From Think Markets:
Oliver Blanchard tells us “Where Danger Lurks” in the macro-finance world.
The big theme is nonlinearity, which is a profoundly conservative move: DSGE modeling is just fine and we don’t need to rethink it at all. We just need to add in some nonlinearities. Blanchard does not tell how to calibrate a model with extreme sensitivity to initial conditions. But if the system is chaotic, it is also unpredictable, so how can you pretend to merely add nonlinearities to DSGE models? It seems like a pretty direct contradiction to me. I mean, you can have the model in a trivial sense, of course. But calibration is an empty exercise that will not let you look around corners.
Blanchard’s second main message is alarming: We do need theoretical innovation, however, in measuring systemic risk. In the modern network literature on financial markets and cascades, one key point is risk externality. My portfolio choice makes your portfolio riskier. We need two things to fix this market failure. First, we need Pigou taxes, which cannot be calculated unless everyone tells the regulator his portfolio so that it can measure systemic risk and calculate a separate Pigou tax for each financial institution. Second, we need to reduce systemic risk. (“[S]teps must be taken to reduce risk and increase distance” from the “dark corners” of the macro-finance system.) In the network literature I suspect Blanchard is alluding to, this is to be done (at least in some of the articles) by having the regulator directly control the portfolios of financial institutions. (Names include: Acharya 2009; Beale et al. 2011; Caccioli et al. 2011; Gai, Haldane,and Kapadia 2011; Haldane and May 2011; and Yellen 2009, 2011)
I take a rather different view of both economic theory and the crisis in my recent IEA Hobart paper From Crisis to Confidence: Macroeconomics after the Crash.