Scribd
Upload
3 views

ECOM165 jan 2022

The document outlines the January Examination Period for the ECOM165 Econometrics I course, detailing the exam format, rules, and submission guidelines. It includes specific instructions for answering questions, using formulas, and submitting handwritten answers. Additionally, it presents two questions related to linear regression models and their properties, requiring the application of Bayesian methods and analysis of estimators under measurement errors.

Uploaded by

fin.collings
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
3 views

ECOM165 jan 2022

The document outlines the January Examination Period for the ECOM165 Econometrics I course, detailing the exam format, rules, and submission guidelines. It includes specific instructions for answering questions, using formulas, and submitting handwritten answers. Additionally, it presents two questions related to linear regression models and their properties, requiring the application of Bayesian methods and analysis of estimators under measurement errors.

Uploaded by

fin.collings
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/ 3

January Examination Period 2021-22

ECOM165 ECONOMETRICS I Duration: 3 hours

Answer ALL questions


If you answer more questions than specified, only the first answers (up to the specified
number) will be marked. Cross out any answers that you do not wish to be marked.

THIS IS AN OPEN BOOK EXAMINATION TO BE CONDUCTED ONLINE. YOU MAY


REFER TO ANY OF THE COURSE MATERIALS, OR ANY OTHER SOURCE OF
INFORMATION. YOU MAY ALSO USE A SPREADSHEET OR CALCULATOR.

YOU MAY SUBMIT HANDWRITTEN ANSWERS

ANSWERS ARE TO BE WRITTEN, SCANNED, AND SUBMITTED TO BOTH QMPLUS &


EMAILED TO: [email protected]

PLEASE ENSURE THAT YOUR WORKING IS CLEARLY SHOWN WITH ALL STEPS
OF YOUR CALCULATION INCLUDED IN YOUR ANSWER DOCUMENT, INCLUDING
ANY FORMULA USED.

If you are typing your formulas, please note the following:

• It is acceptable to use the standard alphabet rather than greek letters. The following
are recommended: m for μ, s for σ, w for ω, r for ρ, d for Δ, b for β.

• For mathematical operators: add +, subtract -, multiply *, and divide /.

• Where appropriate, use an underscore to indicate a subscript, Eg r_f for rf.

• Use the ^ character for power, eg x^2 for x2, x^0.5 for √x.

• As an alternative to x^.5 you may type sqrt(x).

• Use brackets as necessary. To make your answer clearer use different brackets
where appropriate, eg [] {} ().

Examiner: Andrea Carriero

© Queen Mary University of London, 2022


Question 1

I Consider the following linear regression model:

y = X + ;   N (0,  2 IT );

where y, X, ,  are T  1, T  K, K  1 and T  1 matrices respectively and


where the error variance  2 is known. Further assume the following prior beliefs:
 
2 2  2 1  0 s20
|  N ( 0 ,  0 );    , . (1)
2 2

a) Write down the joint prior p.d.f. of  and  2 , and the likelihood of the model.
[10 marks]

b) Use Bayes formula to derive the posterior p.d.f. of  conditional on  2 .


[10 marks]

c) Use Bayes formula to derive the posterior p.d.f. of  2 . [10 marks]

d) Show that the prior and posterior distributions of  and  2 in this model can
be equally obtained by expressing the prior in (1) using dummy variables.
Why is this the case? [10 marks]

e) What is the posterior distribution of |y and how can we obtain it? (no
formal proof is required) [10 marks]

>>>>>>>>>> continues on next page >>>>>>>>>>

1
Question 2
The true Data Generating Process (DGP) for the variable yt is:

yt = xt + t , (2)

where t  iidN (0,  2 ) and is independent from xt , which is a random variable with
finite moments. However there are measurement errors, i.e. the econometrician
can only observe xt and yt :

xt = xt + vt , yt = yt + wt ,

where the measurement errors are distributed as

vt  iidN (0,  2v ), wt  iidN (0,  2w ),

and are mutually independent and independent from xt and yt . Consider the linear
regression model:
yt = xt + t , (3)
and define ̂ as the OLS estimator of . In answering the following questions,
always provide a proof with a short discusion.

a) Assume that  2v = 0 and  2w 6= 0. Is ̂ an unbiased and consistent estimator


of ? [10 marks]

b) Assume that  2v = 0 and  2w 6= 0. Is the OLS estimator ̂ e¢cient when


compared to the (unfeasible) estimator obtainable in absence of measurement
error? [10 marks]

c) Assume now that  2v 6= 0 but  2w = 0. Is ̂ an unbiased estimator of ? [10


marks]

d) Still assume that  2v 6= 0 but  2w = 0. Compute the probability limit of ̂


under these assumptions. Is ̂ a consistent estimator for ? [10 marks]

e) Still assume that  2v 6= 0 but  2w = 0. Consider regressing xt on yt :

xt = yt +  t .

Obtain the probability limit of the OLS 


estimator
 of  in the regression model
above. Show that  falls between p lim 1̂ and p lim(̂). [10 marks]

––––––––End of Examination––––––––

You might also like