Saturday, September 19, 2009

Levine Responds to Krugman

David Levine writes an open letter to Paul Krugman on the state of Macroeconomics. At one point he notes:
our models don't just fail to predict the timing of financial crises - they say that we cannot. Do you believe that it could be widely believed that the stock market will drop by 10% next week? If I believed that I'd sell like mad, and I expect that you would as well. Of course as we all sold and the price dropped, everyone else would ask around and when they started to believe the stock market will drop by 10% next week - why it would drop by 10% right now. This common sense is the heart of rational expectations models. So the correct conclusion is that our - and your - inability to predict the crisis confirms our theories. I feel a little like a physicist at the cocktail party being assured that everything is relative. That isn't what the theory of relativity says: it says that velocity is relative. Acceleration is most definitely not. So were you to come forward with the puzzling discovery that acceleration is not relative...
Levine is being coy here. Without mentioning it, he is trying to outline Krugman's own model of first-generation exchange rate collapses. But not exactly. For in Krugman's model there is a fundamental that is driving the currency collapse. Excessive money growth is driving down reserves, so in Krugman's rational expectations model, investors do not wait for all reserves to be evaporated before selling the currency, they do so at the first moment that an attack is feasible. Krugman's model ties down the timing of the attack exactly, even though it is based on rational expectations (indeed that is the novel point). Thus, I suppose one could argue that with better models of fundamentals we would know more about the timing of crashes.

I should have linked earlier to John Cochrane's response to Krugman. It is long, but well worth reading.

Meanwhile, Gilles Saint-Paul provides a modest defense of the economics profession.

No comments: