Comment of the Day: Charles Steindel: Olivier Blanchard on Role of DSGE:
On the whole, as one would expect, Olivier's exposition was masterful. The rub is, as you say, at the end...
What are good microfoundations?
As he notes, information asymmetries, behavioral economics considerations, and the like, make the crude "micro" backbones of standard DSGE models laughably at odds with reality. But these same considerations pose serious problems for regular old microeconomic analysis. (I found it semi-hysterical to teach traditional production theory to MBAs this summer, noting to people who were mostly working in complex service sector firms that the textbook models seem to be applicable to single-product factory, and when attempts are made to interject reality the whole thing becomes incredibly complex. Then when I asked them if they thought their CEOs were being paid the value of their marginal product...).
It may be the real intellectual crisis is micro, but since microeconomists don't seem ever to be asked to make forecasts with a date attached there isn't the same visible angst.