This is Higher that Theoretical value→ Problem should be solved by Sandwich Matrix. No
need to do accurate calculations except the question asks to compute final answer .
(𝑿′ 𝑿)−𝟏 𝑿′ 𝜺𝜺′ 𝑿 (𝑿′𝑿)−𝟏

0.046 -0.106 0.000 0.030


𝛀 = 𝜺𝜺′ = -0.106 0.247 0.000 -0.069
0.000 0.000 0.000 0.000
0.030 -0.069 0.000 0.019

0.046 -0.106 0.000 0.030


0.41925466 −0.16459627 𝟓 𝟑 𝟏 𝟑 -0.106 0.247 0.000 -0.069
× ×
−0.16459627 0.06832298 𝟏𝟏 𝟕 𝟔 𝟖 0.000 0.000 0.000 0.000
𝟓 𝟏𝟏 0.030 -0.069 0.000 0.019
0.41925466 −0.16459627
×𝟑 𝟕 ×
𝟏 𝟔 −0.16459627 0.06832298
𝟑 𝟖

No need to calculate final answer as not mentioned in the question to calculate it

(5) Autocorrelation:
𝜀𝑡 𝜀𝑡−1 𝜀𝑡 . 𝜀𝑡−1
-0.21429 ---- ---
0.496894 -0.21429 -0.106479415
-0.00093 0.496894 -0.000462111
-0.13944 -0.00093 0.000129679

𝑉𝑎𝑟( 𝜀𝑡 ) = 0.102404
𝑉𝑎𝑟( 𝜀𝑡−1 ) = 0.133189
∑𝜀𝑡 . 𝜀𝑡−1 0.106811847
𝐶𝑜𝑣 (𝜀𝑡 , 𝜀𝑡−1 ) = =− = −0.035603949
𝑁 3

0.035603949
𝜌̂𝑡 = − = −0.30486422
√0.102404 × 0.133189

𝜌̂𝑡2 = 0.09294
𝑄𝐵𝑃 = 3 × 0.09294 = 𝟎. 𝟐𝟕𝟖𝟖
This is Lower than theoretical value, no autocorrelation problem.
Data is Time series; any corrections should be through Newey West technique.

(6) Results are not logic nor credible:


a. The Constant of the model removed
b. R2 may be questioned
c. Multicollinearity is not tested and if tested very high (0.9) which affect the
significance of the coefficients
Question (3)

The change is not expected to affect in any way the relationship between X and Y, therefore the
slope should not change. In fact, X and Y scatter plot won’t change if I use same weights for the
two variables (their mean and standard deviation)

Assume:
X Y
5 11
3 7
1 6
3 8

X and Y normal Linear OLS:


.8
.6
.4
.2

1 2 3 4 5
x1

y Fitted values

X and Y after changing the scale in a constant Way:


-.8
-1
-1.2
-1.4
-1.6

-1 0 1 2 3
nx

ny Fitted values

No change in the Slope However the scale X and Y changed and the constant also expected
to change.

You might also like