Question 1 (1 mark)

If a change in variable x causes a change in variable y, variable x is called the:

(a) dependent variable.

(b) explained variable.

(c) explanatory variable.

(d) response variable.

Question 2 (1 mark)
Data on the income of Monash graduates collected at the same year is:

(a) panel data.

(b) experimental data.

(c) time series data.

(d) cross-sectional data.

Question 3 (1 mark)
A change in the unit of measurement of the dependent variable in a model does not lead to a change
in:

(a) the variance of the residuals.

(b) the sum of squared residuals of the regression.

(c) the R squared of the regression.

(d) the confidence intervals of the regression.

Question 4 (1 mark)
In the following equation, GDP refers to gross domestic product, and FDI refers to foreign direct
investment:
log(GDP) = 2.65 + 0.527 log(bank credit) + 0.222 FDI.
(0.13) (0.022) (0.017)

Which of the following formulates the hypothesis that given foreign direct investment, bank credit
is not an important predictor of log gross domestic product?

(a) β̂log(bank credit) = 0.

(b) βlog(bank credit) = 0.

(c) β̂log(bank credit) = β̂FDI = 0.

(d) βbank credit = βFDI = 0.

Question 5 (1 mark)
In the following equation, GDP refers to gross domestic product, and FDI refers to foreign direct
investment:
log(GDP) = 2.65 + 0.527 log(bank credit) + 0.222 FDI.
(0.13) (0.022) (0.017)

Include a dummy variable ‘POSTGFC’ in the equation above that equals one if GDP is measured
after the global financial crisis and zero otherwise. Which equation can be estimated:

(a) Including an intercept and POSTGFC.

Page 1 of 5
 (b) Including POSTGFC and (1-POSTGFC) but no intercept.

(c) Including an intercept, POSTGFC, and (1-POSTGFC).

(d) Both the equations in (a) and (b).

Question 6 (1 mark)
In the following equation, GDP refers to gross domestic product, and FDI refers to foreign direct
investment:
log(GDP) = 2.65 + 0.527 log(bank credit) + 0.222 FDI.
(0.13) (0.022) (0.017)

Which of the following statements is correct?

(a) If GDP increases by 1%, bank credit increases on average by 0.527%, the level of FDI remaining
constant.

(b) If bank credit increases by 1%, GDP increases on average by 0.527%, the level of FDI remaining
constant.

(c) If GDP increases by 1%, bank credit increases on average by log(0.527)%, the level of FDI
remaining constant.

(d) If bank credit increases by 1%, GDP increases on average by log(0.527)%, the level of FDI
remaining constant.

Question 7 (1 mark)
Which of the following correctly identifies a limitation of logarithmic transformations of variables?

(a) Taking the logarithm of variables makes OLS estimates more sensitive to extreme values in
comparison to variables taken in level.

(b) Logarithmic transformations cannot be used if a variable takes on zero or negative values.

(c) Logarithmic transformations of variables are likely to lead to heteroskedasticity.

(d) Taking the logarithm of a variable often expands its range which can cause inefficient estimates.

Question 8 (1 mark)
The Gauss-Markov theorem will not hold if

(a) the error term has the same variance given any values of the explanatory variables.

(b) the error term has an expected value of zero given any values of the independent variables.

(c) the independent variables have exact linear relationships among them.

(d) the regression model relies on the method of random sampling for collection of data.

Question 9 (1 mark)
Which of the following indicates homoskedasticity?

(a) E[ui ] = 0.

(b) Var(ui ) = σi2 .

(c) Cov(ui , uj ) = 0 for all i 6= j.

(d) ui ∼ N (0, σ 2 ).

Question 10 (1 mark)
If OLS is used in the presence of autocorrelation, which of the following will be likely consequences?

Page 2 of 5
 (a) The R squared may be close to zero.

(b) Coefficient estimates may be misleading.

(c) Forecasts made from the model could be biased.

(d) Hypothesis tests could reach the wrong conclusions.

Question 11 (1 mark)
Consider the estimated ARDL(2,1) model given by

dt ) = −0.7 + 0.5 log(yt−1 ) + 0.3 log(yt−2 ) + 1.4 log(xt ) − 1.2 log(xt−1 ).

Which of the following statements is correct?

(a) A 1% increase in x is predicted to increase y by 1.4% on impact and by 1.2% in the long run.

(b) A 1% increase in x is predicted to increase y by 14% on impact and by 12% in the long run.

(c) A 1% increase in x is predicted to increase y by 1.4% on impact and to decrease y by 1.2% in

(d) A 1% increase in x is predicted to increase y by 1.4% on impact and by 1% in the long run.

Question 12 (1 mark)
Including relevant lagged values of the dependent variable on the right hand side of a regression
equation leads to which one of the following?

(a) Biased but consistent coefficient estimates.

(b) Biased and inconsistent coefficient estimates.

(c) Unbiased but inconsistent coefficient estimates.

(d) Unbiased and consistent but inefficient coefficient estimates.

Page 3 of 5
Question 13 (16 marks)
We have estimated the following model with standard errors in parentheses for 50 states in the USA:
d = 0.452 − 0.128 age − 0.127 north − 0.217 east − 0.166 south,
birth
(0.374) (0.013) (0.054) (0.068) (0.051)

with birth rate in the state (in births per woman), difference median age in the state from median
age across the USA (age), and dummies indicating whether the state is in the North, East, South of
the USA (north, east, south). The R2 equals 0.826. If we omit age from the model, the R2 decreases
to 0.463.

13.a. Test the hypothesis that an increase of the median age with one year decreases the birth rate
with 0.1 births per woman, controlling for region, against the alternative that an increase of
the median age with one year decreases the birth rate with more than 0.1 births per woman.
(4 marks)

13.b. Calculate the 95% confidence interval for the average birth rate in the west of the USA,
controlling for median age. (4 marks)

13.c. Suppose you have to test the hypothesis that the coefficient for east is twice the coefficient for
north. Write down the parameter restriction on the model, and explain which model you have
to estimate to run a t-test for this hypothesis. (4 marks)

13.d. We have also estimated the following model

with the R2 equal to 0.920. Test the hypothesis that age has no effect on birth, controlling for
region, in this model, against the alternative that age does have an effect. (4 marks)

Question 14 (16 marks)

dbyieldi = β0 + β1 diratei + ui ,

where dbyield is the monthly change in bond yield, dirate the monthly change in the short-term
interest rate, and i = 1, . . . , 600.

14.a. Suppose that we obtain a R2 equal to 0.3 and an adjusted R2 equal to 0.28 from the linear
regression model. The R2 increases to 0.32 and the adjusted R2 to 0.30 if we add a lag of
dirate to the model. What can you conclude from the R2 s and adjusted R2 s? Explain. (4
marks)

14.b. The null-hypothesis that H0 : E[u2i |diratei ] = σ 2 for i = 1, . . . , n can be tested by a White
test. There are two forms of the White test that use different auxiliary regressions. Write
down the auxiliary regression for both forms and explain why they will both give the exact
same result in the particular model under consideration. (4 marks)

14.c. Denote the OLS residuals from the linear regression model by u
bi . Suppose that we obtain a
2
R equal to 0.15 from the following regression model

b2i = γ0 + γ1 zi + vi .
u (1)

What conclusion would you draw with respect to the behaviour of the error term ui ? Briefly
explain. (4 marks)

Page 4 of 5
14.d. What are the consequences of heteroskedasticity in the errors for the OLS estimator, the usual
OLS standard error reported by statistical packages, and the standard t-test for the parameter?
(4 marks)

Question 15 (16 marks)

yt = φ0 + φ1 yt−1 + ut , (2)

with t = 1, . . . , n.

15.a. Write down the assumptions on the coefficients and the error terms of the AR(1) model, if the
model is stationary and ut ∼ W N (0, σ 2 ). (4 marks)

15.b. We have estimated the following equation.

ût = γ0 + γ1 yt−1 + γ2 ût−1 + γ3 ût−2 + et , (3)

with n = 100 and R2 = 0.850, and where ût is the residual of the estimated AR(1) model.
Test the hypothesis that the errors of the AR(1) model are not serially correlated against the
alternative that they are autoregressive of order 2. Remember that you need to state all steps
of hypothesis testing to obtain full marks. (4 marks)

15.c. We have the following estimated equation:

ybt = −0.800 + 0.700yt−1 − 0.300yt−2 .

We observe the values y98 = 5.800, y99 = 6.785, and y100 = 7.500. Use the estimated model
together with these observed values to construct the residual in time period t = 100, and
predict the value of yt in time periods t = 101 and t = 102. (4 marks)

15.d. We have the following estimated equation for the ARDL(2,2) model:

ybt = −1.200 + 0.563yt−1 − 0.100yt−1 + 1.300xt − 0.734xt−1 .

Calculate the contemporaneous and long-term effect on y of a one unit increase in x at time
t. (4 marks)

Page 5 of 5