EC 510/472 Time Series Analysis and Forecasting

Winter 2007, 12:00-1:50pm TTH (CH296)
Prof. K.-P. Lin (CH 241G, 725-3931)
Office Hours: 4:00-4:30 TTH & by appointment

This course covers the methodology and applications of econometric time series analysis and forecasting, with focus on issues and problems of predicting the U.S. economy.

Basic understanding of econometric analysis is required (EC 370, 570 or equivalent). Knowledge of calculus, algebra, probability theory and statistics are essential for this course. Familiar with computer programming and econometric packages will be useful. Stata 9 for Windows will be used throughout the course.

Textbook and Software

• W. Enders, Applied Econometric Time Series, 2nd ed., Willey, 2004.
• Stata 9, StataCorp, 2005. A version of Small Stata may be used for the class.

Topics

1. Reviews of Basic Econometrics
   1. Simple and Multiple Regression
   2. Model Evaluation
   3. Model Selection Criteria
2. Least Squares Prediction
   1. Forecasting with Classical Regression Models
   2. Forecasting with Autocorrelation
   3. Forecasting with Lagged Dependent Variable
   4. Forecast Error Statistics and Evaluation
3. Time Series Analysis I: Introduction
   1. Covariance Stationarity
   2. Trend in Time Series
   3. Unit Root Problem: Estimation and Testing
4. Time Series Analysis II: ARIMA Models
   1. Identification
   2. Estimation and Diagnostic Checking
   3. Forecasting
   4. Extension: Transfer Function Models
5. Time Series Analysis III: Advanced Topics
   1. ARMA Analysis of Regression Residuals
   2. ARCH and GARCH Model Estimation
   3. Multi-Equation Time Series Models

Case Study and Homeworks

Expectation

1. There will be a mid-term (February 8, in class) and a final exam (March 22, Thursday 10:15-12:05). In addition, 3 or 4 homeworks will be assigned periodically (due every 2 weeks in average).
2. Grade distribution of this course looks like this:

   Mid-Term	40%
   Final	40%
   Homeworks	20%