A lot of economic theory is empirically empty. Most of what is “under the hood” in an economic model makes no difference whatsoever from the point of view of the “empirical implications.” Here is an example to illustrate what I mean.

There are two ways I could present to you the standard theory of bidding behavior. The first way, and the way we actually do it, starts with some primitive hypotheses and derives the theory from them. So I might start by specifying the rules of the auction, then state some assumptions like that the bidders have certain payoff functions, they are maximizing utility, they have some specific information, and they play an equilibrium. All of these assumptions together tell you what bids you can expect.

The second way is simpler. I just write down the bids you can expect. That is my theory. I hypothesize a certain (joint distribution of) bidding behavior. Now, this theory is incredibly boring. But from an empirical point of view; i.e. in terms of the observable predictions it makes, it is equivalent to the first version of the theory.

(You might be thinking that the first theory makes more hypotheses and these hypotheses can be tested independently of their joint implications. I have two responses. A) that’s probably not true. Utilities, information, optimization, and equilibrium are all separately vacuous assumptions. They have implications only in conjunction with each other. B) from the empirical point of view that’s only a stronger argument that the simpler theory is preferred. If we can match the data on bids, does it matter whether the underlying hypotheses are rejected?)

If the point of economic theory was to generate empirical predictions, most of economic theory would be a waste of time. But it’s not the point.

There is value in a theory that describes the route from parameters to conclusions even if those parameters could never even be measured. And that is because the users of economic theory have a source of data that is unavailable to empirical researchers, and would not be permissible anyway: personal experience and intuition.

If you are about to compete in an auction, you look around you and see your competitors, the context, and the object up for sale. This is a unique situation because every situation is unique. You have a sense that one guy is an insider, one guy is bankrolled, and one guy is his competitor. You know that all of these things matter for how you should bid but you want to know how they matter. A theory is a recipe for translating your completely subjective description of the situation into an educated hypothesis about what is going to happen.

(Drawing:  Summer Off-site from www.f1me.net)