EDA Final Exam Question Paper

Assessment 3
Econometric and Data Analysis
Instructions:
1. This assessment will be graded formally and contribute 40% toward your final marks for this course.
2. All questions are compulsory.
3. Total Marks: 50 (Marks assigned to each question is given in parenthesis.
4. Duration: 4 hours till 3:00PM.
5. Upload or mail your answers before cutoff timing. Penalty in form of grade deduction will be imposed
for late submission. One grade will be deducted if submission gets late by more than 15 minutes.
1. 1. Comment upon the statistical versus economic significance, and choice of significance level in hy-
pothesis testing. (4 )
2. What are AIC and SIC criteria? What are various uses of these? Explain (4)
3. In CLRM, suppose X matrix shows high collinearity. In order to solve the problem of high collinear-
ity , a researcher double the sample by getting two observation on Y , rather than one at each raw
of the original X.
(a) What happen to the degree of collinearity?
(b) What happen to the variance of OLS coefficients?
(c) Comment upon the claim that more data is no remedy for the multicollinearity problem if the
data are simply more of the same. (4)
4. Suppose the true regression model is given as,
y = Xβ + zγ +
where the two parts have k and 1 columns respectively. If we regress y on X without the including
the relevant variable , z, the the estimator is,
b = β + (X0 X)−1 X0 zγ + (X0 X)−1 X0
Note γ is relevant variable, γ 6= 0. Taking the expectation , we see that unless X0 z = 0, b is biased.
The well known result is the omitted variable formula:
E[b|X, z] = β + pXz γ
where,
pXz = (X0 X)−1 X0 z
The vector pXz is the column of slopes in the least squares regression if z on x. The result above
shows that if pXz is nonzero and γ is nonzero, then regression of y on X alone produces a biased and
inconsistent estimator β. Suppose the objective is to forecast y, not to estimate the parameters.
Consider regression of y on X alone to estimate β with b (which is biased). Is the forecast on y
computed using Xb also biased? Assume that E[z|X] is linear function of X. Discuss your finding
generally. What are the implication for the prediction when variables are omitted from regression?
(5)
2. 1. Discuss strengths and weaknesses of the White-Hetrocedasticity test, the Goldfled –Quandt test,
and the BPG test for detecting hetrocedastcity problem. What kind of robust standard error would
you use if you find groupwise hetrocedasticity in the regression model? (5)
2. Do you agree with the statement that “Hetrocedasticity has never been a reason to throw out an
otherwise good model”? (3)
3. Carefully explain how to undertake a Chow test making all assumptions explicit. What are the
other tests that we can use as an alternative of Chow test? Discuss validity of these test. (5)
4. A government study found that people who eat chocolate frequently weigh less than people who
don’t. Researcher questioned 1000 individuals from California between the age 20 and 85 about
their eating habit, and measure their weight and height. On average, participants ate chocolate
twice a week and had a body mass index (BMI) of 28. There was an observed difference of five
to seven pounds in weight between those who ate chocolate five times a week and those who did
not eat any chocolate at all, chocolate eater weighing less on average. Frequent chocolate eater
also consumed more calorie, on average, than people who consumed less chocolate. Based on this
summary, would you recommend that American’s who do not presently eat chocolate, consider
eating chocolate up to five times a week if they want to lose weight? Why or why not? Explain.
(5)
3. A regression results given in Annexure 1 examined the efficiency of health care delivery measured in
disability adjusted life expectancy (DALE) to major inputs like education and per capita health care
expenditures using data set of 191 countries-almost all the population in the world for year 1997. De-
scription of explanatory variables included in the model are as follows:
• HEXP =Per capita health expenditure (in dollars),

• HC3 =Educational attainment (in years),
• OECD =Dummy variable for OECD country, 1 if country OECD, 0 otherwise
• GINI =Gini coefficient for income inequality,
• GEFF =World bank measure of government effectiveness ( Index values),
• VOICE =World bank measure of democratization of the political process (index values)
• TROPICS =Dummy variable for tropical location,
• POPDEN =Population density,
• PUBTHE =Proportion of health expenditure paid by public authorities (in dollars)
• GDPC = per capita GDP (in dollars).
1. Interpret the regression results( In interpretation atleast include interpretation of coefficients values
of education, health expenditure, and OECD variables, you may avoid interpretation of other
coefficients values). Do you suspect any problem(s) in the regression model? List out all those
problems and explain why you think so.
2. Carefully examine all diagnostic tests given after the regression result, and propose what kind of
remedial measures or combination of remedial measures could be adopted by you to get desired
results. (15)
(Hint- use information given in descriptive statistics along with other regression results)
Page 2

EDA Final Exam Question Paper

Uploaded by

Copyright:

Available Formats

EDA Final Exam Question Paper

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

EDA Final Exam Question Paper

Uploaded by

Copyright:

Available Formats

Assessment 3

Econometric and Data Analysis

b = β + (X0 X)−1 X0 zγ + (X0 X)−1 X0

• HEXP =Per capita health expenditure (in dollars),

You might also like

EDA Final Exam Question Paper

Uploaded by

Copyright:

Available Formats

EDA Final Exam Question Paper

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

EDA Final Exam Question Paper

Uploaded by

Copyright:

Available Formats

Assessment 3

Econometric and Data Analysis

b = β + (X0 X)−1 X0 zγ + (X0 X)−1 X0 

• HEXP =Per capita health expenditure (in dollars),

You might also like

b = β + (X0 X)−1 X0 zγ + (X0 X)−1 X0