Econometrics I

 

Economics 7800

David Kiefer, KDGB 307
telephone 581-7481
e-mail: kiefer@economics.utah.edu

course website: webct.utah.edu

           

            This course focuses on regression analysis, the widely used technique of statistical curve fitting. The successes and failures of the regression technique are illustrated by empirical problem sets making extensive use of the computer. The regression method can be generalized and extended to cover a variety of applied problems, including nonspherical error distributions, the use of prior information and qualitative data.

            The course assignments and term project involve computer work. These may be carried out on any available machine, with any available software. Some possible statistics programs are Stata, Limdep, S-Plus, SAS, Shazam, RATS, e-Views and SPSS. My personal favorites are Shazam and Excel.

            The texts for this course are Jeffrey Wooldridge, Introductory Econometrics, Thompson, 3rd edition, and Peter Kennedy, A Guide to Econometrics, 5th edition, MIT press. They may be purchased from the bookstore. Some readings outside this text will also be assigned.

            The grading scheme is:

·      Homework assignments                                        40%

·      Term project                                                         30%

·      Final examination                                                  30%

Late assignments lose points; copies and exact duplicates are unacceptable. The exam must be taken at the scheduled time. It is comprehensive in coverage and open book. Incompletes are not generally given for nonmedical reasons.

            The term project is to be an econometric project of the student's own design. It could be an exercise in applying econometric techniques to some economic, social or financial issue amenable to empirical testing. Alternatively, it might be a Monte Carlo study of some problem in econometric methodology.

            Your final report should be and follow conventional footnoting and bibliographic rules. It should be about 8 pages long, typewritten double-spaced; papers more than 10 pages lose points. Your paper should briefly review the relevant literature. It should define measurable versions of the variables of interest and fit them into an econometric specification. It should apply appropriate estimation techniques, reporting the results clearly and concisely; please do not include raw computer output. Finally, it should discuss the inferences that are justified from your results.

            The written version of your project is due two weeks before the end of classes. During the last two weeks of the semester the students will take turns orally presenting their research; plan a 10-minute discussion of your project. Dates will be arranged in class.

 

Topic Outline and Reading List

 

1.     Introduction

poor reputation of econometrics

economic theories

economic data

causation and correlation

Kennedy, chapter 1

Wooldridge, chapter 1

2.     Ordinary least squares and classic regression

simple and multiple regression

random cross-sectional data

OLS formulas and matrix algebra

six assumptions

interpretation of OLS

goodness-of-fit

omitted and irrelevant variables

multicollinearity

Kennedy, chapter 3

Wooldridge, chapters 2 and 3, appendices D and E

3.     Small sample properties

expected value and variance of OLS

bias and variance

Gauss-Markov theorem

the normality assumption

maximum likelihood

Kennedy, chapter 2

Wooldridge, chapters 3 and 4, appendix B

4.     Monte Carlo experiments

generating random numbers

drawing samples

experimental design

Russell Davidson and James G. MacKinnon, Estimation and Inference in Econometrics, 1993, chapter 21

5.     Inference

confidence intervals and hypothesis testing

t tests, p-values and standard errors

multiple hypotheses and confidence ellipses

tests for normality

Kennedy, chapter 4

Wooldridge, chapter 4, appendix C

6.     Large sample properties

consistency

asymptotic normality

asymptotic efficiency

Kennedy, chapter 9, 20 and appendix C

Wooldridge, chapter 5, appendix C

7.     Modeling

nonlinearities

adjusted  R2 and model selection

prediction and residual analysis

Kennedy, chapters 5, 6 and 14

Peter Kennedy, “Sinning in the basement: what are the rules? Ten commandments of applied econometrics,” Applied Econometrics, 2002: 569-589

Wooldridge, chapter 6

8.     Qualitative Information

dummy independent variables

interaction effects

dummy dependent variable: the linear probability model

Wooldridge, chapter 7

9.     Generalized errors: heteroscedasticity

nonspherical disturbances and inefficiency of OLS

White’s test

generalized least squares (GLS)

weighted least squares

linear probability model

Wooldridge, chapter 8

10.  Specification and data problems

functional form: the RESET test

measurement errors

missing data

influential observations

Wooldridge, chapter 9

11.  Instrumental variables

omitted and endogenous variables

instrumental variables and two-stage least squares

testing for endogeneity

measurement errors

Kennedy, chapters 9 and 11

Heckelman, Jac C. and Berument, Hakan, 1998, Political Business Cycles and Endogenous Elections, Southern Economic Journal 64, 987-1000.

Wooldridge, chapter 9

12.  Review and conclusion