The Identification Zoo - Meanings of Identification in
Econometrics
Arthur Lewbel
Boston College
January 2016 - This is a draft of an article commissioned by the Journal of Economic Literature.
Abstract
Over two dozen different terms for identification ap
Econometrics 2: Midterm Exam
Prof. Jrg Stoye, Spring 2016
This exam has 10 questions (partitioned 4-3-3 and not of equal length or
diculty). Each question is worth 10 points. Good luck!
Throughout the exam, if you think you know whats going on but
dont ha
ECON 6200: Section XIII
TA: Nahim Bin Zahur
April 29, 2016
1. Doing the bootstrap
a. Manual bootstrap: Bootstrap the distribution of the sample mean from the sample (2, 4, 6).
Then test the one-sided hypothesis test = 2 at the 5% significance. Speculate o
ECON 6200: Section I
TA: Nahim Bin Zahur
January 29, 2016
1. OLS and constant term
Consider a regression with a constant. Let X = [J X2 ], where X2 is a n k matrix, and J is a n 1
vector whose elements are all 1. Unknown parameters and its OLS estimator b
ECON 6200: Section X
TA: Nahim Bin Zahur
April 07, 2016
NOTE: I have added some discussion on 1.4, 2.3 and 2.4.
1. Asymptotic normality (adapted from Spring 2015 final)
Assume all the extremum estimator assumptions reproduced below. (Not all will be used.
Econometrics 1: Midterm Exam
Prof. Jrg Stoye, Spring 2013
This exam has 9 questions (partitioned 4-3-2 and not of equal length or
diculty). Each question is worth 10 points. Good luck!
1 Consider the following assumptions about a model characterized by ou
Econometrics 2: Final Exam
Prof. Jrg Stoye, Spring 2016
This exam consists of 15 questions, not of equal length or diculty, grouped
into 3 larger questions, plus a bonus question. You can get 10 points per question
for a total of 150. Good luck!
1 Conside
2.3
GMM with Conditional Homoskedasticity
Up to this point, we did not impose homoskedasticity. This is important because you will very often
want to allow for heteroskedasticity. But GMM specializes in interesting ways when homoskedasticity
is imposed.
A
1
Econometrics II: Final Exam
Prof. Jrg Stoye, Spring 2012
This exam consists of 12 questions grouped into 3 blocks. Each question
carries equal weight. Good luck!
1. Consider the model
= 0 + 1 + 2 2 +
where is i.i.d. with E() = 0 and () = 2 .
1.1 Assum
Econometrics 2: Midterm Exam
Prof. Jrg Stoye, Spring 2014
This exam has 10 questions (partitioned 5-2-3 and not of equal length or
diculty). Each question is worth 10 points. Good luck!
1. Consider the model characterized by
= 0 + 1 +
E( |1 ) = 0
E 2 |1
1
Econometrics II: Final Exam
Prof. Jrg Stoye, Spring 2014
This exam consists of 12 questions, not of equal length or diculty but of
equal weight. They are grouped into three larger questions (7-2-3). Good luck!
1 ML Estimation of the Exponential Distribu
ECON 6200: Section VII
TA: Nahim Bin Zahur
March 11, 2016
NOTE: I have included suggested solutions for the 2nd question.
1. Panel data with instruments
Consider the following system of two equations, over two time periods:
(1) yi1 = + zi1 + wi + i + i1
(
Econometrics 2: Final Exam
Prof. Jrg Stoye, Spring 2015
This exam consists of 15 questions, not of equal length or diculty, grouped
into 3 larger questions. You can get 10 points per question for a total of 150.
Good luck!
1 Consider the model
Pr(
= ) =
=
ECON 6200: Section III
TA: Nahim Bin Zahur
February 12, 2016
NOTE: I have included solutions for Question 2.2 since we did not have time to fully cover it in
section.
1. Introduction to IV Estimation
Consider the following model:
(E1) yi = + zi + i
(E2) z
ECON 6200: Section IV
TA: Nahim Bin Zahur
February 19, 2016
1. Writing down moment conditions (Spring 2014, Midterm 1)
Consider an agent who makes a choice yi between exactly two options, a and b. Each options utility
is a random variable modelled by:
u(a
capture log close
cd "C:\Users\l\Dropbox\Academics\Cornell Courses\2016-17 Spring\ECON
6200\Teaching notes resources\Card 1995\"
log using ".\Log\Card_1995_`c(current_date)'.txt",replace text
use "card.dta", clear
*
/*
The purpose of this do-file is to ca
6
Additional Worked Examples
6.1
Maximum of a Uniform Distribution
Let be i.i.d. uniformly distributed on [0 ]. (The left bound at 0 is for simplicity.) The aim is
to estimate . We briefly note that E = 2 and so
b = 2 is a GMM estimator. Of course, we
un
ECON 6200: Section V
TA: Nahim Bin Zahur
February 26, 2016
NOTE: I will discuss Question 1.6 at the start of next weeks section.
1. OLS, 2SLS and efficient GMM (based on Wooldridges (2001) discussion of Card (1995)
We wish to estimate the returns to educa
Econometrics II: Final Exam
Prof. Jrg Stoye, Spring 2011
This exam consists of 12 questions grouped into 2 blocks. Each question
carries equal weight. Good luck!
1 Let the random variable be distributed acording to a Poisson distribution
with unknown para
ECON 6200: Section XII
TA: Nahim Bin Zahur
April 22, 2016
NOTES: I have included some notes on Q2(d) and 2(e) which we did not have time to fully cover
in section.
1. Selection
We are going to consider a variation of the selection model from Tuesdays lect
1
Econometrics II: Final Exam
Prof. Jrg Stoye, Spring 2013
This exam consists of 12 questions, not of equal length or diculty but of
equal weight. They are grouped into three large questions. Good luck!
1. Let the random variable be distributed i.i.d. Poi
Econometrics 2: Midterm Exam
Prof. Jrg Stoye, Spring 2015
This exam has 10 questions (partitioned 5-5 and not of equal length or
diculty). Each question is worth 10 points. Good luck!
1. This question is about linear GMM. Suppose that there are = 8 moment
3
Extremum Estimators
3.1
A More Formal Take on Identification
As a prelude to extremum estimators, we take a more formal look at identification. In linear models, identification is usually verified by checking some rank and order conditions. But in model
2.4
Multiple-Equation GMM
To handle multiple equations, we need to extend our notation and adapt our assumptions. In general,
the assumptions will not become stronger.
Assumption 1: Linear Model
= z0 + = 1
Notice that we do not restrict the joint distri
2
Generalized Method of Moments
The Generalized Method of Moments (GMM) organizes many tools that you will have seen before,
including anything preceding in this lecture, and many more that you will encounter eventually, e.g.
in time series. Many economet
Gretl Users Guide
Gnu Regression, Econometrics and Time-series Library
Allin Cottrell
Department of Economics
Wake Forest University
Riccardo Jack Lucchetti
Dipartimento di Economia
Universit Politecnica delle Marche
December, 2015
Permission is granted t
JPART 17:379404
E-Government and Bureaucracy: Toward
a Better Understanding of Intranet
Implementation and Its Effect on Red Tape
Eric W. Welch
University of Illinois at Chicago
Sanjay K. Pandey
University of Kansas
ABSTRACT
This article examines the inte
JPART 17:379404
E-Government and Bureaucracy: Toward
a Better Understanding of Intranet
Implementation and Its Effect on Red Tape
Eric W. Welch
University of Illinois at Chicago
Sanjay K. Pandey
University of Kansas
ABSTRACT
This article examines the inte