Sample Question Econometrics
Sample Question Econometrics
School of Economics
Yi 0 1 X i i (1)
2
Denote 1 to be the OLS estimator of 1 and se( 1 ) to be .
x
2
i
For each of the following statements indicate whether it is justified and explain
your reasons.
(a) Multicollinearity raises the standard error of 1 and hence the t-test based on
and se( ) is invalid.
1 1 (6 marks)
(c) Serial correlation leads to an unbiased estimator of 1 but the t-test based on
is always invalid.
1 (6 marks)
2. Suppose you are given the following results for a time series of 23 annual
Australian aggregate economic data (standard errors in parentheses):
(b) Test the null hypotheses that the coefficients of wt , pt , at are individually
indifferent from zero respectively at a 5% significance level using a one-
sided test. (6 marks)
(c) Interpret R 2 . Specify the hypothesis to test the overall significance of the
regression model. Use the value of R 2 to test this hypothesis. (7 marks)
(d) What can you conclude from (b) and (c)? In this case, what are the
properties of the OLS estimator? Do you need to be concerned with the
problem from the perspective of hypothesis testing? Why or why not? (7 marks)
(e) Explain how to use an auxiliary regression to detect the problem you
identified in (d). Suggest a way to solve the problem. (7 marks)
(f) If we now measure the dependent variable and all the independent
variables by millions of A$ instead of by billions of A$, without
re-estimating the model, what are the new estimated coefficients? (7 marks)
C t 8.94 0.61 pt
(1.67) (0.20) ..(2)
R 2 0.8
(g) Which model, (1) or (2), do you prefer? Give your reasoning. (6 marks)
where
Y= annual salary of a junior college teacher (in thousand S$)
X= years of teaching experience
D1i = 0 if female
= 1 otherwise
D2i = 0 if non-Chinese
= 1 otherwise
(a) The term D1i D2i represents the interaction effect. Is this term a dummy
variable? If so, what does this term means? (4 marks)
(b) Show how equation (3) can be used to find the mean salary for a non-Chinese
female. (2 marks)
(c) Show how equation (3) can be used to find the mean salary for a Chinese
female. (2 marks)
(d) Show how equation (3) can be used to find the mean salary for a non-Chinese
male. (2 marks)
(e) Show how equation (3) can be used to find the mean salary for a Chinese
male. (2 marks)
Eviews is used to produce the following regression results (with four t-statistics, one
standard error and one p-value removed) based on cross-sectional data on 1000 junior
college teachers in Singapore:
Dependent Variable: Y
Method: Least Squares
Date: 10/25/05 Time: 11:40
Sample: 1 1000
Included observations: 1000
Variable Coefficient Std. Error t-Statistic Prob.
C 32.0 4.0 A1 0.000
X 1.2 0.2 A2 0.000
D1 2.326 1.0 A3 A6
D2 1.645 1.0 A4 0.10
D1*D2 0.65 A5 6.5 0.000
R-squared 0.686082 Mean dependent var 103.7238
Adjusted R-squared 0.684820 S.D. dependent var 8.951610
S.E. of regression 5.025510 Akaike info criterion 6.071918
Sum squared resid 25129.47 Schwarz criterion 6.096457
Log likelihood -3030.959 F-statistic 543.6556
Durbin-Watson stat 2.060918 Prob(F-statistic) 0.000000
(g) Find the numerical values for A1-A6 in the above Eviews output.
(4 marks)
(h) Construct the 95% confidence interval for the slope parameter of X and
interpret it. (4 marks)
xi y i X i Y i - n X Y
1 = = , Y = 0 + 1 X
xi2
X i2 - n X 2
ei2 2 X i2
= se( 0 ) =
n - K 1 n xi2
se( 1 ) = TSS = ESS + RSS
xi2
ESS RSS /( n K 1)
2
R = , R 2 1 ,
TSS TSS /( n 1)
2
F = R /K
(1 - R2 ) / (n - K 1 )
d n
h (1 - )
2 1 - n[Var( )]