EC303 MST 2022 Solution - R
EC303 MST 2022 Solution - R
SHORT TEST
Semester II 2022
Instructions:
1. This paper has 2 sections:
a. Section I: Conceptual Questions - 20 marks
b. Section II: Analytical Questions - 30 marks
2. This paper has Ten questions, answer all the questions in the answer
booklet provided
3. This test comprises 15% of coursework.
4. There are 5 pages to this examination.
5. This is a closed book examination.
6. Materials allowed (calculator).
1
SECTION I: CONCEPT TYPE QUESTIONS [50 MARKS]
Please answer all questions and show your work. Clearly state your answer to each
problem. Answers without justification/explanation will not be given credit.
QUESTION 1
List all the assumptions of the simple linear regression model (7 Marks)
1. The regression model is linear in the parameters.
2. The explanatory variable is uncorrelated with the disturbance term.
3. Given the value of X i , the expected value of the disturbance term is zero. That is,
E(u|X i )=0
.
ui
4. The variance of each is constant, or homoscedastic.
5. There is no correlation between two error terms, that is, no autocorrelation.
Algebraically, this assumption can be written as
cov (ui , u j )=0 i≠ j .
And, if u ' s are uncorrelated, the Y ' s will be uncorrelated as well.
6. The regression model is correctly specified. That is, there is no specification bias or
specification error in the model used in empirical analysis, which implies that we have
included all necessary variables in the model.
QUESTION 2
1. The error term may represent the influence of those variables that are not explicitly included
in the model.
2. Even if we included all relevant variables determining Y, some intrinsic randomness in Y is
bound to occur that cannot be explained no matter how hard we try. Thus, u may reflect this
inherent randomness in Y.
3. u may also represent errors of measurement.
4. Keep the model as simple as possible. That is, even if we know what other variables might
affect Y, their combined influence on Y may be so small and non-systematic that we can
incorporate it in the random term u. If we truly want to build reality into a model it may be
too unwieldy to be of any practical use.
QUESTION 3
2
Briefly explain the concept of multicollinearity. What are the consequences of multicollinearity?
Outline the measures of detecting multicollinearity and remedial measures to address the
problem. (7 Marks)
The presence of multicollinearity does not affect the property of minimum variance for
the OLS estimators, i.e. they still have minimum variance amongst all linear unbiased
estimators. But just because they provide the smallest variances, does not imply that they
will provide small variances, they could still be the smallest and yet quite large. The
larger the variance the less precise the estimator
Detecting multicollinearity
There are several methods that can be used:
One good way of detecting whether multicollinearity is present in the data is to carry on with the
OLS estimation anyway and use the regression results as a guide to detection. It was stated above
that one of the consequences of multicollinearity is to have few significant t values even though
the variables are jointly significant, i.e. a high goodness of fit. Hence, if you notice that this is the
case, then it may imply that some of your explanatory variables are strongly related to each other.
We can look at the correlation coefficients between some of these variables to see how strongly
related they are. Of course, we can only look at correlation between two variables at a time, so if
you had three explanatory variables in your model you could check the correlation between (X,
Y), (X, Z) and (Y, Z).
We could run what are called auxiliary regressions. Here we take one explanatory variable and
regress it on the other Xs, and we do this for each of the explanatory variables. So if your model
contains 5 X variables, you run 5 auxiliary regressions. These regressions will tell you which
variables are related to the rest by the size and statistical significance of the R2 value from each of
the estimated regressions.
Remedial measures
Consider the situation where we discover that multicollinearity exists in our model. Is
there anything that we can do to remedy, or at least help, the situation?
You could consider dropping one or more of the problem variables, i.e. some of those that are
collinear. This may get rid of a multicollinearity problem, but unfortunately it could cause
another problem. This is due to that fact that in formulating our econometric models we include
variables that economic theory states to be important. By excluding variables from the equation
we are in effect misspecifying the model. This in itself causes the estimates of the remaining
variables to be biased. So, you are stuck between a rock and a hard place! If the variable is
important from an economic point of view, even though from an econometric standpoint we
would like to get rid of it – don’t.
Given that multicollinearity is a sample problem, it may be eradicated if we use a different
sample of data. Of course, we are likely to be very restricted in this respect as good data can be
hard to come by. But if it is possible to increase the sample size by either increasing the number
of years in the sample or, in the case of cross-sectional data, including more individuals or
countries in the analysis, then this could reduce the scale of the problem.
It is possible that changing the functional form of the model could help, e.g. there may be
multicollinearity present in a log-linear or semilog model that does not appear in a purely linear
form.
If the empirical study that you are interested in is the focus of previous literature, then it may be
possible to use results from these studies to help with a multicollinearity problem. For example, if
empirical studies have already been done in your chosen area of research then you could use the
relevant estimated values from them. By replacing the coefficients of the problem variables with
3
these previously estimated values, it should be possible to estimate the remaining parameters in
your model with more precision. The problem of course is that the information you utilize may
itself be incorrect. It may hold under the sample of observations used in that study but not
relevant for yours.
QUESTION 4
Only m-1 dummy variables need to be created and used in a regression to represent m
categories. It is wrong to create m dummy variables for an experiment with m outcomes
since it will lead to exact collinearity problem.
QUESTION 5
Using a sample of 1801 Fijian resident, the following earnings (Y) function has been estimated:
ln (Y )=37.059+0.147ln(edu)+0.49experience-0.201female
se=(0.135 ) (0.008 ) (0.007 ) (0.036)
R2 =0.179 n= 1801
Where Y is yearly earnings (dollar, in thousands), education (edu) and experience is measured in
years and female is a dummy variable (1 for female, 0 otherwise).
Female on average is likely to earn 20.1% less relative to male, holding other things constant.
(Give full marks for full answer and partial marks can also be awarded for partial answers)
1% increase in education will increase earnings by .147% on average, holding other things
constant. (Give full marks for full answer and partial marks can also be awarded for partial
answers)
One year increase in experience will increase earnings by 49% on average holding other things
constant. (Give full marks for full answer and partial marks can also be awarded for partial
answers)
4
The model explains 17.9% of the variation in earnings. (Give full marks for full answer and
partial marks can also be awarded for partial answers, give full mark as long as student are able to
explain r2)
e.) Test the hypothesis that there is no difference in expected earnings between Fijian men and
women, using 5% significance level. (6 marks)
Test for the significance of lnI at the 5% significance level. (6 marks)
H0: 1
β =0 Ha:
β 1≠0 (2 Marks)
α = 0.05
0 .21−0
t= =5 .83
0 . 036 (2 marks)
t0.025,1797 = 1.98. Reject H0 if observed t statistic > critical t value. ( 1 Marks)
Since 5.83 >1.98 (a bell-shaped graph would be very helpful here), H0 is rejected at the
5% significance level. (1 marks)
Conclusion: There is statistical evidence significant in earning between male and female.
(Answer in the previous needs to be considered in allocating mark to the current step. If current
step is correct but answer is incorrect due to mistake in previous step then full mark for current
step)
f.) Test that experience has a positive impact on ln(Y) at the 5% significance level. (6 marks)
ln (Y )=6.059+0.151ln(edu)
se=(0.182) (0.12)
R2 =0.153 n= 1801
Are experience and female jointly significant in the original equation at the5% significance level?
(5 Marks)
Ho: B3=B4=0
H1: atleast 1 is not equal to 0 (1 Marks)
5
( R2ur −R 2r )/ q ( 0 . 0 .179−0 .153 / 2
F= = =28. 4
2
( 1−Rur )/ df ur ( 1−0 .179 )/ 1797
(2 marks)
F 0.2, 05
1979 ≈2 . 99 (1 mark)
QUESTION 6
b.) Calculate to find out the average impact of a coup on GDP per capita. (2 marks)
^y
Expected y for a year without a coup no−coup = 1060+0.27k-0.21*0
^y
Expected y for a year with a coup coup = 1060+0.27k-0.21*1
^y ^y
Assuming k remain the same in two years, the cost of a coup = coup - no−coup = -0.21. (This
figure is obviously too small, but makes sense if the dependent variable is lny.)
THE END
6
Appendix: Formulas
R̄2 =1−
[( n−1
n−k−1 )
( 1−R2 ) ]
R 2 /k
F= ~ F k , n−k−1
(1−R 2 )/(n−k−1 )
a
i
ii
iii
iv
v
vi
F Table for the 5% Significance Level
vii
d.f. for numerator
1 2 3 4 5 6 7
d.f. for denominator
4 7.71 6.94 6.59 6.39 6.26 6.16 6.09
5 6.61 5.79 5.41 5.19 5.05 4.95 4.88
13 4.67 3.81 3.41 3.18 3.03 2.92 2.83
14 4.60 3.74 3.34 3.11 2.96 2.85 2.76
15 4.54 3.68 3.29 3.06 2.90 2.79 2.71
16 4.49 3.63 3.24 3.01 2.85 2.74 2.66
17 4.45 3.59 3.20 2.96 2.81 2.70 2.61
18 4.41 3.55 3.16 2.93 2.77 2.66 2.58
25 4.26 3.40 3.01 2.78 2.62 2.51 2.42
26 4.23 3.37 2.98 2.74 2.59 2.47 2.39
27 4.21 3.35 2.96 2.73 2.57 2.46 2.37
28 4.20 3.34 2.95 2.71 2.56 2.45 2.36
29 4.18 3.33 2.93 2.70 2.55 2.43 2.35
30 4.17 3.32 2.92 2.69 2.53 2.42 2.33
40 4.08 3.23 2.84 2.61 2.45 2.34 2.25
60 4.00 3.15 2.76 2.53 2.37 2.25 2.17
120 3.92 3.07 2.68 2.45 2.29 2.18 2.09
∞ 3.84 3.00 2.60 2.37 2.21 2.10 2.01
viii