Ken Smith
Fall, 1993
Department of Economics

Course Description

Textbook: Robert S. Pindyck and Daniel L. Rubinfeld, Econometric Models & Economic Forecasts

This course will essentially be treated as a continuous 15 week course. The initial three week preparatory course, however, will essentially be devoted to fundamental statistical review. The 12 week fall semester will deal primarily with cross-sectional econometric analysis. Cross-sectional analysis (analysis of statistical relationships at a set point in time) is the fundamental tool of microeconometric study. A brief introduction to time-series analysis, the fundamental tool of macroeconometric analysis, will also be presented at the end of the 12 week semester.

3 Week Preparatory Course

Note: All reading assignments refer to Pindyck and Rubinfeld unless otherwise noted.

Statistical Review - Chapter 2: An introduction to the analysis of random variables. Topics covered include; population statistics, using sample statistics to estimate population statistics, and tests on the reliability of estimated sample statistics.

Introduction to Probability - Hogg and Tanis (pgs. 1-56): A fundamental introduction to probability theory. Topics with direct application to regression analysis will be emphasized.

Introduction to Regression - Chapter 1: A basic introduction to the linear regression model and estimation of regression parameters.

Two Variable Regression - Chapter 3: Analysis of the basic two variable linear regression model. Emphasis is placed on derivation of Ordinary Least-Squares (OLS) as the Best Linear Unbiased Estimator (BLUE) of a regression and on testing the reliability of OLS estimates.

--Two problem sets will be handed out (not to be graded) to provide practice for the exam.

--the final exam for the initial 3 week course will be on Sat. Sept. 18. The specific time and room assignments will be announced.

12 Week Econometrics Course

The Multiple Regression Model - Chapter 4: Extension of the analysis in Chapter 3 to cases where more than one independent variable is used as a regressor.

Using the Multiple Regression Model - Chapter 5: Here practical application of the linear regression model will be extended. Included topics are; the use of dummy variables, joint hypothesis testing of regression parameter estimates, and tests of the stability of parameter estimates when different data sets are used to estimate the same model.

Heteroscedasticity - Chapter 6.1: In Chapter 6 we begin looking at regression analysis when the assumptions that make OLS estimates BLUE are no longer valid. Here we examine the case of non-constant error variances that will make OLS estimates inefficient. Included topics will be testing for heteroscedasticity and potential solutions to the heteroscedasticity problem.

Instrumental Variables and Model Specification - Chapter 7: Continues the theme of the breakdown of assumptions making OLS estimates BLUE. In particular we will look at the problems of model misspecification (omitted variables and including irrelevant variables) and measurement error in data.

Simultaneous Equation Estimation - Chapter 11: The final topic on the breakdown of assumptions making OLS estimates BLUE. Here we look at the frequent case of regressors being dependent on the variable we wish to use as our dependent variable.

--With the remainder of our time WQ will concentrate on time series analysis - regression on random variables over time.

Serial Correlation - Chapter 6.2: The time-series analogue of heteroscedasticity. Here we look at cases where our regression errors are correlated over time. Included topics are; tests for serial correlation, the most basic correction for serial correlation, and the special case of regression with a lagged dependent variable (special as ordinary tests for serial correlation will not be valid when a lagged dependent variable is used as a regressor).

Simple Forecasting Models - Chapter 8: Using the single equation regression model to forecast future events. We will include topics in unconditional forecasting, conditional forecasting, and forecasting in the presence of serially correlated errors.

Linear Time Series Models - Chapter 16: With the time remaining we will be concerned only with the most basic time-series models. These are the autoregressive model and the moving average model.

--Problem sets will be handed out periodically throughout the 12 week course. These will be graded. The problem sets for the twelve week course will emphasise the use of computer analysis in regression. We will use several different data sets, with real world data, and perform extensive analysis of them using the statistical package Micro-TSP.

--I anticipate giving two exams, a midterm and a final. Dates will be determined as soon as possible.

CRC-Curriculum Resource Center
CEU Budapest, Hungary
Modified: May, 1996


Back to the Economics Syllabi List Back to the Syllabi Collection

Back to the CRC Homepage Back to the CEU Homepage