May 25, 2010
The econometric debate continues
The new JEP has a nice symposium on the state of empirical economics, picking up where Rigotnomics left off. It starts off with the mostly harmless econometrics guys who talk about a credibility revolution which primary improvement has been a focus on the quality of empirical research designs and random assignment. But the consensus seems to be against them. Ed Leamer, who's classic argued for more sensitivity analysis almost 30 years ago, still stand on his position. Keane suggests economists could draw lessons from the field of marketing, where the structural paradigm is dominant and where there is great emphasis on external validation. Sims warns that the mostly harmless guys make overbroad claims for their favored methodologies. Nevo and Whinston also argue that credible analysis can come in many guises, so experimental designs seem like a very narrow and dogmatic approach to empirical work. They then explain why empirical analysis in industrial organization differs in such striking ways from that in fields such as labor and development, which have recently emphasized the methods favored by Angrist and Pischke. Finally, Stock repeats what others have said, mainly that "design-based research" misses an important part of the story.