Scribd
Upload
83 views4 pages

Case Project Econometrics

In this solution, I have tried to solve the econometrics assignment, week 7, of Econometrics: Methods and Analysis

Uploaded by

Garima
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
83 views4 pages

Case Project Econometrics

In this solution, I have tried to solve the econometrics assignment, week 7, of Econometrics: Methods and Analysis

Uploaded by

Garima
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 4

CASE PROJECT

ECONOMETRICS
Week-7
Solutions-
(a) n: 240
Sample size: 240

Variable Coeff Coeff Coeff Coeff


Constant 0.693 0.812 0.636 0.729
li1(-1) x -0.34 x -0.372
li2(-1) x x -0.087 -0.12
Loglikehood -152.763 -139.747 -149.521 -134.178

Li1 : LR = -2 (-149.521 – (134.178)) = 30.69 x^2 (1) α =5%


Li2 : LR = -2 (-139.474 – (-134.178)) = 11.14 (α>3.8)
Li1,Li2 :LR = -2(152.763 – (-134.178)) = 37.17 x^2 (2) (α=6.0)

Thus, we can say that Li1 and Li2 are both individually significant and also jointly significant.

(b) n: 240
Variable Coeff Coeff Coeff Coeff
Constant 0.729 0.731 0.746 0.749
li1(-1) -0.372 -0.366 - -
li1(-2) - - -0.429 0.421
li2(-1) -0.12 - -0.131 -
li2(-2) -0.121 - - -0.129
Loglikehood -134.178 -134.126 -130.346 -130.461

Since explained variable and the number of coefficients are the same, we use Mc Fadden
R^2 Value which is calculated as follows:
Mc Fadden R^2 = {1 – log [(b)]}/ log [(b1)] * log[(b1)] = 152.763
R^2 Vlaue = 1-(-134.178/-152.763) = 1- 0.8783 = 0.1217

Li2(-1) Li2(-2)
Li1(-1) 0.1217 0.1220
Li1(-2) 0.1462 0.1460

0.1462 => The model corresponding to this value is the best model according to the
mentioned criteria.
(c) Pgrown = e^0.746 – 0.425 x Li1(-2) – 0.131 x Li1(-1)
1+e^0.746 – 0.429 x Li1(-2) –0.131 x Li2(-1)

Predicted Total

0 1

Realization:0 5 2 7

Realization:1 3 10 13

Total: 8 12 20

Predicted Total

0 1

Realization:0 5/2 2/20 7/20


Realization:1 3/20 10/20 13/20

Total: 8/20 12/20 20/20

The hit rate = sum of diagonal elements (top left to bottom right)

= 5/20 + 10/20 = 0.75

This means that in 75 percent of the cases, this model will give the correct prediction.

(d) n: 240

Dependent variable: GrowthRate

Growth Rate = ΔlogGDP

Coefficient Standard error t-value

Constant 0.090 0.037 2.400

LOGGDP (-1) -0.019 0.050 -2.371

GrowthRate (-1) 0.613 2.9 X 10^-5 12.236

On comparing the critical value of Augmented Dickey Fuller test with constant and trend,

α = -3.8

Thus, H0 of unit not rejected

log GPD is non-stationary.

(e) Li2(-1) Li2(-2)


Li1(-1) 0.5080 0.5077
Li1(-2) 0.4772 0.4771

0.5080 – The model corresponding to this value is the best performing model.

Dependent variable : GrowthRate


n:240

Coefficient Standard error t-value

Constant 0.001737 0.000320 5.433

Growth Rate (-1) 0.461579 0.048302 9.556

Li1 (-1) -0.001023 0.000130 -7.880

Li2 (-1) -0.000149 6.421 X 10^-5 -2.326


(f) Dependent variable: Residuals (written as RESID)
Residuals from model in (e) with k1 = k2 = 1
Sample size : 239

Coefficient
C -0.001
GrowthRate (-1) 0.0291
Li1(-1) 3.14 x 10^-5
Li2(-1) 2.77 x 10^-6
RESID (-1) -0.500
R^2 0.0010

Breusch-Godfrey test statistic = BG

BG = 240 x 0.0010 = 0.24 x^2 (1) (d = 3.8) d= 5%

Thus, H0 not rejected (of no residual serial correlation).

(g) RMSE = 0.008156 i.e., it is 0.3% on quarterly loans.

Graph shows that the predictions are accurate to much extent.

You might also like