|
|
|
|
Name | Prof. Allan W. Gregory | Prof. Chris Ferrall | Jason Hwang |
Office | MC A424 | MC A519 | MC A504 |
Hours | Wed 10:00-12:00
Thur 08:30-10:00 by appointment |
Wed 11-12 Thur 11-12 & by appointment |
Mon 4:00 – 5:30
Tue 3:30 – 5:00 (in Dunning 350) |
Phone | 533-2299 | 533-6658 |
On-Campus: 74030
|
awg@qed | ferrall@qed | hwangj@qed | |
|
The description above was written many years ago and has not been
updated. Despite this heritage it is surprisingly accurate, but a
more precise and detailed discussion of our specific objectives is warranted.:
Part I
In Economics 351, you studied the classical linear regression model. One key assumption of this model was a fixed X (non-stochastic regressors). For time series models, this assumption is untenable. Researchers are now compelled to understand the econometrics of time dependent processes. While the tools of classical econometrics are extremely useful in advancing in this direction, there are many important departures in time series that require additional machinery.
The intention of Part I is to learn and to use the tools of applied time series analysis. This modelling technique is atheoretical.
Our emphasis will be on application; that is, taking the concepts developed in the lectures and employing them in two time series projects. However, there will inevitably be a great deal of theory in the lectures given the nature of the topic.
Notes are bound and sold to you at cost from the AMS Publishing office.These notes are intended to serve as the basis of all the theory. No other book is required.
Part II
As with part one, Economics 351 is our departure point. The first project gives you the opportunity to learn by doing the first steps in the art and practice of applied micro-econometrics.
The second project in Part II gives you the opportunity to learn
by doing the most important class of statistical model beyond multiple
regression for analyzing survey data sets, namely limited dependent
variable (LDV) models of various sorts (logit, probit, and tobit).
The same process is required in the second project as the first, but now
much more attention is paid to applying techniques appropriately and correctly
interpreting the results.
Part II: the second six weeks of both sections, taught by Professor
Ferrall.
|
|||
|
|||
|
|
|
|
I |
|
|
|
|
|
|
|
|
|
|
|
Notes: Each project should be no longer than
10 typed double spaced pages. These projects are viewed as separate
two or three week assignments, and the deadline is taken as given. No exceptions.
In the second half of the second weekly lecture (For Section A -Wednesday
and Section B- Thursday), I will give you some time series
exercise to do on the computer using the built-in STATA programs or
the ones I have written. The real intention here is to help you get familiar
with the STATA code so that you can do your projects.
At the end of 6 weeks, I will make an overall assessment of this work,
as well as your efforts in class and assign a grade out of 10. These weekly
in-class assignments are marked gently and much of the grade is awarded
simply for giving it the old college try.
|
|||
II |
|
||
Estimation of a Limited-Dependent Variable Model | |||
In-class Exercises and Participation | |||
Notes: Each project should be no longer than 10 typed double-spaced pages, excluding tables and appendices. Projects must be posted on the students' qlink WWW homepage Each project MUST be done in groups of two or three. Partners can be from both sections, and they can be different for each assignment. (Prior approval for a solo project is possible only with an explanation.) The in-class mark will be a function of student performance on in-class exercises and general level of constructive participation in class. This includes not only asking and answering questions, but also politiely listening to other students, pointing out problems or confusions in material, etc. |
|||
Total |
|
||
|