Scribd
Upload
21 views2 pages

850 Midterm 2015

Uploaded by

Ybnias Grijalva
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
21 views2 pages

850 Midterm 2015

Uploaded by

Ybnias Grijalva
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/ 2

Page 1 of 2 pages

Queen’s University
School of Graduate Studies and Research
Department of Economics
Economics 850 Fall, 2015
Professor James MacKinnon

Midterm Examination
October 29, 2015. Time Limit: 80 minutes

Please answer all questions. The marks for each question are shown in square
brackets. All answers should be as short as possible. In many cases, a formula, a
single sentence, or a few ordered points will suffice.

1. Consider the linear regression model

y = X1 β1 + X2 β2 + u, u ∼ IID(0, σ 2 I), (1)

where there are n observations and k = k1 + k2 regressors, with the regressors in


X1 exogenous and the ones in X2 predetermined. Under standard assumptions
about X1 , X2 , and u, by what power of n must each of the following expressions
be multiplied in order to turn it into a quantity that is either O(1) or Op (1)? In
each case, indicate whether the final quantity is asymptotically deterministic – that
is, O(1) – or stochastic – that is, Op (1). As usual, MX means I − X(X⊤X)−1X⊤,
and MX1 means I − X1 (X1⊤X1 )−1X1⊤.

a) X1⊤X2

b) X2⊤u

c) β̂1 − β10 , where β̂1 denotes the OLS estimator of β1 in equation (1), and β10
denotes the true value of β1 .

d) y⊤MX y

e) y⊤MX y − (n − k)σ 2

f) y⊤MX1 X2 (X2⊤MX1 X2 )−1 X2⊤MX1 y when β2 = 0

g) y⊤MX1 X2 (X2⊤MX1 X2 )−1 X2⊤MX1 y when β2 ̸= 0


[42]

Continued on next page . . .


Page 2 of 2 pages

2. Consider the linear regression model

y = Xβ + u, u ∼ NID(0, σ 2 I), (2)

where the k regressors are all exogenous and there are n observations.
a) Write down the sum of squared residuals (SSR) as a function of the quantities
on the right-hand side of equation (2).
b) How is the scalar SSR/σ 2 distributed?
c) Explain how you would use the result of part b) to test the hypothesis that
σ = 0.5. Is this an exact test or an asymptotic one?
d) Explain how you would use the result of part b) to construct a 95% confidence
interval for σ. Will this interval be symmetric?
[28]

3. An econometrician is trying to estimate the linear regression model

yt = β1 + β2 xt + β3 xt−1 + ut (3)

using a sample of just 20 observations. The econometrician knows that the xt are
exogenous, serially correlated, and skewed to the right. She believes that the error
terms are IID but suspects that they are severely skewed to the right. The object
is to test the null hypothesis that β2 + β3 = 1.
a) How could you perform an asymptotic test of this hypothesis by running only
one least squares regression and obtaining the test statistic directly from the
regression output?
b) How would you generate 999 bootstrap datasets using a residual bootstrap
DGP, and how would you use them to calculate bootstrap test statistics?
c) Write down algebraic expressions for two different bootstrap P values that
could be obtained from the test statistic of part a) and the bootstrap test
statistics of part b). Which one would you use in this case?
d) What would you conclude about the hypothesis β2 + β3 = 1 if the test statistic
from part a) were 2.42, and numbers 25 and 975 in the list of bootstrap test
statistics sorted from smallest to largest were −1.72 and 2.61, respectively?
[30]

You might also like

pFad - Phonifier reborn

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.


Alternative Proxies:

Alternative Proxy

pFad Proxy

pFad v3 Proxy

pFad v4 Proxy