Ivo Welch, Sep 1, 2009
In Fall 2009, am willing to sponsor a very specific independent study. (I do not sponsor other independent studies.)
The goal of the project is to produce 4 consumption series at monthly frequencies for each of the 50 US states. We want to calculate durable and non-durable consumption, seasonally and non-seasonally adjusted. This should be at high frequency (monthly) as far back as possible. The end result should be 200 data series (4*50) and a report describing how the data were created, what the problems were, and how they reconcile into the aggregate US data series.
The raw data itself is at ICPSR. (The BLS page is here.) The project entails learning what the data is, how it is structured, programming it intelligently, learning how seasonal adjustment is done, and learning how these 50 state data combine to make the aggregate US consumption data. The data begins in 1972, but may have holes.
The deadline for understanding the data is Oct 1. The deadline for producing the data series is Nov 1. The deadline for the analysis (how the four equivalent US data series come about from the 50 states) is Dec 1. The deadline for the full writeup and report is Dec 15. When it comes to grading, I will take timeliness into account.
Background requirements are good programming and econometric skills. The computer language itself is not important, although I would suggest that the use of R or perl would make the project easier.
Please realize that the grade is not an automatic A. You have to produce something good to earn a grade.