Simple Ideas and Complicated Models

My comment on a VC post.

Thank you for linking to the Hakes article. It rings true.

The problem it concerns is not exactly poor writing, though. Rather, it is the problem of the brilliant but simple new idea. If you explain it simply, people say, “That’s obvious”, even if they would never have thought of it in a million years. Thus, you have to math it up, or disguise it in a convoluted hypothetical, or quote in foreign languages to get it published.

When I visited Chicago back in 1989 I was amazed at what George Stigler and Gary Becker could do in seminars. They’d ask a simple one-sentence question (“Have you thought about X?”) and entirely demolish (or expand) somebody’s paper. Unfortunately, many of us refuse to recognize that kind of Feynmannian brilliance, which, indeed, is distinct from IQ.

I was just thinking of this topic today because I have what might be a novel idea on why deposit insurance is useful. It’s so simple that I could write the paper in words, with no equations. But I wonder whether a top journal in economics would publish it then, even if the editor believed that the idea was worthy of a Nobel Prize.

What’s the solution? One is to wrap up the idea in unnecessary math modelling. Another, which is often used, is to have two parts to the paper. The first part is to explain the idea in words or with a numerical example. Everybody reads that part, and referees decide whether to accept the paper based on it. The second part is the general model, all mathed up. Referees require that part, but they don’t really read it.

I should add that I am a conventional modern economist, constantly building math models and believing that they are utterly necessary for most economic research. But not every idea needs a math model.