It's that time of year again - classes are over and I'm grading exams and projects. This term I was teaching an introductory Economic Statistics course and an elective grad. Econometrics course. I've graded the latter exams., but I'm buried in about 200 scripts for the undergraduate course. I've also got the term projects for the graduate course to go through. They look really good!
About two thirds of our Masters students elect to take the second course in econometrics - the one I've just finished teaching. It's compulsory for Ph.D. students. Some of them then go on to take additional econometrics courses. This particular elective course is titled "Themes in Econometrics". When I'm teaching it I tend to cover three thematic approaches to econometrics - this time these were MLE (and the related testing principles), method of moments (and GMM), and Instrumental Variables estimation.
The idea is to present the students with the "big picture" - to show that that there are some thematic connections between the various specific tools that we use in our work. I don't want them to leave our clutches thinking that they have to dream up a new solution to every new problem that they subsequently encounter.
It's a theory course - for the most part. However, we include some applied material in the classes and, most importantly the students have weekly computing lab. classes where they get their hands dirty.
There's also a term project that everyone has to complete. Students generally work on an individual project, but I allow them to work as a team (with a maximum of two students per team), if they wish. More is expected of a team than an individual.
At the start of the term I give them some broad guidelines and spell out what form the final reports should take. When it comes to choosing a topic, my main requirement is that the work must exhibit the application of new knowledge acquired during the course..
Although I offer a list of "illustrative topics", this isn't a list for the students to choose from. They have to come up with their own topic, and get the "O.K." from me as to its suitability, before they get underway.
The projects tend to fall into one of two categories. First, there are ones that are an empirical application, involving "real data". Then, I allow them to undertake a project that involves a simulation experiment that investigates the properties of some test or estimator, in a "non-standard" situation.
Students are free to use whatever software they wish, and this group chose to use EViews, MATLAB, R, and STATA.
It's always fun to see what they come up with, and this term was no exception.
Projects in the first category included:
- A count data analysis of cholera cases.
- A model of consumer credit delinquency in Canada.
- The determinants of market work and schooling in Bangladesh.
- Goodness of fit for the multi-fractal asset model.
- Savings and Inflation in Canada.
- Three that investigated aspects of modelling volatility in returns on financial assets.
The topics for projects of the second type included:
- The finite-sample properties of the t-test in the context of a Box-Cox model, estimated by MLE.
- The finite-sample bias of the MLE for the zero-inflated Poisson model under model mis-specification.
- The risk of the James-Stein estimator under an "inverted normal" loss structure.
- An extended investigation of the Wald test for non-linear restrictions.
All of the students have done a terrific job. This certainly augers well for their future research - whatever specialization they choose!
© 2012, David E. Giles