Wednesday, June 1, 2011

Friedman`s schizophrenic legacy in economic methodology

Brad DeLong had an unexpected take on Friedman`s methodological legacy in economics, highlighting his desire to stay close to data when theorizing rather than his defense of "as-if" theorizing. In DeLong`s words, Friedman was a (pragmatic) Marshallian rather than a (purist) Walrasian:

are the theoretical mechanisms we are studying things that we can see? Are their predictions consistent with the gross features of reality? Supply curves slope up: if we say that demand has changed and pushed us along a supply curve, is it in fact the case that both quantities and prices have risen (or fallen)? Demand curves slope down: if we say that supply has changed and pushed us along a demand curve, is it in fact the case that quantities have risen and prices have fallen (or fallen and risen)?

If the first-order predictions of our theories are not visible in the first-order movements of the data--quantities, prices, asset values, and expectations--then, Friedman (and Marshall) would say, our theory is broken and we need to fix it.

I`ve often been puzzled by examples Friedman`s  pragmatic, close-to-the-data, uncover-the-actual-mechanisms approach and its mismatch with the message economists took away from his essay on methodology. In a footnote in Hausman's book on "the inexact and separate science of economics" he mentions that Lee Hansen
recalls economists in the 1950s reacting to Friedman`s essay with a sense of liberation. They could now get on with the job of exploring and applying their models without bothering with objections to the realism of their assumptions.
More recently, Nathan Berg and Gerd Gigerenzer wrote a paper where they set up the "as if" methodology associated with Friedman as the great big flaw of behavioral as well as neoclassical economics:
For a research program that counts improved empirical realism among its primary goals, it is startling that behavioral economics appears, in many cases, indistinguishable
from neoclassical economics in its reliance on as-if arguments to justify ―psychological models that make no pretense of even attempting to describe the psychological processes that underlie human decision making.
This image of Friedman as the staunchest defend of absurdly speculative rational choice fiction always seemed at odds with other stories about the man`s research. As I understand it, he pored through meeting minutes from the Fed together with Anna Schwartz to understand why the Fed did what it did during the Great Depression, and he was sceptical of data-fitting and overly complex theoretical models. Also, when the Economic Journal had a 100 year anniversary issue (January 1991, vol 101 no 404) and asked a number of famous economists for their predictions about the "next 100 years" of our discipline, Friedman went back to the early issues to actually see what (if anything) had changed. As far as I remember, the other contributions I read were mainly
economists saying that in the future the discipline would finally move
towards what they themselves had been doing for a long time. Friedman concluded that the core subjects of the late 1800s would still be present, some new topics (e.g., property rights, crime, public choice) would probably be present, along with some new topics. The methods would be an updated but recognizable mix of pure theory, descriptive statistics and econometrics. And to conclude he quoted a conclusion Ashley had made after a similar exercise in 1907:
When one looks back on a century of economic teaching and writing, the chief lesson should, I feel, be one of caution and modesty, and especially when we approach the burning issues of our own day. We economists...have been so often in the wrong!