Barron's
October 19, 1998
Oh, Those Assumptions!
An engineer, a chemist and an economist are marooned on an island ... .
Maybe you've heard that classic joke, maybe you haven't. But either way, it's worth reviewing, since it bears direct relevance to the work of the 64-year-old Indian economist Amartya Sen, formerly of Harvard, currently at Trinity College, Cambridge, England, who last week won the 1998 Nobel Memorial Prize in Economic Science.
So these three professionals are marooned on the island, and they find a can of tuna, but have no way to open it.
Says the engineer: "Let's find a rock and crush the can open."
Suggests the chemist: "That's impractical. A better idea is to find some chemicals and blast it open."
Quoth the economist: "You people are truly misguided. There's only one way: Let's assume we have a can-opener ... ."
Ah, those assumptions. In an unintentionally funny 1966 article called "Labour Allocation in a Cooperative Enterprise," the future Nobel laureate Amartya Sen sets forth the "assumptions underlying the analysis," some of which, he assures us, are "serious but not especially odd in this branch of economics." They include: "well-behaved utility and production functions; automatic fulfillment of the second order conditions of welfare maximization and of equilibrium; no uncertainty; perfectly competitive markets; homogeneity of labour . . . ." -- a whole set of can-openers truly dizzying for the colossal naivete they reveal.
All right, that article was published years ago, so maybe professor Sen's stuff has gotten better since then. But based on a review of four books of his that I got from my local Borders (two collections, two full-length studies) I see scant evidence of the brilliance that so impressed the Royal Swedish Academy of Sciences and far more of the mind-rot that is so characteristic of academic economists.
Predictably enough, in response to Wednesday's announcement, the New York Times gushed about how great it was that the Nobel should go to a scholar who dared to examine broad social issues, while The Wall Street Journal groused that the award had been copped by an "establishment leftist." I actually see some merit in both those positions, but what both miss is that Sen's myopic approach places him squarely in the academic mainstream. If he's "muddleheaded," as the Journal states, then he came by it honestly.
What undermines so much of Sen's work is the fetish for math-as-method that is born of the mainstream's pathetic desire to ape the physical sciences. Those assumptions quoted above are silly beyond belief, but Sen was quite right to defend them as textbook. And they're really about something else.
Your typical economist assumes such absurdities as "no uncertainty" in markets, even though uncertainty is any market's middle name, and "well-behaved utility and production functions" (read: unchanging consumer tastes and production processes), even though such good behavior is the rare exception rather than the rule -- and he assumes so for only one reason: Because that way, he gets to use a lot of math. Try asking him to make realistic assumptions, and he's forced to use words, which means he might lose the mantle of "scientist."
And the point is, once this dreadful intellectual process is set in motion, he falls off a cliff. When you start imagining that in some deep, theoretical sense nothing ever does change in the economy -- where, to top things off, labor actually can be homogenous, meaning that one worker's skills are the same as another'sthen the appeal of central planning under socialism really does become valid. Now that all those inconvenient problems that capitalism is uniquely able to deal with have been brushed aside, you can concentrate on what top-down control does best, like distribute income more fairly.
In that regard, I hasten to add that it's unfair to single out Sen. When you consider the history of our nearly total misunderstanding of what was happening in socialist economies around the world -- or recall the mindless parroting of socialist "growth rates," which our pundits found so intimidating -- you wonder just where mainstream economists were at the time. Answer: Intellectually out-to-lunch, consuming all those equations.
But of course, Sen deserves to get bashed a bit, because he did, after all, win this year's Nobel. Even at his best, he can be at his worst. Consider one of the few areas in which he said something important-his work on the history of famines. On that topic, I don't for a moment question his passion or compassion, especially since he lived through the 1943 Bengal famine at the age of 9. But I do occasionally question his good sense.
Sen's finding was that the starvation associated with these awful events has mainly been due to lack of access to food, rather than lack of its availability; the food, he says, was there all along, except the access was denied. That's a pretty simple and stark distinction, but amazingly, Sen dresses it up with an absurd formal apparatus.
At the beginning of his 1981 journal article, "Ingredients of Famine Analysis" (one hopes he used that title ironically), we confront a graph with lots of lines and curves on it, in which the horizontal axis is labeled "Food" and the vertical axis "NonFood" and underneath it we read, "With a price ratio p, and a minimum food requirement OA, the starvation set Si is given by the region OAB. If the endowment vector is xi, the person is in a position to avoid starvation... ." and on and on in that vein.
All this is in the service of drawing that simple distinction between access and availability before he goes on to the actual evidence.
And even regarding his actual analysis of famines, it's hard to completely trust his judgment. By solely grinding the socialist axe of access to supply over supply availability, he gives the latter issue a short shrift that is inappropriate.
For instance, he writes sweepingly that the food security of people in the Western capitalist economies "is not the result of any guarantee that the market or profit-maximization has provided, but rather due to the social security the state has offered."
Really? Could food security have nothing to do with the enormous gains in productivity that the market has brought to the farm? Even Sen himself writes of "deaths on a very large scale" because of famine conditions in China during 1959-61. The way agriculture was then organized probably had something to do with that awful tragedy, just as Soviet collectivism has a lot do with the dreadful prospect of food insecurity currently faced by the Russian people.
But getting back to how he combines those "tools from economics and philosophy," in the 1997 expanded edition of his book, On Economic Inequality, Sen writes, "Even for limited application of the merit principle-giving more than the 'norm' to the specially meritorious but not less than the 'norm' to the demented -- it can be argued that the measure of merit is culture-specific."
He then goes on to dismiss the merit principle thus: "While many of us may be content to live in a society which values the ability to lecture more than it values, say, the ability to make loud, shrill noises by blowing sharply through one's nose, we might be perfectly able to give long lectures about possible societies in which the latter quality would be the more desired virtue."
Quite an assumption-but there's your can-opener again. And with words as tortured as those, maybe I'll take the math instead.
