1

MULTIPLE CHOICE (7 points).

1) In the regression model: 𝐼𝑛𝑐𝑜𝑚𝑒 = 𝛽 + 𝛽 𝐴𝑔𝑒 + 𝛽 𝐸𝑑𝑢 + 𝑢, where Income is measured in
thousands of Vietnamese dong and Age is measured by numbers of years. Holding all other factors
being constant, 𝒕𝒉𝒆 𝒔𝒍𝒐𝒑𝒆 𝒄𝒐𝒆𝒇𝒇𝒊𝒄𝒊𝒆𝒏𝒕 𝛽
A. when multiplied with the value of age and edu, will give you the predicted income.
B. represents the elasticity of income with respect to age.
C. indicates by how much income change given a one-year increase in age.
D. indicates by how much income changes, given a one percent increase in age.
2) In the regression model: 𝑳𝒐𝒈_𝑤𝑎𝑔𝑒 = 𝛽 + 𝛽 𝐴𝑔𝑒 + 𝛽 𝐸𝑑𝑢 + 𝑢, where Log_wage is the
natural log of monthly wage measured in thousands of Vietnamese dong and Edu is measured by
numbers of formal schooling years. Holding all other factors being constant,
𝒕𝒉𝒆 𝒔𝒍𝒐𝒑𝒆 𝒄𝒐𝒆𝒇𝒇𝒊𝒄𝒊𝒆𝒏𝒕 𝛽
A. when multiplied with the value of edu, will give you the predicted log of wage
B. represents the elasticity of wage with respect to edu.
C. indicates by how much the absolute value of wage change if education increase by one
year.
D. indicates by the percentage change in wage if education increases by one year
3) In the regression model: 𝑳𝒐𝒈_𝑝𝑟𝑜𝑓𝑖𝑡 = 𝛽 + 𝛽 𝐿𝑜𝑔_𝑐𝑎𝑝𝑖𝑡𝑎𝑙 + 𝑢, where Log_wage is the
natural log of profits measured in millions of VND and Log_capital is the natural log of total
capital measured in millions of VND. Holding all other factors being constant,
𝒕𝒉𝒆 𝒔𝒍𝒐𝒑𝒆 𝒄𝒐𝒆𝒇𝒇𝒊𝒄𝒊𝒆𝒏𝒕 𝛽
A. when multiplied with the value of capital, will give you the predicted log of profit.
B. represents the elasticity of profit with respect to capital.
C. indicates by how much the absolute value of probit change if capital increases by one
million VND.
D. indicates by the percentage change in profit if capital increases by one percent
E. Both B and D
4) In the regression model: 𝑊𝑎𝑔𝑒 = 𝛽 + 𝛽 𝐸𝑑𝑢𝑐𝑎𝑡𝑖𝑜𝑛 + 𝛽 𝐴𝑔𝑒 + 𝑢, the intercept 𝛽
A. shows the mean wage at the mean level of education and age
B. shows the mean wage for those with the lowest level of education and age
C. represents the mean wage when education and age are set to zero.

1
 2

D. represents the mean wage for those with the highest level of education and age
5) In the regression model: 𝐼𝑛𝑐𝑜𝑚𝑒 = 𝛽 + 𝛽 𝑒𝑑𝑢 + 𝑢, the assumption Cov (𝑢|𝑒𝑑𝑢) = 0
implies that
A. there is a zero correlation between unobservable factors and the 𝑒𝑑𝑢 variable.
B. the sample variance of the 𝑒𝑑𝑢 variable equals the sample variance of the errors.
C. the residuals are correlated with the 𝑒𝑑𝑢 variable.
D. the error has a constant variance, given any values of edu variable
6) In the regression model: 𝑝𝑟𝑜𝑑𝑢𝑐𝑡𝑖𝑣𝑖𝑡𝑦 = 𝛽 + 𝛽 𝑙𝑎𝑛𝑑 𝑞𝑢𝑎𝑙𝑖𝑡𝑦 + 𝑢, the assumption
E(𝑢|𝑙𝑎𝑛𝑑 𝑞𝑢𝑎𝑙𝑖𝑡𝑦) = 0 means that
E. the sample mean of the 𝑙𝑎𝑛𝑑 𝑞𝑢𝑎𝑙𝑖𝑡𝑦 variable equals the sample mean of the errors.
F. given any values of 𝑙𝑎𝑛𝑑 𝑞𝑢𝑎𝑙𝑖𝑡𝑦 variable, the expected value of the error is zero
G. dividing the error by the 𝑙𝑎𝑛𝑑 𝑞𝑢𝑎𝑙𝑖𝑡𝑦 variable results in a zero (on average).
H. there is no correlation between the residuals and the 𝑙𝑎𝑛𝑑 𝑞𝑢𝑎𝑙𝑖𝑡𝑦 variable.
7) Heteroskedasticity or the error variance is unconstant when
A. there is a high level of correlation among explanatory variables.
B. the variance of Y is a function of Xs
C. the residuals are independent of explanatory variables
D. the error is not normally distributed

8) If assumptions 1-4 are satisfied, then the OLS estimator is still unbiased when
A. the error has unconstant variance 5
B. the variance of Y is a function of X 5
C. the error is not normally distributed 6
D. there is high correlation between explanatory variables
E. all of A, B, C and D
9) Which of the following assumptions is NOT necessary for OLS estimators to be BLUE?
A. the error term is normally distributed. 6
B. there is no exact linear relationship among explanatory variables
C. the dependent variable has constant variance given any value of the explanatory variables.
D. the error has a zero mean given any value of the explanatory variables.

2
 3

10) You are given a regression model that satisfies OLS assumptions 1 from 3. For assumption 4,
you find that Cov (𝑢, 𝑥) = 0. Your OLS estimators will be
A. consistent and biased consistance chỉ cần mẫu lớn
B. consistent and unbiased
C. biased and inconsistent.
D. biased and imprecise
11) If there is a high correlation among regressors in a regression model, có tg quan giữa các biến
A. the OLS estimator is less precise trong model
B. there is a decrease in the variance of the error
C. this violates one of the classical assumptions for the OLS estimator
D. this increases the variation in the values of regressors
12) Which of the following assumptions is NOT necessary for OLS estimators to be
Unbiased? chỉ cần 1-4
A. the data is from a random sample.
B. the error has a zero mean given any value of Xs
C. the variance of Y is constant given any value of Xs. 5
D. there is no perfect multicollinearity among explanatory variables.
13) There is an omitted variable bias if
A. a model omitted a variable that has zero partial effect on Y and is negatively correlated
with other explanatory variables.
B. a model did not include X2, while both the coefficients on X and X2 are statistically
significant.
C. a model omitted a variable that has a negative effect on Y and is uncorrelated with other
explanatory variables.
D. a model omitted a variable that has a positive effect on Y and is uncorrelated with other
explanatory variables.
14) An omitted variable bias will not occur when
A. a model omitted a variable that has non zero partial effect on Y and is positively correlated
with other explanatory variables.
B. a model omitted a variable that has a positive effect on Y and is uncorrelated with other
explanatory variables. khi làm model
chỉ cần nói biến X đó có tg quan đến Y nhưng uncorreclated
with X1 3
 4

C. a model omitted a variable that has a negative effect on Y and is negatively correlated with
other explanatory variables.
D. a model omitted a variable that has a non zero partial effect on Y and is negatively
correlated with other explanatory variables.
15) Consider the following regression model, where 𝑟𝑒𝑔𝑖𝑜𝑛 is a dummy variable (1=North;
0=South), 𝑙𝑎𝑛𝑑 is a continuous variable and the interaction term is 𝑟𝑒𝑔𝑖𝑜𝑛 ∗ 𝑙𝑎𝑛𝑑
𝑅𝑖𝑐𝑒 𝑜𝑢𝑡𝑝𝑢𝑡 = 𝛽 + 𝛽 𝑙𝑎𝑛𝑑 + 𝛽 𝑟𝑒𝑔𝑖𝑜𝑛 + 𝜷𝟑 𝑟𝑒𝑔𝑖𝑜𝑛 ∗ 𝑙𝑎𝑛𝑑 + 𝛽 𝑙𝑎𝑏𝑜𝑢𝑟 + 𝑢
In this model, 𝜷𝟑 is
A. the difference in mean outputs between North and South farms
B. the effect of 𝑙𝑎𝑛𝑑 on 𝑜𝑢𝑡𝑝𝑢𝑡𝑠 for both regions
C. the difference in the effect of 𝑙𝑎𝑛𝑑 on 𝑜utputs between the two regions
D. the difference in the intercepts between the two regions
E. the difference in the slope coefficients on the land variable between the two regions
F. Both C and E
16) Consider the following regression model, where 𝑔𝑒𝑛𝑑𝑒𝑟 is a dummy variable (1=male;
0=female, the base group), 𝑎𝑔𝑒 is a continuous variable and the interaction term is 𝑔𝑒𝑛𝑑𝑒𝑟 ∗ 𝑎𝑔𝑒
𝑤𝑎𝑔𝑒 = 𝛽 + 𝛽 𝑎𝑔𝑒 + 𝛽 𝑔𝑒𝑛𝑑𝑒𝑟 + 𝛽 𝑔𝑒𝑛𝑑𝑒𝑟 ∗ 𝑎𝑔𝑒 + 𝑢
Which of the following statements about the regression coefficients are correct?
A. 𝛽 shows the effect of 𝑎𝑔𝑒 on wage for female workers.
B. 𝛽 + 𝛽 shows the effect of 𝑎𝑔𝑒 on wage for male workers
C. 𝛽3 shows the difference in the wage effect of age (or the difference in the slope coefficients
on the age variable) between male and female workers
D. 𝛽2 shows the difference in the intercepts between male and female workers
E. All of A, B, C, 𝐷
17) In a multiple linear regression model, the precision of the slope coefficients will reduce if
F. we have fewer observations
G. there is a lower correlation between explanatory variables
H. there are less measurement errors in the dependent variable.
I. there is a high variation in the values of explanatory variable
18) In a multiple linear regression model, the slope coefficients will be more precise if
F. there is a high level of correlation among explanatory variables.

4
 5

G. there are more measurement errors in the dependent variable.

H. the sample size is larger.
I. there is low variation in the values of explanatory variables
II. CALCULATION (3 points)
19) Consider the following model for factors affecting income per capita (3 POINTS)
𝐿𝑜𝑔_𝑤𝑎𝑔𝑒 = 8.05 + 0.10𝑟𝑒𝑔𝑖𝑜𝑛 + 0.20𝑔𝑒𝑛𝑑𝑒𝑟 + 0.15 𝑟𝑒𝑔𝑖𝑜𝑛 ∗ 𝑔𝑒𝑛𝑑𝑒𝑟 − 0.04𝑒𝑥𝑝𝑒𝑟
+ 0.005𝑒𝑥𝑝𝑒𝑟
Where 𝐿𝑜𝑔_𝑤𝑎𝑔𝑒 is the natural log of monthy wage per worker; region is a dummy variable
(1=urban; 0=rural); gender is a dummy variable (=1 male; =0 female); exper is a continuous
variable (years) and 𝑒𝑥𝑝𝑒𝑟 is the squared term of 𝑒𝑥𝑝𝑒𝑟.
A) What is the wage gap between:
-Male workers in urban areas and female workers in rural areas
-female worker in urban areas and female workers in rural areas
-male workers in rural areas and female workers in rural areas

B) Holding all other factors fixed, how much will wage change if exeper increases from 21
to 22 years?

C) What is the turning point for the effect exper on wage? In this case, does the quadratic
relationship is a diminishing or increasing effect?

20) Consider the following model for factors affecting income per capita (3 POINTS)
𝐼𝑛𝑐𝑜𝑚𝑒 = 2000 + 200 𝑅𝑒𝑔𝑖𝑜𝑛 + 120 𝐸𝑑𝑢 + 90 𝑅𝑒𝑔𝑖𝑜𝑛 ∗ 𝐸𝑑𝑢 + 30𝑒𝑥𝑝𝑒𝑟 − 0.30𝑒𝑥𝑝𝑒𝑟
Where household income per capita (per month) is measured in thousands of VND; region is a
dummy variable (1=urban; 0=rural); Edu is a continuous variable (years of education);
Region*Edu is an interaction term; exper is a continuous variable measured in years and 𝑒𝑥𝑝𝑒𝑟
is the squared term of 𝑒𝑥𝑝𝑒𝑟.
(A) Please interpret the effect of education on income for urban and rural households. (1
point)

5
 6

B) Holding all other factors being constant, how much will income change if the experience
increases from 15 to 16 years? (1 point)

C) What is the turning point for the effect of experience on income? In this case, does the
quadratic relationship reveal a diminishing or increasing effect? (1 point)