I was a PhD student in Economics at NYU in 1986-1987. I remember thinking each week during that school year that that was going to be the week that I learned something new in economics. Unfortunately, it never happened. I was not taught a single theory in the PhD program that I had not already learned as an undergraduate.
Oh, I learned a lot of new math. Math that was abstract and taught in order to prove correct the theoretical models that I already knew. And, by proof I mean social science proof, where many assumptions and axioms are used to “prove” that a particular policy or strategy would have a particular effect when everyone acted in a rational, self-interested way. Of course, much of the math that is learned in economics in graduate school is not useful and often misleading, especially when merely mastering the math gives smart but naive 20-something mathematicians who have little worldly experience a lot of confidence that their economic theories are correct because their models work.
So, I had a laugh when I saw the video below on Harvard Econ professor Greg Mankiw’s blog. It is a cartoon of an Econ PhD student explaining to his mother over Christmas break what he is learning (I clipped the larger video, which can be found on Mankiw’s blog, to get to the essence of the cartoon):
The second news item that prompted this post is not so amusing. Kenneth Arrow and the Austrian School of Economics are right. Economics needs to be de-mathematized. If it were, I would finish the PhD (I bailed on it and finished with an MA as soon as I found a job). Graduate school in economics should instead rely much more on experimentation like in Kahneman and Tversky; focus less on math and more on statistical measurement, econometrics, and inference. Then, economics PhDs could once again add value. An excerpt:
How to Save Economics
first, economists should resist overstating what they actually know. The quest for certainty, as philosopher John Dewey called it in 1929, is a dangerous temptress. In anxious times like the present, experts can gain great favor in society by offering a false resolution of uncertainty. Of course when the falseness is later unmasked as snake oil, the heroic reputation of the expert is shattered. But that tends to happen only after the damage is done.
Second, economists have to recognize the shortcomings of high-powered mathematical models, which are not substitutes for vigilant observation. Nobel laureate Kenneth Arrow saw this danger years ago when he exclaimed, “The math takes on a life of its own because the mathematics pushed toward a tendency to prove theories of mathematical, rather than scientific, interest.”
Financial-market models, for instance, tend to be constructed with building blocks that assume stable and anchored expectations. But the long history of financial crises over the past 200 years belies that notion. As far back as 1921, Frank Knight of the University of Chicago made the useful distinction between measurable risk and “unknown unknowns,” which he called radical uncertainty. Knight’s point was that in a period of radical uncertainty, expectations couldn’t be anchored because they have nothing to latch onto. Financial theories and regulatory designs that hinge on the assumption of stable and anchored expectations are not resilient enough to meet the challenges presented by real financial markets in radically uncertain times.
The third remedy for repairing economics is to reintroduce context. More research on economic history and evidence-based studies are needed to understand the economy and overcome the mechanistic bare-bones models the students at Harvard objected to being taught.
Update 1/25/12: After writing the above I decided to do some research on Econ PhD programs and I came across a paper co-authored by the Chairman of the Economics Department at the University of Virginia. The paper on game theory was published in the American Economic Review in 2001. The following excerpt was taken from the introduction:
The rationality assumptions that underlie (game theory) are often preceded by persuasive adjectives like “perfect,” “intuitive,” and “divine.” If any noise in decision making is admitted, it is eliminated in the limit in a process of “purification.” It is hard not to notice parallels with theology, and the highly mathematical nature of the developments makes this work about as inaccessible to mainstream economists as medieval treatises on theology would have been to the general public.
Brilliant. Charles Holt’s whole paper, a nice overview of game theory–especially one-shot games–and its limits, can be found here: