Monday, May 2, 2011

The Graduate Curriculum

Macro navel-gazing turns (again) to the graduate curriculum....

Brad DeLong:
The fact is that we need fewer efficient-markets theorists and more people who work on microstructure, limits to arbitrage, and cognitive biases. We need fewer equilibrium business-cycle theorists and more old-fashioned Keynesians and monetarists. We need more monetary historians and historians of economic thought and fewer model-builders. We need more Eichengreens, Shillers, Akerlofs, Reinharts, and Rogoffs – not to mention a Kindleberger, Minsky, or Bagehot.

Yet that is not what economics departments are saying nowadays.

Perhaps I am missing what is really going on. Perhaps economics departments are reorienting themselves after the Great Recession in a way similar to how they reoriented themselves in a monetarist direction after the inflation of the 1970’s. But if I am missing some big change that is taking place, I would like somebody to show it to me.
As if it weren't bad enough that some assistant professors are writing blogs, a Michigan grad student named Noah Smith is blogging, too.  After a description of his first-semester macro class, he says:
This course would probably have given Brad DeLong the following reasons for complaint:

1. It contained very little economic history. Everything was math, mostly DSGE math.
2. It was heavily weighted toward theories driven by supply shocks; demand-based theories were given extremely short shrift.
3. The theories we learned had almost no frictions whatsoever (the two frictions we learned, labor search and menu costs, were not presented as part of a full model of the business cycle). Other than Q-theory, there was nothing whatsoever about finance* (Though we did have one midterm problem, based on the professor's own research, involving an asset price shock! That one really stuck with me.).

At the time I took the course, I didn't yet know enough to have any of these objections... 
Paul Krugman:
[M]odern graduate-level macroeconomics has managed to bury and forget what earlier generations knew, so that what was billed as intellectual progress ended up being, in crucial ways, intellectual regress. 
Hmmm.... yes and no...

Grad school is trying to produce "productive" scholars, good practitioners of what Thomas Kuhn would call "normal science", who generate publications in academic journals.  As such, much of it, especially the first year, is about learning techniques and terminology.  Most of what I remember from my first year is doing alot of algebra - and that's not a bad thing, being at least somewhat good at algebra turns out to be pretty important, as is knowing about things like Kuhn-Tucker conditions, Hamiltonians, and Bellman equations.

In that sense, I don't think economics PhD programs do such a bad job.  My first year graduate course was mostly growth theory and dynamic consumption theory, which were good vehicles for exposing us to the nuts and bolts of macroeconomic models.  What is missing, is a sense of perspective and context.  Contemporary academic models are grounded in "microfoundations" - the optimization problems of forward-looking agents - and, as such are very different from the models taught to undergraduates.  First-year graduate students learn quickly that macroeconomics is very different from what they expected (i.e., its "micro with time subscripts"), but they don't know why.  Time is precious in putting together a course, but I think a prologue which develops the motivation behind contemporary methods - principally the Lucas Critique and rational expectations revolution - would be time well spent (this article by Greg Mankiw is a good place to start).

DeLong and Krugman lament the lingering prominence of the "saltwater" Real Business Cycle (RBC) paradigm, and they have pointed out some startling statements by prominent true believers.  However, there is good reason for nonbelievers to learn these models, as the methods used by them are also at the core of many state-of-the-art New Keynesian models.

If a PhD program can get its students through some growth theory and consumption theory, which come together in the Ramsey-Cass-Koopmans model, and take the small step from there to RBC models in the first year, those students would be well-prepared to study models with frictions and market failures more relevant to current problems in the second year.

Previously, I have also argued for including history of economic thought in the graduate curriculum.  The trick would be to get people to take it seriously.  This would help recover some of the insight that Krugman (rightly) believes have been obscured.  Moreover, this might get students thinking more broadly, which would improve the likelihood that they might actually be in a position to re-think and change existing paradigms, rather than just being "productive" within them.

No comments: