2022 Test
2022 Test
2022 Test
24/03/2022
Name: Surname 1: Surname 2:
INSTRUCTIONS
• Fill in your identification data, both on this exam statement sheet and on the answer sheet.
• You are allowed a maximum of 60 minutes to answer this test. If you finish earlier, you can hand in the test and start
with the second part of the exam. In total, the maximum exam time allowed is 120 minuts.
• In each question there is only one correct answer. Correct answers are worth one point, wrong answers detract 1/3
points and questions left unanswered are worth 0.
1. We have estimated a Tobit model in which the only explanatory variable, apart form the constant term, is
ln(X). Which of the following Stata instructions allow us to calculate the marginal effect of a change in X on
the dependent variable (Y ) as an elasticity of the conditional expected value of Y (E[Y |Y > 0]) with respect
to X, after having estimated the Tobit model through maximum likelihood?
(a) margins, eyex(*)
(b) margins, dydx(*)
( c ) margins, eydx(*)
(d) None of the other answers is correct. We should specify predict ystar too!
2. We have estimated a wage equation through the Heckman procedure, in order to take into account potential
sample selection problems affecting this equation, given that wages are not observed for those individuals
who do not work. Let X be an explanatory variable that appears both in the wage equation and in the
participation equation and Z an explanatory variable that appears in the participation equation but not in
the wage equation. Which of the following statements is correct?
(a) None of the other answers is correct.
(b) The marginal effect of X on the unconditional expected value of the wage depends on the effect of X on
the probability that the wage is observed.
( c ) The marginal effect of Z on the unconditional expected value of the wage depends on the effect of Z on
the probability that the wage is observed.
(d) The marginal effect of X on the unconditional expected value of the wage is equal to the coefficient of
X in the wage equation.
Remember that the estimated parameters are the betas of the population line
3. Consider a multinomial Logit model with four transport alternatives. One of the explanatory variables is
the cost in Euro (variable whose value varies depending on the alternative) and its estimated coefficient is
negative. If the cost of all four alternatives increases by 10%, which of the following statements is correct?
(a) The predicted probability of the alternative with the highest cost will decrease.
(b) The predicted probabilities of the four alternatives will not change.
( c ) It is not possible to know how predicted probabilities will change with the available information.
(d) The predicted probabilities of all four alternatives will decrease by approximately 10%.
If all travelcosts increase the same units, nothing happens. But if they change a percentage then
the alternative with the higher cost suffers a decrease in the probability of being chosen while
the alternative with the lower cost experiences an increase.
Permuta n. 1 – Pag. 1
4. Which of the following statements about the lineal regression model is correct?
(a) The mean of the predicted probabilities of choosing alternative 1 is equal to the number of observations
in the sample that have chosen alternative 1. n times the mean is equal to that
(b) The sum of the observed values of the dependent variable is equal to the proportion of observations in
the sample that have chosen alternative 1. the sum of observed values over n is the proportion
( c ) All of the other answers are correct.
(d) The marginal effect of a variable on the probability of choosing alternative 1 is the same for all individuals.
5. We have estimated a Tobit model with a single explanatory variable (X) and a constant term. The estimated
coefficients for the constant term and the variable X are 3 and 1, respectively. Let us consider two individuals:
individual A with X = 10 and individual B with X = 9. Which of the following statements regarding the
marginal effect of a change in X is correct?
(a) The marginal effect on the expected value of the dependent variable is larger for individual A than for
individual B
(b) The marginal effect on the expected value of the dependent variable is different for the two individuals,
but we cannot know for whom the effect is larger.
( c ) The marginal effect on the probability that the dependent variable takes a positive value is smaller for
individual A than for individual B. The marginal effect on the sample is betaxNormal Cdf. For individual
(d) Answers a) i c) are correct. A the CDF part is larger as the CDF is increasing. The marginal effect
on the probit is beta*normal density. The normal density gets smaller
as x_i'beta diverges from 0, and therefore for individual B this is larger
6. In a Poisson model for count data, one of the explanatory variables (X) has an estimated coefficient of −0.1.
When the variable X increases by one unit, which of the following statements is correct?
(a) The variance of the dependent variable gets multiplied by e−0.1 .
(b) The expected value of the dependent variable decreases by approximately 10%.
( c ) Answers a) i b) are correct.
(d) The logarithm of the expected value of the dependent variable decreases by approximately 0.1%.
7. In a regression model with panel data and individual random effects, under the usual assumptions, the error
terms are characterized by:
(a) Heteroskedasticity.
(b) Autocorrelation.
( c ) All covariances between the error terms of the different observations being equal to 0 and the variances
being constant.
(d) Answers a) i b) are correct.
8. In a multinomial Logit model with 4 alternatives, where the baseline alternative is alternative 1, we know
that when the variable X increases by one unit, the ratio P2 /P1 gets multiplied by e0.3 ; the ratio P3 /P2 by
e0.2 and the ratio P4 /P2 by e−0.6 , where Pk is the probability of choosing alternative k. Considering this
information, which are the estimated coefficients for the variable X in the alternatives 2, 3 and 4?
(a) β2 = 0.3 β3 = 0.5 β4 = −0.3
(b) It is not possible to know, with this information.
( c ) β2 = 0.3 β3 = 0.5 β4 = 0
(d) β2 = 0.3 β3 = 0.2 β4 = −0.6
Permuta n. 1 – Pag. 2
9. We want to choose between two Probit models. Both have a constant term, but in one case the only explana-
tory variable is X and in the other the only explanatory variable is Z. Which of the following statistics can
we use?
(a) Likelihood ratio test.
10. We have estimated a binary Logit model with one explanatory variable (X) apart from the constant term.
If the coefficient estimated for this variable is −0.1, which of the following statements is correct when the
variable X increases by one unit?
(a) The ratio between the probability of choosing alternative 1 and the probability of choosing alternative 0
will multiply approximately by 1.1.
(b) The probability of choosing alternative 1 will decrease approximately by 10%. If you compute the % change
in probability (Pnew-Pold)/Pold
( c ) The probability of choosing alternative 0 will increase by 0.1 units. it gives around 10% decrease
(d) None of the other answers is correct.
11. In a regression model with panel data, where each individual is observed over several years, for which of the
following variables will we be able to estimate the coefficient if we use the within-group transformation?
(a) Region of birth.
(b) Year of the observation. The time invariant variables are eliminated in the within
( c ) None of the other answers is correct.
group transformation
12. In a linear regression model, we have specified a qualitative variable that has three categories by including
dummies for categories 2 and 3 and leaving category 1 as the reference. The estimated coefficients for the
two dummy variables are 0.6 and 0.9, respectively. What would the estimated coefficients β1 , β2 and β3 be,
if we included dummies for all three categories, imposing the restriction that the sum of the three coefficients
is zero?
(a) β1 = −0.5 β2 = 0.1 β3 = 0.4
(b) One cannot know this with the available information.
( c ) β1 = −0.3 β2 = 0.3 β3 = 0.6
Simply take into account that the diferrence between beta_1
(d) β1 = −0.3 β2 = 0 β3 = 0.3 and beta_2 should be -0.6, between beta_2 and beta_3 should
be -0.3 and the one between beta_1 and beta_3 is -0.9. Besides
the sum is 0.
Permuta n. 1 – Pag. 3
ECONOMETRICS III FINAL EXAM DNI / NIE: NIA:
24/03/2022
Name: Surname 1: Surname 2:
INSTRUCTIONS
• Fill in your identification data, both on this exam statement sheet and on the answer sheet.
• You are allowed a maximum of 60 minutes to answer this test. If you finish earlier, you can hand in the test and start
with the second part of the exam. In total, the maximum exam time allowed is 120 minuts.
• In each question there is only one correct answer. Correct answers are worth one point, wrong answers detract 1/3
points and questions left unanswered are worth 0.
1. We have estimated a wage equation through the Heckman procedure, in order to take into account potential
sample selection problems affecting this equation, given that wages are not observed for those individuals
who do not work. Let X be an explanatory variable that appears both in the wage equation and in the
participation equation and Z an explanatory variable that appears in the participation equation but not in
the wage equation. Which of the following statements is correct?
(a) The marginal effect of X on the unconditional expected value of the wage depends on the effect of X on
the probability that the wage is observed.
(b) The marginal effect of Z on the unconditional expected value of the wage depends on the effect of Z on
the probability that the wage is observed.
( c ) None of the other answers is correct.
(d) The marginal effect of X on the unconditional expected value of the wage is equal to the coefficient of
X in the wage equation.
2. We have estimated a Tobit model in which the only explanatory variable, apart form the constant term, is
ln(X). Which of the following Stata instructions allow us to calculate the marginal effect of a change in X on
the dependent variable (Y ) as an elasticity of the conditional expected value of Y (E[Y |Y > 0]) with respect
to X, after having estimated the Tobit model through maximum likelihood?
(a) margins, eyex(*)
(b) margins, dydx(*)
( c ) margins, eydx(*)
(d) None of the other answers is correct.
3. In a linear regression model, we have specified a qualitative variable that has three categories by including
dummies for categories 2 and 3 and leaving category 1 as the reference. The estimated coefficients for the
two dummy variables are 0.6 and 0.9, respectively. What would the estimated coefficients β1 , β2 and β3 be,
if we included dummies for all three categories, imposing the restriction that the sum of the three coefficients
is zero?
(a) β1 = −0.3 β2 = 0 β3 = 0.3
Permuta n. 2 – Pag. 1
4. In a regression model with panel data, where each individual is observed over several years, for which of the
following variables will we be able to estimate the coefficient if we use the within-group transformation?
(a) Year of birth.
(b) Year of the observation.
5. Which of the following statements about the lineal regression model is correct?
(a) The mean of the predicted probabilities of choosing alternative 1 is equal to the number of observations
in the sample that have chosen alternative 1.
6. In a regression model with panel data and individual random effects, under the usual assumptions, the error
terms are characterized by:
(a) Heteroskedasticity.
(b) Autocorrelation.
( c ) All covariances between the error terms of the different observations being equal to 0 and the variances
being constant.
(d) Answers a) i b) are correct.
7. We want to choose between two Probit models. Both have a constant term, but in one case the only explana-
tory variable is X and in the other the only explanatory variable is Z. Which of the following statistics can
we use?
(a) Likelihood ratio test.
(b) Hausman test.
( c ) Akaike Information Criterion.
8. We have estimated a binary Logit model with one explanatory variable (X) apart from the constant term.
If the coefficient estimated for this variable is −0.1, which of the following statements is correct when the
variable X increases by one unit?
(a) The probability of choosing alternative 1 will decrease approximately by 10%.
(b) The probability of choosing alternative 0 will increase by 0.1 units.
( c ) None of the other answers is correct.
(d) The ratio between the probability of choosing alternative 1 and the probability of choosing alternative 0
will multiply approximately by 1.1.
Permuta n. 2 – Pag. 2
9. In a multinomial Logit model with 4 alternatives, where the baseline alternative is alternative 1, we know
that when the variable X increases by one unit, the ratio P2 /P1 gets multiplied by e0.3 ; the ratio P3 /P2 by
e0.2 and the ratio P4 /P2 by e−0.6 , where Pk is the probability of choosing alternative k. Considering this
information, which are the estimated coefficients for the variable X in the alternatives 2, 3 and 4?
(a) β2 = 0.3 β3 = 0.5 β4 = −0.3
10. We have estimated a Tobit model with a single explanatory variable (X) and a constant term. The estimated
coefficients for the constant term and the variable X are 3 and 1, respectively. Let us consider two individuals:
individual A with X = 10 and individual B with X = 9. Which of the following statements regarding the
marginal effect of a change in X is correct?
(a) The marginal effect on the expected value of the dependent variable is larger for individual A than for
individual B
(b) The marginal effect on the expected value of the dependent variable is different for the two individuals,
but we cannot know for whom the effect is larger.
( c ) The marginal effect on the probability that the dependent variable takes a positive value is smaller for
individual A than for individual B.
(d) Answers a) i c) are correct.
11. Consider a multinomial Logit model with four transport alternatives. One of the explanatory variables is
the cost in Euro (variable whose value varies depending on the alternative) and its estimated coefficient is
negative. If the cost of all four alternatives increases by 10%, which of the following statements is correct?
(a) The predicted probabilities of all four alternatives will decrease by approximately 10%.
(b) The predicted probabilities of the four alternatives will not change.
( c ) It is not possible to know how predicted probabilities will change with the available information.
(d) The predicted probability of the alternative with the highest cost will decrease.
12. In a Poisson model for count data, one of the explanatory variables (X) has an estimated coefficient of −0.1.
When the variable X increases by one unit, which of the following statements is correct?
(a) The variance of the dependent variable gets multiplied by e−0.1 .
(b) The expected value of the dependent variable decreases by approximately 10%.
( c ) Answers a) i b) are correct.
(d) The logarithm of the expected value of the dependent variable decreases by approximately 0.1%.
Permuta n. 2 – Pag. 3
ECONOMETRICS III FINAL EXAM DNI / NIE: NIA:
24/03/2022
Name: Surname 1: Surname 2:
INSTRUCTIONS
• Fill in your identification data, both on this exam statement sheet and on the answer sheet.
• You are allowed a maximum of 60 minutes to answer this test. If you finish earlier, you can hand in the test and start
with the second part of the exam. In total, the maximum exam time allowed is 120 minuts.
• In each question there is only one correct answer. Correct answers are worth one point, wrong answers detract 1/3
points and questions left unanswered are worth 0.
1. In a Poisson model for count data, one of the explanatory variables (X) has an estimated coefficient of −0.1.
When the variable X increases by one unit, which of the following statements is correct?
(a) The variance of the dependent variable gets multiplied by e−0.1 .
(b) The expected value of the dependent variable decreases by approximately 10%.
( c ) Answers a) i b) are correct.
(d) The logarithm of the expected value of the dependent variable decreases by approximately 0.1%.
2. In a linear regression model, we have specified a qualitative variable that has three categories by including
dummies for categories 2 and 3 and leaving category 1 as the reference. The estimated coefficients for the
two dummy variables are 0.6 and 0.9, respectively. What would the estimated coefficients β1 , β2 and β3 be,
if we included dummies for all three categories, imposing the restriction that the sum of the three coefficients
is zero?
(a) β1 = −0.3 β2 = 0.3 β3 = 0.6
(b) β1 = −0.3 β2 = 0 β3 = 0.3
3. We have estimated a wage equation through the Heckman procedure, in order to take into account potential
sample selection problems affecting this equation, given that wages are not observed for those individuals
who do not work. Let X be an explanatory variable that appears both in the wage equation and in the
participation equation and Z an explanatory variable that appears in the participation equation but not in
the wage equation. Which of the following statements is correct?
(a) The marginal effect of Z on the unconditional expected value of the wage depends on the effect of Z on
the probability that the wage is observed.
(b) The marginal effect of X on the unconditional expected value of the wage depends on the effect of X on
the probability that the wage is observed.
( c ) None of the other answers is correct.
(d) The marginal effect of X on the unconditional expected value of the wage is equal to the coefficient of
X in the wage equation.
Permuta n. 3 – Pag. 1
4. We have estimated a Tobit model with a single explanatory variable (X) and a constant term. The estimated
coefficients for the constant term and the variable X are 3 and 1, respectively. Let us consider two individuals:
individual A with X = 10 and individual B with X = 9. Which of the following statements regarding the
marginal effect of a change in X is correct?
(a) The marginal effect on the expected value of the dependent variable is larger for individual A than for
individual B
(b) The marginal effect on the expected value of the dependent variable is different for the two individuals,
but we cannot know for whom the effect is larger.
( c ) The marginal effect on the probability that the dependent variable takes a positive value is smaller for
individual A than for individual B.
(d) Answers a) i c) are correct.
5. In a regression model with panel data, where each individual is observed over several years, for which of the
following variables will we be able to estimate the coefficient if we use the within-group transformation?
(a) Region of birth.
(b) None of the other answers is correct.
6. In a regression model with panel data and individual random effects, under the usual assumptions, the error
terms are characterized by:
(a) Heteroskedasticity.
(b) Autocorrelation.
( c ) All covariances between the error terms of the different observations being equal to 0 and the variances
being constant.
(d) Answers a) i b) are correct.
7. We want to choose between two Probit models. Both have a constant term, but in one case the only explana-
tory variable is X and in the other the only explanatory variable is Z. Which of the following statistics can
we use?
(a) Likelihood ratio test.
(b) Hausman test.
8. We have estimated a binary Logit model with one explanatory variable (X) apart from the constant term.
If the coefficient estimated for this variable is −0.1, which of the following statements is correct when the
variable X increases by one unit?
(a) The ratio between the probability of choosing alternative 1 and the probability of choosing alternative 0
will multiply approximately by 1.1.
Permuta n. 3 – Pag. 2
9. Consider a multinomial Logit model with four transport alternatives. One of the explanatory variables is
the cost in Euro (variable whose value varies depending on the alternative) and its estimated coefficient is
negative. If the cost of all four alternatives increases by 10%, which of the following statements is correct?
(a) The predicted probabilities of all four alternatives will decrease by approximately 10%.
(b) The predicted probabilities of the four alternatives will not change.
( c ) It is not possible to know how predicted probabilities will change with the available information.
(d) The predicted probability of the alternative with the highest cost will decrease.
10. In a multinomial Logit model with 4 alternatives, where the baseline alternative is alternative 1, we know
that when the variable X increases by one unit, the ratio P2 /P1 gets multiplied by e0.3 ; the ratio P3 /P2 by
e0.2 and the ratio P4 /P2 by e−0.6 , where Pk is the probability of choosing alternative k. Considering this
information, which are the estimated coefficients for the variable X in the alternatives 2, 3 and 4?
(a) β2 = 0.3 β3 = 0.2 β4 = −0.6
(b) β2 = 0.3 β3 = 0.5 β4 = 0
( c ) It is not possible to know, with this information.
11. We have estimated a Tobit model in which the only explanatory variable, apart form the constant term, is
ln(X). Which of the following Stata instructions allow us to calculate the marginal effect of a change in X on
the dependent variable (Y ) as an elasticity of the conditional expected value of Y (E[Y |Y > 0]) with respect
to X, after having estimated the Tobit model through maximum likelihood?
(a) None of the other answers is correct.
(b) margins, eydx(*)
( c ) margins, eyex(*)
(d) margins, dydx(*)
12. Which of the following statements about the lineal regression model is correct?
(a) The sum of the observed values of the dependent variable is equal to the proportion of observations in
the sample that have chosen alternative 1.
(b) All of the other answers are correct.
( c ) The mean of the predicted probabilities of choosing alternative 1 is equal to the number of observations
in the sample that have chosen alternative 1.
(d) The marginal effect of a variable on the probability of choosing alternative 1 is the same for all individuals.
Permuta n. 3 – Pag. 3
ECONOMETRICS III FINAL EXAM DNI / NIE: NIA:
24/03/2022
Name: Surname 1: Surname 2:
INSTRUCTIONS
• Fill in your identification data, both on this exam statement sheet and on the answer sheet.
• You are allowed a maximum of 60 minutes to answer this test. If you finish earlier, you can hand in the test and start
with the second part of the exam. In total, the maximum exam time allowed is 120 minuts.
• In each question there is only one correct answer. Correct answers are worth one point, wrong answers detract 1/3
points and questions left unanswered are worth 0.
1. In a linear regression model, we have specified a qualitative variable that has three categories by including
dummies for categories 2 and 3 and leaving category 1 as the reference. The estimated coefficients for the
two dummy variables are 0.6 and 0.9, respectively. What would the estimated coefficients β1 , β2 and β3 be,
if we included dummies for all three categories, imposing the restriction that the sum of the three coefficients
is zero?
(a) β1 = −0.3 β2 = 0 β3 = 0.3
(b) One cannot know this with the available information.
2. We have estimated a binary Logit model with one explanatory variable (X) apart from the constant term.
If the coefficient estimated for this variable is −0.1, which of the following statements is correct when the
variable X increases by one unit?
(a) The ratio between the probability of choosing alternative 1 and the probability of choosing alternative 0
will multiply approximately by 1.1.
(b) The probability of choosing alternative 0 will increase by 0.1 units.
( c ) None of the other answers is correct.
(d) The probability of choosing alternative 1 will decrease approximately by 10%.
3. We have estimated a Tobit model with a single explanatory variable (X) and a constant term. The estimated
coefficients for the constant term and the variable X are 3 and 1, respectively. Let us consider two individuals:
individual A with X = 10 and individual B with X = 9. Which of the following statements regarding the
marginal effect of a change in X is correct?
(a) The marginal effect on the expected value of the dependent variable is larger for individual A than for
individual B
(b) The marginal effect on the expected value of the dependent variable is different for the two individuals,
but we cannot know for whom the effect is larger.
( c ) The marginal effect on the probability that the dependent variable takes a positive value is smaller for
individual A than for individual B.
(d) Answers a) i c) are correct.
Permuta n. 4 – Pag. 1
4. We have estimated a Tobit model in which the only explanatory variable, apart form the constant term, is
ln(X). Which of the following Stata instructions allow us to calculate the marginal effect of a change in X on
the dependent variable (Y ) as an elasticity of the conditional expected value of Y (E[Y |Y > 0]) with respect
to X, after having estimated the Tobit model through maximum likelihood?
(a) None of the other answers is correct.
5. Which of the following statements about the lineal regression model is correct?
(a) The mean of the predicted probabilities of choosing alternative 1 is equal to the number of observations
in the sample that have chosen alternative 1.
(b) All of the other answers are correct.
( c ) The marginal effect of a variable on the probability of choosing alternative 1 is the same for all individuals.
(d) The sum of the observed values of the dependent variable is equal to the proportion of observations in
the sample that have chosen alternative 1.
6. In a multinomial Logit model with 4 alternatives, where the baseline alternative is alternative 1, we know
that when the variable X increases by one unit, the ratio P2 /P1 gets multiplied by e0.3 ; the ratio P3 /P2 by
e0.2 and the ratio P4 /P2 by e−0.6 , where Pk is the probability of choosing alternative k. Considering this
information, which are the estimated coefficients for the variable X in the alternatives 2, 3 and 4?
(a) β2 = 0.3 β3 = 0.2 β4 = −0.6
(b) β2 = 0.3 β3 = 0.5 β4 = 0
( c ) It is not possible to know, with this information.
7. In a regression model with panel data, where each individual is observed over several years, for which of the
following variables will we be able to estimate the coefficient if we use the within-group transformation?
(a) None of the other answers is correct.
(b) Year of birth.
8. Consider a multinomial Logit model with four transport alternatives. One of the explanatory variables is
the cost in Euro (variable whose value varies depending on the alternative) and its estimated coefficient is
negative. If the cost of all four alternatives increases by 10%, which of the following statements is correct?
(a) The predicted probabilities of all four alternatives will decrease by approximately 10%.
(b) The predicted probability of the alternative with the highest cost will decrease.
( c ) The predicted probabilities of the four alternatives will not change.
(d) It is not possible to know how predicted probabilities will change with the available information.
Permuta n. 4 – Pag. 2
9. In a regression model with panel data and individual random effects, under the usual assumptions, the error
terms are characterized by:
(a) Heteroskedasticity.
(b) Autocorrelation.
( c ) All covariances between the error terms of the different observations being equal to 0 and the variances
being constant.
(d) Answers a) i b) are correct.
10. We want to choose between two Probit models. Both have a constant term, but in one case the only explana-
tory variable is X and in the other the only explanatory variable is Z. Which of the following statistics can
we use?
(a) Likelihood ratio test.
(b) Hausman test.
( c ) Akaike Information Criterion.
11. In a Poisson model for count data, one of the explanatory variables (X) has an estimated coefficient of −0.1.
When the variable X increases by one unit, which of the following statements is correct?
(a) The variance of the dependent variable gets multiplied by e−0.1 .
(b) The expected value of the dependent variable decreases by approximately 10%.
( c ) Answers a) i b) are correct.
(d) The logarithm of the expected value of the dependent variable decreases by approximately 0.1%.
12. We have estimated a wage equation through the Heckman procedure, in order to take into account potential
sample selection problems affecting this equation, given that wages are not observed for those individuals
who do not work. Let X be an explanatory variable that appears both in the wage equation and in the
participation equation and Z an explanatory variable that appears in the participation equation but not in
the wage equation. Which of the following statements is correct?
(a) The marginal effect of X on the unconditional expected value of the wage depends on the effect of X on
the probability that the wage is observed.
(b) None of the other answers is correct.
( c ) The marginal effect of X on the unconditional expected value of the wage is equal to the coefficient of
X in the wage equation.
(d) The marginal effect of Z on the unconditional expected value of the wage depends on the effect of Z on
the probability that the wage is observed.
Permuta n. 4 – Pag. 3