Scribd
Upload
29 views6 pages

Sheet 5 Hetero

The document is a worksheet for a course on Econometrics, focusing on the concept of heteroscedasticity. It includes true/false statements, multiple-choice questions, and problem-solving exercises related to heteroscedasticity and regression models. The content is structured into three parts, addressing definitions, tests, and practical applications in econometric analysis.

Uploaded by

amiraahmedg122
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
29 views6 pages

Sheet 5 Hetero

The document is a worksheet for a course on Econometrics, focusing on the concept of heteroscedasticity. It includes true/false statements, multiple-choice questions, and problem-solving exercises related to heteroscedasticity and regression models. The content is structured into three parts, addressing definitions, tests, and practical applications in econometric analysis.

Uploaded by

amiraahmedg122
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/ 6

Faculty of Economics and Political Science

Economics Department
Third Year (Arabic Section)
Econometrics I
TA. Heba Mohamed

SHEET 5
HETEROSCEDASTICITY

Part (1): State with reason whether the following statements are true or false:

1. Homoscedasticity assumption means ( ) where


2. Heteroscedasticity means that the errors have non-zero mean.
3. If the heteroscedasticity is present, this means that OLS estimators will be
biased but still efficient.
4. Heteroscedasticity is more common in cross-sectional data than in time
series data.
5. If the heteroscedasticity is present, the usual F and T tests will be invalid.
6. The test statistics for the white test is (R2 * n), where R2 is the coefficient of
determination of the original model.
7. The null hypothesis in testing for heteroscedasticity is the existence of the
problem.
8. Logarithmic transformation may be used to reduce heteroscedasticity.

Part (2): Choose the correct answer:

1. Given the following model where Y is the average compensation in


thousands of dollars, and X is average productivity in thousands of dollars,
over 9 employment size of the establishment.
̂i = 1992.3452 + 0.2329Xi
se = (936.4791) (0.0998)
t = (2.1275) (2.333) R2 = 0.4375

1|Page
And for a certain reason the researcher decides to regress the estimated
residuals from the previous model on Xi and obtain the following results:

̂ ̂ = 35.817 − 2.8099 ln Xi
se = (38.319) (4.216)
t = (0.934) (−0.667) R2 = 0.0595

1. The second model represents the formula of which test:


a. White test b. Glejser test c. Park test

d. Goldfeld-Quandt test

2. The researcher performed the second regression in order to test for:


a. Heteroscedasticity b. Multicollinearity. c. Autocorrelation

d. Specification error

3. Does average productivity seem to have a statistically significant effect on the


estimated residuals at 5% significance level?
a. Yes b. No c. Can’t decide

4. Does the problem that the researcher is worried about exist?


a. Yes b. No c. Can’t decide
5. Consider that in the second model the researcher decide to regress the absolute
values of the residuals on , in this case he used:
√

a. Goldfeld-Quandt test b. Glejser test c. Park test d. White test

6. The model in (5) can’t be estimated using OLS:


a. Yes b. No c. Can’t decide

2|Page
7. In (5), the model can give us satisfactory results if the sample size is small
a. Yes b. No c. Can’t decide

2. From cross-sectional data on 41 countries, Stephen Lewis estimated the


following regression model: ln Yi = β1 + β2 ln X2i + β3 ln X3i + ui, where Y =
ratio of trade taxes to total government revenue, X2 = ratio of the sum of
exports plus imports to GNP, and X3 = GNP per capita. Since the data are
cross-sectional involving heterogeneity of countries, the researcher applies a
test for heteroscedasticity and the following results were obtained:
2
̂ = −5.8417 + 2.5629 ln Tradei + 0.6918 ln GNPi−0.4081 (ln Tradei) −
0.0491(ln GNPi) 2 + 0.0015 (ln Tradei) (ln GNPi) R 2 = 0.1148

1. This test is called:


a. Spearman test b. Breusch – Pagan - Godfrey test c. White’s test
a. Glesjer test

2. The null hypothesis in testing for heteroscedasticity is:


a. All parameters equal to zero in the original regression.
b. All slope coefficients equal to zero in the original regression.
c. All parameters equal to zero in the auxiliary regression.
d. All slope coefficients equal to zero in the auxiliary regression.

3. The test statistics of this test follows:


a. Chi2 distribution b. F distribution c. t distribution

3|Page
e. Standard Normal Z test

4. The test statistic for this test is:


a. 4.7068 b. 0.1148 c. 192.9788 d. 0.7351

5. Given that the critical chi2 value is 11.075, we can conclude that:
a. We reject H0 so the errors are homoscedastic.
b. We reject H0 so the errors are heteroscedastic.
c. We can’t reject H0 so the errors are homoscedastic.
d. We can’t reject H0 so the errors are heteroscedastic.

3. All of the approaches below are plausible approaches to deal with a model
that exhibits heteroscedasticity except:
a. Taking logarithms of each of the variable
b. Using a generalized Least Squares Method
c. Add lagged values of the variables to the regression

4. In the following regression model: , suppose


that a researcher interested in conducting White’s heteroscedasticity test
using the residuals from an estimation of the model above, what would be the
most appropriate form for the auxiliary regression?
a. ̂ ̂
b. ̂
c. ̂ ̂
d. ̂

5. If the error variance is not known, but you suspect that it is proportional to
,so you have to:
a. divide the model by
b. multiply the model by
c. divide the model by
d. none of the above

4|Page
Part (3): Solve the following problems:

1. If you have the following estimated regression model during the period
from ( 1965 to 1996):

̂
(se) (2.73) (0.0060) (0.0736)

Where GNP: gross national product.


: Private consumption.
: Military expenditure.

And for a certain reason, the researcher has estimated the following regression
model.
̂
( )
(se) (2.22) (0.006) (0.0597)

a) What is the reason which makes the researcher to transform from the first
model to the second one?
b) What are the consequences of the presence of this problem?
c) What is the main assumption which makes the researcher to transform from
the first model to the second one?
d) What are the main methods to detect this problem?
e) How can we know that the problem has been solved in the second regression
model?
f) Are the values of the two regressions model comparable or not?

2. If you have the following regression model:

Where ( )
( )
( ) {
( )

5|Page
Find efficient estimators of β1 and β2 by using the following data

yi 3 5 7.5 10 13 16.5
Xi 1 2 3 4 5 6

6|Page

You might also like