Section 1: Multiple Choice Questions (1 X 12) Time: 50 Minutes
Section 1: Multiple Choice Questions (1 X 12) Time: 50 Minutes
Section 1: Multiple Choice Questions (1 X 12) Time: 50 Minutes
14. Explain the significance of disturbance term in a stochastic specification of sample regression
function given by Y 1 X1 2 X 2 3 X 3 u ? (3)
15. Explain the difference between Conditional Mean, Unconditional Mean and Arithmetic
Mean? (2)
***********************
BITS-Pilani Dubai Campus
ECON F241 ECONOMETRIC METHODS
Semester I, 2019 Quiz 1 (Closed Book)
Time: 20 minutes; Multiple Choice Questions (1X10)
Note: Encircle the most appropriate option. All the notations have usual meanings unless otherwise defined!
1 Suppose you are interested in examining the factors that determine the Inflation rate (Yt) of a
country. Theoretically, the main factors which may affect inflation include factors like demand
(measured by money supply M), supply (measure by supply shocks S) and inflation expectations
(approximated by inflation rate in previous period Yt-1,). Now suppose that you estimated a
regression model:
Yt = βˆ1 + βˆ2 St + βˆ3 Yt-1 + Ut. instead of Yt = βˆ1 + βˆ2 Mt + βˆ3 St + βˆ4 Yt-1 + Ut.
Which assumption(s) of CLRM you may have violated
a E(Ui)=0 b) Cov(Yi, Ui)=0 c) Homoscedasticity
d No-Autocorrelation e) both a and b f) None
2 We know that a country with high GDP today is going to have higher GDP next year. Suppose you
are interested in examining the factors that determine the GDP level (Y) of a country. Theoretically,
the main factors which may effect GDP include investments on Land (L) Labor (N), Capital(C) and
Entrepreneurship (E). And now suppose that you used a regression model: Yt = βˆ1 + βˆ2 Lt + βˆ3
Nt + βˆ4 Ct + βˆ5 Et + Ut instead of: Yt = βˆ1 + βˆ2 Lt + βˆ3 Nt + βˆ4 Ct + βˆ5 Et + βˆ6 Yt-1 + Ut.
Which assumption(s) of CLRM you suspect might be violated
a) Linearity b) No-Autocorrelation c) Homoscedasticity
b) Multicollinearity e) All
3 Assume that you estimated a sample regression function Yi = β1 + β2 Xi + Ui,where Yi measures
punctuality (office arrival time) of employees and X denotes various salary levels. Suppose that
salary and punctuality are positively related. Then, the variance of Ui is expected to be
a) Homoscedastic b) Heteroskedastic c) time invariant
4 The numerical property that the sample regression line generated with estimators obtained from
OLS pass through sample means of Yi and Xi implies:
a That OLS estimators are exact/perfect approximation of population parameters
b That they do not vary with the change in sample
c That the expected value of error term is zero
5 Assume a multivariate CLRM, Y = βˆ1 + βˆ2X1 + βˆ3X2 +Ui, where Y denotes the level of
investment and X1 and X2 denote income and saving levels ,respectively, across different
individuals in UAE. Then, in estimating this regression model using OLS, you may face an issue of
a) Autocorrelation b) Multicollinearity c) both d) none
6 The violation of the assumption that no-correlation between ui and estimated values of Yi implies
that
a) ui is purely random b) ui is systematically generated c) Yi is purely random d) both a and c
7. The mean value of predicted/estimated value of Y (obtained from OLS method) is always going to be
equal to mean value of actual Y. ( T / F)
8. The technique of Ordinary Least Squares used to estimate the unknown parameters in a regression
model is based on the principle of minimizing (Yi – Y^i). ( T / F)
9. In estimation of sample regression function, suppose you found that the sum of residual term is equal
zero. Then it implies that the OLS estimates are perfect/exact estimators (but not the approximation) of
the true population estimators. ( T / F)
10. The regression model yi 2 xi ui is an equal/same representation Yi 1 2 X i U i . ( T / F )
Name:…………………………………………. ID No:…………………………..
BITS-Pilani Dubai Campus
ECON F241 ECONOMETRIC METHODS
Semester I, 2019 Test-2 (Open Book)
Section 1: Multiple Choice Questions (1X14)
Note: Encircle the most appropriate option and return this paper with answer book!
All the notations have usual meanings unless otherwise defined!
1 Suppose you estimated a model for interest rate given by
it 1 2 ( t * ) 3 ( yt y* ) 4 Et et for India. Also, it is noted that the interest rate
process is highly persistent (related with its past values). Then, in et 1et 1 2et 2 ut , you
expect
a 1 0; 2 0 b) 1 0; 2 0 c) cov(et , et j ) 0
2 Using OLS technique, suppose estimated interest rate model:
it 1.7 0.4( t * ) 0.3( yt y* ) 0.2Et et , where p-values associated with
1 , 2 , 3 , 4 are respectively given by 0.01, 0.04, 0.32, 0.05. Then these results imply that
a all inflation-deviation( t * ), output-gap ( yt y * )and exchange rate ( Et ) determine the
changes in interest rate ( it )
b only inflation-deviation( t * ) and output-gap ( yt y * )determine the changes in interest
rate ( it )
c only inflation-deviation( t * ) and exchange rate ( Et ) determine the changes in interest
rate ( it )
3 the magnitude of the estimate of coefficient of determination that is R2 is used to evaluate the
a explanatory power of a regression model b) predictive power of a regression model
c relevance of independent variables d) All a, b , c
4 The coefficient of determination measures a proportion of total variation in dependent
variable attributed to changes in
a relevant independent variables included in the model
b relevant independent variables and error term included in the model
c changes in all possible independent variable
5 Suppose you are testing for presence of autocorrelation in your regression model. Then which
value of the Durbin-Watson(DW) statics given below indicate that there is no autocorrelation
in your model
a DW value equal to 0 b) DW value equal to 1 c) DW value equal to 2
6 Suppose, I formulate a null hypothesis that “all crows are black”. Based on the sample test
statistic, suppose, I reject this null hypo. Then which type of error is committed
Type I b) Type II c) both a and b d) Neither a nor b
7 Suppose you estimated a regression model xt 1 2 t et ; and the OLS estimates are
given by xt 5 0.3 t et , where p-values associated with 1 , 2 are as 0.01, 0.34,
respectively. These results indicate that the
a probably of rejecting the null hypo, which is actually true, is 1% and 34% respectively for
1 , 2
b probably of rejecting the null hypo, which is actually true, is 99% and 66% respectively for
1 , 2
c probably of accepting the null hypo, which is actually true, is 1% and 34% respectively for
1 , 2 d) All a, b, c
8 We know that the consumption of food greater or less than the optimal-quantity will
deteriorate our fitness levels. Suppose you are interested in examining the relationship
between quantity of food taken and fitness levels among various individuals in UAE. Which
of the following hypothesis test is more appropriate to use
a One-tail test b) two-tail test c) both a and b are OK
9 If sample data provides significant evidence against the null hypothecs, then the t-value is
expected to have
a higher magnitude b) lower magnitude c) will remain same
10 The relatively lower s.e. of an OLS estimate is considered as an evidence in favor of the view
that the
a sample estimate is true population value b) sample estimate is not true population value
c true population value is much lower than the sample estimate
11 Suppose, you are interested in examine the relationship between price (P) and quantity
demand(Q) of a particular commodity, which is given by Qt 1 2 Pt ut . Then, in testing
the null-hyphothesis (H0) 2 0 , the alternative-hypothesis H1 is of
a simple nature- unidirectional b) composite nature- bidirectional c) both a and b
12
Suppose the confidence interval
Pr 2* t se( ˆ2 ) ˆ2 2* t se( ˆ2 ) 1
2 2
under Ho: β2=0.30 gives 0.21 and 0.38 as the lower and upper critical values, respectively.
Then which of the following inference is most appropriate
a βˆ2=0.39 should be considered statistically same as β2=0.30
b βˆ2=0.24 should be considered statistically same as β2=0.30
c βˆ2=0.37 should not be considered statistically same as β2=0.30
d βˆ2=0.22 should not be considered statistically same as β2=0.30
e All a, b, c and d are true
13 For a normally distributed residual term, the skewness and kurtosis are expected to be
a 0, 2 respectively b) 0, 3 respectively c) 0, 4 respectively d) 0, 5 respectively
14 Assume that you want to estimate the relationship between time spend in gym (T) by
different students and their Fitness (F). Also assume that there exists some maximum time-
limit one can spent in gym and anymore time spend beyond that gives more muscle stress and
injuries, and thereby reduces fitness levels. As this relationship is inverted U-shaped and can
be specified as F = β1 +β2T - β3 T 2+ u. In this case, what would you expect r and r-square to
be
a r-square > 0 and r = 0 b) r-square = 1 and r = 0
c r-square = 0 and r > 0 d) r-square < 1 and r < 0
15. Define hypothesis. Illustrate with an appropriate example the formulation of null and
alternative hypothesis. What determines your choice about the type of test (one-tail/two-tail) (2)
16. State and explain the three estimates which are used to evaluate the precision of OLS
estimates.(2)
17. Define forecasting error. Explain the techniques of in-sample and out-of-sample forecasting.
(2)
BITS-Pilani Dubai Campus
F241 ECONOMETRIC METHODS
Semester I, 2019 Comprehensive Examination (Closed Book)
Time Allowed: 03:00 hrs Marks: 60 Date: 19/12/2019
Very Short Answer type questions: Answer all the questions (5X2)
6. State and explain the properties of various types of time series data generating processes
commonly used to approximate the real world economic variables.
7. According to the market mechanism, the quantity demand (Dt) of a commodity should be
equal to the quantity supplied (St) of it over period of time. Assume that both these variables
follow I(1) process and you want to model the equilibrium relationship between Dt and St.
Discuss and explain the appropriate econometric technique which you will use to estimate this
relationship.
8. State and explain the distinguishing features of various economic time series variable.
9. We know that in an economy three time series variables: Consumption spending (Ct), Income
(Yt), Saving (St) are related in such a way that a change in any variable causes a change in every
other variable. Model these variables in VAR (Vector Autoregression model) framework and
interpret the coefficients, and also explain how this framework is used to infer causality
between variables and endogeneity/exogeneity of variables.
10. Assume a linear regression model Yi 0 1 yi ui which is estimated using OLS method.
State and explain the properties of 0 and 1 under the assumption that ui follows a normal
distribution ui N (0, 2 ) .
11. Suppose you want to examine the relationship between student grades (Gi) and their
attendance (Ai) of a particular course in a given semester. Frame a null and alternative
hypothesis for this relationship. State and explain the various steps (stages) involved in
modelling this relationship.
12. Suppose that you are interested in examining the factors that determine sales (St) of a
particular company (say coca cola) overtime. Suppose you estimated a regression model:
St 0 1 Pt 2 J t 3 Dt 4 At ut using OLS method; where, P is price of coke, J is price of
Pepsi, D is seasonal dummy variable and A measures advertising cost. The results from this
model are summarized in the table given below.
(i) Based on the given results, evaluate each of the following null-hypothesis:
0 0; 1 0; 2 0; 3 0; 4 0 .
(ii) What does these results suggest about the relationship between variables S and P, J, D, A.
Interpret the coefficients, and examine whether the signs of these coefficients are consistent
with theory? How are S and J related?
(iii) Which factor (among; its own Price, seasonality, Price of Pepsi, and Advertising outlay) is
more important in determining the sales of this commodity?
(v) How will you infer whether this regression estimation is appropriate?
13. Suppose, you are interested in forecasting the heart rate (Ht, at some frequency) of a
particular individual. Illustrate and discuss the various steps involved in univariate modelling of
heart rate using Box-Jenkins method?