Economics 583: Econometric Theory I

R Hints

Discussion Board

Course Description

Winter 2013
Office hours: MW 11-12

Eric Zivot
348 Savery Hall

This is a general course in econometric theory for second year Ph.D. students in economics. The main goal of the course is to develop the econometric and statistical tools necessary for reading and understanding the current literature in econometrics. A secondary goal is to introduce some current research topics in econometrics. The main focus of the course will be on estimation and inference in econometric models using the generalized method of moments and maximum likelihood techniques.

Textbooks and Other Background Material

The main required textbooks are:

  1. Hayashi, F. (2000). Econometrics. Princeton University Press. Check out Hayshi's homepage with errata, answers to analytical exercises and review problems and other supplemental material.
  2. Hall, A.R. (2005). Generalized Method of Moments. Advanced Texts in Econometrics, Oxford University Press.

Most of the lectures are based on material in the main textbooks.

Supplemental optional textbooks are:

  1. Matyas, L.  (ed.) (1999), Generalized Method of Moments, Cambridge University Press: Cambridge, UK.
  2. Gourieroux, C.  and Monfort, A. (1995). Statistics and Econometric Models, Vols 1 and 2, Cambridge University Press.
  3. Hamilton, J (1994). Time Series Analysis. Princeton University Press
  4. Wooldridge, J. (2010). Econometric Analysis of Cross Section and Panel Data, Second Edition. MIT Press.
  5. Cameron, C., and Travedi, P. (2005). Microeconometrics: Methods and Applications. Cambridge University Press.

The supplemental textbooks provide optional background material.

I will also assign readings from various journal articles and working papers. See the class syllabus page for details.

Course Prerequisites

Prerequisites for course are Economics 580 - 582 (or equivalents). I will assume that you are familiar with the classical linear regression model and its extensions, and have been exposed to maximum likelihood estimation and method of moments estimation. In addition, I will assume that you are familiar with Eviews or Stata and at least one matrix programming language (GAUSS, Matlab, Ox, R or S-PLUS).

Some of the homework assignments will utilize the R programming language ( I highly recommend using RStudio ( or Revolution R (  as an integrated development environment for R. There are many docoments on the web for learning R. Here are some good ones:

Many of the UseR books from Springer-Verlag are available as free ebooks from the UW library. A link to the UserR series that is accessible from the UW library is here.  Recommended books for this class are

Course Requirements

  1. Weekly homework assignments (20%)
  2. GMM replication paper (40%)
  3. Take home Final exam (40%)

The homework assignments will make use of various statistics and econometrics programs (Eviews, Matlab, R,  Stata). The research paper requirement is fairly general and may involve an applied project, theoretical project or simulation experiment. However, the paper cannot be used to satisfy a paper requirement for another course.

The GMM replication paper must involve a replication of an existing empirical study (working paper or published paper) that uses GMM estimation in an empirical study. It is important that you choose a topic for your paper early in the quarter. Once you have picked a paper to replicate, you must obtain my approval to use it for class. Everyone must have a paper topic no later than Friday, February 22, 2013. A final version of the paper must be submitted either in hardcopy or electronically in pdf format by 6.00pm on Friday, March 22, 2013.

Replication Paper Guidelines

The following must be addressed in your GMM analysis

  • Clear description of underlying economic/statistical model
  • Presentation of moment conditions (they need not be fully derived) used for estimation
  • Discussion of identification issues if any
  • Discussion of choice of weight matrix
    • HC vs HAC estimator of S
  • Discussion of type of GMM estimator (inefficient, 2-step efficient, iterative, CU etc)
  • Presentation of estimation results with discussion
  • J test of overidentification conditions (if applicable)
  • Tests of other hypotheses (if applicable)