University of Illinois
Fall 2014
Department of Economics
Roger Koenker
Economics 508
Lecture 24
Treatment Eects, Matching and Propensity Scores
In randomized experiments we are assured that the treatment indicator,
say Di , is assigned independently of bo
University of Illinois
Fall 2014
Department of Economics
Roger Koenker
Economics 508
Lecture 13
Specication Tests in Simultaneous Equations Models
In this lecture we will consider the problem of testing the validity of
certain hypotheses concerning the ex
University of Illinois
Fall 2014
Department of Economics
Roger Koenker
Economics 508
Lecture 10
Introduction to Simultaneous Equation
Econometric Models
1. Review of Linear Seemingly Unrelated Regressions
The simplest example of simultaneous equation mode
Department of Economics
Fall 2013
Econ 508
Lecture 11
Binary Treatment Models, Randomization and Errors in
Variables
1. Binary Treatment and Randomization
The simplest experimental treatment model is the following
yi = + Di + ui
where Di is 1 if the subje
University of Illinois
Fall 2013
Department of Economics
Roger Koenker
Economics 472
Lecture 9
Introduction to Non-Stationary Time Series
Consider a univariate time series cfw_yt
t= . We say that cfw_yt is (strictly) stationary if the joint
distribution
University of Illinois
Fall 2013
Department of Economics
Roger Koenker
Economics 508
Lecture 8
The Regression Fallacy
1. Galtons Regression to Mediocrity
Arguably, the most important statistical graphic ever produced is Galtons (1885)
gure illustrating re
University of Illinois
Fall 2014
Department of Economics
Roger Koenker
Economics 508
Lecture 7
Introduction to Specication Testing in Dynamic
Econometric Models
In this lecture I want to briey describe some techniques for evaluating dynamic econometric mo
cv.new.m
This is a Mathetica notbook to illustrate the computations for the
welfare analysis of the gasoline demand problem set.
The long run form of the demand model is:
b=cfw_-7.733521, 2.798471, -2.53209,
-0.21095,
0.6953243
q[p_,y_]:=Exp[b[1]+b[2] Log
University of Illinois
Fall 2014
Econ 508
Department of Economics
Roger Koenker
Lecture 6
Some Welfare Econometrics of Empirical Demand Analysis
The second problem set investigates U.S. demand for gasoline. There
are at least 3 rationales for this:
(1) Th
University of Illinois
Fall 2013
Department of Economics
Roger Koenker
Economics 508
Lecture 12
Introduction to Dynamic Simultaneous Equation Models
1. Structural Equation Models An Overview
p
0
5
10
15
20
The classical problem in economics, and therefore
University of Illinois
Fall 2014
Department of Economics
Roger Koenker
Economics 508
Lecture 14
Estimation of Systems of Simultaneous Equation Model
In this brief lecture we try to introduce estimation methods for simultaneous equation models which apply
University of Illinois
Fall 2014
Department of Economics
Roger Koenker
Economics 508
Lecture 15
Optimization
Iterative methods of numerical optimization play an important role throughout economics, but are
especially crucial in modern econometrics. In thi
University of Illinois
Fall 2014
Department of Economics
Roger Koenker
Economics 508
Lecture 23
Introduction to Non-Parametrics
1. Introduction
In this lecture I would like to introduce some basic ideas about nonparametrics and also to reenforce some prin
University of Illinois
Fall 2014
Department of Economics
Roger Koenker
Economics 508
Lecture 22
Duration Models
There is considerable interest, especially among labor-economists in models of duration. These models originated in biomedical applications, in
University of Illinois
Fall 2013
Department of Economics
Roger Koenker
Economics 508
Lecture 21
Counts, Tobit, Sample Selection, and Truncation
The simplest of this general class of models is Tobins (1958) model for durable demand
yi = xi + ui
ui iid F
yi
University of Illinois
Fall 2013
Department of Economics
Roger Koenker
Economics 508
Lecture 20
Binary Response Models
Lets begin with a model for an observed proportion, or frequency. We
would like to explain variation in the proportion pi as a function
University of Illinois
Fall 2014
Econ 508
Department of Economics
Roger Koenker
Lecture 19
Inference about tail-behavior and measures of inequality
An important topic, which is only rarely addressed in econometrics
courses, is the measurement of inequalit
Quantile Regression: A Gentle Introduction
Roger Koenker
CEMMAP and University of Illinois, Urbana-Champaign
LSE: 16 May 2011
Roger Koenker (CEMMAP & UIUC)
Introduction
LSE: 16.5.2011
1 / 63
Overview of the Course
The Basics: What, Why and How?
Inference
University of Illinois
Fall 2013
Department of Economics
Roger Koenker
Economics 508
Lecture 17 Panel Data
Consider a model of the form
yit = xit + zi + i + uit
i = 1, . . . , n t = 1, . . . T
(1)
where, for example,
yit = log wage of person i at time t.
University of Illinois
Fall 2013
Econ 508
Department of Economics
Roger Koenker
Lecture 16
Introduction to Quantile Regression
Quantile regression is a statistical technique intended to estimate, and conduct inference about, conditional
quantile functions
University of Illinois
Fall 2013
Econ 508
Department of Economics
Roger Koenker
Lecture 5
The -Method and the Bootstrap
Introduction to Nonlinear Inference
Let me begin with a very simple inference problem which has a personal
attraction to me, because it
University of Illinois
Fall 2013
Department of Economics
Roger Koenker
Economics 508
Lecture 4
Model Selection
and
Fishing for Signicance
1. Model Selection
Classical hypothesis testing plays a central role in econometrics, but in
many applied problems we