Sunday, May 16, 2010

The House of Econometrics

Last week Russ Roberts interviewed Ed Learner about econometrics. They noted, quite correctly, that any given econometric study has a lot of arbitrariness of it. the author of any work can jiggle the model until s/he gets the desired results. This is a problem and every economist knows it. And because we know it, it takes a lot of studies saying the same thing to be convinced of anything. A brick is too small to build up a house, so we get a lot of bricks. A single study is too precarious to hang a major conclusion on, so you get a lot of studies.

This is crucial because Roberts often points out that whenever he asks another economist to point to an econometric study which changed their opinion on economics, he doesn't get an answer. Of course he doesn't get an answer; no study is good enough! It's like asking which brick of a house holds its roof up. All of them do. But unless you've specialized in the field, you can't remember all of the studies. You probably haven't read them all (hence it appears that studies just confirm people's belief).

Which is why Roberts' question is a bit of a red herring. Intellectual houses are built over decades--it takes that long to get enough studies done. That's why the process of adoption is so slow and why it looks like it doesn't convince anyone. Yesterday's studies convince tomorrow's economist.

No comments: