UNANSWERED Issue 3 In Gujarati Econometrics By Example numbe

UNANSWERED Issue #3 In Gujarati, Econometrics By Example number:

5.5: Refer to Tbale 5.5. Assume that the error variance is related to the square of income instead of to the square of ABORTIONF. Transform the original abortion rate function replacing ABORTIONF by income and compare your results with those given in Table 5.5. A priori, would you expect a different conclusion about the presence of heteroscedasticity? Why or why not. Show the necessary calculations

Table 5.5 Transformed Eq. (5.1)

Dependent Variable: ABORTION/ABORTIONF

Method: Least Squares

Sample: 1 50

Included observations: 50

Coefficient

Std. Error

t-Statistic

Prob.

1/ABORTIONF

12.81786

11.22852

1.141545

0.2601

RELIGION/ABORTIONF

0.066088

0.068468

0.965239

0.3400

PRICE/ABORTIONF

-0.051468

0.017507

-2.939842

0.0053

LAWS/ABORTIONF

-1.371437

1.819336

-0.753812

0.4552

FUNDS/ABORTIONF

2.726181

3.185173

0.855897

0.3969

EDUC/ABORTIONF

-0.228903

0.147545

-1.551408

0.1283

INCOME/ABORTIONF

0.002220

0.000481

4.616486

0.0000

PICKET/ABORTIONF

-0.082498

0.031247

-2.640211

0.0116

R-squared

0.074143

Mean Dep. Var.

1.011673

Adj. R^2

-0.080166

S.D. Dep. Var.

0.334257

S.E. of Regression

0.347396

Akaike info criterion

0.868945

Sum squared resid.

5.068735

Schwarz criterion

1.174869

Log likelihood

-13.72363

Durbin-Watson Stat

2.074123

Eq. 5.1: ABR i = B1 + B2 Rel i + B3 Price i + B4 Laws i + B5 Funds i + B6 Educ i + B7 Income i + B8 Picket i + ui     i = 1,2,…,50

Table 5.5 Transformed Eq. (5.1)
Dependent Variable: ABORTION/ABORTIONF
Method: Least Squares
Sample: 1 50
Included observations: 50
	Coefficient	Std. Error	t-Statistic	Prob.
1/ABORTIONF	12.81786	11.22852	1.141545	0.2601
RELIGION/ABORTIONF	0.066088	0.068468	0.965239	0.3400
PRICE/ABORTIONF	-0.051468	0.017507	-2.939842	0.0053
LAWS/ABORTIONF	-1.371437	1.819336	-0.753812	0.4552
FUNDS/ABORTIONF	2.726181	3.185173	0.855897	0.3969
EDUC/ABORTIONF	-0.228903	0.147545	-1.551408	0.1283
INCOME/ABORTIONF	0.002220	0.000481	4.616486	0.0000
PICKET/ABORTIONF	-0.082498	0.031247	-2.640211	0.0116
R-squared	0.074143	Mean Dep. Var.	1.011673
Adj. R^2	-0.080166	S.D. Dep. Var.	0.334257
S.E. of Regression	0.347396	Akaike info criterion	0.868945
Sum squared resid.	5.068735	Schwarz criterion	1.174869
Log likelihood	-13.72363	Durbin-Watson Stat	2.074123
Eq. 5.1: ABR i = B1 + B2 Rel i + B3 Price i + B4 Laws i + B5 Funds i + B6 Educ i + B7 Income i + B8 Picket i + ui     i = 1,2,…,50

Solution

The multiple regression model after fitting the data and model is given below,

ABR i = 12.81786 + 0.066088 * Rel i -0.051468 * Price i -1.371437 * Laws i + 2.726181 * Funds i - 0.228903 * Educ i + 0.002220 * Income i - 0.082498 * Picket i ; i = 1,2,…,50

among the independent variable, Price, Income and Picket variables have significance at 0.05 level of significane and rest are non significance. R- square value is 0.074143 which is indicate that model is not explaine much better. they only approx 7 % explaine the dependent variable ABR. its means residuals sum of square is near about the total sum of square, thats\'s why R- square is less and Adj. R^2 is negative. Another way residuals sum of square is more, variance is not constant. if variance is not constant than independent variable may be correlated to each other it is a case of hetroskeductisity . so need to test the hetroskeductisity using White\'s Test and Breusch-Pagan then we say any thing about the hetroskeductisity is presentsor not.