John Rust's take on the debate is available on the website below.
link here
comments are a response to a paper by Michael Keane. the Keane paper makes points that are similar to the Deaton "randomisation in the tropics" article that created so much debate earlier this year. The Keane paper contains some interesting discussions of the role of quasi-experiments in economics, the extent to which "atheoretical" exploration of relationships has led to progress in other sciences, the importance of validation and related issues.
link here
I read this years ago. Its a bit of a whinge but very interesting nonetheless. While there is no harm in a bit of Levitt-bashing (a growing industry, inevitably) its important not to confuse one person's work, however influential or flawed, with a general approach to econometrics.
ReplyDeletethe keane paper is more subtle kevin and raises some very interesting issues.
ReplyDeleteI may be somewhat biased in this debate, but I think it's important to read the response of Guido Imbens.
ReplyDelete