Tuesday, August 9, 2011

The real divide between academic economists and practicing economists

Steve Williamson (who I think is just as shrill as Krugman and myself) makes the following and I think accurate point. He responds to the following comment by Larry Meyer (whom I would label as an economic forecaster):

My views would be considered outrageous in the academic community, but I feel very strongly about them. Those models [modern macro models] are a diversion. They haven’t been helpful at all at understanding anything that would be relevant to a monetary policymaker or fiscal policymaker. So we’d better come back to, and begin with as our base, these classic macro-econometric models.

Williamson writes:
The gulf there is not a problem for modern macro, it's a problem for Meyer, who has not taken the time and trouble to understand what macroeconomists are doing and how he can make use of it.

What are modern macroeconomists doing? Coming from a school that used a lot of DSGE models, the focus is not on forecasting but in understanding the economy. What are its relevant components and how important are they? Just because these models do not make forecasts as macroeconometric models does not imply that they are not useful.

The real divide is that DSGE models are unable to provide real time short term policy guidance but I view that more as a data and measurement failure than a model or content failure. But I would also think that large-scale macroeconometric models have not added that much value in the crisis either. Williamson’s remarks were in also in response to Mark Thoma (ht: Econbrowser):

How much confidence would you have in the medical profession if the teaching faculty in medical schools had very little experience actually treating patients, and very little connection to-- even a lack of respect for-- the practitioners in the field? Would your confidence be improved if medical research had little to do with the questions that are important to the doctors trying to serve patients?

Unfortunately, that's a pretty good description of how economics has been practiced....

My sense is that this is certainly not true or at least not now although I don’t really know what is referred to as practitioners. If he means central bankers then I would think that the past few years have seen a lot more interaction than before - I assume that Ben Bernanke in the central bank would certainly use models as well as all the tools that he knows. There have also been an inflow into various regional Feds - the Minneapolis Fed is one that comes to mind (although I would tend to discount them somewhat). Moreover the use of DSGE models seems to have made some inroads into central banks, for instance the use of the Smets-Wouters model at the ECB.

However, it is unclear to me whether DSGE models are really that suitable for policy work, or as a path for future modern macroeconomics either. While more familiar with its predecessor, the RBC model my sense is that DSGE models in general are still very reliant on exogenous (technology) shocks. This has implications for identification of shock propagation and the interpretation of how these shocks will affect the economy and hence policy responses.

For instance, consider the collapse of a bubble. There is no shock that can be identified that can cause a bubble to collapse. Of course, we can assume a shock under the guise of regime shift in expectations (shocks) but in my mind that only assumes the results. But a DSGE practitioner might say that the shock to the financial sector itself is not of interest but how it affects the economy at large, thus for practical purposes the endogeneity of the shock in the financial sector itself is not of interest. The shock to the economy can be assumed to be exogenous for the part of the economy that we are interested in. But then I say that it would leave out the most interesting parts of the what we are interested in.

In any case, DSGE models have become so complex that it is probably true that a clever enough economist (academic or practicing) can model anything he or she wants and get the desired result by assuming that a shock that fits the observed economy. DSGE models don't seem to be able to get away from shocks and it has begged the question since the days of RBC models - what are shocks?

For their part, ‘practitioners’ of economics and finance have not had a stellar record either. Assuming that these are financial economists and forecasters, they neither predicted the crisis but some also abetted in them by misapplying or perhaps misusing and abusing financial models like VAR (which were perhaps developed by academic economists). Or maybe we are all pots and kettles calling each other black.

No comments: