ECON 614 Econometrics II
Professor Martin D. D. Evans
Fall
2009
This is the second course in the
Econometrics sequence. Students should have completed Econometrics I before
taking this course. The main text I will be using is Fumio Hayashi’s
“Econometrics”. (His web page with the homework datasets is http://www.e.u-tokyo.ac.jp/~hayashi/). Datasets
for the empirical exercises can be downloaded from here. Know typos are listed here.
Requirements
I will assign homework on a regular basis. Some of the problems will be practical requiring computer work. For this purpose, you will need to use Gauss. An introduction to Gauss can be found at Mark Watson's GAUSS tutorial. This covers the absolute basics. More extensive documentation can be found at the Gauss website. My overheads on structured programming are available here.
Grades
Grades for the course will be based on homework (15%), a mid-term (35%) and final exam (50%). Both the mid-term and final exams will have an in-class portion (for traditional analytic questions) and a take-home portion (for applied questions) requiring programming.
Reaching
Me
My office is 568 in the Economics
Department; phone, x7-1570, and email: evansm1@georgetown.edu. I will announce
my office hours in the first class. The TA for the course is Alberto Fuertes
Mendoza, Email: af258@georgetown.edu.
He will hold office hours in ICC 550 on Wednesdays from 2:30 to 3:30 pm.
Topic
Outline
1. Review: Finite Sample Properties of OLS
2. Problem Set 2 Questions (from Hayashi Chapter 1)
Data
for Problem Set 2: Data.asc
3. Asymptotic Theory
a. Limit Theorems
b. Sationarity/Ergodicity
c. Hypothesis Testing
d. Implications of heteroskedasticity and serial correlation
4. Introduction to GMM
a. Large Sample properties in single equation models
b. Testing
c. Implications of Conditional Heteroskedasticity
5. GMM for systems for equations
a. Technical Conditions
b. FIVE, 3SLS and SUR
6. Serial Correlation
a. ARMA processes
b. Vector Processes
c. GMM and serial correlation
7. Extremum Estimators
a. Maximum likelihood
a. NLS
b. Linear and Nonlinear GMM
c.
Hypothesis testing
Gauss Optimization Procedure (From
Numerical Recipes) MLE/GMM
Problem Set, MLE/GMM Data,
code
8. Applications of Maximum Likelihood
a. Truncated Regression Models
b. Censored Regression Models
c. Time series models
Mid-terms
2006 Questions,
Finals