An introductory or soft intermediate applied econometrics textbook that discusses cointegration?

July 25, 2006 11:01 AM

Please recommend an introductory or soft-intermediate applied econometrics textbook that discusses cointegration.

I have a basic familiarity with ARCH, GARCH, and ARMA models, but my non-stat math skills are pretty weak. I don't care about proofs, and I don't really care about derivations. Please note the emphasis on "applied." I do not want to become an econometrician-- I merely want to learn to properly test claims such as "the employment level drives new car sales" or "housing price change decelerates six months after interest rates trough."

I've been testing similar claims using The Granger-Engle method, and I'm getting somewhere, but I fear that I know too little to really interpret the tests I am conducting, and I should really learn this a bit better before I present my findings as fact-- especially considering that much of what I know, I taught myself, so my teacher is terrible.

Also, I'd like to avoid books written especially for one computer application.

posted by Kwantsar to education (10 comments total)

On the other hand, if you think it would be better to just buy RATS and CATS, say so; but tell me what I should buy to learn how to use them properly.
posted by Kwantsar at 11:07 AM on July 25, 2006

It's been a while, but I seem to remember Stock and Watson as the go-to introductory econometrics textbook.
posted by diftb at 1:42 PM on July 25, 2006

Many (I've heard) say that the definitive textbook on Econometrics is by Wooldridge.
posted by SeizeTheDay at 2:26 PM on July 25, 2006

If you have access to EViews, take a look at the "Vector Autoregressions" help topic. EViews tends to have very good plain language discussions of the underlying econometric theory with using application-specific terms.
posted by GarageWine at 2:42 PM on July 25, 2006

diftb and SeizeTheDay, do either of those books devote more than a page or two to cointegration?

GarageWine, I don't have access, but I can finagle an academic license for cheap. Can you give me any more commentary?
posted by Kwantsar at 2:54 PM on July 25, 2006

8 pages. Pgs. 647-654 and a couple of examples. Though I hope you can just borrow this from a library or something, because $140 for 8 pages seems like a pretty crappy deal. If you'd like, I could just scan the pages and send them to you, since I have the book in front of me.
posted by SeizeTheDay at 3:13 PM on July 25, 2006

Your email is in your profile, so I'll head to the library this weekend before I put you through the trouble. Nonetheless, you have my gratitude.
posted by Kwantsar at 3:20 PM on July 25, 2006

I just sent you a PDF of the relevant topic.
posted by GarageWine at 4:44 PM on July 25, 2006

The Kennedy Econometrics book is a pretty easy manual without the textbook price, it has about 5 or 6 pages on cointegration.
posted by stratastar at 9:28 AM on July 26, 2006

This post doesn't mention a book but may be useful

R has a unit root and cointegration package urca that will give you GPL licensed software to run your own tests and view the source code, comments and documentation. It may be worth your while to check out their compilation of open source software for Computational Econometrics.
posted by pegstar at 10:40 AM on July 26, 2006