Abstract
This article presents a discussion among leading economists on how to do empirical research in economics. The participants discuss their reasons for starting research projects, data base construction, the methods they use, the role of theory, and their views on the main alternative empirical approaches. The article ends with a discussion of a set of articles which exemplify best practice in empirical work.
Models abound in economics. Testing them is rather natural, at least for some. For instance, structural estimation is a natural research strategy when interested in the job search model, an endeavour for which Jean-Marc Robin just received the Frisch medal. But theory is much less clear when one goes further away from economics and what matters is providing clear facts, i.e. clear and robust causal relationships. The quality of data and the quality of identification have become essential elements in our capacity to produce scientific evidence. There, Josh Angrist has led the pack and pushed all of us in directions that were unanticipated 10 or 20 years ago. Another debate revolves around data and the role of experiments. It was therefore a pleasure to have Armin Falk with us. A recent article by List and Levitt (2005) constitutes a good complement to Armin’s response in our discussion. In particular, List and Levitt discuss strengths and limits of experiments in more detail than is possible here, because our questions are, in a sense, more personal. David Blau and Chris Taber offer balanced perspectives, using broad methodological approaches, even though both are strong believers in economic models. All of our interviewees are excellent econometricians who use the most advanced techniques, or even advance the techniques if they feel this is necessary for their empirical goals. They are all role models for empirical economists, even though we might sometimes disagree with one element or one detail of their research strategy. But, when reading them, we learn from them, even when we might have adopted a slightly different route.
No comments:
Post a Comment