Scribd
Upload
268 views4 pages

Ec 6041 - Econometrics Theory and Practice: T 0 1 1t 2 2t T

The document contains 7 questions related to econometrics theory and practice. Question 1 asks about the importance of the intercept term in a linear regression model, causes of the error term, consequences of violating classical linear regression assumptions, and how violations impact coefficient and variance estimates. Question 2 provides sample data and asks to estimate coefficients, calculate a fitted value, and discuss heteroscedasticity testing. Question 3 asks about a model with heteroscedastic errors, properties of OLS estimators, and data transformation. Question 4 considers autocorrelated errors, properties of OLS, and consequences of ignoring autocorrelation. Questions 5-7 ask about additional econometrics concepts like joint significance testing, multicollinearity, and partitioned

Uploaded by

Kabelo
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
268 views4 pages

Ec 6041 - Econometrics Theory and Practice: T 0 1 1t 2 2t T

The document contains 7 questions related to econometrics theory and practice. Question 1 asks about the importance of the intercept term in a linear regression model, causes of the error term, consequences of violating classical linear regression assumptions, and how violations impact coefficient and variance estimates. Question 2 provides sample data and asks to estimate coefficients, calculate a fitted value, and discuss heteroscedasticity testing. Question 3 asks about a model with heteroscedastic errors, properties of OLS estimators, and data transformation. Question 4 considers autocorrelated errors, properties of OLS, and consequences of ignoring autocorrelation. Questions 5-7 ask about additional econometrics concepts like joint significance testing, multicollinearity, and partitioned

Uploaded by

Kabelo
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 4

EC 6041 - ECONOMETRICS THEORY AND PRACTICE

ASSIGNMENT AND REVISION QUESTIONS

Question 1
Given a linear regression model as specied below:

yt = β0 + β1 x1t + β2 x2t + ... + εt (1)

(i) What is the importance of β0 in the model above?


(ii) Discuss the causes of the error term εt
(iii) The approach to be employed in the estimation of the model largely depends on the on the violation of the
classical linear regression assumptions. Discuss how the classical linear assumptions are violated in the above model.
(iv) What are the consequences of assumption violation to the estimated coecients and variances?

Question 2
Given the following data sampling process (DSP):
yi = β1 + β2 x2i + εi (2)

You are given the following empirical information: y = [4 6 5 1 5]


0

X = [0 1 1 0 1]0
a) Estimate β1 and β2
b) What is the tted equation from these estimates?
c) Calculate the value of thr tted equation when x2i = 6
d) Discuss the statistical properties of the OLS estimator in the presence of heteroscedasticity.
e) Discuss the Goldfeld-Quant test for heteroscedasticity.

Question 3
You are given the following estimation equation

yi = β1 + β2 x2i + εi (3)
√
Where εi = x2i ∗ µi , µi ∼ N (0, 1) and x2i is a non-stochastic variable.
i) Show that this model is heteroscedastic, explain you answer
ii) Discuss the statistical properties of the OLS estimators in this case
iii) Discuss and show how you would transform the data so that you could remove the heteroscedasticity

1
Question 4
Consider the following regression equation
yt = βxt + εt
where
εt = 0.6εt−1 + µt
µt −→ N (0, 1). µ −→ IID
a) Assume that ε −→ N (µε , σε2 ) for every t. Show that µε = 0 and var(εt ) = 25
16 .

b) (i) Show that this model has autocorrelation.


(ii) Calculate the correlation coecient between εt and εt−1
(iii) What does the error covariance matrix look like, given that n = 5?
c) You are given the following empirical information: y = [3 4 5 − 1 4] and X2 = [5 5 10 0 10]0
(i) Estimate β by Least Squares
(ii) Discuss the statistical properties of the OLS estimators in this case.
(iii) Discuss how you would transform the data to remove the autocorrelation.
(iv) What are the consequences of ignoring autocorrelation in the model above?

Question 5
Mauro (1995) provides a highly inuential analysis of the relationship between eco- nomic growth and corruption.In
the Appendix we provide a number of regressions analogous to those run by Mauro. The data are macroeconomic
indicators for a range of countries for the period 1960 to 1985.
(a) Interpret regression 2
(b) Test for the joint signicance of per capita GDP in 1960, secondary school edu- cation and the population growth
rate in that regression.
(c) Does multicollinearity appear to be a problem in that regression? Explain. If yes, what would you do about it?
(d) Do you detect heteroscedasticity in that regression? If yes, discuss how you might deal with it.
Note:
The variables have the following meanings:
growth6085−→ Average country growth rate (of GDP) between 1960 and 1985
bureaue −→ A bureaucratic e¢ ciencyindex. It is a ten point scale going from 1 (least e¢ cient) to 10 (most). In
the data the lowest value is 1.89 (D.R. Congo) to 10 (Netherlands, Switzerland, New Zealand and Singapore). This
index is a summary measure constructed from a red tapeindex, a judicial e¤ectivenessindex and a corruption index.
gdp60 −→(log) GDP per capita in 1960
seced60 −→ Secondary school enrolment (proportion) in 1960
GPOP −→population growth rate
elf −→ethno-linguistic fractionalisation (index). In the data it goes from zero (perfect homogeneity) in Korea to 90
(D.R. Congo)

2
Question 6
You are given the following model of a data sampling process:
y = β1 + β2 X + µ
where each element of the vector ε is independently and identically distributed. Your empirical data is given by:
y = [9 6 8 10 9] and X = [2 1 2 2 1]0
i) show that the DSP satisfy the assumptions of the CLRM

3
ii) Assume that you have estimated this model by OLS ignoring any possible problems that there might be. What
are your estimates of β?
ii) What is the covariance matrix of the OLS estimators β̂ given that σ 2 = 13 ?
iii) Discuss the statistical properties of the OLS estimator for this DSP.

Question 7
Given a partitioned regession model as follows:
y = X1 β1 + X2 β2 + ε where X = [X1 X2 ] and β = [β1 β2 ]
0

1) Using the Frisch-Waugh-Lovell theorem


i) Solve for the least square estimates βˆ1 and βˆ2
ii) what is the signicance of regressing e1 on e2 ?
2) Using the projection and residual maker matrices, proof the following:
i) that the projection and residual maker matrices are idempotent and symmetric
ii) that MX PX = PX MX = 0

You might also like