Theory and Practice of Econometrics
Spring 2010
Jeffrey Parker
Course Information
Course Content
Class Format
Prerequisites
Office Hours
Exams and Other Assignments
Grading
Texts and Software
Course Content
Econometrics is the application of statistical techniques to the analysis of economic data. It has its basis in pure mathematical statistics, but has developed as a distinct branch of the parent subject in order to serve particular data-analysis needs often faced by economists. The best (perhaps the only) way to learn econometrics is to do econometrics. This course will emphasize hands-on application of econometric techniques with natural data, not textbook proofs of statistical theorems. Considerable attention will be paid to the nature and sources of economic data. The principal mission of this course is to provide students with a set of econometric tools that can be applied as needed in economic research.
Class Format
This class meets for a total of four hours per week, 12-1pm on Mondays, Wednesdays, Thursdays, and Fridays. Most weeks, about three of those hours will be a lecture/conference session that will examine econometric techniques and their applications. The remaining time will usually include discussion of the weekly assignments.
Prerequisites
Economics: The only economics prerequisite is Econ 201. In
practice, the more economics you have taken, the more applications of
econometrics you will recognize as we proceed through the course.
Mathematics:
Math 141 or equivalent background in statistical analysis is an
important prerequisite. Students will be expected to have a working
understanding of basic statistics, including the concepts of
probability distributions, expected value and variance, estimators,
unbiasedness, hypothesis tests, and confidence intervals. A very brief
review of these topics is conducted January 20-22 for those whose
background in statistics may be limited or rusty. Matrix and vector
notation will also be used in class at times. If there are sufficient
students with no knowledge of matrix methods, one or more sessions may
be organized through the Quantitative Skills Center during the first
two weeks of the semester.
Office Hours
The instructor will hold office hours in Vollum 229 on Mondays and Wednesdays from 1:30 to 3:00. Students for whom these hours are inconvenient may make appointments at other times by contacting the instructor at extension 7308 or by email to parker@reed.edu.
Exams and Other Assignments
Exams: There will be a mid-term exam and a comprehensive
final exam. Both exams may have in-class and/or take-home parts. There
will be no term paper for the course, though there will be a final
project as discussed below.
Weekly problems: Most
weeks, there will a set of theoretical problems due in class on Friday
mornings. These assignments do not usually involve any estimation and
are to be done individually.
Weekly econometric projects:
An applied project using statistical software to estimate models will
be assigned most weeks. Detailed information about assignments will be
posted weekly on the Web site. The weekly econometric projects will
usually be done by teams of two students, who jointly submit a single
collaborative report. Past experience suggests that teams who work
together on the projects produce better results than those who divide
the work and do the parts individually. However your team decides to
work, both members of the team must participate in both the econometric
estimation and the writing of the report. Both must "sign-off" on the
final product. An appendix to the report should indicate how the work
was shared between team members. The instructor will assign teams; you
will generally work with a different partner each week. Reports are to
be submitted electronically by email no later than 6:00am on Tuesdays.
The instructor will read the reports and send email comments to each
student. We will often devote Wednesday's class to the projects.
Because of the short turnaround time between the submission deadline
and the class discussion, late work will be severely penalized.
Projects received between 6:00am Tuesday and 6:00am on Wednesday will
receive half credit. Projects more than 24 hours late will not be
accepted.
Course Project: We will undertake a course
project during the semester that will attempt to simulate the process
of performing an econometric analysis in a consulting environment.
There will be several groups of students working on separate projects.
Each group will have an external "client" for the research, who may
come from off campus. One "client" will be Jon Rivenburg, Reed's
Director of Institutional Research. The clients will present us with
some questions that they would like us to answer and some data to use
in attempting to formulate answers.
The projects will be
announced midway through the course and students will choose their
groups in a lottery-determined order. Once groups are determined, the
partner pairings for the subsequent weekly projects will allow each
student to work together with each other member of his or her group. As
the semester progresses, groups should think about how the various
econometric methods we study could be applied to their projects. Some
preliminary analysis on the course project, such as data collection and
cleaning, may be done during the semester as well. During the last two
weeks of the semester, work on the course project will replace the
weekly project assignments and groups will work intensively together to
complete their analyses. The outcome of the analysis will be presented
in a written report and an oral presentation during final exam week.
Grading
Grades will be based on all information the instructor has about your level of understanding of econometrics. This includes evidence from exams, homework, projects, class participation, and individual discussions.
Texts and Software
Text: The main text for this course is James Stock and Mark
Watson's Introduction to Econometrics, 2d ed. This is a modern text
that covers fairly recent developments in econometrics. It has
extensive examples of applications and good problems, both theoretical
and applied. Students will be expected to have unlimited access to this
book, which is available in the Reed Bookstore. A couple of copies will
be on reserve. The datasets referred to in the text (and used in the
examples and problems) are available from the publisher's Web site at http://www.aw-bc.com/stock_watson.
The Stock and Watson text is designed for courses at two levels: quite
easy and more challenging. Much of the mathematical analysis has been
segregated into the final two chapters (17 and 18) and appendices
throughout the book. In general, we will cover this more
mathematical material; sections of Chapters 17 and 18 will be
integrated with earlier chapters as noted in reading assignments on the
course calendar.
Other Texts: There are several other
econometrics textbooks on reserve. Some of these are less mathematical
than Stock and Watson; others are more rigorous. Most sections of the
reading list has references to relevant sections of these texts that
cover the same topic. There will be some required reading assignments
from these alternative texts, but most of the time you are simply
encouraged to explore these texts if you find that our presentation is
inappropriately easy or difficult.
Applied Assignments: Ernst Berndt's The Practice of Econometrics is a unique book. It offers chapters covering ten applied areas of econometrics. Each chapter surveys several important empirical studies in the area, then presents a series of empirical assignments (with the requisite data) in which you may replicate or extend the studies. Berndt's book will be the basis for several of the weekly econometric projects during the semester.
Software: Students will use the statistical package Stata for most course assignments. This software is available on the computers in the IRCs and the PPW. If you are interested in having your own copy of Stata for either PC or Mac, you may purchase a one-year student license to the full version of Intercooled Stata 11 for $98.00 from http://www.stata.com/coursegp.html. (The GradPlan ID is NNECON.) A full set of Stata manuals is on reserve in the library.
The econometric package EViews is a more powerful platform for time-series work. You may wish to use it for some of the time-series projects in the second half of the course. Its publisher no longer supports a Mac version, so EViews is only available on the PCs in the PPW.