Are Economics PhDs Learning the Wrong Thing?

Economists rely on models--maybe too much
Photograph by Fotosearch/Getty Images

“Sure, I’ll be your straight man,” says Gene Grossman. He’s the chair of Princeton’s economics department, which U.S. News & World Report ranks in a four-way tie for first with Harvard, MIT, and the University of Chicago. He has rather reasonably agreed to answer a rather rude question: Are the best graduate programs screwing up their economists?

Last week Bloomberg Businessweek ran a short piece on Varoufakis, the former head of the economics PhD program at the University of Athens. Funding has run out for the program, and Varoufakis is now teaching in Austin, Tex. When it was thriving, however, the PhD program in Athens tried to train economists in a new way: It forced them to understand the math of economic modeling but also taught them philosophy and the history of economics.

Varoufakis wanted his students to understand what economic models couldn’t do. “One of the great problems that economics has as a discipline is that we have a tendency to mathematize our theories,” says Varoufakis, “Physicists have done wonders with this, but the problem is, the moment you try to mathematize an economic hypothesis, you end up with mathematical models that are highly indeterminate.” By “indeterminate,” he means that economic models can produce a range of answers. You can fix this by assuming one of two things to be true: Either everyone has the exact same motivations, or time doesn’t exist. “If you introduce both complexity and time in the same model,” says Varoufakis, “anything goes.”

The Rational Expectations hypothesis, for example, predicts that an increase in money supply from a central bank will necessarily create inflation. It assumes that everyone is the same, omniscient person. Varoufakis points to formulas such as Black-Scholes and Value at Risk, which investors use to price derivatives and assess firm-wide risk. “They’re actually very nice,” he says, “I spent many hours happily studying them; they were like a piece of art. And each and every one of them had the same assumption in it—that everyone knows everything in advance.” The problem, says Varoufakis, was not the models themselves, but the way they were taken as articles of faith. Having mastered a model, he says, “you feel that you are one of the masters of the universe, it gives you a false sense of security when, effectively, [the models] are assuming reality away. ”

Journals and universities prefer economists who can present something that looks like science. Economists who enter the financial sector find themselves committed to assumptions that make them lots of money. Abstract models inform concrete government policy. “My extreme worry,” says Varoufakis, “was that the young PhDs coming through the ranks worldwide were utterly illiterate.”

I e-mailed James K. Galbraith, who holds a chair in government and business at the Lyndon B. Johnson School of Public Affairs at the University of Texas, Austin. He suggested that UT hire Varoufakis, so he’s inclined to be sympathetic, but he shares the same perspective. In an e-mail, he writes about the “mysticism” in models, which he defines in two ways: “(1) models that traffic in concepts with no operational meaning—abstractions that will never be observed or measured—and (2) models that frame (mis-frame) problems in ways that lead to extreme and unreasonable policy conclusions. … These are serious issues in the United States, and they do have something to do with the lack of training in economic history and history of economic thought.”

Galbraith’s and Varoufakis’s worry has been around for 20 years, says Grossman at Princeton. “It comes and goes. It particularly comes when economic times are hard, when it appears that the old way has let us down.” Did the old way let us down? “No,” he answers, “which is very different from saying we know everything. … I don’t think medicine has let us down because there are diseases we don’t know how to cure.”

Grossman agrees that models can be abused. An economist is obliged to spell out assumptions, he says, though “just because you write down your assumptions doesn’t mean they’re good assumptions.” Like Varoufakis, he describes interests in the real world that are neither neutral nor academic. “There’s no question that if you’re a trade economist and you write a model that says protection is great,” he says, “the interests will use your model. And they will hire you.”

What economists talk about, he says, matters to pocketbooks—more so than with physicists. But he doesn’t think this should change the way his department trains economists. He is in the business of teaching people who will graduate and go looking for jobs. “You have to teach graduate students the discipline,” he says, “otherwise you’re doing them a disservice.” Princeton has not altered its approach since the financial crisis, other than changes individual professors have made to their core macroeconomics courses.

Harald Uhlig, chair of the economics department at the University of Chicago, another of the No. 1-rated economics graduate schools in the U.S., offers a kind of a model of his own. In an e-mail, he writes: “Above all, I do not believe in central planning. What is true in private markets is true in PhD education as well: It is good to see different places try different approaches, to let the PhD students decide where they want to be educated, and to let the marketplace for future scientists decide what works and what does not. I am sure that if the new PhD program in Athens is successful and produces the top young economics researchers of the next generation, many other PhD programs will take notice.”

It’s not clear whether Uhlig already knew that the department in Athens is on its last legs, so we’ll take his e-mail at face value. It, too, contains an assumption: that the market for grants and hires at economics departments is driven by quality and value to the public alone.

    Before it's here, it's on the Bloomberg Terminal.