NATIONAL ECONOMICS UNIVERSITY ECONOMETRICS

SCHOOL OF BANKING & FINANCE FINAL EXAMINATION

Program Bachelor Financial Investment Code 01
Duration: 90 minutes

Class:__________________________ Name:__________________________
StudentID:______________________ Signature: ______________________

Instructions (read before starting):

 Materials to be supplied:textbook, slide or handout
 Approved calculators are permitted
 Whenever conducting a test, use a 5% significance level unless stated
otherwise. State null and alternative hypotheses, null distribution (including degrees
of freedom), rejection criterion (critical values and rejection region) and outcome.
 Use a two-sided alternative hypothesis for tests, unless stated otherwise.

PART I MULTIPLE – CHOICE [30 POINTS]

(Each question carries 3 pts. There is exactly one correct answer for each
question. Circle your answer)
1. In the equation y=β 0 + β 1 x 1 + β 2 x 2 +u, β 1 is the ......, β 2 is the ......
A. Intercept, slope B. Both slopes
C. Both partial regression cofficients D. Both B and C
2. A linear regression with 2explained variables with n=40 , SST =170 and
SSR=50. The adjusted R2 of this regression is:
A. 0.71 B. 0.69
C. 0.29 D. 0.72
3. With n = 40, the SRF is ^y =0.55+0.3 x 1+ 0.6 x 2 with standard errors are 0.3, 0.12,
0.25, respectively. The confidence interval 95% for regression coefficient of x 1
is
A. (0.06 , 0.54) B. (0.1 , 0.5)
C. (0.11 ,1.09) D. (0.19 , 1.01)
4. With n = 40, the SRF is ^y =0.55+0.3 x 1+ 0.6 x 2 with standard errors are 0.3, 0.12,
 0.25, respectively. Test hypothesis β 1=0.5, the t−stat is
A. 2.5 B. 5.0
C. −1.67 D. 0.4
2 2
5. If Rur =0.65, Rr =0.58, number of restrictions ¿ 2, n=40 and the number of
independent variables is 5, the CORRECT statement is:
A. F−stat=3.40, reject H 0 B. F−stat=12.63, reject H 0
C. F−stat=9.39 , reject H 0 D. F−stat=1.36 , accept H 0
6. Regression with a qualitative variable has ONLY 5 categories as explanatory
variables, the minimum number of dummy variables (use binary variables)
can be generated is
A. 3 B. 4
C. 5 D. None of these above
7. The original regression has k =3 and n=30. The biggest number of coefficients
on an auxiliary regression for the WHITE test is:
A. 7 B. 9
C. 10 D. 11
8. Which of the following economic variable can be represented by dummy
variables directly?
A. Quantity of demand of commodity B. The number of graduation types
M
C. Money supply D. None of these above
9. The estimated results of a production function: ln ^ ( Q ) =1.37+0.58 ln ( K ) +0.63 ln(L)
where Q , K and L are firms’s output, capital and the number of labors. The
standard errors of two slope coefficients are 0.2 and 0.3, respectively. If
cov ( β^ ln ( K ) , β^ ln ( L ) )=−0.05. What is the CORRECT statement at significance level
5%:
A. There is constant returns to scale B. There is increasing returns to scale
C. There is decreasing returns to scale D. There is non-constant returns to
scale
10. With these results from the software, which is the CORRECT statements:
Heteroskedasticity Test: White Prob. F(9,76) 0.0001
Breusch-Godfrey Serial Correlation Test: Prob. F(1,81) 0.0520
A. Homoskedasticity and Auto Correlation B. Heteroskedasticity and No Auto
Correlation
C. Heteroskedasticity and Auto Correlation D. Homoskedasticity and No Auto Correlation

PART II TRUE – FALSE? [30 POINTS]

(Are the following statements true or false? Explain your answer in a maximum
of one or two sentences. Each question carries 5 marks.)
a) In a linear regression without an intercept, R-squared equals Adjusted R-
squared.
b) The bigger confidence level is, the wider the CI range is.
c) If Assumption Homoskedasticity is violated, but all other assumptions are
satisfied, the OLS estimators are still unbiased.
^
d) The estimated results of a production function: ln ( Q ) =1.37+0.63 ln ( K ) +0.45 ln(L)
. When K=L=1, so predicted value of this regression is 3 , 9.
e) In linear regression with only one independent variable, the sign of the slope
is necessarily equal to the sign of the sample correlation coefficient of x and y .
f) When testing H 0 : β j=0, a t−test is always positive.

PART III CASE ANALYSIS [40 POINTS]

Consider the regression model:
fscore=β 0 + β 1 skipped+ β 2 extra+ β 3 female+u

where fscore is score of the final exam (in 4 – point scale), skipped is the number of
skipped lectures, extra is the number of studying hours at home/library and female is a
dummy variable equal to 1 if the individual is female student and 0 otherwise.
The following results have been obtained by ordinary least squares regression
(standard errors in parentheses):
^
fscore=0.71−0.35 skipped +0.19 extra−0.03 female
( 0.6 )( 0.14 ) ( 0.64 ) ( 0.02 )
2
n=315 , R =0.64
a) (5 pts) Provide an interpretation for the R−squared from this regression
b) (5 pts) Interpret the estimated coefficient on female. Is β 3 statistically
significant at the 5% level? Conduct the appropriate test.
c) (5 pts) Is β 2 statistically significant at the 5% level? Conduct the appropriate
test. Then compute the 95% confidence interval for this coefficient.
 d) (5 pts) Re-run the above regression, caught the new R2=0.53, after eliminating
female and extra . What does this mean?
e) (5 pts) The 95% confidence interval of ^β skippied .
f) (5 pts) Regress the squared residuals on all original explanatory variables,
received new R squared = 0.47 , comment on this result?
g) (5 pts) If Minister of Education wants to find out the diffenrence of final
scores between Engineers students and Economics students, how can we
adjust previous model and do the analysis.
h) (5 pts) With these results from the software, state the pair of hypotheseses, test
equation and conclusion:

Heteroskedasticity Test: White (no crossterms) Prob. F 0.0081

Ramsey RESET Test: Prob. F 0.0680

THE END