lOMoARcPSD|19476778

Introductory-econometrics-test-bank 5th edi

Applied Econometrics for Economics and Finance (Deakin University)

Studocu is not sponsored or endorsed by any college or university

Downloaded by K61.FTU ?INH H?NG NG?C (031304000270@ftu.edu.vn)
Test Bank –Introductory Econometrics: A Modern Approach, 5th Edition by
Jeffrey M. Wooldridge

With PERFECT SOLUTION AVAILABLE OF ALL CHAPTERS

Chapter 1

1. Econometrics is the branch of economics that .

a. studies the behavior of individual economic agents in making economic decisions
b. develops and uses statistical methods for estimating economic relationships
c. deals with the performance, structure, behavior, and decision-making
of an economy as a whole
d. applies mathematical methods to represent economic theories and
solve economic problems.

Answer: b
Difficulty: Easy
Bloom’s: Knowledge
A-Head: What is Econometrics?
BUSPROG:
Feedback: Econometrics is the branch of economics that develops and uses
statistical methods for estimating economic relationships.

2. Nonexperimental data is called .

a. cross-sectional data
b. time series data
c. observational data
d. panel data

Answer: b
Difficulty: Easy
Bloom’s: Knowledge
A-Head: What is Econometrics?
BUSPROG:
Feedback:

3. Which of the following is true of experimental data?

a. Experimental data are collected in laboratory environments in the
natural sciences.
b. Experimental data cannot be collected in a controlled environment.
c. Experimental data is sometimes called observational data.
d. Experimental data is sometimes called retrospective data.

Answer: a
Difficulty: Easy

Downloaded by K61.FTU ?INH H?NG NG?C

Bloom’s: Knowledge
A-Head: What is Econometrics?
BUSPROG:
Feedback:

4. An empirical analysis relies on to test a theory.

a. common sense
b. ethical considerations
c. data
d. customs and conventions

Answer: c
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Steps in Empirical Economic Analysis
BUSPROG:
Feedback: An empirical analysis relies on data to test a theory.

5. The term ‘u’ in an econometric model is usually referred to as the .

a. error term
b. parameter
c. hypothesis
d. dependent variable

Answer: a
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Steps in Empirical Economic Analysis
BUSPROG:
Feedback: The term u in an econometric model is called the error term or
disturbance term.

6. The parameters of an econometric model .

a. include all unobserved factors affecting the variable being studied
b. describe the strength of the relationship between the variable under study
and the factors affecting it
c. refer to the explanatory variables included in the model
d. refer to the predictions that can be made using the model

Answer: b
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Steps in Empirical Economic Analysis
BUSPROG:

Downloaded by K61.FTU ?INH H?NG NG?C

Feedback: The parameters of an econometric model describe the direction and
strength of the relationship between the variable under study and the factors
affecting it.

7. Which of the following is the first step in empirical economic analysis?

a. Collection of data
b. Statement of hypotheses
c. Specification of an econometric model
d. Testing of hypotheses

Answer: c
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Steps in Empirical Economic Analysis
BUSPROG:
Feedback: The first step in empirical economic analysis is the specification of the
econometric model.

8. A data set that consists of a sample of individuals, households, firms, cities,

states, countries, or a variety of other units, taken at a given point in time, is
called a(n) .
a. cross-sectional data set
b. longitudinal data set
c. time series data set
d. experimental data set

Answer: a
Difficulty: Easy
Bloom’s: Knowledge
A-Head: The Structure of Economic Data
BUSPROG:
Feedback: A data set that consists of a sample of individuals, households, firms,
cities, states, countries, or a variety of other units, taken at a given point in time, is
called a cross-sectional data set.

9. Data on the income of law graduates collected at different times during the
same year is .
a. panel data
b. experimental data
c. time series data
d. cross-sectional data

Answer: d
Difficulty: Easy

Downloaded by K61.FTU ?INH H?NG NG?C

Bloom’s: Application
A-Head: The Structure of Economic Data
BUSPROG: Analytic
Feedback: A data set that consists of a sample of individuals, households, firms,
cities, states, countries, or a variety of other units, taken at a given point in time, is
called a cross-sectional data set. Therefore, data on the income of law graduates on
a particular year are examples of cross-sectional data.

10. A data set that consists of observations on a variable or several variables

over time is called a data set.
a. binary
b. cross-sectional
c. time series
d. experimental

Answer: c
Difficulty: Easy
Bloom’s: Knowledge
A-Head: The Structure of Economic Data
BUSPROG:
Feedback: A time-series data set consists of observations on a variable or several
variables over time.

11. Which of the following is an example of time series data?

a. Data on the unemployment rates in different parts of a country during a year.
b. Data on the consumption of wheat by 200 households during a year.
c. Data on the gross domestic product of a country over a period of 10 years.
d. Data on the number of vacancies in various departments of an organization on
a particular month.

Answer: c
Difficulty: Easy
Bloom’s: Application
A-Head: The Structure of Economic Data
BUSPROG: Analytic
Feedback: A time-series data set consists of observations on a variable or several
variables over
time. Therefore, data on the gross domestic product of a country over a period of 10
years is an example of time series data.

12. Which of the following refers to panel data?

a. Data on the unemployment rate in a country over a 5-year period
b. Data on the birth rate, death rate and population growth rate in
developing countries over a 10-year period.

Downloaded by K61.FTU ?INH H?NG NG?C

c. Data on the income of 5 members of a family on a particular year.
d. Data on the price of a company’s share during a year.

Answer: b
Difficulty: Easy
Bloom’s: Application
A-Head: The Structure of Economic Data
BUSPROG: Analytic

Feedback: A panel data set consists of a time series for each cross-sectional
member in the data set. Therefore, data on the birth rate, death rate and infant
mortality rate in developing countries over a 10-year period refers to panel data.

13. Which of the following is a difference between panel and pooled cross-
sectional data?
a. A panel data set consists of data on different cross-sectional units over a given
period of time while a pooled data set consists of data on the same cross-
sectional units over a given period of time.
b. A panel data set consists of data on the same cross-sectional units over a
given period of time while a pooled data set consists of data on different cross-
sectional units over a given period of time
c. A panel data consists of data on a single variable measured at a given point in
time while a pooled data set consists of data on the same cross-sectional units
over a given period of time.
d. A panel data set consists of data on a single variable measured at a given point
in time while a pooled data set consists of data on more than one variable at a
given point in time.

Answer: b
Difficulty: Easy
Bloom’s: Knowledge
A-Head: The Structure of Economic Data
BUSPROG:
Feedback: A panel data set consists of data on the same cross-sectional units over a
given period of time while a pooled data set consists of data on the same cross-
sectional units over a given period of time.

14. has a causal effect on .

a. Income; unemployment
b. Height; health
c. Income; consumption
d. Age; wage

Answer: c

Downloaded by K61.FTU ?INH H?NG NG?C

Difficulty: Moderate
Bloom’s: Application
A-Head: Causality and the Notion of Ceteris Paribus in Econometric Analysis
BUSPROG: Analytic
Feedback: Income has a causal effect on consumption because an increase in
income leads to an increase in consumption.

15. Which of the following is true?

a. A variable has a causal effect on another variable if both variables increase
or decrease simultaneously.
b. The notion of ‘ceteris paribus’ plays an important role in causal analysis.
c. Difficulty in inferring causality disappears when studying data at fairly high
levels of aggregation.
d. The problem of inferring causality arises if experimental data is used for analysis.

Answer: b
Difficulty:
Moderate
Bloom’s: Knowledge
A-Head: Causality and the Notion of Ceteris Paribus in Econometric Analysis
BUSPROG:
Feedback: The notion of ‘ceteris paribus’ plays an important role in causal analysis.

16. Experimental data are sometimes called retrospective data.

Answer: False
Difficulty: Easy
Bloom’s: Knowledge
A-Head: What is Econometrics?
BUSPROG:
Feedback: Nonexperimental data are sometimes called retrospective data.

17. An economic model consists of mathematical equations that describe

various relationships between economic variables.

Answer: True
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Steps in Empirical Economic Analysis
BUSPROG:
Feedback: An economic model consists of mathematical equations that describe
various relationships between economic variables.

18. A cross-sectional data set consists of observations on a variable or

several variables over time.

Downloaded by K61.FTU ?INH H?NG NG?C

Answer: False
Difficulty: Easy
Bloom’s: Knowledge
A-Head: The Structure of Economic Data
BUSPROG:
Feedback: A time series data set consists of observations on a variable or several
variables over
time.

19. A time series data is also called a longitudinal data set.

Answer: False
Difficulty: Easy
Bloom’s: Knowledge
A-Head: The Structure of Economic Data
BUSPROG:
Feedback: A time series data is also called a longitudinal data set.

20. The notion of ceteris paribus means “other factors being equal.”

Answer: True
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Causality and the Notion of Ceteris Paribus in Econometric Analysis
BUSPROG:
Feedback: The notion of ceteris paribus means “other factors being equal.”

Chapter 2

1. A dependent variable is also known as a(n) .

a. explanatory variable
b. control variable
c. predictor variable
d. response variable

Answer: d
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Definition of the Simple Regression Model
BUSPROG:
Feedback: A dependent variable is known as a response variable.

Downloaded by K61.FTU ?INH H?NG NG?C

2. 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

Answer: c
Difficulty: Easy
Bloom’s: Comprehension
A-Head: Definition of the Simple Regression Model
BUSPROG:
Feedback: If a change in variable x causes a change in variable y, variable x is
called the independent variable or the explanatory variable.

β0 β1 β0
3. In the equation y = + x + u, is the .
a. dependent variable
b. independent variable
c. slope parameter
d. intercept parameter

Answer: d
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Definition of the Simple Regression Model
BUSPROG:
β0 β1 β0
Feedback: In the equation y = + x + u, is the intercept parameter.

β0 β1 β0
4. In the equation y = + x + u, what is the estimated value of ?

a. ´y−β^1 x´

b. ´y + β 1 ´x
y
yi −
´¿
¿
¿
c. ( x i− ´x) ¿
n

∑¿
i=1
¿

Downloaded by K61.FTU ?INH H?NG NG?C

 n

d. ∑
i=1
xy

Answer: a
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Deriving the Ordinary Least Squares Estimates
BUSPROG:
β0 ´y−β^1 x´
Feedback: The estimated value of is .

β0 β1
5. In the equation c = + i + u, c denotes consumption and i denotes

c5
income. What is the residual for the 5th observation if =$500 and c^5
=$475?
a. $975
b. $300
c. $25
d. $50

Answer: c
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Deriving the Ordinary Least Squares Estimates
BUSPROG:
Feedback: The formula for calculating the residual for the i th observation is
u^i= y i −^y i u^5=c5 −c^5
. In this case, the residual is =$500 -$475=
$25.

^y=β^0 + ^β 1 x
6. What does the equation denote if the regression equation is y =
β0
+ β1x1 + u?
a. The explained sum of squares
b. The total sum of squares
c. The sample regression function
d. The population regression function

Answer: c
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Deriving the Ordinary Least Squares Estimates
BUSPROG:

Downloaded by K61.FTU ?INH H?NG NG?C

 ^y=β^0 + ^β 1 x
Feedback: The equation denotes the sample regression function
of the given regression model.

7. Consider the following regression model: y = β0 + β1x1 + u. Which of the following

is a property of Ordinary Least Square (OLS) estimates of this model and their
associated statistics?
a. The sum, and therefore the sample average of the OLS residuals, is positive.
b. The sum of the OLS residuals is negative.
c. The sample covariance between the regressors and the OLS residuals is positive.
´x y´ ) always lies on the OLS regression line.
d. The point ( ,

Answer: d
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Properties of OLS on Any Sample of Data
BUSPROG:
Feedback: An important property of the OLS estimates is that the point (
x´ ,
´y )

always lies on the OLS regression line. In other words, if

, the predicted value
x=x´

of y is ´y .

y i =β 0 + β 1 x1+ u1
8. The explained sum of squares for the regression function, ,
is defined as .
n

a. ∑(
i=1
y i− ´y )2

b. ∑ ( y −^y )
i=1
i
2

c. ∑ u^
i=1
i

d. ∑ (u )
i=1
i
2

Answer: b
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Properties of OLS on Any Sample of Data

Downloaded by K61.FTU ?INH H?NG NG?C

BUSPROG:
n

Feedback: The explained sum of squares is defined as ∑ ( y −^y )

i
2

i=1

9. If the total sum of squares (SST) in a regression equation is 81, and the
residual sum of squares (SSR) is 25, what is the explained sum of squares (SSE)?
a. 64
b. 56
c. 32
d. 18

Answer: b
Difficulty:
Moderate
Bloom’s: Application
A-Head: Properties of OLS on Any Sample of Data
BUSPROG: Analytic
Feedback: Total sum of squares (SST) is given by the sum of explained sum of
squares (SSE) and residual sum of squares (SSR). Therefore, in this case, SSE=81-
25=56.

10. If the residual sum of squares (SSR) in a regression analysis is 66 and the
total sum of squares (SST) is equal to 90, what is the value of the coefficient of
determination?
a. 0.73
b. 0.55
c. 0.27
d. 1.2

Answer: c
Difficulty:
Moderate
Bloom’s: Application
A-Head: Properties of OLS on Any Sample of Data
BUSPROG: Analytic
Feedback: The formula for calculating the coefficient of determination is

R2=1− 2 66
SSR R =1− =0.27
. In this case, 9
SST

11. Which of the following is a nonlinear regression

model? a. y = β0 + β1x1/2 + u
b. log y = β0 + β1log x +u
c. y = 1 / (β0 + β1x) + u
d. y = β0 + β1x + u

Downloaded by K61.FTU ?INH H?NG NG?C

Answer: c
Difficulty:
Moderate
Bloom’s: Comprehension
A-Head: Properties of OLS on Any Sample of Data
BUSPROG:
Feedback: A regression model is nonlinear if the equation is nonlinear in the
parameters. In this case, y=1 / (β0 + β1x) + u is nonlinear as it is nonlinear in its
parameters.

12. Which of the following is assumed for establishing the unbiasedness of

Ordinary Least Square (OLS) estimates?
a. The error term has an expected value of 1 given any value of the
explanatory variable.
b. The regression equation is linear in the explained and explanatory variables.
c. The sample outcomes on the explanatory variable are all the same value.
d. The error term has the same variance given any value of the
explanatory variable.

Answer: d
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Expected Values and Variances of the OLS Estimators
BUSPROG:
Feedback: The error u has the same variance given any value of the explanatory
variable.

13. The error term in a regression equation is said to exhibit homoskedasticty if

.
a. it has zero conditional mean
b. it has the same variance for all values of the explanatory variable.
c. it has the same value for all values of the explanatory variable
d. if the error term has a value of one given any value of the explanatory variable.

Answer: b
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Expected Values and Variances of the OLS Estimators
BUSPROG:
Feedback: The error term in a regression equation is said to exhibit homoskedasticty
if it has the same variance for all values of the explanatory variable.

14. In the regression of y on x, the error term exhibits heteroskedasticity if .

a. it has a constant variance

Downloaded by K61.FTU ?INH H?NG NG?C

b. Var(y|x) is a function of x
c. x is a function of y
d. y is a function of x

Answer: b
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Expected Values and Variances of the OLS Estimators
BUSPROG:
Feedback: Heteroskedasticity is present whenever Var(y|x) is a function of x
because Var(u|x) = Var(y|x).

15. What is the estimated value of the slope parameter when the
regression equation, y = β0 + β1x1 + u passes through the origin?
n

a. ∑ yi
i=1

y
¿
¿
b.
¿ )
n
∑¿
i=1

∑i=1
xi yi
c. n

∑
i=1
xi 2
n

d. ∑(
i=1
y i− ´y )2

Answer: c
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Regression through the Origin and Regression on a Constant
BUSPROG:

Downloaded by K61.FTU ?INH H?NG NG?C

Feedback: The estimated value of the slope parameter when the regression
n

∑
i=1
xi yi
equation passes through the origin is n .

∑
i=1
xi 2

16. A natural measure of the association between two random variables is

the correlation coefficient.

Answer: True
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Definition of the Simple Regression Model
BUSPROG:
Feedback: A natural measure of the association between two random variables is
the correlation coefficient.

17. The sample covariance between the regressors and the Ordinary Least
Square (OLS) residuals is always positive.

Answer: False
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Properties of OLS on Any Sample of Data
BUSPROG:
Feedback: The sample covariance between the regressors and the Ordinary Least
Square (OLS) residuals is zero.

2
18. R is the ratio of the explained variation compared to the total variation.

Answer: True
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Properties of OLS on Any Sample of Data
BUSPROG:
Feedback: The sample covariance between the regressors and the Ordinary Least
Square (OLS) residuals is zero.

19. There are n-1 degrees of freedom in Ordinary Least Square residuals.

Answer: False

Downloaded by K61.FTU ?INH H?NG NG?C

Difficulty: Easy
Bloom’s: Knowledge
A-Head: Expected Values and Variances of the OLS Estimators
BUSPROG:
Feedback: There are n-2 degrees of freedom in Ordinary Least Square residuals.

20. The variance of the slope estimator increases as the error variance decreases.

Answer: False
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Expected Values and Variances of the OLS Estimators
BUSPROG:
Feedback: The variance of the slope estimator increases as the error variance
increases.

Chapter 3

y=β 0 + β 1 x1 + β2 x2+ u β2
1. In the equation, , is a(n) .

a. independent variable
b. dependent variable
c. slope parameter
d. intercept parameter

Answer: c
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Motivation for Multiple Regression
BUSPROG:
y=β 0 + β 1 x1 + β2 x2+ u β2
Feedback: In the equation, , is a slope parameter.

2. Consider the following regression equation: y=β

β2 1 + x1+ β2 x2+u
. What does β1
imply?
β1 x1 x
a. measures the ceteris paribus effect of on 2.

β1 y on x1
b. measures the ceteris paribus effect of .
β1 x1 y.
c. measures the ceteris paribus effect of on
β1 x1
d. measures the ceteris paribus effect of on u .

Downloaded by K61.FTU ?INH H?NG NG?C

Answer: c
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Motivation for Multiple Regression
BUSPROG:
β1 x1 y.
Feedback: measures the ceteris paribus effect of on

3. If the explained sum of squares is 35 and the total sum of squares is 49, what
is the residual sum of squares?
a. 10
b. 12
c. 18
d. 14

Answer: d
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Mechanics and Interpretation of Ordinary Least Squares
BUSPROG: Analytic
Feedback: The residual sum of squares is obtained by subtracting the explained
sum of squares from the total sum of squares, or 49-35=14.

4. Which of the following is true of R 2?

a. R2 is also called the standard error of regression.
b. A low R2 indicates that the Ordinary Least Squares line fits the data well.
c. R2 usually decreases with an increase in the number of independent variables in
a regression.
d. R2 shows what percentage of the total variation in the dependent variable, Y,
is explained by the explanatory variables.

Answer: d
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Mechanics and Interpretation of Ordinary Least Squares
BUSPROG:
Feedback: R2 shows what percentage of the total variation in Y is explained by the
explanatory variables.

5. The value of R2 always .

a. lies below 0
b. lies above 1
c. lies between 0 and 1

Downloaded by K61.FTU ?INH H?NG NG?C

d. lies between 1 and 1.5

Answer: c
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Mechanics and Interpretation of Ordinary Least Squares
BUSPROG:
Feedback: By definition, the value of R2 always lies between 0 and 1.

6. If an independent variable in a multiple linear regression model is an exact

linear combination of other independent variables, the model suffers from the
problem of
.
a. perfect collinearity
b. homoskedasticity
c. heteroskedasticty
d. omitted variable bias

Answer: a
Difficulty: Easy
Bloom’s: Knowledge
A-Head: The Expected Value of the OLS Estimators
BUSPROG:
Feedback: If an independent variable in a multiple linear regression model is an
exact linear combination of other independent variables, the model suffers from the
problem of perfect collinearity.

7. The assumption that there are no exact linear relationships among the
independent variables in a multiple linear regression model fails if , where n is
the sample size and k is the number of parameters.
a. n>2
b. n=k+1
c. n>k
d. n<k+1

Answer: d
Difficulty: Easy
Bloom’s: Knowledge
A-Head: The Expected Value of the OLS Estimators
BUSPROG:
Feedback: The assumption of no perfect collinearity among independent variables
fails if n<k+1.

8. Exclusion of a relevant variable from a multiple linear regression model leads

to the problem of .

Downloaded by K61.FTU ?INH H?NG NG?C

a. misspecification of the model
b. multicollinearity
c. perfect collinearity
d. homoskedasticity

Answer: a
Difficulty: Easy
Bloom’s: Knowledge
A-Head: The Expected Value of the OLS Estimators
BUSPROG:
Feedback: Exclusion of a relevant variable from a multiple linear regression model
leads to the problem of misspecification of the model.

9. Suppose the variable x2 has been

~ omitted from the following regression equation,
y=β0+ β1 x1+ β2 x2+ u . β
1 is the estimator obtained when x2 is omitted from the
~
equation. The bias in β1 is positive if .
β2
a. >0 and x 1 and x 2 are positively correlated
β2
b. <0 and x 1 and x 2 are positively correlated
β2
c. >0 and x 1 and x 2 are negatively correlated
β2
d. = 0 and x 1 and x 2 are negatively correlated

Answer: a
Difficulty: Easy
Bloom’s: Knowledge
A-Head: The Expected Value of the OLS Estimators
BUSPROG:
Feedback: When the variable x2 is omitted from the regression, the bias ~
β1 is
β2
in positive if >0 and x 1 and x 2 are positively correlated.

10. Suppose the variable x2 has been omitted from the following regression
equation,
y=β0+ β1 x1+ β2 x2+ u .
is the estimator obtained when x2 is omitted
~
β1
~
from the equation. The bias in
β 1 is negative if .

β2
a. >0 and x 1 and x 2 are positively correlated

Downloaded by K61.FTU ?INH H?NG NG?C

 β2
b. <0 and x 1 and x 2 are positively correlated
β2
c. =0 and x 1 and x 2 are negatively correlated
β2
d. =0 and x 1 and x 2 are negatively correlated

Answer: b
Difficulty: Easy
Bloom’s: Knowledge
A-Head: The Expected Value of the OLS Estimators
BUSPROG:
Feedback: When the variable x2 is omitted from the regression, the bias ~
β1 is
β2
in negative if <0 and x 1 and x 2 are positively correlated.

11. Suppose the variable x2 has been omitted from the following regression
equation,
y=β0+ β1 x1+ β2 x2+ u .
is the estimator obtained when x2 is omitted
~
β1
~ ~
from the equation. If E( β1 ) >β1,
β1 is said to .
a. have an upward bias
b. have a downward bias
c. be unbiased
d. be biased toward zero

Answer: a
Difficulty: Easy
Bloom’s: Knowledge
A-Head: The Expected Value of the OLS Estimators
BUSPROG:
Feedback: When the variable x2 is omitted from the following regression equation,
y=β0+ β1 x1+ β2 x2+ u , ,
has an upward bias if E( ) >β1.
~
β1
~
β1

12. High (but not perfect) correlation between two or more independent variables
is called .
a. heteroskedasticty
b. homoskedasticty
c. multicollinearity
d. micronumerosity

Downloaded by K61.FTU ?INH H?NG NG?C

Answer: c

Downloaded by K61.FTU ?INH H?NG NG?C

Difficulty: Easy
Bloom’s: Knowledge
A-Head: The Variance of the OLS Estimators
BUSPROG:
Feedback: High, but not perfect, correlation between two or more independent
variables is called multicollinearity.

13. The term refers to the problem of small sample size.

a. micronumerosity
b. multicollinearity
c. homoskedasticity
d. heteroskedasticity

Answer: a
Difficulty: Easy
Bloom’s: Knowledge
A-Head: The Variance of the OLS Estimators
BUSPROG:
Feedback: The term micronumerosity refers to the problem of small sample size.

14. Find the degrees of freedom in a regression model that has 10 observations
and 7 independent variables.
a. 17
b. 2
c. 3
d. 4

Answer: b
Difficulty: Easy
Bloom’s: Knowledge
A-Head: The Variance of the OLS Estimators
BUSPROG: Analytic
Feedback: The degrees of freedom in a regression model is computed by
subtracting the number of parameters from the number of observations in a
regression model. Since, the number of parameters is one more than the number of
independent variables, the degrees of freedom in this case is 10-(7 + 1) = 2.

15. 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

Downloaded by K61.FTU ?INH H?NG NG?C

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

Answer: c
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Efficiency of OLS: The Gauss-Markov Theorem
BUSPROG:
Feedback: The Gauss-Markov theorem will not hold if the independent variables
have exact linear relationships among them.

16. The term “linear” in a multiple linear regression model means that the
equation is linear in parameters.

Answer: True
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Motivation for Multiple Regression
BUSPROG:
Feedback: The term “linear” in a multiple linear regression model means that the
equation is linear in parameters.

17. The key assumption for the general multiple regression model is that all
factors in the unobserved error term be correlated with the explanatory variables.

Answer: False
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Motivation for Multiple Regression
BUSPROG:
Feedback: The key assumption of the general multiple regression model is that all
factors in the unobserved error term be uncorrelated with the explanatory variables.

18. The coefficient of determination (R2) decreases when an independent variable

is added to a multiple regression model.

Answer: False
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Mechanics and Interpretation of Ordinary Least Squares
BUSPROG:
Feedback: The coefficient of determination (R2) never decreases when an
independent variable is added to a multiple regression model.

Downloaded by K61.FTU ?INH H?NG NG?C

19. An explanatory variable is said to be exogenous if it is correlated with the
error term.

Answer: False
Difficulty: Easy
Bloom’s: Knowledge
A-Head: The Expected Value of the OLS Estimators
BUSPROG:
Feedback: An explanatory variable is said to be endogenous if it is correlated with
the error term.

20. A larger error variance makes it difficult to estimate the partial effect of any
of the independent variables on the dependent variable.

Answer: True
Difficulty: Easy
Bloom’s: Knowledge
A-Head: The Variance of the OLS Estimators
BUSPROG:
Feedback: A larger error variance makes it difficult to estimate the partial effect of
any of the independent variables on the dependent variable.

Chapter 4

1. The normality assumption implies that:

a. the population error u is dependent on the explanatory variables and is
normally distributed with mean equal to one and variance σ2.
b. the population error u is independent of the explanatory variables and is normally
distributed with mean equal to one and variance σ.
c. the population error u is dependent on the explanatory variables and is
normally distributed with mean zero and variance σ.
d. the population error u is independent of the explanatory variables and is normally
distributed with mean zero and variance σ2.

Answer: d
Difficulty:
Moderate
Bloom’s: Knowledge
A-Head: Sampling Distributions of the OLS Estimators
BUSPROG:
Feedback: The normality assumption implies that the population error ‘u’ is
independent of the explanatory variables and is normally distributed with mean
zero and variance σ2.

2. Which of the following statements is true?

Downloaded by K61.FTU ?INH H?NG NG?C

a. Taking a log of a nonnormal distribution yields a distribution that is closer
to normal.
b. The mean of a nonnormal distribution is 0 and the variance is σ2.
c. The CLT assumes that the dependent variable is unaffected by
unobserved factors.
d. OLS estimators have the highest variance among unbiased estimators.

Answer: a
Difficulty:
Moderate
Bloom’s: Knowledge
A-Head: Sampling Distribution of the OLS Estimators
BUSPROG:
Feedback: Transformations such as logs of nonnormal distributions, yields
distributions which are closer to normal.

3. A normal variable is standardized by:

a. subtracting off its mean from it and multiplying by its standard deviation.
b. adding its mean to it and multiplying by its standard deviation.
c. subtracting off its mean from it and dividing by its standard deviation.
d. adding its mean to it and dividing by its standard deviation.

Answer: c
Difficulty:
Moderate
Bloom’s: Knowledge
A-Head: Sampling Distribution of the OLS Estimators
BUSPROG:
Feedback: A normal variable is standardized by subtracting off its mean from it and
dividing by its standard deviation.

4. Which of the following is a statistic that can be used to test hypotheses about
a single population parameter?
a. F statistic
b. t statistic
c. χ2 statistic
d. Durbin Watson statistic

Answer: b
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Testing Hypotheses about a Single Population Parameter: The t Test
BUSPROG:
Feedback: The t statistic can be used to test hypotheses about a single population
parameter.

Downloaded by K61.FTU ?INH H?NG NG?C

5. Consider the equation, Y = β1 + β2X2 + u. A null hypothesis, H0: β2 = 0 states that:
a. X2 has no effect on the expected value of β2.
b. X2 has no effect on the expected value of Y.
c. β2 has no effect on the expected value of Y.
d. Y has no effect on the expected value of X2.

Answer: b
Difficulty:
Moderate
Bloom’s: Comprehension
A-Head: Testing Hypotheses about a Single Population Parameter: The t Test
BUSPROG:
Feedback: In such an equation, a null hypothesis, H0: β2 = 0 states that X2 has
no effect on the expected value of Y. This is because β2 is the coefficient
associated with X2.

6. The significance level of a test is:

a. the probability of rejecting the null hypothesis when it is false.
b. one minus the probability of rejecting the null hypothesis when it is false.
c. the probability of rejecting the null hypothesis when it is true.
d. one minus the probability of rejecting the null hypothesis when it is true.

Answer: c
Difficulty:
Moderate
Bloom’s: Knowledge
A-Head: Testing Hypotheses about a Single Population Parameter: The t Test
BUSPROG:
Feedback: The significance level of a test refers to the probability of rejecting
the null hypothesis when it is in fact true.

7. The general t statistic can be written as:

Hypothesized value
a. t = Standard e rror
estimate – hypothesized value
b. t = ¿ )

(estimate – hypothesized value )

c. t = variance

(estimate – hypothesized value)

d. t = standard error

Answer: d
Difficulty: Easy

Downloaded by K61.FTU ?INH H?NG NG?C

Bloom’s: Knowledge
A-Head: Testing Hypotheses about a Single Population Parameter: The t Test
BUSPROG:
Feedback: The general t statistic can be written as t = (estimate – hypothesized
value)/standard error.

8. Which of the following statements is true of confidence intervals?

a. Confidence intervals in a CLM are also referred to as point estimates.
b. Confidence intervals in a CLM provide a range of likely values for the
population parameter.
c. Confidence intervals in a CLM do not depend on the degrees of freedom of
a distribution.
d. Confidence intervals in a CLM can be truly estimated when heteroskedasticity
is present.

Answer: b
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Confidence Intervals
BUSPROG:
Feedback: Confidence intervals provide a range of likely values for the population
parameter and are not point estimates. Estimation of confidence intervals depends
on the degrees of freedom of the distribution and cannot be truly estimated when
heteroskedasticity is present.

9. Which of the following statements is true?

a. When the standard error of an estimate increases, the confidence interval for
the estimate narrows down.
b. Standard error of an estimate does not affect the confidence interval for
the estimate.
c. The lower bound of the confidence interval for a regression coefficient, say βj,
^ ^β
is given by β J - [standard error × ( J )]

d. The upper bound of the confidence interval for a regression coefficient, say β j,
^ ^β
is given by β J + [Critical value × standard error ( J )]

Answer: d
Difficulty:
Moderate
Bloom’s: Knowledge
A-Head: Confidence Intervals
BUSPROG:

Downloaded by K61.FTU ?INH H?NG NG?C

Feedback: The upper bound of the confidence interval for a regression coefficient,
^β
say βj, is given by J + [Critical value × standard error ( )].
J

^β

10. Which of the following tools is used to test multiple linear restrictions?
a. t test
b. z test
c. F test
d. Unit root test

Answer: c
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Testing Multiple Linear Restrictions: The F test
BUSPROG:
Feedback: The F test is used to test multiple linear restrictions.

11. Which of the following statements is true of hypothesis testing?

a. The t test can be used to test multiple linear restrictions.
b. A test of single restriction is also referred to as a joint hypotheses test.
c. A restricted model will always have fewer parameters than its unrestricted model.
d. OLS estimates maximize the sum of squared residuals.

Answer: c
Difficulty:
Moderate
Bloom’s: Knowledge
A-Head: Testing Multiple Linear Restrictions: The F test
BUSPROG:
Feedback: A restricted model will always have fewer parameters than its
unrestricted model.

12. Which of the following correctly defines F statistic if SSR r represents sum of
squared residuals from the restricted model of hypothesis testing, SSR ur represents
sum of squared residuals of the unrestricted model, and q is the number of
restrictions placed?
(SSRur − SSRr )/ q
a. F = SSR ur /( n−k −1)

( SSRr − SSRur )/
b. F q SSR ur /(n−k
=
−1)

(SSR ur − SSRr )/
c. F = q SSR r /(
n−k−1)
Downloaded by K61.FTU ?INH H?NG NG?C
 (SSRur −SSR r )/(
d. F = n−k −1)
SSR ur / q

Answer: b
Difficulty:
Moderate
Bloom’s: Knowledge
A-Head: Testing Multiple Linear Restrictions: The F test
BUSPROG:
(SSR r −SSRur )/
Feedback: The F statistic is given by, F = q SSR ur /(n−k
−1)

13. Which of the following statements is true?

a. If the calculated value of F statistic is higher than the critical value, we reject
the alternative hypothesis in favor of the null hypothesis.
b. The F statistic is always nonnegative as SSRr is never smaller than SSRur.
c. Degrees of freedom of a restricted model is always less than the degrees
of freedom of an unrestricted model.
d. The F statistic is more flexible than the t statistic to test a hypothesis with
a single restriction.

Answer: b
Difficulty:
Moderate
Bloom’s: Comprehension
A-Head: Testing Multiple Linear Restrictions: The F test
BUSPROG:
Feedback: The F statistic is always nonnegative as SSR r is never smaller than SSRur.

14. If R2 ur= 0.6873, R2 = 0.5377, number of restrictions = 3, and n – k – 1 = 229, F

statistic equals:
a. 21.2
b. 28.6
c. 36.5
d. 42.1

Answer: c
Difficulty: Hard
Bloom’s: Application
A-Head: Testing Multiple Linear Restrictions: The F test
BUSPROG: Analytic
Feedback: The F statistic can be calculated as F = [(R 2ur – R2r)/q] / [(1-R2ur)/n – k – 1].
Here, q represents the number of restrictions. In this case it is equal to [(0.6873 –
0.5377)/3] / [(1 – 0.6873)/229] = [0.04986/0.001365] = 36.5.
Downloaded by K61.FTU ?INH H?NG NG?C
15. Which of the following correctly identifies a reason why some authors prefer
to report the standard errors rather than the t statistic?
a. Having standard errors makes it easier to compute confidence intervals.
b. Standard errors are always positive.
c. The F statistic can be reported just by looking at the standard errors.
d. Standard errors can be used directly to test multiple linear regressions.

Answer: a
Difficulty: Medium
Bloom’s: Comprehension
A-Head: Reporting Regression
Results BUSPROG:
Feedback: One of the advantages of reporting standard errors over t statistics is
that confidence intervals can be easily calculated using stand errors.

16. Whenever the dependent variable takes on just a few values it is close to
a normal distribution.

Answer: False
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Sampling Distribution of the OLS Estimators
BUSPROG:
Feedback: Whenever the dependent variable takes on just a few values it cannot
have anything close to a normal distribution. A normal distribution requires the
dependent variable to take up a large range of values.

17. If the calculated value of the t statistic is greater than the critical value, the
null hypothesis, H0 is rejected in favor of the alternative hypothesis, H1.

Answer: True
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Testing Hypotheses about a Single Population Parameter: The t Test
BUSPROG:
Feedback: If the calculated value of the t statistic is greater than the critical value,
H0 is rejected in favor of H1.

18. H1: βj ≠ 0, where βj is a regression coefficient associated with an explanatory

variable, represents a one-sided alternative hypothesis.

Answer: False
Difficulty: Easy

Downloaded by K61.FTU ?INH H?NG NG?C

Bloom’s: Knowledge
A-Head: Testing Hypotheses about a Single Population Parameter: The t Test
BUSPROG:
Feedback: H1: βj ≠ 0, where βj is a regression coefficient associated with an
explanatory variable, represents a two-sided alternative hypothesis.

^β ^β
19. If 1 and 2 are estimated values of regression coefficients associated
with two explanatory variables in a regression equation, then the standard error (

^β
1 –

^β ^β ^β
2 ) = standard error ( ) – standard error (
1 ).
2

Answer: False
Difficulty: Easy
Bloom’s: Comprehension
A-Head: Testing Hypotheses about a Single Linear Combinations of the Parameters
BUSPROG:
^β ^β
Feedback: If 1 and 2 are estimated values of regression coefficients

associated with two explanatory variables in a regression equation, then the

^β ^β ^
standard error ( 1 – ) ≠ standard error ( β
2 1 ) – standard error ( ).
2

^β

20. Standard errors must always be positive.

Answer: True
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Testing Hypotheses about a Single Linear Combinations of the Parameters
BUSPROG:
Feedback: Standard errors must always be positive since they are estimates of
standard deviations.
Chapter 5

1. Which of the following statements is true?

a. The standard error of a regression, σ^ , is not an unbiased estimator for σ ,

the standard deviation of the error, u, in a multiple regression model.

b. In time series regressions, OLS estimators are always unbiased.
c. Almost all economists agree that unbiasedness is a minimal requirement for
an estimator in regression analysis.
d. All estimators in a regression model that are consistent are also unbiased.

Downloaded by K61.FTU ?INH H?NG NG?C

Answer: b

Downloaded by K61.FTU ?INH H?NG NG?C

Difficulty: Moderate
Bloom’s: Knowledge
A-Head: Consistency
BUSPROG:
Feedback: The standard error of a regression is not an unbiased estimator for the
standard deviation of the error in a multiple regression model.

^β β , is consistent, then the:

2. If j, an unbiased estimator of j

^β β
a. distribution of j becomes more and more loosely distributed around j as
the sample size grows.
^β β
b. distribution of j becomes more and more tightly distributed around j as
the sample size grows.
^β
c. distribution of j tends toward a standard normal distribution as the
sample size grows.
^β
d. distribution of j remains unaffected as the sample size grows.

Answer: b
Difficulty:
Medium
Bloom’s: Knowledge
A-Head:
Consistency
BUSPROG:
^β β , is consistent, then the distribution
Feedback: If j, an unbiased estimator of j

^β β
of j becomes more and more tightly distributed around j as the sample
size grows.

^β β , is also a consistent estimator of β ,

3. If j, an unbiased estimator of j j

then when the sample size tends to infinity:

^β
a. the distribution of j collapses to a single value of zero.
^β
b. the distribution of j diverges away from a single value of zero.
^β β .
c. the distribution of j collapses to the single point j

^β β .
d. the distribution of j diverges away from j

Answer: c

Downloaded by K61.FTU ?INH H?NG NG?C

Difficulty: Easy

Downloaded by K61.FTU ?INH H?NG NG?C

Bloom’s: Knowledge
A-Head:
Consistency
BUSPROG:
^β β , is also a consistent estimator of
Feedback: If j, an unbiased estimator of j

β , then when the sample size tends to infinity the distribution of ^β

j j

collapses to the single point

β .
j

4. In a multiple regression model, the OLS estimator is consistent if:

a. there is no correlation between the dependent variables and the error term.
b. there is a perfect correlation between the dependent variables and the
error term.
c. the sample size is less than the number of parameters in the model.
d. there is no correlation between the independent variables and the error term.

Answer: d
Difficulty:
Moderate
Bloom’s: Knowledge
A-Head:
Consistency
BUSPROG:
Feedback: In a multiple regression model, the OLS estimator is consistent if there is
no correlation between the explanatory variables and the error term.

5. If the error term is correlated with any of the independent variables, the
OLS estimators are:
a. biased and consistent.
b. unbiased and inconsistent.
c. biased and inconsistent.
d. unbiased and consistent.

Answer: c
Difficulty:
Easy
Bloom’s: Knowledge
A-Head:
Consistency
BUSPROG:
Feedback: If the error term is correlated with any of the independent variables, then
the OLS estimators are biased and inconsistent.

6. If δ1 = Cov(x1/x2) / Var(x1) where x1 and x2 are two independent variables in

a regression equation, which of the following statements is true?

Downloaded by K61.FTU ?INH H?NG NG?C

a. If x2 has a positive partial effect on the dependent variable, and δ1 > 0, then the
inconsistency in the simple regression slope estimator associated with x1 is
negative.
b. If x2 has a positive partial effect on the dependent variable, and δ1 > 0, then the
inconsistency in the simple regression slope estimator associated with x1 is
positive.
c. If x1 has a positive partial effect on the dependent variable, and δ1 > 0, then the
inconsistency in the simple regression slope estimator associated with x1 is
negative.
d. If x1 has a positive partial effect on the dependent variable, and δ1 > 0, then the
inconsistency in the simple regression slope estimator associated with x1 is
positive.

Answer: b
Difficulty:
Moderate
Bloom’s: Knowledge
A-Head:
Consistency
BUSPROG:
Feedback: Given that δ1 = Cov(x1/x2)/Var(x1) where x1 and x2 are two independent
variables in a regression equation, if x2 has a positive partial effect on the
dependent variable, and δ1 > 0, then the inconsistency in the simple regression
slope estimator associated with x1 is positive.

7. If OLS estimators satisfy asymptotic normality, it implies that:

a. they are approximately normally distributed in large enough sample sizes.
b. they are approximately normally distributed in samples with less than
10 observations.
c. they have a constant mean equal to zero and variance equal to σ2.
d. they have a constant mean equal to one and variance equal to σ.

Answer: a
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Asymptotic Normality and Large Sample Inference
BUSPROG:
Feedback: If OLS estimators satisfy asymptotic normality, it implies that they are
approximately normally distributed in large enough sample sizes.

8. In a regression model, if variance of the dependent variable, y, conditional on

an explanatory variable, x, or Var(y|x), is not constant, .
a. the t statistics are invalid and confidence intervals are valid for small
sample sizes
b. the t statistics are valid and confidence intervals are invalid for small
sample sizes
c. the t statistics confidence intervals are valid no matter how large the sample

Downloaded by K61.FTU ?INH H?NG NG?C

size is

Downloaded by K61.FTU ?INH H?NG NG?C

c. the t statistics and confidence intervals are both invalid no matter how large the
sample size is

Answer: d
Difficulty:
Moderate
Bloom’s: Knowledge
A-Head: Asymptotic Normality and Large Sample Inference
BUSPROG:
Feedback: If variance of the dependent variable conditional on an explanatory
variable is not a constant the usual t statistics confidence intervals are both invalid
no matter how large the sample size is.

^β
9. If j is an OLS estimator of a regression coefficient associated with one of
the explanatory variables, such that j= 1, 2, …., n, asymptotic standard error of

^β
j

will refer to the:

^β
a. estimated variance of j when the error term is normally distributed.

b. estimated variance of a given coefficient when the error term is not

normally distributed.
c. square root of the estimated variance β^ j when the error term is normally
of distributed.
d. square root of the estimated variance β^ j when the error term is not normally
of distributed.
Answer: d
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Asymptotic Normality and Large Sample Inference
BUSPROG:
Feedback: Asymptotic standard error refers to the square root of the
^β
estimated variance of j when the error term is not normally distributed.

10. A useful rule of thumb is that standard errors are expected to shrink at a
rate that is the inverse of the:
a. square root of the sample size.
b. product of the sample size and the number of parameters in the model.
c. square of the sample size.
d. sum of the sample size and the number of parameters in the model.

Downloaded by K61.FTU ?INH H?NG NG?C

Answer: a
Difficulty:
Moderate
Bloom’s: Knowledge
A-Head: Asymptotic Normality and Large Sample Inference
BUSPROG:
Feedback: Standard errors can be expected to shrink at a rate that is the inverse of
the square root of the sample size.

11. An auxiliary regression refers to a regression that is used:

a. when the dependent variables are qualitative in nature.
b. when the independent variables are qualitative in nature.
c. to compute a test statistic but whose coefficients are not of direct interest.
d. to compute coefficients which are of direct interest in the analysis.

Answer: c
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Asymptotic Normality and Large Sample Inference
BUSPROG:
Feedback: An auxiliary regression refers to a regression that is used to compute a
test statistic but whose coefficients are not of direct interest.

12. The n-R-squared statistic also refers to the:

a. F statistic.
b. t statistic.
c. z statistic.
d. LM statistic.

Answer: d
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Asymptotic Normality and Large Sample Inference
BUSPROG:
Feedback: The n-R-squared statistic also refers to the LM statistic.

13. The LM statistic follows a:

a. t distribution.
b. f distribution.
c.
χ 2 distribution.

d. binomial distribution.

Answer: c
Difficulty: Easy

Downloaded by K61.FTU ?INH H?NG NG?C

Bloom’s: Knowledge
A-Head: Asymptotic Normality and Large Sample Inference
BUSPROG:
Feedback: The LM statistic follows a
χ 2 distribution.

14. Which of the following statements is true?

a. In large samples there are not many discrepancies between the outcomes of
the F test and the LM test.
b. Degrees of freedom of the unrestricted model are necessary for using the
LM test.
c. The LM test can be used to test hypotheses with single restrictions only
and provides inefficient results for multiple restrictions.
d. The LM statistic is derived on the basis of the normality assumption.

Answer: a
Difficulty:
Moderate
Bloom’s: Knowledge
A-Head: Asymptotic Normality and Large Sample Inference
BUSPROG:
Feedback: In large samples there are not many discrepancies between the F test
and the LM test because asymptotically the two statistics have the same probability
of a Type 1 error.

15. Which of the following statements is true under the Gauss-Markov assumptions?
a. Among a certain class of estimators, OLS estimators are best linear unbiased,
but are asymptotically inefficient.
b. Among a certain class of estimators, OLS estimators are biased
but asymptotically efficient.
c. Among a certain class of estimators, OLS estimators are best linear unbiased
and asymptotically efficient.
d. The LM test is independent of the Gauss-Markov assumptions.

Answer: c
Difficulty:
Moderate
Bloom’s: Knowledge
A-Head: Asymptotic Efficiency of OLS
BUSPROG:
Feedback: Under the Gauss-Markov assumptions, among a certain class of
estimators, OLS estimators are best linear unbiased and asymptotically efficient.

Downloaded by K61.FTU ?INH H?NG NG?C

16. If variance of an independent variable in a regression model, say x1, is greater
^β
than 0, or Var(x1) > 0, the inconsistency in 1 (estimator associated with x1)
is negative, if x1 and the error term are positively related.

Answer: False
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Consistency
BUSPROG:
Feedback: If variance of an independent variable, say x 1, is greater than 0, the
^β
inconsistency in 1 (estimator associated with x1) is positive if x1 and the
error term are positively related.

17. Even if the error terms in a regression equation, u1, u2,….., un, are not normally
distributed, the estimated coefficients can be normally distributed.

Answer: False
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Asymptotic Normality and Large Sample Inference
BUSPROG:
Feedback: Even if the error terms in a regression equation, u1, u2,….., un, are not
normally distributed, the estimated coefficients cannot be normally distributed.

18. A normally distributed random variable is symmetrically distributed about its

mean, it can take on any positive or negative value (but with zero probability), and
more than 95% of the area under the distribution is within two standard
deviations.

Answer: True
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Asymptotic Normality and Large Sample Inference
BUSPROG:
Feedback: A normally distributed random variable is symmetrically distributed about
its mean, it can take on any positive or negative value (but with zero probability),
and more than 95% of the area under the distribution is within two standard
deviations.

19. The F statistic is also referred to as the score statistic.

Answer: False
Difficulty: Easy

Downloaded by K61.FTU ?INH H?NG NG?C

Bloom’s: Knowledge
A-Head: Asymptotic Normality and Large Sample Inference
BUSPROG:
Feedback: The LM statistic is also referred to as the score statistic.

20. The LM statistic requires estimation of the unrestricted model only.

Answer: False
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Asymptotic Normality and Large Sample Inference
BUSPROG:
Feedback: The LM statistic requires estimation of the restricted model only.
Chapter 6

1. A change in the unit of measurement of the dependent variable in a model

does not lead to a change in:
a. the standard error of the regression.
b. the sum of squared residuals of the regression.
c. the goodness-of-fit of the regression.
d. the confidence intervals of the regression.

Answer: c
Difficulty:
Moderate
Bloom’s: Knowledge
A-Head: Effects of Data Scaling on OLS Statistics
BUSPROG:
Feedback: Changing the unit of measurement of the dependent variable in a model
does not lead to a change in the goodness of fit of the regression.

2. Changing the unit of measurement of any independent variable, where log of

the dependent variable appears in the regression:
a. affects only the intercept coefficient.
b. affects only the slope coefficient.
c. affects both the slope and intercept coefficients.
d. affects neither the slope nor the intercept coefficient.

Answer: a
Difficulty:
Moderate
Bloom’s: Comprehension
A-Head: Effects of Data Scaling on OLS Statistics
BUSPROG:

Downloaded by K61.FTU ?INH H?NG NG?C

Feedback: Changing the unit of measurement of any independent variable, where
log of the independent variable appears in the regression only affects the intercept.
This follows from the property log(ab) = log(a) + log(b).

3. A variable is standardized in the sample:

a. by multiplying by its mean.
b. by subtracting off its mean and multiplying by its standard deviation.
c. by subtracting off its mean and dividing by its standard deviation.
d. by multiplying by its standard deviation.

Answer: c
Difficulty:
Moderate
Bloom’s: Knowledge
A-Head: Effects of Data Scaling on OLS Statistics
BUSPROG:
Feedback: A variable is standardized in the sample by subtracting off its mean and
dividing by its standard deviation.

4. Standardized coefficients are also referred to as:

a. beta coefficients.
b. y coefficients.
c. alpha coefficients.
d. j coefficients.

Answer: a
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Effects of Data Scaling on OLS Statistics
BUSPROG:
Feedback: Standardized coefficients are also referred to as beta coefficients.

5. If a regression equation has only one explanatory variable, say x1, its
standardized coefficient must lie in the range:
a. -2 to 0.
b. -1 to 1.
c. 0 to 1.
d. 0 to 2.

Answer: b
Difficulty: Easy
Bloom’s: Comprehension
A-Head: Effects of Data Scaling on OLS Statistics
BUSPROG:

Downloaded by K61.FTU ?INH H?NG NG?C

Feedback: If a regression equation has only one explanatory variable, say x 1, its
standardized coefficient is the correlation coefficient between the dependent
variable and x1, and must lie in the range -1 to 1.

6. In the following equation, gdp refers to gross domestic product, and FDI refers
to foreign direct investment.

log(gdp) = 2.65 + 0.527log(bankcredit) + 0.222FDI

(0.13) (0.022) (0.017)

Which of the following statements is then true?

a. If gdp increases by 1%, bank credit increases by 0.527%, the level of
FDI remaining constant.
b. If bank credit increases by 1%, gdp increases by 0.527%, the level of
FDI remaining constant.
c. If gdp increases by 1%, bank credit increases by log(0.527)%, the level of
FDI remaining constant.
d. If bank credit increases by 1%, gdp increases by log(0.527)%, the level of
FDI remaining constant.

Answer: b
Difficulty:
Moderate
Bloom’s: Application
A-Head: More on Functional Form
BUSPROG:
Feedback: The equation suggests that if bank credit increases by 1%, gdp increases
by 0.527%. This is known from the value of the coefficient associated with bank
credit.

7. In the following equation, gdp refers to gross domestic product, and FDI refers
to foreign direct investment.
log(gdp) = 2.65 + 0.527log(bankcredit) + 0.222FDI
(0.13) (0.022) (0.017)

Which of the following statements is then true?

a. If FDI increases by 1%, gdp increases by approximately 22.2%, the amount
of bank credit remaining constant.
b. If FDI increases by 1%, gdp increases by approximately 26.5%, the amount
of bank credit remaining constant.
c. If FDI increases by 1%, gdp increases by approximately 24.8%, the amount
of bank credit remaining constant.
d. If FDI increases by 1%, gdp increases by approximately 52.7%, the amount
of bank credit remaining constant.

Downloaded by K61.FTU ?INH H?NG NG?C

Answer: c
Difficulty: Hard
Bloom’s: Application
A-Head: More on Functional Form
BUSPROG:
Feedback: The equation suggests that if FDI increases by 1%, gdp increases by
100(exp(0.222) – 1)%. This equals (1.24857 -1) = 24.8% approx.

8. Which of the following statements is true when the dependent variable, y > 0?
a. Taking log of a variable often expands its range.
b. Models using log(y) as the dependent variable will satisfy CLM assumptions
more closely than models using the level of y.
c. Taking log of variables make OLS estimates more sensitive to extreme values.
d. Taking logarithmic form of variables make the slope coefficients more
responsive to rescaling.

Answer: b
Difficulty:
Moderate
Bloom’s: Knowledge
A-Head: More on Functional Form
BUSPROG:
Feedback: Models using log(y) as the dependent variable will satisfy CLM
assumptions more closely than models using the level of y. This is because taking
log of a variable gets it closer to a normal distribution.

9. Which of the following correctly identifies a limitation of

logarithmic transformation of variables?
a. Taking log of variables make 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 log of a variable often expands its range which can cause
inefficient estimates.

Answer: b
Difficulty:
Moderate
Bloom’s: Comprehension
A-Head: More on Functional Form
BUSPROG:
Feedback: Logarithmic transformations cannot be used if a variable takes on zero or
negative values.

Downloaded by K61.FTU ?INH H?NG NG?C

10. Which of the following models is used quite often to capture decreasing
or increasing marginal effects of a variable?
a. Models with logarithmic functions
b. Models with quadratic functions
c. Models with variables in level
d. Models with interaction terms

Answer: b
Difficulty: Easy
Bloom’s: Knowledge
A-Head: More on Functional Form
BUSPROG:
Feedback: Models with quadratic functions are used quite often to capture
decreasing or increasing marginal effects of a variable

11. Which of the following correctly represents the equation for adjusted R2?
a. R 2
= 1 – [SSR/(n –1)]/[SST/(n+1)]
´
2
= 1 – [SSR/(n –k – 1)]/[SST/(n+1)]
b. R
2
= 1 – [SSR/(n –k – 1)]/[SST/(n – 1)]
´
2
= 1 – [SSR]/[SST/(n – 1)]
c. R
´

d. R
´

Answer: c
Difficulty: Easy
Bloom’s: Knowledge
A-Head: More on Goodness-of-Fit and Selection of Regressors
BUSPROG:
Feedback: R´ 2 = 1 – [SSR/(n –k – 1)]/[SST/(n – 1)]

12. Which of the following correctly identifies an advantage of using adjusted

R2 over R2?
a. Adjusted R2 corrects the bias in R2.
b. Adjusted R2 is easier to calculate than R2.
c. The penalty of adding new independent variables is better understood
through adjusted R2 than R2.
d. The adjusted R2 can be calculated for models having logarithmic functions
while R2 cannot be calculated for such models.

Answer: c
Downloaded by K61.FTU ?INH H?NG NG?C
Difficulty:
Moderate
Bloom’s: Knowledge

Downloaded by K61.FTU ?INH H?NG NG?C

A-Head: More on Goodness-of-Fit and Selection of Regressors
BUSPROG:
Feedback: The penalty of adding new independent variables is better understood
through adjusted R2 than R2 since its calculation is directly dependent on the
number of independent variables included.

13. Two equations form a nonnested model when:

a. one is logarithmic and the other is quadratic.
b. neither equation is a special case of the other.
c. each equation has the same independent variables.
d. there is only one independent variable in both equations.

Answer: b
Difficulty: Easy
Bloom’s: Knowledge
A-Head: More on Goodness-of-Fit and Selection of Regressors
BUSPROG:
Feedback: Two equations form a nonnested model when neither equation is a
special case of the other.

14. A predicted value of a dependent variable:

a. represents the difference between the expected value of the dependent
variable and its actual value.
b. is always equal to the actual value of the dependent variable.
c. is independent of explanatory variables and can be estimated on the basis of
the residual error term only.
d. represents the expected value of the dependent variable given particular
values for the explanatory variables.

Answer: d
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Prediction and Residual Analysis
BUSPROG:
Feedback: A predicted value of a dependent variable represents the expected value
of the dependent variable given particular values for the explanatory variables.

15. Residual analysis refers to the process of:

a. examining individual observations to see whether the actual value of
a dependent variable differs from the predicted value.
b. calculating the squared sum of residuals to draw inferences for the consistency
of estimates.

Downloaded by K61.FTU ?INH H?NG NG?C

c. transforming models with variables in level to logarithmic functions so as to
understand the effect of percentage changes in the independent variable on
the dependent variable.
d. sampling and collection of data in such a way to minimize the squared sum
of residuals.

Answer: a
Difficulty:
Moderate
Bloom’s: Knowledge
A-Head: Prediction and Residual Analysis
BUSPROG:
Feedback: Residual analysis refers to the process of examining individual
observations to see whether the actual value of a dependent variable differs from
the predicted value.

16. Beta coefficients are always greater than standardized coefficients.

Answer: False
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Effects of Data Scaling on OLS Statistics
BUSPROG:
Feedback: Beta coefficients the same as standardized coefficients.

17. If a new independent variable is added to a regression equation, the adjusted

R2 increases only if the absolute value of the t statistic of the new variable is
greater than one.

Answer: True
Difficulty: Easy
Bloom’s: Knowledge
A-Head: More on Goodness-of-Fit and Selection of Regressors
BUSPROG:
Feedback: If a new independent variable is added to a regression equation, the
adjusted R2 increases only if the absolute value of the t statistic of the new variable
is greater than one in absolute value.

18. F statistic can be used to test nonnested models.

Answer: False
Difficulty: Easy
Bloom’s: Knowledge
A-Head: More on Goodness-of-Fit and Selection of Regressors
BUSPROG:

Downloaded by K61.FTU ?INH H?NG NG?C

Feedback: F statistic can be used only to test nested models.

19. Predictions of a dependent variable are subject to sampling variation.

Answer: True
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Prediction and Residual Analysis
BUSPROG:
Feedback: Predictions of a dependent variable are subject to sampling variation
since they are obtained using OLS estimators.

20. To make predictions of logarithmic dependent variables, they first have to

be converted to their level forms.

Answer: False
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Prediction and Residual Analysis
BUSPROG:
Feedback: It is possible to make predictions of dependent variables when they are in
their logarithmic form. It is not necessary to convert them into their level forms.
Chapter 7

1. A variable is used to incorporate qualitative information in a

regression model.
a. dependent
b. continuous
c. binomial
d. dummy

Answer: d
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Describing Qualitative Information
BUSPROG:
Feedback: A dummy variable or binary variable is used to incorporate qualitative
information in a regression model.

2. In a regression model, which of the following will be described using a

binary variable?
a. Whether it rained on a particular day or it did not
b. The volume of rainfall during a year
c. The percentage of humidity in air on a particular day

Downloaded by K61.FTU ?INH H?NG NG?C

d. The concentration of dust particles in air

Answer: a
Difficulty: Medium
Bloom’s: Comprehension
A-Head: Describing Qualitative Information
BUSPROG:
Feedback: A binary variable is used to describe qualitative information in regression
model. Therefore, such a variable will be used to describe whether it rained on a
particular day or it did not.

3. Which of the following is true of dummy variables?

a. A dummy variable always takes a value less than 1.
b. A dummy variable always takes a value higher than 1.
c. A dummy variable takes a value of 0 or 1.
d. A dummy variable takes a value of 1 or 10.

Answer: c
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Describing Qualitative Information
BUSPROG:
Feedback: A dummy variable takes a value of 0 or 1.

The following simple model is used to determine the annual savings of an individual
on the basis of his annual income and education.
Savings = β0+∂0 Edu + β1Inc+u
The variable ‘Edu’ takes a value of 1 if the person is educated and the variable ‘Inc’
measures the income of the individual.

4. Refer to the model above. The inclusion of another binary variable in this
model that takes a value of 1 if a person is uneducated, will give rise to the
problem of
.

a. omitted variable bias

b. self-selection
c. dummy variable trap
d. heteroskedastcity

Answer: c
Difficulty: Medium
Bloom’s: Application
A-Head: Describing Qualitative Information
BUSPROG: Analytic

Downloaded by K61.FTU ?INH H?NG NG?C

Feedback: The inclusion of another dummy variable in this model would introduce
perfect collinearity and lead to a dummy variable trap.

The following simple model is used to determine the annual savings of an individual
on the basis of his annual income and education.
Savings = β0+∂0 Edu + β1Inc+u
The variable ‘Edu’ takes a value of 1 if the person is educated and the variable ‘Inc’
measures the income of the individual.

5. Refer to the model above. The benchmark group in this model is .

a. the group of educated people
b. the group of uneducated people
c. the group of individuals with a high income
d. the group of individuals with a low income

Answer: b
Difficulty:
Moderate
Bloom’s: Application
A-Head: A Single Dummy Independent Variable
BUSPROG: Analytic
Feedback: The benchmark group is the group against which comparisons are made.
In this case, the savings of a literate person is being compared to the savings of an
illiterate person; therefore, the group of illiterate people is the base group or
benchmark group.

The following simple model is used to determine the annual savings of an individual
on the basis of his annual income and education.
Savings = β0+∂0 Edu + β1Inc+u
The variable ‘Edu’ takes a value of 1 if the person is educated and the variable ‘Inc’
measures the income of the individual.

6. Refer to the above model. If ∂0 > 0, .

a. uneducated people have higher savings than those who are educated
b. educated people have higher savings than those who are not educated
c. individuals with lower income have higher savings
d. individual with lower income have higher savings

Answer: b
Difficulty:
Moderate
Bloom’s: Application
A-Head: A Single Dummy Independent Variable
BUSPROG: Analytic

Downloaded by K61.FTU ?INH H?NG NG?C

Feedback: The coefficient ∂0 measures the impact of education on an individual’s
annual savings. If it has a positive impact, as in this case, educated people should
have higher savings.

7. The income of an individual in Budopia depends on his ethnicity and several

other factors which can be measured quantitatively. If there are 5 ethnic groups in
Budopia, how many dummy variables should be included in the regression equation
for income determination in Budopia?
a. 1
b. 5
c. 6
d. 4

Answer: d
Difficulty:
Moderate
Bloom’s: Application
A-Head: Using Dummy Variables for Multiple Categories
BUSPROG: Analytic
Feedback: If a regression model is to have different intercepts for, say, g groups or
categories, we need to include g -1 dummy variables in the model along with an
intercept. In this case, the regression equation should include 5-1=4 dummy
variables since there are 5 ethnic groups.

8. The quarterly increase in an employee’s salary depends on the rating of his

work by his employer and several other factors as shown in the model below:
Increase in salary= β0+∂0Rating + other factors. The variable ‘Rating’ is a(n)
variable.
a. dependent variable
b. ordinal variable
c. continuous variable
d. Poisson variable

Answer: b
Difficulty:
Moderate
Bloom’s: Application
A-Head: Using Dummy Variables for Multiple Categories
BUSPROG: Analytic
Feedback: The value of the variable ‘Rating’ depends on the employer’s rating
of the worker. Therefore, it incorporates ordinal information and is called an
ordinal variable.

9. Which of the following is true of Chow test?

a. It is a type of t test.
b. It is a type of sign test.

Downloaded by K61.FTU ?INH H?NG NG?C

c. It is only valid under homoskedasticty.
d. It is only valid under heteroskedasticity.

Answer: c
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Interactions Involving Dummy Variables
BUSPROG:
Feedback: Since the Chow test is just an F test, it is only valid under
homoskedasticity.

10. Which of the following is true of dependent variables?

a. A dependent variable can only have a numerical value.
b. A dependent variable cannot have more than 2 values.
c. A dependent variable can be binary.
d. A dependent variable cannot have a qualitative meaning.

Answer: c
Difficulty: Easy
Bloom’s: Knowledge
A-Head: A Binary Dependent Variable: The Linear Probability Model
BUSPROG:
Feedback: A dependent variable is binary if it has a qualitative meaning.

11. In the following regression equation, y is a binary variable:

y= β0+β1x1+…βk xk+ u
In this case, the estimated slope ^β
1 measures .
coefficient,
a. the predicted change in the value of y when x1 increases by one unit, everything
else remaining constant
b. the predicted change in the value of y when x1 decreases by one unit, everything
else remaining constant
c. the predicted change in the probability of success when x1 decreases by one
unit, everything else remaining constant
d. the predicted change in the probability of success when x1 increases by one
unit, everything else remaining constant

Answer: d
Difficulty: Easy
Bloom’s: Knowledge
A-Head: A Binary Dependent Variable: The Linear Probability Model
BUSPROG:
Feedback: A binary dependent variable is used when a regression model is used to
explain a qualitative event. The dependent variable takes a value of 1 when the

Downloaded by K61.FTU ?INH H?NG NG?C

event takes place (success) and it takes a value of zero when the event does not
take place. The coefficient of an independent variable in this case measures the
predicted change in the probability of success when the independent variable
increases by one unit.

12. Consider the following regression equation: y = β 0+β1x1+…βk xk+

u In which of the following cases, the dependent variable is binary?
a. y indicates the gross domestic product of a country
b. y indicates whether an adult is a college dropout
c. y indicates household consumption expenditure
d. y indicates the number of children in a family

Answer: b
Difficulty: Easy
Bloom’s: Application
A-Head: A Binary Dependent Variable: The Linear Probability Model
BUSPROG: Analytic
Feedback: The dependent variable, y is binary if it is used to indicate a qualitative
outcome.

13. Which of the following Gauss-Markov assumptions is violated by the

linear probability model?
a. The assumption of constant variance of the error term.
b. The assumption of zero conditional mean of the error term.
c. The assumption of no exact linear relationship among independent variables.
d. The assumption that none of the independent variables are constants.

Answer: a
Difficulty: Easy
Bloom’s: Knowledge
A-Head: A Binary Dependent Variable: The Linear Probability Model
BUSPROG:
Feedback: The linear probability model violates the assumption of constant variance
of the error term.

14. Which of the following problems can arise in policy analysis and
program evaluation using a multiple linear regression model?
a. There exists homoscedasticity in the model.
b. The model can produce predicted probabilities that are less than zero and
greater than one.
c. The model leads to the omitted variable bias as only two independent factors
can be included in the model.
d. The model leads to an overestimation of the effect of independent variables
on the dependent variable.

Downloaded by K61.FTU ?INH H?NG NG?C

Answer: b
Difficulty: Easy
Bloom’s: Knowledge
A-Head: More on Policy Analysis and Program Evaluation
BUSPROG:
Feedback: The model can produce predicted probabilities that are less than zero and
greater than one.

15. Consider the following regression equation: y = β 0+β1x1+…βk xk+

u In which of the following cases, is ‘y’ a discrete variable?
a. y indicates the gross domestic product of a country
b. y indicates the total volume of rainfall during a year
c. y indicates household consumption expenditure
d. y indicates the number of children in a family

Answer: d
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Interpreting Regression Results with Discrete Dependent Variables
BUSPROG:
Feedback: The number of children in a family can only take a small set of integer
values. Therefore, y is a discrete variable if it measures the number of children in a
family.

16. A binary variable is a variable whose value changes with a change in

the number of observations.

Answer: False
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Describing Qualitative Information
BUSPROG:
Feedback: A binary variable is one whose value depends on the event taking place.

17. A dummy variable trap arises when a single dummy variable describes a
given number of groups.

Answer: False
Difficulty: Easy
Bloom’s: Knowledge
A-Head: A Single Dummy Independent Variable
BUSPROG:

Downloaded by K61.FTU ?INH H?NG NG?C

Feedback: A dummy variable trap arises when too many dummy variables describe
a given number of groups.

18. The dummy variable coefficient for a particular group represents the
estimated difference in intercepts between that group and the base group.

Answer: True
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Using Dummy Variables for Multiple Categories
BUSPROG:
Feedback: The dummy variable coefficient for a particular group represents the
estimated difference in intercepts between that group and the base group.

19. The multiple linear regression model with a binary dependent variable is
called the linear probability model.

Answer: True
Difficulty: Easy
Bloom’s: Knowledge
A-Head: A Binary Dependent Variable: The Linear Probability Model
BUSPROG:
Feedback: The multiple linear regression model with a binary dependent variable is
called the linear probability model.

20. A problem that often arises in policy and program evaluation is that
individuals (or firms or cities) choose whether or not to participate in certain
behaviors or programs.

Answer: True
Difficulty: Easy
Bloom’s: Knowledge
A-Head: More on Policy Analysis and Program Evaluation
BUSPROG:
Feedback: A problem that often arises in policy and program evaluation is that
individuals (or firms or cities) choose whether or not to participate in certain
behaviors or programs and their choice depends on several other factors. It is not
possible to control for these factors while examining the effect of the programs.

Chapter 8

1. Which of the following is true of heteroskedasticity?

a. Heteroskedasticty causes inconsistency in the Ordinary Least Squares estimators.
b. Population R2 is affected by the presence of heteroskedasticty.

Downloaded by K61.FTU ?INH H?NG NG?C

c. The Ordinary Least Square estimators are not the best linear unbiased
estimators if heteroskedasticity is present.
d. It is not possible to obtain F statistics that are robust to heteroskedasticity of
an unknown form.

Answer: c
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Consequences of Heteroskedasticity for OLS
BUSPROG:
Feedback: The Ordinary Least Square estimators are no longer the best linear
unbiased estimators if heteroskedasticity is present in a regression model.

2. Consider the following regression model: yi=β0+β1 xi+ui. If the first four Gauss-
Markov assumptions hold true, and the error term contains heteroskedasticity,
then
.
a. Var(ui|xi) =0
b. Var(ui|xi) =1
c. Var(ui|xi) = σ i 2
d. Var(ui|xi) =σ

Answer: c
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Heteroskedasticity-Robust Inference after OLS Estimation
BUSPROG:
Feedback: If the first four Gauss-Markov assumptions hold and the error term
contains heteroskedasticity, then Var(ui|xi) = σi2.

3. The general form of the t statistic is .

estimate−hypothesized value
t= standard error
a

hypothesized value−estimate
t= standard error
b

standard error
t = estimate−hypothesized value
c

d. t =estimate−hypothesized value

Answer: a
Difficulty: Easy
Bloom’s: Knowledge

Downloaded by K61.FTU ?INH H?NG NG?C

A-Head: Heteroskedasticity-Robust Inference after OLS Estimation
BUSPROG:
estimate−hypothesized value
t=
Feedback: The general form of the t statistic is .
standard

4. Which of the following is true of the OLS t statistics?

a. The heteroskedasticity-robust t statistics are justified only if the sample size
is large.
b. The heteroskedasticty-robust t statistics are justified only if the sample size
is small.
c. The usual t statistics do not have exact t distributions if the sample size is large.
d. In the presence of homoscedasticity, the usual t statistics do not have exact t
distributions if the sample size is small.

Answer: a
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Heteroskedasticity-Robust Inference after OLS Estimation
BUSPROG:
Feedback: The heteroskedasticity-robust t statistics are justified only if the sample
size is large.

5. The heteroskedasticity-robust is also called the heteroskedastcity-

robust Wald statistic.
a. t statistic
b. F statistic
c. LM statistic
d. z statistic

Answer: b
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Heteroskedasticity-Robust Inference after OLS Estimation
BUSPROG:
Feedback: The heteroskedasticity-robust F statistic is also called the
heteroskedastcity-robust Wald statistic.

6. Which of the following tests helps in the detection of heteroskedasticity?

a. The Breusch-Pagan test
b. The Breusch-Godfrey test
c. The Durbin-Watson test
d. The Chow test

Downloaded by K61.FTU ?INH H?NG NG?C

Answer: a
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Testing for Heteroskedasticity
BUSPROG:
Feedback: The Breusch-Pagan test is used for the detection of heteroskedasticity in
a regression model.

7. What will you conclude about a regression model if the Breusch-Pagan test
results in a small p-value?
a. The model contains homoskedasticty.
b. The model contains heteroskedasticty.
c. The model contains dummy variables.
d. The model omits some important explanatory factors.

Answer: b
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Testing for Heteroskedasticity
BUSPROG:
Feedback: The Breusch-Pagan test results in a small p-value if the regression model
contains heteroskedasticty.

8. A test for heteroskedasticty can be significant if .

a. the Breusch-Pagan test results in a large p-value
b. the White test results in a large p-value
c. the functional form of the regression model is misspecified
d. the regression model includes too many independent variables

Answer: c
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Testing for Heteroskedasticity
BUSPROG:
Feedback: A test for heteroskedasticty can be significant if the functional form of the
regression model is misspecified.

9. Which of the following is a difference between the White test and the
Breusch- Pagan test?
a. The White test is used for detecting heteroskedasticty in a linear
regression model while the Breusch-Pagan test is used for detecting
autocorrelation.
b. The White test is used for detecting autocorrelation in a linear regression
model while the Breusch-Pagan test is used for detecting heteroskedasticity. .

Downloaded by K61.FTU ?INH H?NG NG?C

c. The number of regressors used in the White test is larger than the number
of regressors used in the Breusch-Pagan test.
d. The number of regressors used in the Breusch-Pagan test is larger than
the number of regressors used in the White test.

Answer: c
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Testing for Heteroskedasticity
BUSPROG:
Feedback: The White test includes the squares and cross products of all
independent variables. Therefore, the number of regressors is larger for the White
test.

10. Which of the following is true of the White test?

a. The White test is used to detect the presence of multicollinearity in a
linear regression model.
b. The White test cannot detect forms of heteroskedasticity that invalidate the
usual Ordinary Least Squares standard errors.
c. The White test can detect the presence of heteroskedasticty in a linear
regression model even if the functional form is misspecified.
d. The White test assumes that the square of the error term in a regression model
is uncorrelated with all the independent variables, their squares and cross
products.

Answer: d
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Testing for Heteroskedasticity
BUSPROG:
Feedback: The White test assumes that the square of the error term in a regression
model is uncorrelated with all the independent variables, the squares of
independent variables and all the cross products.

11. Which of the following is true?

a. In ordinary least squares estimation, each observation is given a different weight.
b. In weighted least squares estimation, each observation is given an
identical weight.
c. In weighted least squares estimation, less weight is given to observations with
a higher error variance.
d. In ordinary least squares estimation, less weight is given to observations with
a lower error variance.

Answer: c
Difficulty: Easy

Downloaded by K61.FTU ?INH H?NG NG?C

Bloom’s: Knowledge
A-Head: Weighted Least Squares Estimation
BUSPROG:
Feedback: In weighted Least Squares estimation, less weight is given to
observations with a higher error variance.

12. Weighted least squares estimation is used only when .

a. the dependent variable in a regression model is binary
b. the independent variables in a regression model are correlated
c. the error term in a regression model has a constant variance
d. the functional form of the error variances is known

Answer: d
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Weighted Least Squares Estimation
BUSPROG:
Feedback: Weighted Least Squares estimation is used only when the functional form
of the error variances is known.

y=β 0 + β 1 x1 +u
13. Consider the following regression equation: . Which of
the following indicates a functional form misspecification in E(y|x)?
a. Ordinary Least Squares estimates equal Weighted Least Squares estimates.
b. Ordinary Least Squares estimates exceed Weighted Least Squares estimates by
a small magnitude.
c. Weighted Least Squares estimates exceed Ordinary Least Squares estimates by
a small magnitude.
d. Ordinary Least Square estimates are positive while Weighted Least
Squares estimates are negative.

Answer: d
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Weighted Least Squares Estimation
BUSPROG:
Feedback: If Ordinary Least Square estimates are positive while Weighted Least
Squares estimates are negative, the functional form of a regression equation is said
to be misspecified.

14. Which of the following tests is used to compare the Ordinary Least
Squares (OLS) estimates and the Weighted Least Squares (WLS) estimates?
a. The White test
b. The Hausman test

Downloaded by K61.FTU ?INH H?NG NG?C

c. The Durbin-Watson test
d. The Breusch-Godfrey test

Answer: b
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Weighted Least Squares Estimation
BUSPROG:
Feedback: The Hausman test can be used to formally compare the OLS and WLS
estimates to see if they differ by more than sampling error suggests they should.

15. The linear probability model contains heteroskedasticity unless .

a. the intercept parameter is zero
b. all the slope parameters are positive
c. all the slope parameters are zero
d. the independent variables are binary

Answer: c
Difficulty: Easy
Bloom’s: Knowledge
A-Head: The Linear Probability Model Revisited
BUSPROG:
Feedback: The linear probability model contains heteroskedasticity unless all the
slope parameters are zero.

16. The interpretation of goodness-of-fit measures changes in the presence

of heteroskedasticity.

Answer: False
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Consequences of Heteroskedasticity for OLS
BUSPROG:
Feedback: The interpretation of goodness-of-fit measures is unaffected by the
presence of heteroskedasticty.

17. Multicollinearity among the independent variables in a linear regression

model causes the heteroskedasticity-robust standard errors to be large.

Answer: True
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Heteroskedasticity-Robust Inference after OLS Estimation
BUSPROG:

Downloaded by K61.FTU ?INH H?NG NG?C

Feedback: Multicollinearity among the independent variables in a linear regression
model causes the heteroskedasticity-robust standard errors to be large.
18. If the Breusch-Pagan Test for heteroskedasticity results in a large p-value,
the null hypothesis of homoskedasticty is rejected.

Answer: False
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Testing for Heteroskedasticity
BUSPROG:
Feedback: If the Breusch-Pagan Test for heteroskedasticity results in a large p-value,
the null hypothesis of heteroskedasticty is rejected.

19. The generalized least square estimators for correcting heteroskedasticity

are called weighed least squares estimators.

Answer: True
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Weighted Least Squares Estimation
BUSPROG:
Feedback: The generalized least square estimators for correcting heteroskedasticity
are called weighed least squares estimators.

20. The linear probability model always contains heteroskedasticity when the
dependent variable is a binary variable unless all of the slope parameters are
zero.

Answer: True
Difficulty: Easy
Bloom’s: Knowledge
A-Head: The Linear Probability Model Revisited
BUSPROG:
Feedback: The linear probability model always contains heteroskedasticity when the
dependent variable is a binary variable unless all of the slope parameters are zero.
Chapter 9

1. Consider the following regression model: log(y) = β0 + β1x1 + β 2x 2

+ β3x3 + u.
This model will suffer from functional form misspecification if .
a. β0 is omitted from the model
b. u is heteroskedastic
c. x 21 is omitted from the model
d. x3 is a binary variable

Answer: c

Downloaded by K61.FTU ?INH H?NG NG?C

Difficulty: Easy
Bloom’s: Comprehension
A-Head: Functional Form Misspecification
BUSPROG:
Feedback: The model suffers from functional form misspecification if x 12 is omitted
from the model since it is a function of x1 which is an observed explanatory variable.

2. A regression model suffers from functional form misspecification if .

a. a key variable is binary.
b. the dependent variable is binary.
c. an interaction term is omitted.
d. the coefficient of a key variable is zero.

Answer: c
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Functional Form Misspecification
BUSPROG:
Feedback: A regression model suffers from functional form misspecification if an
interaction term is omitted.

3. Which of the following is true?

a. A functional form misspecification can occur if the level of a variable is used
when the logarithm is more appropriate.
b. A functional form misspecification occurs only if a key variable is
uncorrelated with the error term. .
c. A functional form misspecification does not lead to biasedness in the
ordinary least squares estimators.
d. A functional form misspecification does not lead to inconsistency in the
ordinary least squares estimators.

Answer: a
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Functional Form Misspecification
BUSPROG:
Feedback: A functional form misspecification can occur if the level of a variable is
used when the logarithm is more appropriate.

4. Which of the following is true of Regression Specification Error Test (RESET)?

a. It tests if the functional form of a regression model is misspecified.
b. It detects the presence of dummy variables in a regression model.
c. It helps in the detection of heteroskedasticity when the functional form of
the model is correctly specified.

Downloaded by K61.FTU ?INH H?NG NG?C

d. It helps in the detection of multicollinearity among the independent variables in
a regression model.

Answer: a
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Functional Form Misspecification
BUSPROG:
Feedback: It tests if the functional form of a regression model is misspecified.

5. A proxy variable .
a. increases the error variance of a regression model
b. cannot contain binary information
c. is used when data on a key independent variable is unavailable
d. is detected by running the Davidson-MacKinnon test

Answer: c
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Using Proxy Variables for Unobserved Explanatory Variables
BUSPROG:
Feedback: A proxy variable is used when data on a key independent variable is
unavailable.

6. Which of the following assumptions is needed for the plug-in solution to

the omitted variables problem to provide consistent estimators?
a. The error term in the regression model exhibits heteroskedasticity.
b. The error term in the regression model is uncorrelated with all the
independent variables.
c. The proxy variable is uncorrelated with the dependent variable.
d. The proxy variable has zero conditional mean.

Answer: b
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Using Proxy Variables for Unobserved Explanatory Variables
BUSPROG:
Feedback: The error term in the regression model is uncorrelated with the proxy
variable.

7. Which of the following is a drawback of including proxy variables in a

regression model?
a. It leads to misspecification analysis.
b. It reduces the error variance.

Downloaded by K61.FTU ?INH H?NG NG?C

c. It increases the error variance.
d. It exacerbates multicollinearity.

Answer: d
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Using Proxy Variables for Unobserved Explanatory Variables
BUSPROG:
Feedback: The inclusion of a proxy variable in a regression model exacerbates
multicollinearity.

8. Consider the following equation for household consumption expenditure:

Consmptn= β0+ β1Inc + β2Consmptn-1+ u
where ‘Consmptn’ measures the monthly consumption expenditure of a
household, ‘Inc’ measures household income and ‘Consmptn -1’ is the consumption
expenditure in the previous month. Consmptn-1 is a variable.
a. exogenous
b. binary variable
c. lagged dependent
d. proxy variable

Answer: c
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Using Proxy Variables for Unobserved Explanatory Variables
BUSPROG:
Feedback: ‘Consmptn-1’ is a lagged dependent variable in this model.

9. A measurement error occurs in a regression model when .

a. the observed value of a variable used in the model differs from its actual value
b. the dependent variable is binary
c. the partial effect of an independent variable depends on unobserved factors
d. the model includes more than two independent variables

Answer: a
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Properties of OLS under Measurement Error
BUSPROG:
Feedback: A measurement error occurs in a regression model when the observed
value of a variable used in the model differs from its actual value.

10. The classical errors-in-variables (CEV) assumption is that .

Downloaded by K61.FTU ?INH H?NG NG?C

a. the error term in a regression model is correlated with all observed
explanatory variables
b. the error term in a regression model is uncorrelated with all observed
explanatory variables
c. the measurement error is correlated with the unobserved explanatory variable
d. the measurement error is uncorrelated with the unobserved explanatory variable

Answer: d
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Properties of OLS under Measurement Error
BUSPROG:
Feedback: The classical errors-in-variables (CEV) assumption is that the
measurement error is uncorrelated with the unobserved explanatory variable.

11. Which of the following is true of measurement error?

a. If measurement error in a dependent variable has zero mean, the ordinary
least squares estimators for the intercept are biased and inconsistent.
b. If measurement error in an independent variable is uncorrelated with
the variable, the ordinary least squares estimators are unbiased.
c. If measurement error in an independent variable is uncorrelated with
other independent variables, all estimators are biased.
d. If measurement error in a dependent variable is correlated with the
independent variables, the ordinary least squares estimators are unbiased.

Answer: b
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Properties of OLS under Measurement Error
BUSPROG:
Feedback: If measurement error in an independent variable is uncorrelated with the
variable, the ordinary least squares estimators are unbiased.

12. Sample selection based on the dependent variable is called .

a. random sample selection
b. endogenous sample selection
c. exogenous sample selection
d. stratified sample selection

Answer: b
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Missing Data, Nonrandom Samples, and Outlying Observations
BUSPROG:

Downloaded by K61.FTU ?INH H?NG NG?C

Feedback: Sample selection based on the dependent variable is called endogenous
sample selection.

13. The method of data collection in which the population is divided

into nonoverlapping, exhaustive groups is called .
a. random sampling
b. stratified sampling
c. endogenous sampling
d. exogenous sampling

Answer: b
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Missing Data, Nonrandom Samples, and Outlying Observations
BUSPROG:
Feedback: The method of data collection in which the population is divided into
nonoverlapping, exhaustive groups is called stratified sampling.

14. Which of the following types of sampling always causes bias or inconsistency
in the ordinary least squares estimators?
a. Random sampling
b. Exogenous sampling
c. Endogenous sampling
d. Stratified sampling

Answer: c
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Missing Data, Nonrandom Samples, and Outlying Observations
BUSPROG:
Feedback: Endogenous sampling always causes bias in the OLS estimators. If the
sample is based on whether the dependent variable is above or below a given
value, bias always occurs in OLS in estimating the population model.

15. Which of the following is a difference between least absolute deviations

(LAD) and ordinary least squares (OLS) estimation?
a. OLS is more computationally intensive than LAD.
b. OLS is more sensitive to outlying observations than LAD.
c. OLS is justified for very large sample sizes while LAD is justified for
smaller sample sizes.
d. OLS is designed to estimate the conditional median of the dependent
variable while LAD is designed to estimate the conditional mean.

Answer: b

Downloaded by K61.FTU ?INH H?NG NG?C

Difficulty: Easy
Bloom’s: Knowledge
A-Head: Least Absolute Deviations Estimation
BUSPROG:
Feedback: OLS is more sensitive to outlying observations than LAD.

16. An explanatory variable is called exogenous if it is correlated with the

error term.

Answer: False
Difficulty: Easy
Bloom’s: Knowledge
A-Head:
BUSPROG:
Feedback: An explanatory variable is called endogenous if it is correlated with the
error term.

17. A multiple regression model suffers from functional form misspecification

when it does not properly account for the relationship between the dependent and
the observed explanatory
variables.

Answer: True
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Functional Form Misspecification
BUSPROG:
Feedback: A multiple regression model suffers from functional form misspecification
when it does not
properly account for the relationship between the dependent and the observed
explanatory
variables.

18. The measurement error is the difference between the actual value of a
variable and its reported value.

Answer: True
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Properties of OLS under Measurement Error
BUSPROG:
Feedback: The measurement error is the difference between the actual value of a
variable and its reported value.

Downloaded by K61.FTU ?INH H?NG NG?C

19. Studentized residuals are obtained from the original OLS residuals by
dividing them by an estimate of their standard deviation.

Answer: True
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Missing Data, Nonrandom Samples, and Outlying Observations
BUSPROG:
Feedback: Studentized residuals are obtained from the original OLS residuals by
dividing them by an estimate of their standard deviation.

20. The Least Absolute Deviations (LAD) estimators in a linear model minimize
the sum of squared residuals.

Answer: False
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Least Absolute Deviations Estimation
BUSPROG:
Feedback: The Least Absolute Deviations (LAD) estimators in a linear model
minimize the sum of the absolute values of the residuals.

Chapter 10

1. Which of the following correctly identifies a difference between cross-

sectional data and time series data?
a. Cross-sectional data is based on temporal ordering, whereas time series data
is not.
b. Time series data is based on temporal ordering, whereas cross-sectional data
is not.
c. Cross-sectional data consists of only qualitative variables, whereas time
series data consists of only quantitative variables.
d. Time series data consists of only qualitative variables, whereas cross-
sectional data does not include qualitative variables.

Answer: b
Difficulty: Easy
Bloom’s: Knowledge
A-Head: The Nature of Time Series Data
BUSPROG:
Feedback: Time series data is based on temporal ordering, whereas cross sectional
data is not.

2. A stochastic process refers to a:

Downloaded by K61.FTU ?INH H?NG NG?C

a. sequence of random variables indexed by time.
b. sequence of variables that can take fixed qualitative values.
c. sequence of random variables that can take binary values only.
d. sequence of random variables estimated at the same point of time.

Answer: a
Difficulty: Easy
Bloom’s: Knowledge
A-Head: The Nature of Time Series Data
BUSPROG:
Feedback: A stochastic process refers to a sequence of random variables indexed by
time.

3. The sample size for a time series data set is the number of:
a. variables being measured.
b. time periods over which we observe the variables of interest less the number
of variables being measured.
c. time periods over which we observe the variables of interest plus the number
of variables being measured.
d. time periods over which we observe the variables of interest.

Answer: d
Difficulty: Easy
Bloom’s: Knowledge
A-Head: The Nature of Time Series Data
BUSPROG:
Feedback: The sample size for a time series data set is the number of time periods
over which we observe the variables of interest.

4. The model: Yt = β0 + β1ct + ut, t = 1,2,…….n, is an example of a(n):

a. autoregressive conditional heteroskedasticity model.
b. static model.
c. finite distributed lag model.
d. infinite distributed lag model.

Answer: b
Difficulty: Medium
Bloom’s: Comprehension
A-Head: Examples of Time Series Regression Models
BUSPROG:
Feedback: The model: yt = β0 + β1ct + ut, t = 1,2,…….,n, is an example of a static
model.

5. A static model is postulated when:

Downloaded by K61.FTU ?INH H?NG NG?C

a. a change in the independent variable at time ‘t’ is believed to have an effect
on the dependent variable at period ‘t + 1’.
b. a change in the independent variable at time ‘t’ is believed to have an effect
on the dependent variable for all successive time periods.
c. a change in the independent variable at time ‘t’ does not have any effect on
the dependent variable.
d. a change in the independent variable at time ‘t’ is believed to have an
immediate effect on the dependent variable.

Answer: d
Difficulty:
Moderate
Bloom’s: Knowledge
A-Head: Examples of Time Series Regression Models
BUSPROG:
Feedback: A static model is postulated when a change in the independent variable
at time ‘t’ is believed to have an immediate effect on the dependent variable.

6. Refer to the following model.

yt = α0 + β0st + β1st-1 + β2st-2 + β3st-3 + ut
This is an example of a(n):
a. infinite distributed lag model.
b. finite distributed lag model of order 1.
c. finite distributed lag model of order 2.
d. finite distributed lag model of order 3.

Answer: d
Difficulty:
Moderate
Bloom’s: Comprehension
A-Head: Examples of Time Series Regression Models
BUSPROG:
Feedback: The model: yt = α0 + β0st + β1st-1 + β2st-2 + β3st-3 + ut, is an example of
a finite distributed lag model of order 3.

7. Refer to the following model.

yt = α0 + β0st + β1st-1 + β2st-2 + β3st-3 + ut
β0 + β1 + β2 + β3 represents:
a. the short-run change in y given a temporary increase in s.
b. the short-run change in y given a permanent increase in s.
c. the long-run change in y given a permanent increase in s.
d. the long-run change in y given a temporary increase in s.

Answer: c
Difficulty:
Moderate
Bloom’s: Comprehension

Downloaded by K61.FTU ?INH H?NG NG?C

A-Head: Examples of Time Series Regression Models
BUSPROG:
Feedback: In the model, the sum of the coefficients on current and lagged z, β 0 + β1
+ β2 + β3 represents the long-run change in y given a permanent change in s.

8. Which of the following is an assumption on which time series regression is based?

a. A time series process follows a model that is nonlinear in parameters.
b. In a time series process, no independent variable is a perfect linear
combination of the others.
c. In a time series process, at least one independent variable is a constant.
d. For each time period, the expected value of the error ut, given the
explanatory variables for all time periods, is positive.

Answer: b
Difficulty:
Moderate
Bloom’s: Knowledge
A-Head: Finite Sample Properties of OLS under Classical Assumptions
BUSPROG:
Feedback: Time series regression is based on the assumption that no independent
variables are constant.

9. Under the assumptions of time series regression, which of the

following statements will be true of the following model: yt = α0 + α1dt +
ut?
a. d can have a lagged effect on y.
b. ut can be correlated with past and future values of d.
c. Changes in the error term cannot cause future changes in d.
d. Changes in d cannot cause changes in y at the same point of time.

Answer: c
Difficulty:
Moderate
Bloom’s: Knowledge
A-Head: Finite Sample Properties of OLS under Classical Assumptions
BUSPROG:
Feedback: Under the assumptions of time series regression, changes in the error
term cannot cause future changes in d, in the given model.

10. If an explanatory variable is strictly exogenous it implies that:

a. changes in the lag of the variable does not affect future values of the
dependent variable.
b. the variable is correlated with the error term in all future time periods.
c. the variable cannot react to what has happened to the dependent variable in
the past.
d. the conditional mean of the error term given the variable is zero.

Downloaded by K61.FTU ?INH H?NG NG?C

Answer: c
Difficulty:
Moderate
Bloom’s: Knowledge
A-Head: Finite Sample Properties of OLS under Classical Assumptions
BUSPROG:
Feedback: If an explanatory variable is strictly exogenous it implies that the variable
cannot react to what has happened to the dependent variable in the past.

11. A study which observes whether a particular occurrence influences

some outcome is referred to as a(n):
a. event study.
b. exponential study.
c. laboratory study.
d. comparative study.

Answer: a
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Functional Form, Dummy Variables, and Index Numbers.
BUSPROG:
Feedback: A study which observes whether a particular occurrence influences some
outcome is referred to as an event study.

12. With base year 1990, the index of industrial production for the year 1999 is
112. What will be the value of the index in 1999, if the base year is changed to
1982 and the index measured 96 in 1982?
a. 112.24
b. 116.66
c. 85.71
d. 92.09

Answer: b
Difficulty:
Moderate Bloom’s:
Apply
A-Head: Functional Form, Dummy Variables, and Index Numbers.
BUSPROG: Analytic
Feedback: If the base year is changed to 1982, the new index of industrial
production for 1999 will equal 100(112/96) = 116.67.

13. Which of the following statements is true?

a. The average of an exponential time series is a linear function of time.
b. The average of a linear sequence is an exponential function of time.
c. When a series has the same average growth rate from period to period, it can
be approximated with an exponential trend.

Downloaded by K61.FTU ?INH H?NG NG?C

d. When a series has the same average growth rate from period to period, it can
be approximated with a linear trend.

Answer: c
Difficulty:
Moderate
Bloom’s: Knowledge
A-Head: Trends and Seasonality
BUSPROG:
Feedback: When a series has the same average growth rate from period to period, it
can be approximated with an exponential trend.

14. Adding a time trend can make an explanatory variable more significant if:
a. the dependent and independent variables have similar kinds of trends, but
movement in the independent variable about its trend line causes movement in
the dependent variable away from its trend line.
b. the dependent and independent variables have similar kinds of trends and
movement in the independent variable about its trend line causes movement in
the dependent variable towards its trend line.
c. the dependent and independent variables have different kinds of trends and
movement in the independent variable about its trend line causes movement in
the dependent variable towards its trend line.
d. the dependent and independent variables have different kinds of trends, but
movement in the independent variable about its trend line causes movement in
the dependent variable away from its trend line.

Answer: d
Difficulty: Hard
Bloom’s: Knowledge
A-Head: Trends and Seasonality
BUSPROG:
Feedback: Adding a time trend can make an explanatory variable more significant if
the dependent and independent variables have different kinds of trends and
movement in the independent variable about its trend line causes movement in the
dependent variable away from its trend line.

15. A seasonally adjusted series is one which:

a. has had seasonal factors added to it.
b. has seasonal factors removed from it.
c. has qualitative explanatory variables representing different seasons.
b. has qualitative dependent variables representing different seasons.

Answer: b
Difficulty:
Moderate
Bloom’s: Knowledge

Downloaded by K61.FTU ?INH H?NG NG?C

A-Head: Trends and Seasonality
BUSPROG:
Feedback: A seasonally adjusted series is one which has seasonal factors removed
from it.

16. Economic time series are outcomes of random variables.

Answer: True
Difficulty: Easy
Bloom’s: Knowledge
A-Head: The Nature of Time Series Data
BUSPROG:
Feedback: Economic time series are outcomes of random variables.

17. In a static model, one or more explanatory variables affect the

dependent variable with a lag.

Answer: False
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Examples of Time Series Regression Models
BUSPROG:
Feedback: In a finite distributed lag model, one or more explanatory variables affect
the dependent variable with a lag. In a static model, no lags are included.

18. Time series regression is based on series which exhibit serial correlation.

Answer: False
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Finite Sample Properties of OLS under Classical Assumptions
BUSPROG:
Feedback: One of the assumptions of time series regression is that there should be
no serial correlation in the concerned series.

19. Price indexes are necessary for turning a time series measured in real value
into nominal value.

Answer: False
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Functional Forms, Dummy Variables, and Index Numbers.
BUSPROG:

Downloaded by K61.FTU ?INH H?NG NG?C

Feedback: Price indexes are necessary for turning a time series measured in
nominal value into real value.

20. Dummy variables can be used to address the problem of seasonality

in regression models.

Answer: True
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Trends and Seasonality
BUSPROG:
Feedback: Dummy variables can be used to account for seasonality in the
dependent variable, the independent variables, or both and thus, address the
problem of seasonality in regression models.

Chapter 11

1. A process is stationary if:

a. any collection of random variables in a sequence is taken and shifted ahead by h
time periods; the joint probability distribution changes.
b. any collection of random variables in a sequence is taken and shifted ahead by h
time periods, the joint probability distribution remains unchanged.
c. there is serial correlation between the error terms of successive time periods
and the explanatory variables and the error terms have positive covariance.
d. there is no serial correlation between the error terms of successive time
periods and the explanatory variables and the error terms have positive
covariance.

Answer: b
Difficulty:
Moderate
Bloom’s: Knowledge
A-Head: Stationary and Weakly Dependent Time Series
BUSPROG:
Feedback: A process is stationary if any collection of random variables in a
sequence is taken and shifted ahead by h time periods; the joint probability
distribution remains unchanged.

2. A stochastic process {xt: t = 1,2,….} with a finite second moment [E(x t2) < ∞] is
covariance stationary if:
a. E(xt) is variable, Var(xt) is variable, and for any t, h ≥ 1, Cov(xt, xt+h) depends only
on ‘h’ and not on ‘t’.
b. E(xt) is variable, Var(xt) is variable, and for any t, h ≥ 1, Cov(xt, xt+h) depends only
on ‘t’ and not on h.

Downloaded by K61.FTU ?INH H?NG NG?C

c. E(xt) is constant, Var(xt) is constant, and for any t, h ≥ 1, Cov(xt, xt+h) depends
only on ‘h’ and not on ‘t’.
d. E(xt) is constant, Var(xt) is constant, and for any t, h ≥ 1, Cov(xt, xt+h) depends
only on ‘t’ and not on ‘h’.

Answer: c
Difficulty:
Moderate
Bloom’s: Knowledge
A-Head: Stationary and Weakly Dependent Time Series
BUSPROG:
Feedback: A stochastic process {xt: t = 1,2,….} with a finite second moment [E(x 2t )
< ∞] is covariance stationary if E(xt) is constant, Var(xt) is constant, and for any t, h
≥ 1, Cov(xt, xt+h) depends only on ‘h’ and not on ‘t’.

3. A covariance stationary time series is weakly dependent if:

a. the correlation between the independent variable at time ‘t’ and the
dependent variable at time ‘t + h’ goes to ∞ as h → 0.
b. the correlation between the independent variable at time ‘t’ and the
dependent variable at time ‘t + h’ goes to 0 as h → ∞.
c. the correlation between the independent variable at time ‘t’ and the
independent variable at time ‘t + h’ goes to ∞ as h → 0.
d. the correlation between the independent variable at time ‘t’ and the
independent variable at time ‘t + h’ goes to 0 as h → ∞.

Answer: d
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Stationary and Weakly Dependent Time Series
BUSPROG:
Feedback: A covariance stationary time series is weakly dependent if the
correlation between the independent variable at time ‘t’ and the independent
variable at time ‘t + h’ goes to 0 as h → ∞.

4. The model yt = et + β1et – 1 + β2et – 2 , t = 1, 2, ….. , where et is an i.i.d.

sequence with zero mean and variance σ2e represents a(n):
a. static model.
b. moving average process of order one.
c. moving average process of order two.
d. autoregressive process of order two.

Answer: c
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Stationary and Weakly Dependent Time Series

Downloaded by K61.FTU ?INH H?NG NG?C

BUSPROG:
Feedback: The model yt = et + β1et – 1 + β2et – 2 , t = 1, 2, ….. , where et is an i.i.d.
sequence with zero mean and variance σ 2e, represents an moving average process
of order two.

5. The model xt = α1xt – 1 + et , t =1,2,…. , where et is an i.i.d. sequence with

zero mean and variancee
σ2 represents a(n):
a. moving average process of order one.
b. moving average process of order two.
c. autoregressive process of order one.
d. autoregressive process of order two.

Answer: c
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Stationary and Weakly Dependent Time Series
BUSPROG:
Feedback: The model xt = α1xt – 1 + et , t =1,2,…. , where et is an i.i.d. sequence with
zero mean and variance σ2 ,e represents an autoregressive process of order one.

6. Which of the following is assumed in time series regression?

a. There is no perfect collinearity between the explanatory variables.
b. The explanatory variables are contemporaneously endogenous.
c. The error terms are contemporaneously heteroskedastic.
d. The explanatory variables cannot have temporal ordering.

Answer: a
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Asymptotic Properties of OLS
BUSPROG:
Feedback: One of the assumptions of time series regression is that there should be
no perfect collinearity between the explanatory variables.

7. Suppose ut is the error term for time period ‘t’ in a time series regression
model the explanatory variables are xt = (xt1, xt2 …., xtk). The assumption that the
errors are contemporaneously homoskedastic implies that:
a. Var(ut|xt) = √σ.
b. Var(ut|xt) = ∞.
c. Var(ut|xt) = σ2.
d. Var(ut|xt) = σ.

Answer: c
Difficulty:
Moderate

Downloaded by K61.FTU ?INH H?NG NG?C

Bloom’s: Knowledge
A-Head: Asymptotic Properties of OLS
BUSPROG:
Feedback: If ut is the error term for time period ‘t’ and xt is a matrix consisting of all
independent variables for time ‘t’, the assumption of contemporaneously
homoskedasticity implies that Var(ut|xt) = σ2.

8. Which of the following statements is true?

a. A model with a lagged dependent variable cannot satisfy the strict
exogeneity assumption.
b. Stationarity is critical for OLS to have its standard asymptotic properties.
c. Efficient static models can be estimated for nonstationary time series.
d. In an autoregressive model, the dependent variable in the current time
period varies with the error term of previous time periods.

Answer: a
Difficulty:
Moderate
Bloom’s: Knowledge
A-Head: Asymptotic Properties of OLS
BUSPROG:
Feedback: A model with a lagged dependent variable cannot satisfy the strict
exogeneity assumption. When explanatory variables are correlated with the past,
strict exogeneity does not hold.

9. Consider the model: yt = α0 + α1rt1 + α2rt2 + ut. Under weak dependence, the
condition sufficient for consistency of OLS is:
a. E(rt1|rt2) = 0.
b. E(yt |rt1, rt2) = 0.
c. E(ut |rt1, rt2) = 0.
d. E(ut |rt1, rt2) = ∞.

Answer: c
Difficulty:
Moderate
Bloom’s: Knowledge
A-Head: Asymptotic Properties of OLS
BUSPROG:
Feedback: If a time series model is weakly dependent, the condition sufficient for
consistency of OLS is E(ut|rt1, rt2) = 0.

10. The model yt = yt – 1 + et, t = 1, 2, … represents a:

a. AR(2) process.
b. MA(1) process.
c. random walk process.
d. random walk with a drift process.

Downloaded by K61.FTU ?INH H?NG NG?C

Answer: c
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Using Highly Persistent Time Series in Regression Analysis
BUSPROG:
Feedback: The model yt = yt – 1 + et, t = 1, 2, … represents a random walk process.

11. Which of the following statements is true?

a. A random walk process is stationary.
b. The variance of a random walk process increases as a linear function of time.
c. Adding a drift term to a random walk process makes it stationary.
d. The variance of a random walk process with a drift decreases as an
exponential function of time.

Answer: b
Difficulty:
Moderate
Bloom’s: Knowledge
A-Head: Using Highly Persistent Time Series in Regression Analysis
BUSPROG:
Feedback: The variance of a random walk process increases as a linear function of
time. This is because the variance of the dependent variable is equal to σ 2t.

12. If a process is said to be integrated of order one, or I(1), .

a. it is stationary at level
b. averages of such processes already satisfy the standard limit theorems
c. the first difference of the process is weakly dependent
d. it does not have a unit root

Answer: c
Difficulty:
Moderate
Bloom’s: Knowledge
A-Head: Using Highly Persistent Time Series in Regression Analysis
BUSPROG:
Feedback: If a process is said to be integrated of order one, or I(1), the first
difference of the process is weakly dependent.

13. Which of the following statements is true of dynamically complete models?

a. There is scope of adding more lags to the model to better forecast the
dependent variable.
b. The problem of serial correlation does not exist in dynamically complete models.
c. All econometric models are dynamically complete.
d. Sequential endogeneity is implied by dynamic completeness..

Downloaded by K61.FTU ?INH H?NG NG?C

Answer: b
Difficulty:
Moderate
Bloom’s: Knowledge
A-Head: Dynamically Complete Models and the Absence of Serial Correlation
BUSPROG:
Feedback: The problem of serial correlation does not exist in dynamically complete
models.

14. In the model yt = α0 + α1xt1 + α2xt2 + ….. + αkxtk + ut, the explanatory variables, xt
= (xt1, xt2 …., xtk), are sequentially exogenous if:
a. E(ut|xt , xt-1, ……) = E(ut) = 0, t = 1,2, ….
b. E(ut|xt , xt-1, ……) ≠ E(ut) = 0, t = 1,2, ….
c. E(ut|xt , xt-1, ……) = E(ut) > 0, t = 1,2, ….
d. E(ut|xt , xt-1, ……) = E(ut) = 1, t = 1,2, ….

Answer: a
Difficulty:
Moderate
Bloom’s: knowledge
A-Head: Dynamically Complete Models and the Absence of Serial Correlation
BUSPROG:
Feedback: In the given model, the explanatory variables are sequentially exogenous
if E(ut|xt , xt-1, ……) = E(ut) = 0, t = 1,2, ….

15. If ut refers to the error term at time ‘t’ and yt – 1 refers to the dependent
variable at time ‘t – 1’, for an AR(1) process to be homoskedastic, it is required
that:
a. Var(ut|yt – 1) > Var(yt|yt-1) = σ2.
b. Var(ut|yt – 1) = Var(yt|yt-1) > σ2.
c. Var(ut|yt – 1) < Var(yt|yt-1) = σ2.
d. Var(ut|yt – 1) = Var(yt|yt-1) = σ2.

Answer: d
Difficulty:
Moderate
Bloom’s: Knowledge
A-Head: The Homoskedasticity Assumption for Time Series Models
BUSPROG:
Feedback: If ut refers to the error term at time ‘t’ and yt – 1 refers to the
dependent variable at time ‘t – 1’, for an AR(1) model to be homoskedastic, it is
required that Var(ut|yt – 1) = Var(yt|yt-1) = σ2.

16. Covariance stationarity focusses only on the first two moments of a

stochastic process.

Answer: True

Downloaded by K61.FTU ?INH H?NG NG?C

Difficulty: Easy

Downloaded by K61.FTU ?INH H?NG NG?C

Bloom’s: Knowledge
A-Head: Stationary and Weakly Dependent Time Series
BUSPROG:
Feedback: Covariance stationarity focusses only on the first two moments of a
stochastic process: the mean and variance, which are constant over time.

17. Under adaptive expectations, the expected current value of a variable does
not depend on a recently observed value of the variable.

Answer: False
Difficulty: Moderate
Bloom’s: Knowledge
A-Head: Asymptotic Properties of OLS
BUSPROG:
Feedback: Under adaptive expectations, the expected current value of a variable
adapts to a recently observed value of the variable.

18. Weakly dependent processes are said to be integrated of order zero.

Answer: True
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Stationary and Weakly Dependent Time Series
BUSPROG:
Feedback: Weakly dependent processes are said to be integrated of order zero, or
I(0).

19. Sequential exogeneity is implied by dynamic completeness.

Answer: True
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Dynamically Complete Models and the Absence of Serial Correlation
BUSPROG:
Feedback: Sequential exogeneity is implied by dynamic completeness.

20. The homoskedasticity assumption in time series regression suggests that

the variance of the error term cannot be a function of time.

Answer: True
Difficulty: Easy
Bloom’s: Knowledge
A-Head: The Homoskedasticity Assumption for Time Series Models
BUSPROG:

Downloaded by K61.FTU ?INH H?NG NG?C

Feedback: The homoskedasticity assumption in time series regression suggests that
the variance of the error term cannot be a function of time. Homeskedasticity
implies that the variance of the error term is constant and hence cannot be a
function of time.
Chapter 12

1. In the presence of serial correlation:

a. estimated standard errors remain valid.
b. estimated test statistics remain valid.
c. estimated OLS values are not BLUE.
d. estimated variance does not differ from the case of no serial correlation.

Answer: c
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Properties of OLS with Serially Correlated Errors
BUSPROG:
Feedback: As the Gauss-Markov Theorem requires both homoscedasticity and
serially uncorrelated errors, OLS in no longer BLUE in the presence of serial
correlation.

2. When a series is stationary, weakly dependent, and has serial correlation:

a. the adjusted R2 is inconsistent, while R2 is a consistent estimator of the
population parameter.
b. the adjusted R2 is consistent, while R2 is an inconsistent estimator of the
population parameter.
c. both the adjusted R2 and R2 are inconsistent estimators of the population
parameter.
d. both the adjusted R2 and R2 are consistent estimators of the population
parameter.

Answer: d
Difficulty:
Moderate
Bloom’s: Knowledge
A-Head: Properties of OLS with Serially Correlated Errors
BUSPROG:
Feedback: When a series is stationary, weakly dependent, and has serial correlation
both the adjusted R2 and R2 are consistent estimators of the population parameter
as the calculation of R2 and adjusted R2 is based on the variance of the dependent
variable and the error term, which do not change over time.

3. Which of the following is a test for serial correlation in the error terms?
a. Johansen test
b. Dickey Fuller test

Downloaded by K61.FTU ?INH H?NG NG?C

c. Durbin Watson test
d. White test

Answer: c
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Testing for Serial Correlation
BUSPROG:
Feedback: The Durbin Watson test can be used to test for serial correlation in error
terms.

4. For a given significance level, if the calculated value of the Durbin Watson
statistic lies between the lower critical value and the upper critical value, .
a. the hypothesis of no serial correlation is accepted
b. the hypothesis of no serial correlation is rejected
c. the test is inconclusive
d. the hypothesis of heteroskedasticity is accepted

Answer: c
Difficulty:
Moderate
Bloom’s: Knowledge
A-Head: Testing for Serial Correlation
BUSPROG:
Feedback: For a given significance level, if the calculated value of the Durbin
Watson statistic lies between the lower critical value and upper critical value, the
test is inconclusive.

5. Which of the following statements is true?

a. When explanatory variables are not strictly exogenous, the t test for
serial correlation is valid.
b. When explanatory variables are not strictly exogenous, the Durbin Watson
test for serial correlation is valid.
c. Breusch-Godfrey test can be used to check for second order serial correlation.
d. White test can be used to check for second order serial correlation.

Answer: c
Difficulty:
Moderate
Bloom’s: Knowledge
A-Head: Testing for Serial Correlation
BUSPROG:
Feedback: Breusch-Godfrey test can be used to check for second order serial
correlation.

6. The Breusch-Godfrey test statistic follows a:

Downloaded by K61.FTU ?INH H?NG NG?C

a. χ2distribution.
b. t distribution.
c. normal distribution.
d. F distribution.

Answer: a
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Testing for Serial Correlation
BUSPROG:
Feedback: The Breusch-Godfrey test statistic follows a χ2distribution.

7. In a model based on a weakly dependent time series with serial correlation

and strictly exogenous explanatory variables, .
a. the feasible generalized least square estimates are unbiased
b. the feasible generalized least square estimates are BLUE
c. the feasible generalized least square estimates are asymptotically more
efficient than OLS estimates
d. the feasible generalized least square estimates are asymptotically less
efficient than OLS estimates

Answer: c
Difficulty:
Moderate
Bloom’s: Knowledge
A-Head: Correcting for Serial Correlation with Strictly Exogenous Regressors
BUSPROG:
Feedback: In a model based on a weakly dependent time series with serial
correlation and strictly exogenous explanatory variables the feasible generalized
least square estimates are asymptotically more efficient than OLS estimates.

8. Which of the following is an example of FGLS estimation?

a. Dickey-Fuller estimation
b. Vector error correction estimation
c. Prais-Winsten estimation
d. OLS estimation.

Answer: c
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Correcting for Serial Correlation with Strictly Exogenous Regressors
BUSPROG:
Feedback: Prais-Winsten estimation is a type of FGLS estimation.

Downloaded by K61.FTU ?INH H?NG NG?C

9. Which of the following is the reason why standard errors measured by OLS
differ from standard errors measured through Prais-Winsten transformation?
a. OLS standard errors account for serial correlation, whereas Prais-
Winsten estimations do not.
b. Prais-Winsten standard errors account for serial correlation, whereas
OLS estimations do not.
c. Prais-Winsten standard errors account for heteroskedasticity, whereas
OLS estimations do not.
d. OLS standard errors account for heteroskedasticity, whereas Prais-
Winsten estimations do not.

Answer: b
Difficulty:
Moderate
Bloom’s: Knowledge
A-Head: Correcting for Serial Correlation with Strictly Exogenous Regressors
BUSPROG:
Feedback: The standard errors measured by OLS differ from the standard errors
measured by Prais-Winsten transformation because Prais-Winsten standard errors
account for serial correlation, whereas OLS estimations do not.

10. Which of the following identifies an advantage of first differencing a time-series?

a. First differencing eliminates most of the serial correlation.
b. First differencing eliminates most of the heteroskedastcicty.
c. First differencing eliminates most of the multicollinearity.
d. First differencing eliminates the possibility of spurious regression.

Answer: a
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Differencing and Serial Correlation
BUSPROG:
Feedback: First differencing of a time-series helps eliminate most of the serial
correlation.

11. Which of the following is a limitation of serial correlation-robust standard errors?

a. The serial correlation-robust standard errors are smaller than OLS standard
errors when there is serial correlation.
b. The serial correlation-robust standard errors can be poorly behaved when there
is substantial serial correlation and the sample size is small.
c. The serial correlation-robust standard errors cannot be calculated
for autoregressive processes of an order greater than one.
d. The serial correlation-robust standard errors cannot be calculated after
relaxing the assumption of homoskedasticity.

Downloaded by K61.FTU ?INH H?NG NG?C

Answer: b
Difficulty:
Moderate
Bloom’s: Knowledge
A-Head: Serial Correlation-Robust Inference after OLS
BUSPROG:
Feedback: The serial correlation-robust standard errors can be poorly behaved when
there is substantial serial correlation and the sample size is small.

12. Which of the following statements is true?

a. Prais-Winsten and Cochrane-Orcutt transformations are consistent when
explanatory variables are not strictly exogenous.
b. The SC-robust standard errors cannot be estimated in models with
lagged dependent variables.
c. The SC-robust standard errors work better after quasi-differencing a time
series that is expected to be serially correlated.
d. Estimation of SC-robust standard errors is independent of the sample size.

Answer: c
Difficulty:
Moderate
Bloom’s: Knowledge
A-Head: Serial Correlation-Robust Inference after OLS
BUSPROG:
Feedback: The SC-robust standard errors work better after quasi-differencing a
time series that is expected to be serially correlated. Quasi-differencing helps limit
serial correlation.

13. In the presence of heteroskedasticity, the usual OLS estimates of:

a. standard errors are valid, whereas the t statistics and F statistics are invalid.
b. t statistics are valid, but the standard errors and F statistics are invalid.
c. F statistics are valid, but the standard errors and t statistics are invalid.
d. standard errors, t statistics, and F statistics are invalid.

Answer: d
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Heteroskedasticity in Time Series Regression
BUSPROG:
Feedback: In the presence of heteroskedasticity, the usual OLS estimates of
standard errors, t statistics, and F statistics are invalid.

14. Which of the following tests can be used to test for heteroskedasticity in a
time series?
a. Johansen test
b. Dickey-Fuller test

Downloaded by K61.FTU ?INH H?NG NG?C

c. Breusch-Pagan test
d. Durbin’s alternative test

Answer: c
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Heteroskedasticity in Time Series Regression
BUSPROG:
Feedback: The Breusch-Pagan test can be used to test for heteroskedasticicty in a
time series.

15. The equation u2 = α + α u2 + v is an autoregressive model in .

t 0 1 t –1
a. ut
b. u2 t
c. vt
d. ut – 1

Answer: b
Difficulty:
Moderate
Bloom’s: Knowledge
A-Head: Heteroskedasticity in Time Series Regression
BUSPROG:
Feedback: The model u2t = α0 + α1u2 t – 1+ vt is an autoregressive model in u2 .

16. In presence of serial correlation, the OLS variance formula accurately

estimates the true variance of the OLS estimator.

Answer: False
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Properties of OLS with Serially Correlated Errors
BUSPROG:
Feedback: In presence of serial correlation, the OLS variance formula either
understates or overstates the true variance of the OLS estimator.

17. Durbin’s alternative test is valid even if the explanatory variables are
strictly exogenous.

Answer: True
Difficulty: easy
Bloom’s: Knowledge
A-Head: Testing for Serial Correlation
BUSPROG:
Feedback: Durbin’s alternative test is valid even if the explanatory variables are
strictly exogenous.

Downloaded by K61.FTU ?INH H?NG NG?C

18. Consistency of feasible generalized least square estimators requires the
error term to be correlated with lags of the explanatory variable.

Answer: False
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Testing for Serial Correlation
BUSPROG:
Feedback: Consistency of feasible generalized least square estimators requires
the error term to be uncorrelated with lags of the explanatory variable. Correlation
will lead to inconsistent estimates.

19. FGLS estimates are efficient when explanatory variables are not
strictly exogenous.

Answer: False
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Testing for Serial Correlation
BUSPROG:
Feedback: FGLS estimates are inefficient when explanatory variables are not strictly
exogenous.

20. In time series regressions, it is advisable to check for serial correlation

first, before checking for heteroskedasticity.

Answer: True
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Heteroskedasticity in Time Series Regression
BUSPROG:
Feedback: Serial correlation invalidates heteroskedasticity tests. Hence, it is
advisable to check for serial correlation first, before checking for heteroskedasticity
in time series regressions.
Chapter 13

1. Which of the following is a reason for using independently pooled cross sections?
a. To obtain data on different cross sectional units
b. To increase the sample size
c. To select a sample based on the dependent variable
d. To select a sample based on the independent

variable Answer: b

Downloaded by K61.FTU ?INH H?NG NG?C

Difficulty: Easy
Bloom’s: Knowledge
A-Head: Pooling Independent Cross Sections across Time
BUSPROG:
Feedback: One reason for using independently pooled cross sections is to increase
the sample size. By pooling random samples drawn from the same population, but
at different points in time, we can get more precise estimators and test statistics
with more power.

2. Pooling independent cross sections across time is useful in providing

precise estimators if .
a. the relationship between the dependent variable and at least some of
the independent variables remains constant over time
b. the relationship between the dependent variable and at least some of
the independent variables is linear
c. the relationship between the dependent variable and at least one of
the independent variables changes over time
d. the relationship between the dependent variable and at least one of
the independent variables is positive

Answer: a
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Pooling Independent Cross Sections across Time
BUSPROG:
Feedback: Pooling independent cross sections across time is useful in providing
precise estimators if the relationship between the dependent variable and at least
some of the independent variables remains constant over time.

3. A Chow test .
a. is used to test the presence of heteroskedasticty in a regression model.
b. is used to determine how multiple regression differs across two groups.
c. cannot detect changes in the slope coefficients of dependent variables over time.
d. cannot be computed for more than two time periods.

Answer: b
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Pooling Independent Cross Sections across Time
BUSPROG:
Feedback: A Chow test is used to determine how multiple regression differs across
two groups.

4. Which of the following is true of a natural experiment?

Downloaded by K61.FTU ?INH H?NG NG?C

a. A natural experiment occurs when some endogenous event changes
the environment in which individuals, families, firms, or cities operate.
b. The control group in a natural experiment is randomly chosen.
c. The treatment group in a natural experiment is randomly chosen.
d. Control and treatment groups in a natural experiment arise due to an
exogenous event.

Answer: d
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Policy Analysis with Pooled Cross Sections
BUSPROG:
Feedback: Control and treatment groups in a natural experiment arise due to an
exogenous event.

5. The average treatment effect measures .

a. the effect of a policy or program on the dependent variable
b. the effect of unobserved factors on observed explanatory variables
c. the effect of a change in an explanatory variable on other explanatory variables
d. the effect of unobserved factors on the average outcome of the
dependent variable

Answer: a
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Policy Analysis with Pooled Cross Sections
BUSPROG:
Feedback: The average treatment effect measures the effect of a policy or program
on the dependent variable.

6. Idiosyncratic error is the error that occurs due to .

a. incorrect measurement of an economic variable
b. unobserved factors that affect the dependent variable and change over time
c. unobserved factors that affect the dependent variable and do not change
over time
d. correlation between the independent variables

Answer: b
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Two-Period Panel Data Analysis
BUSPROG:
Feedback: Idiosyncratic error is the error that occurs due to unobserved factors that
affect the dependent variable and change over time.

Downloaded by K61.FTU ?INH H?NG NG?C

7. A regression model exhibits unobserved heterogeneity if .
a. there are unobserved factors affecting the dependent variable that change
over time
b. there are unobserved factors affecting the dependent variable that do not
change over time
c. there are unobserved factors which are correlated with the observed
independent variables
d. there are no unobserved factors affecting the dependent variable

Answer: b
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Two-Period Panel Data Analysis
BUSPROG:
Feedback: A regression model exhibits unobserved heterogeneity if there are
unobserved factors affecting the dependent variable but they do not change over
time.

8. Composite error is the error that occurs due to .

a. the incorrect measurement of explanatory variables
b. the inclusion of too many explanatory variables in the model
c. correlation between the independent variables
d. unobserved factors affecting a dependent variable

Answer: d
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Two-Period Panel Data Analysis
BUSPROG:
Feedback: Composite error is the error that occurs due to all unobserved factors
affecting a dependent variable.

9. Ordinary least squares estimation is subject to heterogeneity bias if .

a. the regression model exhibits heteroskedasticty
b. the unobserved effect is correlated with the observed explanatory variables
c. the regression model includes a lagged dependent variable
d. the explanatory variables do not change over time

Answer: b
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Two-Period Panel Data Analysis
BUSPROG:

Downloaded by K61.FTU ?INH H?NG NG?C

Feedback: Ordinary least estimation is subject to heterogeneity bias if the
unobserved effect is correlated with the observed explanatory variables.

10. Which of the following assumptions is required to obtain a first-

differenced estimator in a two-period panel data analysis?
a. The explanatory variable does not change over time for any cross-sectional unit.
b. The explanatory variable changes by the same amount in each time period.
c. The variance of the error term in the regression model is not constant.
d. The idiosyncratic error at each time period is uncorrelated with the
explanatory variables in both time periods.

Answer: d
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Two-Period Panel Data Analysis
BUSPROG:
Feedback: The assumption that the idiosyncratic error at each time period is
uncorrelated with the explanatory variables in both time periods is required to
obtain a first-differenced estimator.

11. The assumption of “strict exogeneity” in a regression model means that

a. the dependent variable is binary
b. all explanatory variables change over time
c. the model does not include a lagged dependent variable as a regressor
d. the model includes all relevant explanatory variables

Answer: c
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Differencing with More Than Two Time Periods
BUSPROG:
Feedback: The assumption of “strict exogeneity” in a regression model means that
the model does not include a lagged dependent variable as a regressor.

12. A data set is a balanced panel if it .

a. consists of a sample of individuals, households, firms, cities, states, countries,
or a variety of other units, taken at a given point in time.
b. consists of observations on a variable or several variables over time
c. consists of data for each cross sectional unit over the same time period
d. consists of time-demeaned data for different cross sectional units

Answer: c
Difficulty: Easy
Bloom’s: Knowledge

Downloaded by K61.FTU ?INH H?NG NG?C

A-Head: Differencing with More Than Two Time Periods
BUSPROG:
Feedback: A data set is a balanced panel if it consists of data for different cross
sectional units over the same time period.

13. Which of the following assumptions is needed for the usual standard errors to
be valid when differencing with more than two time periods?
a. The regression model exhibits heteroskedasticty.
b. The differenced idiosyncratic error or
∆ u is uncorrelated over time.
it

c. The unobserved factors affecting the dependent variable are time-constant.

d. The regression model includes a lagged independent variable.

Answer: b
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Differencing with More Than Two Time Periods
BUSPROG:
Feedback: The differenced idiosyncratic error is uncorrelated over time.

14. First-differenced estimation in a panel data analysis is subject to serious

biases if .
a. key explanatory variables vary significantly over time
b. the explanatory variables do not change by the same unit in each time period
c. one or more of the explanatory variables are measured incorrectly
d. the regression model exhibits homoscedasticity

Answer: c
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Differencing with More Than Two Time Periods
BUSPROG:
Feedback: First-differenced estimation is subject to serious biases if one or more of
the explanatory variables are measured incorrectly.

15. The general approach to obtaining fully robust standard errors and test
statistics in
the context of panel data is known as .
a. confounding
b. differencing
c. clustering
d. attenuating

Answer: c

Downloaded by K61.FTU ?INH H?NG NG?C

Difficulty: Easy
Bloom’s: Knowledge
A-Head: Appendix 13A
BUSPROG:
Feedback: The general approach to obtaining fully robust standard errors and test
statistics in
the context of panel data is known as clustering.

16. If a random sample is drawn at each time period, pooling the resulting
random samples gives us a panel data set.

Answer: False
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Pooling Independent Cross Sections across Time
BUSPROG:
Feedback: If a random sample is drawn at each time period, pooling the resulting
random samples gives us an independently pooled cross section.

17. A natural experiment occurs when an endogenous event changes

the environment in which individuals, families, firms, or cities operate.

Answer: False
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Pooling Independent Cross Sections across Time
BUSPROG:
Feedback: A natural experiment occurs when an exogenous event changes the
environment in which individuals, families, firms, or cities operate.

18. One way of organizing two periods of panel data is to have only one record
per cross-sectional unit.

Answer: True
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Two-Period Panel Data Analysis
BUSPROG:
Feedback: One way of organizing two periods of panel data is to have only one
record per cross-sectional unit.

19. Two-period panel data is used for program evaluation and policy analysis.

Answer: True

Downloaded by K61.FTU ?INH H?NG NG?C

Difficulty: Easy
Bloom’s: Knowledge
A-Head: Policy Analysis with Two-Period Panel Data
BUSPROG:
Feedback: Two-period panel data is used for program evaluation and policy
analysis. In the simplest program evaluation setup, a sample of individuals, firms,
cities, and so on is obtained in the first time period. Some of these units, those in
the treatment group, then take part in a particular program in a later time period;
the ones that do not are the control group.

20. First-differenced estimation gives unbiased estimators if the regression

model includes a lagged dependent variable.

Answer: False
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Differencing with More Than Two Time Periods
BUSPROG:
Feedback: First-differenced estimation is subject to serious biases if the regression
model includes a lagged dependent variable.
Chapter 14

1. 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 strictly exogenous.
d. The unobserved effect is correlated with the explanatory variables.

Answer: c
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Fixed Effects Estimation
BUSPROG:
Feedback: Under a strict exogeneity assumption on the explanatory variables, the
fixed effects estimator is unbiased.

2. A pooled OLS estimator that is based on the time-demeaned variables is

called the .
a. random effects estimator
b. fixed effects estimator
c. least absolute deviations estimator
d. instrumental variable estimator

Downloaded by K61.FTU ?INH H?NG NG?C

Answer: b
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Fixed Effects Estimation
BUSPROG:
Feedback: A pooled OLS estimator that is based on the time-demeaned variables is
called the fixed effects estimator.

3. What should be the degrees of freedom (df) for fixed effects estimation if the
data set includes ‘N’ cross sectional units over ‘T’ time periods and the
regression model has ‘k’ independent variables?
a. N-kT
b. NT-k
c. NT-N-k
d. N-T-k

Answer: c
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Fixed Effects Estimation
BUSPROG:
Feedback: If the data set includes N cross sectional units over T time periods, the
total number of observations is NT. Since the regression model includes k
independent variables, the model should have NT-k degrees of freedom. However,
for each cross-sectional observation, we lose one df because of the time-
demeaning. Therefore, the appropriate degrees of freedom is NT - N – k.

4. 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

Answer: d
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Fixed Effects Estimation
BUSPROG:
Feedback: A fixed effects model cannot include a time-constant independent
variable.

5. Which of the following is a property of dummy variable regression?

Downloaded by K61.FTU ?INH H?NG NG?C

a. This method is best suited for panel data sets with many cross-
sectional observations.
b. The R-squared obtained from this method is lower than that obtained
from regression on time-demeaned data.
c. The degrees of freedom cannot be computed directly with this method.
d. The major statistics obtained from this method are identical to that obtained
from regression on time-demeaned data.

Answer: d
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Fixed Effects Estimation
BUSPROG:
Feedback: The major statistics obtained from this method are identical to that
obtained from regression on time-demeaned data.

6. Which of the following is a difference between a fixed effects estimator and

a first-difference estimator?
a. The fixed effects estimators are always larger than the first difference
estimators in a two-period panel data analysis.
b. The fixed effects estimator is more efficient than the first-difference
estimator when the idiosyncratic errors are serially uncorrelated.
c. The first difference estimator is more sensitive to nonnormality
and heteroskedasticity.
d. The bias in the first difference estimator depends on the time period (T)
of analysis while the bias in the fixed effect does not depend on T.

Answer: b
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Fixed Effects Estimation
BUSPROG:
Feedback: The fixed effects estimator is more efficient than the first-difference
estimator when the idiosyncratic errors are serially uncorrelated.

7. Which of the following assumptions is required for obtaining unbiased

random effect estimators?
a. The idiosyncratic errors are heteroskedastic.
b. The unobserved effect is independent of all explanatory variables in all
time periods.
c. The idiosyncratic errors are serially correlated.
d. The unobserved effect is correlated with the explanatory variables.

Answer: b

Downloaded by K61.FTU ?INH H?NG NG?C

Difficulty: Easy
Bloom’s: Knowledge
A-Head: Random Effects Models
BUSPROG:
Feedback: The unobserved effect is independent of all explanatory variables in all
time periods.

8. The estimator obtained through regression on quasi-demeaned data is called the

.
a. random effects estimator
b. fixed effects estimator
c. hetroskedasticity-robust OLS estimator
d. instrumental variables estimator

Answer: a
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Random Effects Models
BUSPROG:
Feedback: The estimator obtained through regression on quasi-demeaned data is
called the random effects estimator.

9. The random effects approach .

a. cannot be used if the key explanatory variable is constant over time.
b. is preferred to pooled OLS because RE is generally more efficient.
c. is suitable if the Hausman test rejects the assumption that the unobserved
effect is uncorrelated with the explanatory variables.
d. is more convincing than fixed effects for policy analysis using aggregate data.

Answer: b
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Random Effects Models
BUSPROG:
Feedback: RE is preferred to pooled OLS because RE is generally more efficient.

10. The random effects estimate is identical to the fixed effects estimate if the
estimated transformation parameter ,
^ , in generalized least squares estimation
θ

that eliminates serial correlation between error terms is, .

a. less than zero
b. equal to zero
c. equal to one
d. greater than one

Downloaded by K61.FTU ?INH H?NG NG?C

Answer: c
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Random Effects Models
BUSPROG:
Feedback: The random effects estimate is identical to the fixed effects estimate if

the estimated transformation parameter,

θ^ , in generalized least squares

estimation that eliminates serial correlation in error terms, is equal to one.

11. Which of the following is true of the correlated random effects approach (CRE)?
a. The CRE approach assumes that the unobserved effect is uncorrelated with
the observed explanatory variables.
b. The CRE approach cannot be used if the regression model includes a
time- constant explanatory variable.
c. The CRE approach considers that the unobserved effect is correlated with
the average level of explanatory variables.
d. The CRE estimate equals the random effects estimate.

Answer: c
Difficulty: Easy
Bloom’s: Knowledge
A-Head: The Correlated Random Effects Approach
BUSPROG:
Feedback: The CRE approach considers that the unobserved effect is correlated with
the average level of explanatory variables.

12. Which of the following is a reason for using the correlated random
effects approach?
a. It provides unbiased and consistent estimators when the idiosyncratic errors
are serially correlated.
b. It provides unbiased and consistent estimators when the idiosyncratic errors
are heteroskedastic.
c. It provides a more efficient estimate than the fixed effects approach.
d. It provides a way to include time-constant explanatory variables in a fixed
effects analysis.

Answer: d
Difficulty: Easy
Bloom’s: Knowledge
A-Head: The Correlated Random Effects Approach
BUSPROG:

Downloaded by K61.FTU ?INH H?NG NG?C

Feedback: It provides a way to include time-constant explanatory variables in a
fixed effects analysis.

13. In the correlated random effects approach, the regression model includes .
a. time averages as separate explanatory variables
b. at least one dummy variable
c. more than one endogenous explanatory variable
d. an instrumental variable

Answer: a
Difficulty: Easy
Bloom’s: Knowledge
A-Head: The Correlated Random Effects Approach
BUSPROG:
Feedback: In the correlated random effects approach, the regression model includes
time averages as separate explanatory variables.

14. An economist wants to study the effect of income on savings. He collected

data on 120 identical twins. Which of the following methods of estimation is the
most suitable method, if income is correlated with the unobserved family effect?
a. Random effects estimation
b. Fixed effects estimation
c. Ordinary least squares estimation
d. Weighted Least squares estimation

Answer: b
Difficulty: Easy
Bloom’s: Application
A-Head: The Correlated Random Effects Approach
BUSPROG: Analytic
Feedback: Fixed effects estimation is the most suitable method, if income is
correlated with the unobserved family effect. The key requirement for using the
random effects estimation is that income is uncorrelated with the unobserved family
effect and ordinary least squares estimation will provide unbiased estimators if
income is uncorrelated with the unobserved family effect.

15. Which of the following statements is true?

a. Fixed effects estimation is not suitable when the unobserved cluster effect
is correlated with one or more explanatory variables.
b. Fixed effects approach is not applicable if the key explanatory variables
change only at the level of the cluster.
c. The ordinary least squares standard errors are incorrect when there is
cluster effect.

Downloaded by K61.FTU ?INH H?NG NG?C

d. Random effects estimation can be applied to a cluster sample only if the
unobserved cluster effect is correlated with one or more explanatory
variables.

Answer: c
Difficulty: Easy
Bloom’s: Knowledge
A-Head: The Correlated Random Effects Approach
BUSPROG:
Feedback: The ordinary least squares standard errors are incorrect when there is
cluster effect.

16. A data set is called an unbalanced panel if it has missing years for at least
some cross-sectional units in the sample.

Answer: True
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Fixed Effects Estimation
BUSPROG:
Feedback: A data set is called an unbalanced panel if it has missing years for at
least some cross-sectional units in the sample.

17. In a random effects model, we assume that the unobserved effect is

correlated with each explanatory variable.

Answer: False
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Random Effects Models
BUSPROG:
Feedback: In a random effects model, we assume that the unobserved effect is
uncorrelated with each explanatory variable.

18. The value of the estimated transformation parameter in generalized least

square estimation that eliminates serial correlation in error terms indicates whether
the estimates are likely to be closer to the pooled OLS or the fixed effects
estimates.

Answer: True
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Random Effects Models
BUSPROG:
Feedback: The value of the estimated transformation parameter in generalized least
squares estimation that eliminates serial correlation in error terms indicates

Downloaded by K61.FTU ?INH H?NG NG?C

whether the estimates are likely to be closer to the pooled OLS or the fixed effects
estimates.

19. The correlated random effects approach cannot be applied to models with
many time-varying explanatory variables.

Answer: False
Difficulty: Easy
Bloom’s: Knowledge
A-Head: The Correlated Random Effects Approach
BUSPROG:
Feedback: The correlated random effects approach can be applied to models with
many time-varying explanatory variables.

20. Pooled ordinary least squares estimation is commonly applied to cluster

samples when eliminating a cluster effect via fixed effects is infeasible or
undesirable.

Answer: True
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Applying Panel Data Methods to Other Data Structures
BUSPROG:
Feedback: Pooled ordinary least squares estimation is commonly applied to cluster
samples when eliminating a cluster effect via fixed effects is infeasible or
undesirable.

Chapter 15

1. Consider the following simple regression model: y = β0 + β1x1 + u. In order to

obtain consistent estimators of β0 and β1, when x and u are correlated, a new
variable z is introduced into the model which satisfies the following two
conditions: Cov(z,x)≠0 and Cov (z,u)=0. The variable z is called a(n) variable.
a. dummy
b. instrumental
c. lagged dependent variable
d. random

Answer: b
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Motivation: Omitted Variables in a Simple Regression Model
BUSPROG:

Downloaded by K61.FTU ?INH H?NG NG?C

Feedback: Variable z is called an instrumental variable since Cov(z,x)≠0 and
Cov(z,u)=0.

2. Consider the following simple regression model: y = β0 + β1x1 + u. Suppose z is

an instrument for x. Which of the following conditions denotes instrument
exogeneity?
a. Cov(z,u) > 0
b. Cov(z,x) > 0
c. Cov(z,u) = 0
d. Cov(z,x) = 0

Answer: c
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Motivation: Omitted Variables in a Simple Regression Model
BUSPROG:
Feedback: The condition Cov(z,u) = 0 denotes instrument exogeneity in this case.

3. Consider the following simple regression model y=β 0 + β1x1 + u. Suppose z is

an instrument for x. Which of the following conditions denotes instrument
relevance?
a. Cov(z,u) > 0
b. Cov(z,u) < 0
c. Cov(z,x) ≠ 0
d. Cov(z,x)=0

Answer: c
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Motivation: Omitted Variables in a Simple Regression Model
BUSPROG:
Feedback: The condition Cov(z,x) ≠ 0 denotes instrument relevance.

4. Consider the following simple regression model y=β 0 + β1x1 + u. Suppose z is

an instrument for x.
Which of the following statements is true?
a. The condition Cov(z,u) = 0 can be tested statistically.
b. The condition Cov(z,x) ≠ 0 cannot be tested statistically.
c. The instrumental variables estimator is always biased if Cov(x,u)≠0.
d. The ordinary least squares estimator is unbiased if Cov(x,u)≠0.

Answer: c
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Motivation: Omitted Variables in a Simple Regression Model
BUSPROG:

Downloaded by K61.FTU ?INH H?NG NG?C

 Feedback: The instrumental variables estimator is always biased if Cov(x,u)≠0.

5. Consider the following simple regression model y=β 0 + β1x1 + u. Suppose z is an

instrument for x. if Cov(z,u) = 0 and Cov(z,x) ≠ 0, the value of β1 in terms of
population covariances is .
Cov ( z , y )
a.
Cov ( z , x)

Cov( z , u)
b.
Cov ( z , x
)
c.
Cov(z,u)
d.
Cov(z,x)

Answer: a
Difficulty:
Easy
Bloom’s: Knowledge
A-Head: Motivation: Omitted Variables in a Simple Regression Model
BUSPROG:
Cov ( z , y )
Feedback: The value of β1 is Cov ( z , x) .

6. The sampling variance for the instrumental variables (IV) estimator is larger
than the variance for the ordinary least square estimators (OLS) because .
a. R2 >1
b. R2 <0
c. R2 =1
d. R2 <1

Answer: d
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Motivation: Omitted Variables in a Simple Regression Model
BUSPROG:
Feedback: The variance of the OLS estimator differs from the comparable formula
for the IV estimator
in that R2 appears in the denominator of the IV variance. Because an R-squared is
always less than
one, the IV variance is always larger than the OLS variance.

7. Consider the following simple regression model y=β 0 + β1x1 + u. The variable z is
a poor instrument for x if .
a. there is a high correlation between z and x

Downloaded by K61.FTU ?INH H?NG NG?C

b. there is a low correlation between z and x
c. there is a high correlation between z and u
d. there is a low correlation between z and u

Answer: b
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Motivation: Omitted Variables in a Simple Regression Model
BUSPROG:
Feedback: The variable z is a poor instrument for x if there is a low correlation
between z and x.

8. Which of the following assumptions is known as exclusion restrictions?

a. The assumption that an instrumental variable is excluded from a
regression model and is correlated with the error term
b. The assumption that an instrumental variable is excluded from a
regression model and correlated with an exogenous explanatory variable
c. The assumption that an exogenous explanatory variable is excluded from
a regression model and is uncorrelated with the error term
d. The assumption that an endogenous explanatory variable excluded from
a regression model and is uncorrelated with the error term

Answer: c
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Two Stage Least Squares
BUSPROG:
Feedback: The assumption that an exogenous explanatory variable is excluded from
a regression model and is uncorrelated with the error term.

9. Which of the following assumptions is required for two-stage least

squares estimation method?
a. There are perfect linear relationships among the instrumental variables.
b. There is strong correlation between each instrumental variable and the
error term.
c. The conditional variance of the error term depends on an exogenous
explanatory variable.
d. The error term has zero mean.

Answer: d
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Two Stage Least Squares
BUSPROG:

Downloaded by K61.FTU ?INH H?NG NG?C

Feedback: The error term has zero mean.

10. Which of the following is true of two stage least squares estimators?
a. The two stage least squares estimator is equal to the instrumental
variable estimator if R2 is equal to 1.
b. The two stage least squares estimators are biased if the regression
model exhibits multicollinearity.
c. The two stage least squares estimators have lower variance than the
ordinary least squares estimators.
d. The two stage least squares estimators have large standard errors when R2 lies
close to 0.

Answer: b
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Two Stage Least Squares
BUSPROG:
Feedback: The two stage least squares estimators are biased if the regression
model exhibits multicollinearity.

11. The necessary condition for identification of an equation is called the .

a. order condition
b. rank condition
c. condition of instrumental exogeneity
d. the condition of instrumental relevance.

Answer: a
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Two Stage Least Squares
BUSPROG:
Feedback: The necessary condition for identification of an equation is called the
order condition.

12. The order condition for identification of an equation requires that there
should be .
a. at least one exogenous explanatory variable in a structural equation
b. at least as many excluded exogenous explanatory variables as there are
included endogenous explanatory variables
c. at least as many dummy variables in an equation as there are
exogenous explanatory variables
d. as many lagged independent variables in an equation as there are
exogenous explanatory variables

Downloaded by K61.FTU ?INH H?NG NG?C

Answer: b
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Two Stage Least Squares
BUSPROG:
Feedback: The order condition for identification of an equation requires that there
should be at least as many excluded exogenous explanatory variables as there are
included endogenous explanatory variables.

13. The procedure of comparing different instrumental variables estimates of

the same parameter is an example of testing .
a. overidentifying restrictions
b. endogeneity
c. heteroskedasticity
d. serial correlation

Answer: a
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Testing for Endogeneity and Testing Overidentifying Restrictions
BUSPROG:
Feedback: The procedure of comparing different instrumental variables estimates of
the same parameter is an example of testing overidentifying restrictions.

14. The test for overidentifying restrictions is valid if .

a. the regression model exhibits heteroskedasticity
b. the regression model exhibits homoskedasticity
c. the number of instrumental variables are less than the number of
endogenous explanatory variables
d. the number of instrumental variables are just enough for obtaining
consistent estimators.

Answer: b
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Testing for Endogeneity and Testing Overidentifying Restrictions
BUSPROG:
Feedback: The test for overidentifying restrictions is valid if the regression model
exhibits homoskedasticity.

15. Which of the following assumptions is required for two stage least
squares estimation with time series data but not required for two-stage least
squares estimation with cross sectional data?
a. The conditional mean of the error term is zero.

Downloaded by K61.FTU ?INH H?NG NG?C

b. The error term has constant conditional variance.
c. The model includes at least one dummy variable.
d. The error terms are not serially correlated.

Answer: d
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Applying 2SLS to Time Series Equations
BUSPROG:
Feedback: The additional assumption required for two stage least squares
estimation using time-series data is that there is no serial correlation.

16. If the instrumental variable estimator has an upward bias, the ordinary
least square estimator always has a downward bias.

Answer: False
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Motivation: Omitted Variables in a Simple Regression Model
BUSPROG:
Feedback: It is possible for the directions of the asymptotic biases to be different for
IV and OLS but this situation is usually rare in practice.

17. Instrumental variables cannot be used for estimating a regression equation

if the regression model suffers from the measurement error problem.

Answer: False
Difficulty: Easy
Bloom’s: Knowledge
A-Head: IV Solutions to Errors-in-Variables Problems
BUSPROG:
Feedback: The instrumental variables procedure can be used for estimation if the
regression model suffers from the measurement error problem.

18. The two stage least squares estimator is less efficient than the ordinary
least squares estimator when the explanatory variables are exogenous.

Answer: True
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Testing for Endogeneity and Testing Overidentifying Restrictions
BUSPROG:
Feedback: The two stage least squares estimator is less efficient than the ordinary
least squares estimator when the explanatory variables are exogenous.

Downloaded by K61.FTU ?INH H?NG NG?C

19. Increasing the number of overidentifying restrictions can cause severe biases
in two stage least squares estimators.

Answer: True
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Testing for Endogeneity and Testing Overidentifying Restrictions
BUSPROG:
Feedback: Increasing the number of overidentifying restrictions can cause severe
biases in two stage least squares estimators.

20. Two stage least squares estimation cannot be applied to a panel data set.

Answer: False
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Applying 2SLS to Pooled Cross Sections and Panel Data
BUSPROG:
Feedback: Two stage least squares estimation can be applied to a panel data set.
Chapter 16

1. In econometrics, simultaneity arises when:

a. strictly exogenous explanatory variables determine the dependent
variable through a step-by-step process.
b. the error term is correlated with both the dependent variable and
explanatory variables.
c. one or more of the explanatory variables is jointly determined with the
dependent variable.
d. both serial correlation and heteroskedasticity are present in an
hypothesized model.

Answer: c
Difficulty:
Moderate
Bloom’s: Knowledge
A-Head: The Nature of Simultaneous Equations Models
BUSPROG:
Feedback: In econometrics, simultaneity arises when one or more of the explanatory
variables is jointly determined with the dependent variable, typically through an
equilibrium process.

2. The following simultaneous equations describe the demand and supply for
a particular good in a competitive market.
Qi = α1Pi + β1zi1 + ui1

Downloaded by K61.FTU ?INH H?NG NG?C

Qi = α2Pi + β2zi2 + ui2
Which of the following are the endogenous variables in this model?
a. Pi, zi1, and zi2
b. Pi and Qi
c. zi1, and zi2
d. ui1 and ui2

Answer: b
Difficulty: Easy
Bloom’s: Knowledge
A-Head: The Nature of Simultaneous Equations Models
BUSPROG:
Feedback: Pi and Qi are the endogenous variables in the given simultaneous
equation model.

3. Which of the following correctly identifies a characteristic of structural equations?

a. A structural equation should contain equal number of dependent
and independent variables.
b. A structural equation should contain equal number of endogenous and
exogenous variables.
c. A structural equation should have a behavioral, ceteris paribus interpretation
on its own.
d. A structural equation should not contain structural errors.

Answer: c
Difficulty: Easy
Bloom’s: Knowledge
A-Head: The Nature of Simultaneous Equations Models
BUSPROG:
Feedback: A structural equation should have a behavioral, ceteris paribus
interpretation on its own.

4. The following equations represent a simultaneous equations

model: K1 = α1K2 + β1z1 + u1
K2 = α2K1 + β2z2 + u2
The reduced form equation for K2 will express:
a. K1 as a function of K2 and the error terms.
b. K2 as a function of K1 and the error terms.
c. K2 as a function of exogenous variables and the error terms.
d. K1 as a function of exogenous variables, the error terms, and K2.

Answer: c
Difficulty:
Moderate
Bloom’s: Knowledge

Downloaded by K61.FTU ?INH H?NG NG?C

A-Head: Simultaneity Bias in OLS
BUSPROG:
Feedback: For the given simultaneous equations model, the reduced form equation
for K2 will express K2 as a function of exogenous variables and the error terms.

5. The following equations represent a simultaneous equations

model: K1 = α1K2 + β1z1 + u1
K2 = α2K1 + β2z2 + u2
OLS will suffer from simultaneity bias if:
a. u1 is correlated with z1.
b. z1 is correlated with z2.
c. K2 is correlated with u1.
d. K1 is correlated with u1.

Answer: c
Difficulty:
Moderate
Bloom’s: Knowledge
A-Head: Simultaneity Bias in OLS
BUSPROG:
Feedback: In the given simultaneous equation model, OLS will suffer from
simultaneity bias if K2 is correlated with u1.

6. Refer to the simultaneous equations model above. The reduced form error
from the reduced form equation for K2 will be a:
a. quadratic function of u1 and u2, and correlated with z1 and z2.
b. quadratic function of u1 and u2, and uncorrelated with z1 and z2.
c. linear function of u1 and u2, and correlated with z1 and z2.
d. linear function of u1 and u2, and uncorrelated with z1 and z2.

Answer: d
Difficulty:
Moderate
Bloom’s: Knowledge
A-Head: Simultaneity Bias in OLS
BUSPROG:
Feedback: The reduced form error from the reduced form equation for K 2 will be a
linear function of u1 and u2, and uncorrelated with z1 and z2.

7. Exclusion restrictions are said to be imposed in a two-equation

simultaneous equations model if it is assumed that:
a. certain exogenous variables do not appear in the first equation and others
are absent from the second equation.
b. certain endogenous variables do not appear in the first equation and others
are absent from the second equation.
c. the error terms in each equation is uncorrelated with the exogenous variables.

Downloaded by K61.FTU ?INH H?NG NG?C

b. the error terms in each equation is correlated with the exogenous variables.

Answer: a
Difficulty:
Moderate
Bloom’s: Knowledge
A-Head: Identifying and Estimating a Structural Equation
BUSPROG:
Feedback: Exclusion restrictions are said to be imposed in a two-equation
simultaneous equations model if it is assumed that certain exogenous variables do
not appear in the first equation and others are absent from the second equation.

8. Which of the following is a method which can be used for estimation

in simultaneous equations models?
a. Feasible generalized least squares estimation
b. Prais-Winsten transformation
c. Cochrane-Orcutt transformation
d. Two stage least squares estimation

Answer: d
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Identifying and Estimating a Structural Equation
BUSPROG:
Feedback: The two stage least squares estimation can be used in simultaneous
equations models.

9. The rank condition for identification of a structural equation states that the first
equation in a two-equation simultaneous equations model is identified if, and only
if:
a. the first equation contains at least one exogenous variable (with a
nonzero coefficient) that is excluded from the second equation.
b. the first equation contains at least two exogenous variables (with a
nonzero coefficient) that are excluded from the second equation.
c. the second equation contains at least one exogenous variable (with a
nonzero coefficient) that is excluded from the first equation.
d. the second equation contains at least two exogenous variables (with a
nonzero coefficient) that are excluded from the first equation.

Answer: c
Difficulty:
Moderate
Bloom’s: Knowledge
A-Head: Identifying and Estimating a Structural Equation
BUSPROG:
Feedback: The rank condition for identification of a structural equation states that
the first equation in a two-equations simultaneous equations model is identified if,

Downloaded by K61.FTU ?INH H?NG NG?C

and only if, the second equation contains at least one exogenous variable (with a
nonzero coefficient) that is excluded from the first equation.

10. Which of the following statements is true?

a. The use of 2SLS is applicable to simultaneous equations model with at most
two equations.
b. Identification of simultaneous equations with three or more equations is based
on matrix algebra.
c. 2SLS method used in the estimation of simultaneous equations is an example of
a system estimation method.
d. The maximum number of equations permissible for a simultaneous
equations model is four.

Answer: b
Difficulty:
Moderate
Bloom’s: Knowledge
A-Head: Systems with More than Two Equations
BUSPROG:
Feedback: Identification of simultaneous equations with three or more equations is
based on matrix algebra.

11. An equation in the simultaneous equations model satisfies the order

condition for identification if:
a. the number of excluded endogenous variables from the equation is at most
as large as the number of right-hand side exogenous variables.
b. the number of excluded endogenous variables from the equation is at least
as large as the number of right-hand side exogenous variables.
c. the number of excluded exogenous variables from the equation is at most
as large as the number of right-hand side endogenous variables.
d. the number of excluded exogenous variables from the equation is at least
as large as the number of right-hand side endogenous variables.

Answer: d
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Systems with More than Two Equations
BUSPROG:
Feedback: An equation in the simultaneous equations model satisfies the order
condition for identification if the number of excluded exogenous variables from the
equation is at least as large as the number of right-hand side endogenous variables.

12. A predetermined variable in a simultaneous equations model is:

a. a constant.
b. the error term.

Downloaded by K61.FTU ?INH H?NG NG?C

c. a lagged variable.
d. an omitted variable.

Answer: c
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Simultaneous Equations Models with Time Series
BUSPROG:
Feedback: A predetermined variable in a simultaneous equations model is a lagged
variable.

13. Which of the following correctly highlights a limitation of applying

simultaneous equation models to time series data?
a. Most time series have variables with a unit root and 2SLS is complicated
when applied to equations with such variables.
b. 2SLS estimates are inefficient when applied to variables that are not in
their levels but in first differences.
c. It is difficult to form simultaneous equations which satisfy the rank and
order conditions using time series data.
d. The problem of serial correlation greatly limits the efficiency of
simultaneous equation models and leads to inefficient estimations.

Answer: a
Difficulty:
Moderate
Bloom’s: Knowledge
A-Head: Simultaneous Equations Models with Time Series
BUSPROG:
Feedback: Most time series have variables with a unit root and 2SLS is complicated
when applied to equations with such variables.

14. An alternative to using simultaneous equation models with panel data is:
a. to use OLS estimates after first differencing the data.
b. to use fixed effects transformation on the equations and then apply 2SLS.
c. to convert the equations into reduced form and then apply feasible
generalized least squares.
d. to convert the equations into reduced form and then apply OLS.

Answer: b
Difficulty:
Moderate
Bloom’s: Knowledge
A-Head: Simultaneous Equations Model with Panel Data
BUSPROG:
Feedback: An alternative to using simultaneous equation models with panel data is
to use fixed effects transformation on the equations and then apply 2SLS.

Downloaded by K61.FTU ?INH H?NG NG?C

15. Which of the following statements is true?
a. White test can be used to efficiently determine the presence of serial
correlation in panel data.
b. The t statistic is not an efficient test to determine serial correlation in panel data.
c. Instrumental variables for both endogenous and exogenous variables are
required for estimating simultaneous equation models concerned with panel data.
d. 2SLS should be applied to simultaneous equation models with panel data
only after removing the unobserved effects from the equations of interest.

Answer: d
Difficulty:
Moderate
Bloom’s: Knowledge
A-Head: Simultaneous Equations Model with Panel Data
BUSPROG:
Feedback: 2SLS should be applied to simultaneous equation models with panel data
only after removing the unobserved effects from the equations of interest.

16. A simultaneous equations model is suitable whenever two variables

are determined simultaneously.

Answer: False
Difficulty: Easy
Bloom’s: Knowledge
A-Head: The Nature of Simultaneous Equations Models
BUSPROG:
Feedback: Just because two variables are determined simultaneously, it does not
imply that a simultaneous equations model is suitable. The criteria for using a
simultaneous equations model is that each equation in the model should make
sense in isolation from the other equation.

17. OLS is biased and inconsistent when applied to a structural equation in

a simultaneous equations system.

Answer: True
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Simultaneity Bias in OLS
BUSPROG:
Feedback: OLS is biased and inconsistent when applied to a structural equation in a
simultaneous equations system.

18. The instrumental variables in the two stage least squares estimation
method consists of endogenous variables appearing in either equation.

Downloaded by K61.FTU ?INH H?NG NG?C

Answer: False
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Identifying and Estimating a Structural Equation
BUSPROG:
Feedback: The instrumental variables in the two stage least squares method
consists of exogenous variables appearing in either equation.

19. The order condition is a necessary and sufficient condition for identification
of an equation in a simultaneous equations model.

Answer: False.
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Systems with More than Two Equations
BUSPROG:
Feedback: The order condition is a necessary condition for identification of an
equation in a simultaneous equations model. It is not a sufficient condition
for identification.

20. If a structured model contains a time trend—which may capture

exogenous, trending factors that are not directly modeled—then the trend acts
as its own instrumental variable.

Answer: True
Difficulty: Moderate
Bloom’s: Knowledge
A-Head: Simultaneous Equations Models with Time Series
BUSPROG:
Feedback: If a structured model contains a time trend, then the trend acts as its
own instrumental variable.

Chapter 17

1. Which of the following is an example of a binary response model?

a. MA model
b. ARCH model
c. GARCH model
d. Logit model

Answer: d
Difficulty: Easy
Bloom’s: Knowledge

Downloaded by K61.FTU ?INH H?NG NG?C

A-Head: Specifying Logit and Probit Models
BUSPROG:
Feedback: The logit model is an example of a binary response model.

2. The model: G(z) = [exp(z)]/[1 + exp(z)],where G is between zero and one for
all real numbers ‘z’, represents a:
a. logit model.
b. probit model.
c. Tobit model.
d. linear probability model.

Answer: a
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Specifying Logit and Probit Models
BUSPROG:
Feedback: The following model: G(z) = [exp(z)]/[1 + exp(z)], where G is between
zero and one for all real numbers ‘z’, represents a logit model.

3. The model: G(z) = ∫ ø ( v ) dv where ø(z) = (2π)-1/2exp(-z2/2) represents a:

−∞

a. Tobit model.
b. logit model.
c. probit model.
d. linear probability model.

Answer: b
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Specifying Logit and Probit Models
BUSPROG:
z

Feedback: The following model G(z) = ∫ ø ( v ) dv where ø(z) = (2π)-1/2exp(-

z2/2)
−∞

represents a probit model.

4. The likelihood ratio statistic is given by:

a. LR = (log-likelihoodunrestricted + log-likelihoodrestricted)
b. LR = 2 × (log-likelihoodunrestricted + log-likelihoodrestricted)
c. LR = (log-likelihoodunrestricted – log-likelihoodrestricted)
d. LR = 2 × (log-likelihoodunrestricted – log-likelihoodrestricted)

Downloaded by K61.FTU ?INH H?NG NG?C

Answer: d
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Specifying Logit and Probit Models
BUSPROG:
Feedback: The likelihood ratio statistic is given by LR = 2 × (log-
likelihoodunrestricted – log-likelihoodrestricted)

5. The model is designed to model corner solution dependent variables.

a. linear probability
b. logit
c. probit
d. Tobit

Answer: d
Difficulty: Easy
Bloom’s: Knowledge
A-Head: The Tobit Model for Corner Solution Responses
BUSPROG:
Feedback: The Tobit model is designed to model corner solution dependent
variables.

6. The model: y* = β0 + xβ + u, given u|x ~ Normal(0, σ2) and y = max(0, y*)

represents a:
a. ARCH model.
b. GARCH model
c. Tobit model
d. logit model

Answer: c
Difficulty: Easy
Bloom’s: Knowledge
A-Head: The Tobit Model for Corner Solution Responses
BUSPROG:
Feedback: The model: y* = β0 + xβ + u, given u|x ~ Normal(0, σ2) and y = max(0,
y*) represents a Tobit model.

7. Which of the following tests can be used to test hypotheses with

multiple restrictions under a Tobit model?
a. White test
b. Wald test
c. Dickey Fuller test
d. Durbin Watson test

Downloaded by K61.FTU ?INH H?NG NG?C

Answer: b
Difficulty: Easy
Bloom’s: Knowledge
A-Head: The Tobit Model for Corner Solution Responses
BUSPROG:
Feedback: The Wald test can be used to check for multiple restrictions under a Tobit
model.

8. A count variable refers to a dependent variable that can take on:

a. nonnegative integer values.
b. nonnegative fractional values.
c. negative fractional values.
d. negative integer values.

Answer: a
Difficulty: Easy
Bloom’s: Knowledge
A-Head: The Poisson Regression Model
BUSPROG:
Feedback: A count variable refers to a dependent variable that can take on
nonnegative integer values.

9. Which of the following statements is true?

a. Taking logarithmic of a count variable is a suitable way to model it.
b. All standard count data distributions exhibit heteroskedasticity.
c. The nonlinear least squares estimation aims at maximizing R2.
d. Count variables cannot take on the value zero.

Answer: b
Difficulty:
Moderate
Bloom’s: Knowledge
A-Head: The Poisson Regression Model
BUSPROG:
Feedback: All standard count data distributions exhibit heteroskedasticity.

10. The nominal distribution for count data is the:

a. binomial distribution.
b. normal distribution.
c. Poisson distribution.
d. Bernoulli distribution.

Answer: c
Difficulty:
Moderate
Bloom’s: Knowledge

Downloaded by K61.FTU ?INH H?NG NG?C

A-Head: The Poisson Regression Model
BUSPROG:
Feedback: The nominal distribution for count data is the Poisson distribution.

11. Which of the following statements is true?

a. A probit or logit model should be used for corner solution outcomes, and
a Poisson regression model should be used for a binary response.
b. A Poisson regression model should be used for corner solution outcomes, and
a probit or logit model should be used for a binary response.
c. A probit or logit model should be used for count variables, and a
Poisson regression model should be used for a binary response.
d. A Poisson regression model should be used for count variables, and a probit
or logit model should be used for a binary response.

Answer: d
Difficulty:
Moderate
Bloom’s: Knowledge
A-Head: Censored and Truncated Regression Models
BUSPROG:
Feedback: A Tobit model should be used for corner solution outcomes, a
Poisson regression model should be used for count variables, and a probit or
logit model should be used for a binary response.

12. Which of the following statements is true?

a. OLS estimates in censored regression models are consistent estimators of
the population coefficients.
b. In a truncated regression model, the samples are not included randomly from
an underlying population but are based on a given rule.
c. In a censored regression model, units in the sample are taken from a
particular subset of the population.
d. Maximum likelihood estimators are consistent in truncated regression
models even if there is nonnormality or heteroskedasticity in the error terms.

Answer: b
Difficulty:
Moderate
Bloom’s: Knowledge
A-Head: Censored and Truncated Regression Models
BUSPROG:
Feedback: In a truncated regression model, we do not have a random sample from
the underlying population, but we know the rule that was used to include units in
the sample. This rule is determined by whether the dependent variable is above or
below a certain threshold.

13. Duration is a variable that measures:

Downloaded by K61.FTU ?INH H?NG NG?C

a. the time when a certain event occurs.
b. the time before a certain event occurs.
c. the time after a certain event occurs.
d. the appropriate number of lags for a regression model.

Answer: b
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Censored and Truncated Regression Models
BUSPROG:
Feedback: Duration is a variable that measures the time before a certain event
occurs.

14. Which of the following statements is true?

a. A truncated regression is a special case of a random sample selection.
b. Nonrandom sample selection can arise in cases of cross-sectional and time
series data, but not in the case of panel data.
c. The Tobit regression model is based on endogenous sample selection.
d. The censored regression model is based on nonrandom sample selection.

Answer: c
Difficulty:
Moderate
Bloom’s: Knowledge
A-Head: Sample Selection Corrections
BUSPROG:
Feedback: The Tobit regression model is based on endogenous sample selection.

15. Which of the following is a method to correct for sample selection bias for
the problem of incidental truncation?
a. Vector error correction method
b. First differencing method
c. Heckman’s method
d. Johansen method

Answer: c
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Sample Selection Corrections
BUSPROG:
Feedback: Heckman’s method can be used for correcting sample selection bias for
the problem of incidental truncation.

16. The cumulative distribution function for a standard logistic random variable is
a decreasing function.

Downloaded by K61.FTU ?INH H?NG NG?C

Answer: False
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Specifying Logit and Probit Models
BUSPROG:
Feedback: The cumulative distribution function for a standard logistic random
variable is an increasing function.

17. The Tobit model relies crucially on normality and heteroskedasticity in

the underlying latent variable model.

Answer: False
Difficulty: Easy
Bloom’s: Knowledge
A-Head: The Tobit Model for Corner Solution
Responses BUSPROG:
Feedback: The Tobit model relies crucially on normality and homoskedasticity in the
underlying latent variable model. Under heteroskedasticity, using a Tobit model is
inefficient.

18. In the Poisson regression model, the probability distribution is given by P(y =
h| x) = exp[-exp(xβ)][exp(xβ)]h/h!, h = 0, 1, …..

Answer: True
Difficulty: Easy
Bloom’s: Knowledge
A-Head: The Poisson Regression Model
BUSPROG:
Feedback: In the Poisson regression model, the probability distribution is given by
P(y = h|x) = exp[-exp(xβ)][exp(xβ)]h/h!, h = 0, 1, …..

19. When a variable is top coded, its value is known only up to a certain threshold.

Answer: True
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Censored and Truncated Regression Models
BUSPROG:
Feedback: When a variable is top coded, its value is known only up to a certain
threshold. For responses greater than the threshold, it is only known that the
variable is at least as large as the threshold.

20. In case of endogenous sample selection, OLS is unbiased but consistent.

Downloaded by K61.FTU ?INH H?NG NG?C

Answer: False
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Sample Selection Corrections
BUSPROG:
Feedback: In case of endogenous sample selection, OLS is unbiased and
inconsistent.
Chapter 18

1. Let {(yt, zt): t = …, 2, 1, 0, 1, 2, …} be a bivariate time series process.

The model: yt = α + βozt + β1zt – 1 + β2zt – 2 + ….. + ut, where t = …..,-2,-
1,0,1,2,……, represents a(n):
a. moving average model.
b. ARIMA model.
c. finite distributed lag model.
d. infinite distributed lag model.

Answer: d
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Infinite Distributed Lag Models
BUSPROG:
Feedback: The model: yt = α + βozt + β1zt – 1 + β2zt – 2 + ….. + ut, where t = 0,1,2,
……, represents an infinite distributed lag model relating yt to all current and past
values of z.

2. The Koyck distributed lag model is an example of:

a. a moving average model.
b. an autoregressive conditional heteroskedasticity model.
c. an infinite distributed lag model.
d. a finite distributed lag model.

Answer: c
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Infinite Distributed Lag
Models BUSPROG:
Feedback: The Koyck distributed lag model is an example of an infinite distributed
lag model.

3. The model: yt = α0 + γ0zt +ρyt – 1 + γ1zt – 1 +vt, where vt = ut – ρut – 1 represents

a:
a. finite distributed lag model.
b. simultaneous equations model.

Downloaded by K61.FTU ?INH H?NG NG?C

c. rational distributed lag model.
d. vector error correction model.

Answer: c
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Infinite Distributed Lag Models
BUSPROG:
Feedback: The model: yt = α0 + γ0zt +ρyt – 1 + γ1zt – 1 +vt, where vt = ut –
ρut – 1 - represents a rational distributed lag model.

4. In the given AR(1) model, yt = α + ρyt – 1, t = 1,2…… , the Dickey-Fuller

distribution refers to the:
a. asymptotic distribution of the t statistic under the hypothesis ρ – 1 = 0.
b. asymptotic distribution of the F statistic under the hypothesis ρ – 1 = 0.
c. asymptotic distribution of the χ2 statistic under the hypothesis ρ – 1 = 0.
d. asymptotic distribution of the z statistic under the hypothesis ρ – 1 = 0.

Answer: a
Difficulty:
Moderate
Bloom’s: Knowledge
A-Head: Testing for Unit Roots
BUSPROG:
Feedback: In the given model: yt = α + ρyt – 1, t = 1,2…… , the Dickey-
Fuller distribution refers to the asymptotic distribution of the t statistic
under the hypothesis ρ – 1 = 0.

5. Which of the following is used to test whether a time series follows a unit
root process?
a. Wald test
b. White test
c. Augmented Dickey-Fuller test
d. Johansen test

Answer: c
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Testing for Unit Roots
BUSPROG:
Feedback: The augmented Dickey-Fuller test can be used to check for unit root in a
time series

6. A spurious correlation refers to a situation where:

a. two variables are related through their correlation with a third variable.

Downloaded by K61.FTU ?INH H?NG NG?C

b. the correlation coefficient between two variables cannot be estimated.
c. there is direct causal relationship between two variables but tests for
correlations reject this relationship.
d. the correlation between two variables is positive till the sample size reaches
a threshold, and negative after the sample size crosses the threshold.

Answer: a
Difficulty:
Moderate
Bloom’s: Knowledge
A-Head: Spurious Regression
BUSPROG:
Feedback: A spurious correlation refers to a situation where two variables are
related through their correlation with a third variable.

7. A spurious regression refers to a situation where:

a. the direction of the relationship between the dependent variable and
the explanatory variables is uncertain.
b. even though two variables are independent, the OLS regression of one
variable on the other indicates a relationship between them.
c. a few important and necessary explanatory variables are left out of a
regression equation, thus leading to inefficient and inconsistent forecasts.
d. at least one of the variables used in a regression equation does not have a
unit root and the error terms are heteroskedastic.

Answer: b
Difficulty:
Moderate
Bloom’s: Knowledge
A-Head: Spurious Regression
BUSPROG:
Feedback: A spurious regression refers to a situation where even though two
variables are independent, the OLS regression of one variable on the other indicates
a relationship between them.

8. Which of the following statements is true of spurious regressions?

a. The OLS estimates of the population parameters are efficient and unbiased
and the t statistic is valid.
b. Even if the explanatory variables and the dependent variable are
independent times series processes, the R2 can large.
c. Spurious regressions are limited to I(0) processes, and are not possible in case
of I(1) processes.
d. Spurious regressions are limited to I(1) processes, and are not possible in case
of I(0) processes.

Answer: b

Downloaded by K61.FTU ?INH H?NG NG?C

Difficulty: Moderate
Bloom’s: Knowledge
A-Head: Spurious Regression
BUSPROG:
Feedback: In case of a spurious regression, even if the explanatory variables and
the dependent variable are independent times series processes the R 2 can large.

9. Two series are said to be cointegrated if:

a. both series are I(1) but a linear combination of them is I(0).
b. both series are I(0) but a linear combination of them is I(1).
c. both series have the same set of explanatory variables but a different
dependent variable.
d. both series have the same dependent variable but a different set of
explanatory variables.

Answer: b
Difficulty:
Moderate
Bloom’s: Knowledge
A-Head: Cointegration and Error Correction Models
BUSPROG:
Feedback: Two series are said to be cointegrated if both series are I(0) but a linear
combination of them is I(1).

10. Which of the following tests can be used to check for cointegration between
two series?
a. Wald test
b. Breush-Pagan test
c. White test
d. Engle-Granger test

Answer: d
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Cointegration and Error Correction Models
BUSPROG:
Feedback: The Engle-Granger test can be used to check for cointegration between
two series.

11. Which of the following statements is true?

a. The calculated t statistic is valid and efficient in case of a spurious regression.
b. If an explanatory variable or a dependent variable is integrated of the order
one, the OLS estimators are asymptotically normally distributed.

Downloaded by K61.FTU ?INH H?NG NG?C

c. An error correction model can be used to study the short-run dynamics in
the relationship between the dependent variable and the explanatory variables
in a time series model.
d. The Dickey-Fuller test can be used to test for heteroskedasticity in the
error terms.

Answer: c
Difficulty:
Moderate
Bloom’s: Knowledge
A-Head: Cointegration and Error Correction Models
BUSPROG:
Feedback: An error correction model can be used to study the short-run dynamics in
the relationship between the dependent variable and the explanatory variables.

12. If ft denotes the forecast of yt+1 made at time t, then the forecast error is given
by:
a. et+1 = ft/yt+1.
b. et+1 = yt+1/ft.
c. et+1 = yt+1 + ft.
d. et+1 = yt+1 – ft.

Answer: d
Difficulty:
Moderate
Bloom’s: Knowledge
A-Head: Forecasting
BUSPROG:
Feedback: If ft denotes the forecast of yt+1 made at time t, then the forecast error
is given by et+1 = yt+1 – ft.

13. Which of the following is true of squared forecast errors?

a. An error of +4 yields a greater loss than an error of -4.
b. An error of -4 yields a greater loss than an error of +4.
c. An error of -4 or +4 yields the same loss.
d. Loss from positive and negative forecast errors cannot be compared.

Answer: c
Difficulty:
Easy
Bloom’s: Application
A-Head: Forecasting
BUSPROG: Analytic
Feedback: In case of squared forecast errors, an error of -2 or +2 yields the same
loss.

Downloaded by K61.FTU ?INH H?NG NG?C

14. Which of the following statements correctly identifies the difference between
an autoregressive model and a vector autoregressive model?
a. In an autoregressive model, the dependent variable is expressed as a function
of its own lag, whereas in a vector autoregressive model, the dependent variable
is expressed as a function of the lag of an explanatory variable.
b. In an autoregressive model, the dependent variable is expressed as a function
of the lag of an explanatory variable, whereas in a vector autoregressive model,
the dependent variable is expressed as a function of its own lag.
c. In an autoregressive model several series are modelled in terms of their own
past, whereas in a vector autoregressive model only one series is modelled in
terms of its own past.
d. In an autoregressive model one series is modelled in terms of its own past,
whereas in a vector autoregressive model several series are modelled in terms
of their past.

Answer: d
Difficulty:
Easy
Bloom’s: Knowledge
A-Head: Forecasting
BUSPROG:
Feedback: In an autoregressive model one series is modelled in terms of its own
past, whereas in a vector autoregressive model several series are modelled in terms
of their past.

15. In case of forecasts, the root mean squared error is the:

a. average of the forecast errors divided by the variance of the errors.
b. average of the absolute forecast errors.
c. standard deviation of the forecast errors without any degrees of
freedom adjustment.
d. standard deviation of the forecast errors with a degrees of freedom adjustment.

Answer: c
Difficulty:
Easy
Bloom’s: Knowledge
A-Head: Forecasting
BUSPROG:
Feedback: In case of forecasts, the root mean squared error is the standard
deviation of the forecast errors without any degrees of freedom adjustment.

16. If the t statistic for the presence of a unit root in a variable is -7.22 and the
5% critical value is -2.86, there is strong evidence against a unit root in the
variable.

Answer: True
Difficulty: Easy

Downloaded by K61.FTU ?INH H?NG NG?C

Bloom’s: Application
A-Head: Testing for Unit Roots
BUSPROG: Analytic
Feedback: Since the absolute value of the t statistic for the presence of a unit root
in a variable is greater than the absolute critical value, there is strong evidence
against a unit root in the variable.

17. The R2 calculated in a spurious regression is a valid and efficient estimate of

the goodness-of-fit of the regression equation.

Answer: False
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Spurious Regression
BUSPROG:
Feedback: The R2 calculated in a spurious regression is not a valid and efficient
estimate of the goodness-of-fit of the regression equation as the calculated value
can be very high even if there is no relationship between the dependent variable
and the explanatory variables.

18. Exponential smoothing is a forecasting method where the weights on the

lagged dependent variable decline to zero exponentially.

Answer: True
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Forecasting
BUSPROG
Feedback: Exponential smoothing is a forecasting method where the weight on the
lagged dependent variable decline to zero exponentially.

19. In calculation of squared forecast errors, an error of +3 yields a loss three

times greater than an error of -1.

Answer: False
Difficulty: Moderate
Bloom’s: Application
A-Head: Forecasting
BUSPROG: Analytic
Feedback: In calculation of squared forecast errors, an error of +3 yields a loss nine
times an error of -1.

20. Vector autoregressive models should be used for forecasting if the series
being studied are cointegrated.

Downloaded by K61.FTU ?INH H?NG NG?C

Answer: False
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Forecasting
BUSPROG:
Feedback: Error correction models should be used for forecasting if the series being
studied are cointegrated.

Chapter 19

1. Which of the following is a characteristic of a good research question?

a. A good research question should have policy implications.
b. A good research question should be constrained to macroeconomics.
c. A good research question should be backed by available information.
d. A good research question should be subjective and not objective.

Answer: c
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Posing a Question
BUSPROG:
Feedback: A good research question should be backed by available information in
the form of data so that it can be answered.

2. Which of the following is a characteristic of a good research paper?

a. A good research paper should contain a review of relevant literature.
b. A good research paper should not have any published precedents.
c. A good research paper should always be useful for policy implications.
d. A good research paper should have more than one author.

Answer: a
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Literature Review
BUSPROG:
Feedback: A good research paper should contain a review of relevant literature.

3. Which of the following statements is true of an appropriate data set?

a. An appropriate data set should be collected from government
registered websites.
b. An appropriate data set should have enough controls to do a reasonable
ceteris paribus analysis.
c. An appropriate data set should not have time series units.

Downloaded by K61.FTU ?INH H?NG NG?C

d. An appropriate data set should not be based on surveys.

Answer: b
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Data Collection
BUSPROG:
Feedback: An appropriate data set should have enough controls to do a ceteris
paribus analysis.

4. The most flexible way to obtain data in electronic form is as a standard file.
a. PDF
b. WMV
c. text (ASCII)
d. PPTX

Answer: c
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Data Collection
BUSPROG:
Feedback: The most flexible way to obtain data in electronic form is as a standard
text (ASCII) file.

5. Time series data should be stored:

a. with the earliest time period listed as the first observation, and the most
recent time period as the last observation.
b. with the earliest time period listed as the last observation, and the most
recent time period as the first observation.
c. with the time period in which the concerned variable takes the highest value
listed as the first observation, and the time period in which the concerned
variable takes the lowest value as the last observation.
d. with the time period in which the concerned variable takes the lowest value
listed as the first observation, and the time period in which the concerned variable
takes the highest value as the last observation.

Answer: a
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Data Collection
BUSPROG:
Feedback: Time series data should be stored with the earliest time period listed as
the first observation, and the most recent time period as the last observation.

Downloaded by K61.FTU ?INH H?NG NG?C

6. Which of the following correctly identifies an advantage of entering data in
a spreadsheet rather than a text file?
a. Data in various websites is more often available as spreadsheets than as
text files.
b. Different observations are more likely to be run together in spreadsheets than
in text files.
c. Spreadsheets are readable on most computers, whereas text files are not.
d. Spreadsheets allow manipulation of data, whereas text files do not.

Answer: d
Difficulty:
Moderate
Bloom’s: Knowledge
A-Head: Data Collection
BUSPROG:
Feedback: Spreadsheets allow manipulation of data such as calculation of averages,
medians, etc. whereas text files do not.

7. Which of the following statements is true?

a. To carry out OLS estimation, it is necessary that the error terms should
be correlated with the explanatory variables.
b. Measurement error and simultaneity are potential sources of endogeneity.
c. While entering data for a regression analysis, qualitative explanatory
variables should be assigned numerical values.
d. AR models are best suited for ordered responses.

Answer: b
Difficulty:
Moderate
Bloom’s: Knowledge
A-Head: Econometric Analysis
BUSPROG:
Feedback: Measurement error and simultaneity are potential sources of
endogeneity.

8. Which of the following statements is true of stepwise regression?

a. In a stepwise regression, the dependent variable is regressed on each
independent variable individually with an attempt to understand the effect of
each independent variable on the dependent variable.
b. In a stepwise regression, the dependent variable is regressed on different
combinations of the independent variable with an attempt to come up with the
best model.
c. In a stepwise regression, the dependent variable can take up binary values
only while the explanatory variables can take up quantitative values.
d. In a stepwise regression, the dependent variable can take up quantitative
values while the explanatory variables can take up only binary values.

Downloaded by K61.FTU ?INH H?NG NG?C

Answer: b
Difficulty:
Moderate
Bloom’s: Knowledge
A-Head: Econometric Analysis
BUSPROG:
Feedback: In a stepwise regression, the dependent variable is regressed on
different combinations of the independent variable with an attempt to come up with
the best model.

9. Which of the following statements is true?

a. Graphs should be avoided while introducing a research paper’s topic.
b. The introduction to a research paper should be longer than the conclusion to
the research paper.
c. The summary of a research paper can be presented in the introduction of
a research paper.
d. Statistics should be avoided in introductions for research papers.

Answer: c
Difficulty:
Moderate
Bloom’s: Knowledge
A-Head: Writing an Empirical Paper
BUSPROG:
Feedback: The summary of a research paper can be presented in the introduction
of a research paper as it helps grab the reader’s attention.

10. Which of the following statements is true?

a. OLS is a model, whereas feasible GLS is an estimation method.
b. OLS is an estimation method, whereas feasible GLS is a model.
c. Both OLS and feasible GLS are models.
d. Both OLS and feasible GLS are estimation methods.

Answer: d
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Writing an Empirical Paper
BUSPROG:
Feedback: Both OLS and feasible GLS are methods to estimate models and not
models by themselves.

11. Which of the following statements is true?

a. A model shows the estimated parameters of the explanatory variables.
b. A model represents a population relationship.

Downloaded by K61.FTU ?INH H?NG NG?C

c. In an empirical paper, the estimation methods should be discussed
before specifying a model.
d. The methods for estimating a model are same as the model itself.

Answer: b
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Writing an Empirical Paper
BUSPROG:
Feedback: A model represents a population relationship.

12. Which of the following highlights a requirement of a good instrumental variable?

a. It should be included in and endogenous to the equation of interest.
b. It should be included in and exogenous to the equation of interest.
c. It should be omitted from and exogenous to the equation of interest.
d. It should be omitted from and endogenous to the equation of interest.

Answer: c
Difficulty:
Moderate
Bloom’s: Knowledge
A-Head: Writing an Empirical Paper
BUSPROG:
Feedback: For efficient estimation, a good instrumental variable should be omitted
from and exogenous to the equation of interest.

13. Which of the following statements is true?

a. If a model has several explanatory variables and several variations to the
general model are to be presented, it is better to report the results in an equation
form, rather than a tabular form.
b. If a model has several explanatory variables and several variations to the
general model are to be presented, it is better to report the results in a tabular
form, rather than an equation form.
c. The adjusted R-squared should always be listed in the result section of a
research paper, while the R-squared may or may not be listed.
d. The number of observations for each equation should always be listed in the
result section of a research paper, while the R-squared may or may not be
listed.

Answer: b
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Writing an Empirical Paper
BUSPROG:

Downloaded by K61.FTU ?INH H?NG NG?C

Feedback: If a model has several explanatory variables and several variations to
the general model are to be presented, it is better to report the results in a tabular
form, rather than an equation form.

14. Which of the followings statements is true?

a. The title of a research paper should be mentioned only in the
introduction section.
b. The title page should not include any other information except the title of
the research paper.
c. Graphs and tables should be included in the main body only.
d. All equations should begin on a new line and should be numbered consecutively.

Answer: d
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Writing an Empirical Paper
BUSPROG:
Feedback: All equations used in a research paper should begin on a new line and
should be numbered consecutively.

15. Which of the following statements is true?

a. If standard errors are too large or too small, they should be reported in
scientific notation.
b. If coefficients are too large or too small, they should be reported in
scientific notation.
c. While reporting figures, the number of digits after decimals should be limited.
d. The list of references used should be reported in the appendix.

Answer: c
Difficulty:
Moderate
Bloom’s: Knowledge
A-Head: Writing an Empirical Paper
BUSPROG:
Feedback: While reporting figures in research papers, the number of digits after
decimals should be limited so as not to convey a false sense of precision.

16. A good research question should not be backed by time series data.

Answer: False
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Posing a Question
BUSPROG:

Downloaded by K61.FTU ?INH H?NG NG?C

Feedback: A good research question can be backed by time series data. Most of the
important macroeconomic researches are based on time series data.

17. Historical data sets are available only in printed form.

Answer: True
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Data Collection
BUSPROG:
Feedback: Historical data sets are available only in printed form.

18. The practice of data mining is consistent with the assumptions made
in econometric analysis.

Answer: False
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Econometric Analysis
BUSPROG:
Feedback: The practice of data mining is violates the assumptions on which
econometric analysis is based.

19. Instrumental variables can be used to solve misspecification errors related

to omitted variables.

Answer: True
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Econometric Analysis
BUSPROG:
Feedback: Instrumental variables can be used to solve misspecification errors
related to omitted variables.

20. Sensitivity analysis is restricted to natural sciences.

Answer: False
Difficulty: Easy
Bloom’s: Knowledge
A-Head: Econometric Analysis
BUSPROG:
Feedback: Sensitivity analysis can be used for social sciences as well.

Downloaded by K61.FTU ?INH H?NG NG?C