Scribd
Upload
13 views

Mock_exam (17)

économétrie

Uploaded by

dorineish
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
13 views

Mock_exam (17)

économétrie

Uploaded by

dorineish
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 5

Applied Econometrics I

Mock Exam

Barnabé Walheer

2022

Question 1 [2 points]. Argue whether the following statements are correct


or not:

(a) [0.5 point] Heteroskedasticity causes OLS coefficients to be biased in


finite samples.

(b) [0.5 point] The Mills’ ratio is obtained using observed data.

(c) [0.5 point] It is possible that the estimated average marginal effect
has a different sign than the estimated coefficient for probit models.

(d) [0.5 point] The Durbin–Wu–Hausman test tells us whether estimated


coefficients are efficient.

Question 2 [1 points]. Explain the intuition behind the following statis-


tical tests:

(a) [0.5 point] White test.

(b) [0.5 point] LR test.

Question 3 [1 points]. Compare the following concepts:

(a) [0.5 point] Adjusted R2 and Mc Fadden R2 .

(b) [0.5 point] Proxy variable and instrumental variable.

Question 4 [4 points]. Answer the following questions:

(a) [1 point] What is consistency in econometrics ?

1
(b) [1 point] Name four variables or statistics that directly impact the
standard errors of the coefficients in a multiple linear regression. Explain
the intuition of the sign of the impact for each case.

Question 5 [5 points]. Figure 1 presents the POLS estimates of a model


which explains the wage by different individual characteristics. The dependent
variable is the log of wage and the explanatory variables are defined as:

• educ: years of schooling

• age (in years)

• exper (experience) = age - educ - 6

• expersq = exper2

• union: dummy variable which equals 1 if the individual belongs to a


labor union, 0 otherwise

• female: dummy variable which equals 1 if the individual is a female, 0


otherwise

• married: dummy variable which equals 1 if the individual is married, 0


otherwise

• female_married = female × married

• nonwhite: dummy variable which equals 1 if the individual is nonwhite,


0 otherwise

• y85: dummy variable which equals 1 if the individual is observed in


1985, 0 otherwise. The base group is 1978.

(a) [1.5 point] Give an interpretation for the coefficient of "educ". Which
test do you need to assess the significance of the coefficient ? Compute the
test statistics and tell if the coefficient is significant or not.

(b) [0.5 point] The p-value of the Breusch–Pagan test for heteroskedas-
ticity equals 0.02. What can you conclude ?

(c) [1 point] What is the impact of being a white married female on wage.
Compute the exact impact.

(d) [2 point] Do you think the fixed effects estimator could be useful for
this model ? Which variables should disappear from the regression if you
apply the fixed effects estimator ? Why ?

2
Figure 1: Question 5

Question 6 [6 points]. Figure 2 presents the logit estimates of a model


which explains the probability of abusing alcohol by different individual
characteristics. Figure 3 presents the average partial effects (APE) of all
the explanatory variables from the logit regression. The dependent variable
equals 1 if an individual abuses alcohol and the explanatory variables are
defined as:

• age (in years)

• agesq = age2

• educ: years of schooling

• educsq = educ2

• famsize: family size

• beertax: state excise tax, $ per gallon

• married: dummy variable which equals 1 if the individual is married, 0


otherwise

• white: dummy variable which equals 1 if the individual is white, 0


otherwise

• employ: dummy variable which equals 1 if the individual is employed,


0 otherwise

3
(a) [1.5 point] Give an interpretation for the coefficient of famsize. Which
test do you need to assess the significance of the coefficient ? Compute the
test statistics and tell if the coefficient is significant or not.

(b) [0.5 point] What the regression is capturing with "educsq" ? Does it
seems logical ?

(c) [2 point] Let two individuals (A and B) be identical in every way


except that individuals A has 10 years of education while individuals B has
16 years of education. Compute the expected difference in their average
probability to abuse alcohol. Is individual A or B more likely to abuse alcohol
?

(d) [1 point] Write the null hypothesis of the joint F test of significance
of the coefficients of "educ" and "educsq". If the p-value equals 0.001, what
can you conclude ?

(e) [1 point] Explain what is an average partial effect ? Why is it useful


with the logit regression ?

Figure 2: Question 6

4
Figure 3: Question 6

You might also like