Midterm Exam spring 2022-2023

Date: Friday, 03/17/2023.

Course Title: Advanced Econometrics (BA 351)
Duration: 1h and 30m
Groups: Junior & Master.
Number of pages: 6

INSTRUCTIONS

1. Cell phones, books and notes are strictly not permitted.

2. Only non-scientific (no memory) calculators are permitted.

3. Choose one answer; choosing more than one answer is graded zero.

4. Record your answer on the answer sheet provided in page 7.

Name:-----------------------------------------------------------

ID:----------------------------------------------GR-----------------

1
MCQ (100 pts)

1. What are panel data? 5. The main advantage of using panel

data over cross sectional data is that
a. Data containing units measured at it
different time points
b. Data where each unit is measured at a. gives you more observations.
more than one-time point b. allows you to analyze behavior
c. Data containing skewed variable across time but not across entities.
distributions c. allows you to control for some types
d. Data measured at one point in time of omitted variables without actually
observing them.
d. implies that inference based on the
2. Panel data is also called standard normal distribution is
more appropriate.
a. longitudinal data.
b. cross-sectional data.
c. time series data. 6. Which model takes into account
d. experimental data. both the time-varying independent
variables, the time-invariant
independent variables and the
3. What does lagging variables mean? unmeasured time-invariant
variables?
a. That we use previous values of a
variable a. Pooled OLS
b. That we use future values of a b. Between effects
variable c. Fixed effects
c. That we set all values at its mean d. Random effects
across time
d. That we log transform the variable
7. Working with short panels (𝑁 ≫ 𝑇)
is similar to

4. What does it mean when we say that

our panel is balanced? a. Cross section data analysis
b. Time series analysis
a. When we have more time periods c. Pooled regression analysis
than units d. Fixed effect analysis
b. When we have a large sample of units
c. When we have an equal number of
time periods per unit
8. Working with short panels (𝑇 ≫ 𝑁)
d. When we have an unequal number of
time periods per unit is similar to

a. Cross section data analysis

b. Time series analysis
c. Pooled regression analysis
d. Fixed effect analysis

2
9. How is the heteroscedastic b. By running an ordinary least squares
assumption captured in a pooled regression with robust standard
regression model? errors.
c. By taking the mean of each variable
a. By running an ordinary least squares for each unit across time, and
regression. running a regression on the
b. By running an ordinary least squares collapsed dataset of means.
regression with robust standard d. By way of including unit dummy
errors. variables.
c. By taking the mean of each variable
for each unit across time, and
running a regression on the 13. If 𝑋𝑖𝑡 is correlated with 𝑋𝑖𝑠 for
collapsed dataset of means. different values of 𝑠 and 𝑡, then
d. By way of including unit dummy
variables.
a. 𝑋𝑖𝑡 is said to be autocorrelated
b. the OLS estimator cannot be
10. The Fixed Effects regression model
computed
c. statistical inference cannot proceed
a. has n different intercepts.
b. the slope coefficients are allowed to in a standard way even if clustered
differ across entities, but the standard errors are used
intercept is "fixed" (remains d. this is not of practical importance
unchanged). since these correlations are typically
c. has "fixed" (repaired) the effect of weak in applications
heteroskedasticity.
d. in a log-log model may include logs of
the binary variables, which control 14. In panel data, the regression error
for the fixed effects.
a. is likely to be correlated over time
within an entity
11. Time Fixed Effects regression are b. should be calculated taking into
useful in dealing with unobservable account heteroskedasticity but not
variables autocorrelation
c. only exists for the case of T > 2
a. even if you only have a cross-section d. fits all of the three descriptions
of data available. above
b. if these unobservable variables are
constant across entities but vary over
time. 15. The notation for panel data is (𝑋𝑖𝑡 ,
c. when there are more than 100 𝑌𝑖𝑡 ), 𝑖 = 1, . . . , 𝑛 and 𝑡 =
observations. 1, . . . , 𝑇 because
d. if these unobservable variables are
constant across entities but not over
a. we take into account that the
time.
entities included in the panel change
over time and are replaced by
others.
12. How is the fixed effect model b. the X's represent the observed
estimated? effects and the Y the omitted fixed
effects.
a. By running an ordinary least squares c. there are 𝑛 individuals and 𝑇 time
regression. periods.

3
d. n has to be larger than T for the OLS d. remove the time constant
estimator to exist. component in the error term.
19. The fact that panel data varies
across both individuals and time-
16. When you add “Governorate fixed periods should allow us to ___________
effects” to a simple regression model by accounting for the time constant
for Tunisia Governorates over a component in the error term.
certain time period, and the
regression 𝑅 2 increases
significantly, then it is safe to a. generate unbiased parameter
assume that estimates
b. correctly estimate the true
parameter values
a. the included explanatory variables, c. control for autocorrelation
other than “the Governorate fixed d. improve the efficiency of the
effects”, are unimportant. parameter estimates
b. “Governorate fixed effects” account
for a large amount of the variation in
the data. 20. If there is a pooled cross-section
c. the coefficients on the other model with and intercept and five
included explanatory variables will explanatory variables and there are
not change. 100 observations each year for three
d. time fixed effects are unimportant. years

a. the sample size is 100.

17. A pattern in the coefficients of the b. there are 294 degrees of freedom.
time fixed effects variables may c. the sample size is 294.
reveal the following in a study of the d. there are 300 degrees of freedom.
determinants of Governorate
unemployment rates (Tunisia) using
panel data:
21. The "within transform" involves
a. macroeconomic effects, which affect
all Governorates equally in a given a. Taking the average values of the
year. variables
b. attitude differences towards b. Subtracting the mean of each entity
unemployment between away from each observation on that
Governorates. entity
c. there is no economic information c. Estimating a panel data model using
that can be retrieved from these least squares dummy variables
coefficients. d. Using both time dummies and cross-
d. regional effects, which affect all sectional dummies in a fixed effects
Governorates equally, as long as they panel model
are a member of that region.

18. First-differenced models are 22. The fixed effects panel model is also
preferable to pooled cross-section sometimes known as
models because they
a. A seemingly unrelated regression
a. control for potential autocorrelation.
model
b. control for potential
b. The least squares dummy variables
heteroskedasticity.
approach
c. control for the random effects in the
c. The random effects model
data.

4
d. Heteroscedasticity and
autocorrelation consistent

23. Which of the following is a

disadvantage of the fixed effects
approach to estimating a panel
model?

a. The model is likely to be technical to

estimate
b. The approach may not be valid if the
composite error term is correlated
with one or more of the explanatory
variables
c. The number of parameters to
estimate may be large, resulting in a
loss of degrees of freedom
d. The fixed effects approach can only
capture cross-sectional
heterogeneity and not temporal
variation in the dependent variable.

24. Which of the following assumptions

is required for obtaining unbiased
fixed effect estimators?

a. The errors are heteroskedastic

b. The errors are serially correlated
c. The explanatory variables are
exogenous
d. The unobserved effect is correlated
with the explanatory variables

25. Which of the following types of

variables cannot be included in a
fixed effects model?

a. Dummy variable
b. Discrete dependent variable
c. Time-varying independent variable
d. Time-constant independent variable

5
Question 1
Question 2
Question 3
Question 4
Question 5
Question 6
Question 7
Question 8
Question 9
Question 10
Question 11
Question 12
Question 13
Question 14
Question 15
Question 16
Question 17
Question 18
Question 19
Question 20
Question 21
Question 22
Question 23
Question 24
Question 25