View source for Economic Realism (Wonkish)

<!-- you can have any number of categories here -->
[[Category:Assuming Theory Applies To A Real World Example]]
[[Category:Paul Krugman]]
[[Category:Milton Friedman]]
[[Category:Common Fallacies Of Economics]]
<!-- 1 URL must be followed by >= 0 Other URL and Old URL and 1 End URL.-->
{{URL | url = http://krugman.blogs.nytimes.com/2014/03/28/economic-realism-wonkish/}}
<!-- {{Other URL | url = }} -->
<!-- {{Old URL | url = }} -->
{{End URL}}
{{DES | des = A discussion of how realistic economic models should be, and for what purpose.  Includes a good criticism of [[Milton Friedman]]. | show=}}
<!-- insert wiki page text here -->
<!-- DPL has problems with categories that have a single quote in them.  Use these explicit workarounds. -->
<!-- otherwise, we would use {{Links}} and {{Quotes}}  -->
{{List|title=Economic Realism (Wonkish)|links=true}}
{{Quotations|title=Economic Realism (Wonkish)|quotes=true}}

{{Text |
Mark Thoma sends us to Paul Pfleiderer (pdf) on the misuse of economic models. Pfleiderer writes his paper as a condemnation of “chameleons” — of what we might alternatively call three-card-monte economics. Someone markets a theoretical model as something that carries lessons for the real world, but when challenged on the assumptions retreats to claiming that it’s just theory.

It’s a very good paper, but I don’t think it’s mainly about chameleons. What it is, instead, is a more general discussion of where, in the process of modeling, you should start “filtering” by what you know about the world. There’s an extreme position associated with Milton Friedman, which says that you should pay no attention to the realism of assumptions, only to the predictions the model makes (which Friedman was arguing largely in order to dismiss the overwhelming evidence against perfect competition in most markets.) Pfleiderer makes short work of this claim, and argues that you should employ filters all the way, testing the realism of assumptions as well as those of predictions.

He’s right about that. But the hard part is knowing how tight to set the filter. As Pfleiderer acknowledges, any model involves untrue assumptions; how unrealistic are you allowed to be? That’s always a judgment call.

I would add one point that I think Pfleiderer would agree to, but that isn’t explicit: a lot depends on the kind of weight your assumptions are being asked to bear.

Let me give an example from one of my own fields, overlapping a bit with Pfleiderer’s. Here’s the question: is it OK to assume rational expectations in financial markets?

Consider first one of the classic models in international macro, Dornbusch’s overshooting model (pdf) of the exchange rate, which is used to show how exchange rates can be as volatile as they are — basically it’s the combination of fast-moving asset markets where prices react quickly to news and slow-moving goods markets where prices can take years to adjust. To make his model as clean as possible (which is to say, very clean — Rudi was my role model), Dornbusch assumed rational expectations in asset markets, so that the exchange rate immediately reflects news about the money supply and other things. Was this assumption realistic? No — there is in fact a lot of evidence that the specific result that interest differentials are equal to expected future exchange rate changes isn’t right. But that really wasn’t crucial to Dornbusch’s point; he was using rational expectations as a simple way to get fast-moving asset prices, and his story wasn’t too sensitive to the precise correctness of the assumption.

Now compare this to people who say that because asset markets are efficient, there can’t be bubbles. They’re making the same assumption — but they’re putting much more weight on it, weight that it can’t actually bear.

So the question isn’t “Is this assumption realistic?” It isn’t even “Is it realistic enough?” It’s, “Is it realistic enough for what I’m trying to do?”
}}