The Relationship between Economic Theory and Experiments

Let me begin by relating a story. I was recently a member of my university’s promotion and tenure committee, and one of the cases we discussed was that of a young scientist being brought up for tenure by his department. When I read the docket, I was stuck by a sentence in one of the candidate’s letters, which said, in essence, that if this theorist’s work was supported by upcoming experiments, his work would have to be considered a major accomplishment. What struck me about this statement was not what it said about the natural sciences, but rather what it implied about economics, and more speci…cally economic theory. In all of my years of hanging out with economic theorists, I do not think I ever heard one of them say, "I have just written a theoretical paper, and I hope that, when it is tested, it will prove correct, since otherwise all my work will have been for naught". On the contrary, my experience with theorists is one where they say, "I have just written a theoretical paper; is it not beautiful and elegant?" Historically, theorists in economics have not looked to experimentalists for validation, either because they feel that testing is irrelevant, or else that experiments cannot be trusted to reveal the truth. (Read your referee reports. When experiments fail to corroborate a theory in the lab, it somehow always seems to be a failure of the experiment and not the theory.) Most economic theorists, like mathematicians, judge theories by their elegance, not by their potential for empirical validation. This attitude is very di¤erent from that found in physics. Take Einstein and his theory of general relativity, for example. In the history of that theory, experimental results were intimately tied to the validation of the theory and its legitimacy. In fact, in 1920 Einstein is quoted as saying: