Rishon Rishon

Paul Krugman - Conservative?

I just read an article by Paul Krugman (via Gene Expression) about comparative advantage: Ricardo's Difficult Idea. Comparative advantage explains why trade makes all parties wealthier, even when one is better at doing all things (has higher productivity) than the other. To be more precise, the article is not about comparative advantage per se, but about the difficulties in getting the general public to understand it. What is truly breathtaking about the article is that this avowed leftist Bush-hater makes a very eloquent case against the prevailing intellectual zeitgeist, and in favor of - Conservatism!

Modern intellectuals are supposed to be daring innovators, not respecters of tradition. As any publisher will tell you, books about startling new scientific discoveries always sell better than books about known areas of science, even though the things science already knows are in many ways stranger than any of the speculations in the latest cosmological best-seller. Old ideas are viewed as boring, even if few people have heard of them; new ideas, even if they are probably wrong and not terribly important, are far more attractive. And books that say (or seem to say) that the experts have all been wrong are far more likely to attract a wide audience than books that explain why the experts are probably right. Stephen Jay Gould's Wonderful Life (Gould 1989) which to many readers seemed to say that recent discoveries refute Darwinian orthodoxy, attracted far more attention than Richard Dawkins' equally well-written The Blind Watchmaker (Dawkins 1986), which explained the astonishing implications of that orthodoxy.

He even elucidates intellectual fondness for impenetrable ideas, even when they are wrong:

Intellectuals do reserve, both in evolution and economics, a small pedestal for mathematical modelers -- as long as their models are confusing and seem to refute orthodoxy. Call it the "Santa Fe syndrome". At one point in Dennett's book he reports a list of the top ten objections raised to Steven Pinker's theories about the evolution of language; one of them is "Natural selection is irrelevant, because now we have chaos theory". At about the same time I read this passage I had received a barrage of protests over an article that tried, without explicit mathematics, to walk through some simple models of international trade (Krugman 1994); several of the letters insisted that because of nonlinear dynamics it was impossible to reach any meaningful conclusions from simple models. ("Have you ever thought about the implications of increasing returns? You should read the work of Brian Arthur and Paul Romer.")

There are two odd things about the popularity of certain kinds of mathematical modeling among intellectuals who are generally hostile to such models. One is that the preferred models are typically far more difficult and obscure than the standard models in the field. The other is that the supposedly heterodox conclusions of these models are often not heterodox at all. To take a theme common to both evolution and economics: the idea that small random events can under certain conditions set in motion a cumulative process of change is the theme both of "peacock's tail" accounts of sexual selection and of external economy accounts of international specialization, both familiar stories that lie well inside the boundaries of academic orthodoxy, stories that can be and are illustrated with simple models in advanced undergraduate textbooks like Maynard Smith (1989) and Krugman and Obstfeld (1994). Yet many intellectuals believe that this idea was discovered at Santa Fe and challenges the foundations of both fields.

The secret to the popularity of certain mathematical modelers, I suspect, is that they are valued precisely because they seem to absolve intellectuals from the need to understand the models that underpin orthodox views. Hardly anyone tries to understand what the Santa Fe theorists are actually saying; it is the pose of opposition to received wisdom, together with the implication that in a complicated world you can't learn anything from simple models anyway, that is valued, because it seems to say that not knowing what's in the textbooks is OK.

I, myself, have a fondness for both evolution and economics precisely because they are, at heart, so simple. But as they say about the game, Go: A minute to learn - a lifetime to master! So too, evolution and economics. The more you look at them, the more interesting and surprising they get. A good friend of mine who is an economist once told me, he has the feeling that if Adam Smith were to come back to life, he would look around at the field and say, "That's obvious!" I get the same feeling about Darwin, though in his case he did his work with no knowledge of genetics. He would probably say, "Eureka, the missing link! Now I understand everything!"

Krugman lists three non-obvious assumptions that you have to understand before you can understand comparative advantage:

Wages are determined in a national labor market

Constant employment is a reasonable approximation

The balance of payments is not a problem

I would list two:

Economics is not a zero-sum game - in a voluntary transaction everybody wins

Economics is a feedback system - in the long run, secondary effects can be greater than direct effects

These are both hard concepts - not well understood by most people in theory. In practice, people understand them very well when applied to their personal lives: "I have too much corn, and you have too much cloth, let's trade some of my corn for your cloth", or "If you do everything for your children, they'll never learn to stand on their own feet". The first illustrates that economics is not a zero-sum game, the second that secondary effects (teaching your children to stand on their own feet) are sometimes more important than direct effects (doing something for your children). I would call Krugman's three assumptions examples of his "Sante Fe syndrome" - they are hard concepts which are used to supposedly refute the point, but instead merely obscure it - without refuting it.

Posted by David Boxenhorn at November 30, 2004 11:41 AM