Paul Romer on “The Trouble With Macroecomics”

Eminent macro/growth economist Paul Romer denounced business cycle macroeconomics. I strongly advise reading his wonderful paper, which is devastating and funny. I’d also look at Noah Smith’s intelligent discussion. But if you insist, read on.

Here is a fair use of Romer’s abstract

For more than three decades, macroeconomics has gone backwards. The
treatment of identification now is no more credible than in the early 1970s
but escapes challenge because it is so much more opaque. Macroeconomic
theorists dismiss mere facts by feigning an obtuse ignorance about such simple
assertions as “tight monetary policy can cause a recession.” Their models
attribute fluctuations in aggregate variables to imaginary causal forces that
are not influenced by the action that any person takes. A parallel with string
theory from physics hints at a general failure mode of science that is triggered
when respect for highly regarded leaders evolves into a deference to authority
that displaces objective fact from its position as the ultimate determinant of
scientific truth.

Romer argues that Robert Lucas and Edward Prescott decided that monetary policy didn’t matter roughly about the time when Paul Volcker’s FOMC proved that it does. This was crazy. Also the field either followed them (fresh water macroeconmics) or tried to reconcile their assumption of a rational representative consumer with Keynesian policy conclusions and with reality (salt water macroeconomics).

Romer argues that the resulting effort leads to non explanations in which economic fluctuationsa re caused by otherwise unobservable and completely unexplained exogenous shocks. The polemical rhetoric is largely based on the names he chooses “phlogiston, “investment specific” phlogiston, a troll, a gremlin, ether, and caloric.”

The critique is devastating. The point is that the first rational representative consumer models had many testable implications which were rejected by the data — for example in the first model predictable changes in income must not be correlated with predictable changes in consumption. The current models have one free parameter per correlation to be explained. This is exactly what one would expect to happen if the core assumptions are not useful approximations to reality (that is if the approach is totally misguided).

Sorry that he also discusses identification. The point is that modern DSGE macroeconomic models fit a small number of data points with a huge number of parameters. As a result, the final model used for forecasting and policy evaluation is based largely on assumptions. The idea seems to be that these models are superior to old models, because there is a lot of math between the arbitrary assumptions (which Romer calls FWUTV’s for facts with unknown truth value)and the interaction with data

“In practice, what math does is let macroeconomists locate the FWUTV’s farther
away from the discussion of identification. The Keynesians tended to say “Assume P
is true. Then the model is identified.” Relying on a micro-foundation lets an author
can say, “Assume A, assume B, … blah blah blah …. And so we have proven that P is
true. Then the model is identified.”

I think this is universally acknoweldged — I use the phrase “assumptions washed clean in the blood of Bellman” to try to express the point.

I really don’t have much to add to this. I agree entirely. There are more links after that one — as Romer notes, his point has been made by Paul Krugman, O.J. Blanchard, and Canova and Sala. Noah Smith has made similar arguments. Most strikingly, Narayana Kocherlakota, who has been head fresh water macroeconomist as chairman of the U Minnesota econommics department and of the Minneapolis Federal Reserve Bank, recently did too.

Sadly, this post is mostly about how we macroeconomists got here, where we were back in around 1970, and where should we go from here. That’s really two topics, because (dog bites man) I agree with Krugman that we should go back there.

Romer doesn’t just criticize today’s macroeconomics — he also criticizes the macroeconomics “of the early 70s”. Romer didn’t respond to Krugman’s thoughts on Tobin (which were written after he wrote his paper). Could he ? This is tiresome intellectual history (of my childhood years too) but I don’t think he can. I believe that Romer consistently insists on considering only one strand of early 70s macroeconomics — the large non micro founded macro models. He insists that the merits or lack thereof of those models was the important topic of debate in the 1970s.

But what does this have to do with Tobin, Solow and Samuelson ? None of them worked with the large macro models. In my ignorance, I know of no case in which they mentioned the output of those large macro models. Writing such large models was an active (and sterile) area of researhc in the 60s and early 70s, but it wasn’t the only area of research.

Reading Tobin, Solow and Samuelson, one finds small models and a whole lot of words. Often their arguments were qualitative. One of the leaders of the movement towards numerical models, Jakob Marscack anticipated the Lucas critique (and was cited by Lucas in “Econometric Policy Evaluation: A Critique”) so it would be properly called the Marschak critique, except for the fact that many others made the same point. Now it may be that we can and should aim for more than verbal arguments and little stylized models which just illustrate what the words mean (natural scientists have certainly achieved more). But that doesn’t mean Keynes, Hicks, Samuelson, Solow, and Tobin have nothing to teach us. Their arguments should not be dismissed without reading their papers. The main point is that they are not the developers of the large macro models.

I can’t help going back to my complaints about something else Romer wrote “Solow’s Choice”. Here he briefly suggests that old Keynesians share much of the blame for the sorry state of macroeconomics

Bob Lucas, Ed Prescott, and Tom Sargent led the development of post-real
macroeconomics. Prior to 1980, they made important scientific contributions
to macroeconomic theory. They shared experience “in the foxhole” when these
contributions provoked return fire that could be sarcastic, dismissive, and wrongheaded.
As a result, they developed a bond of loyalty that would be admirable and
productive in many social contexts.

First if it were generally true that researchers who face return fire including return fire which is “sarcastic, dismissive and wrongheaded” abandon scientific principles, the scientific enterprise would hav ended centuries ago. Second, as many noted in reply to Solow’s Choice, Romer has a problem with timing and who started it (his example is a discussion of Lucas and Sargent 1979 which is itself highly sarcastic and dismissive and which followed many sarcastic and dismissive comments made by Lucas).

I wasn’t an economist in the 1970s, but I think I can show that the the Old Keynesian reception o Lucas wasn’t sarcastic, dismissive, or wrong headed. The date is 1971 and the old Keynesian is James Tobin who wrote

tobinonlucas1

tobinonlucas2

I assert that “an outstanding original theoretical paper” is not sarcastic or dismissive (although it may have been wrongheaded). I think this may have been the first official old Keynesian response to Lucas’s criticisms of old Keynesian macroeconomics. It predates publication of Lucas’s paper. The conference might, for all I know, be the first time Lucas presented his model outside of Chicago.

Now Tobin went on to argue that he finds the rational expectations assumption implausible. So did Sargent, the claim is (almost) always that it is a useful modelling strategy not that it is literally true. I think Romer is just wrong about which side started the macro wars of the 70s and 80s.

I also think his idea that Lucas Sargent and Prescott are united by a “bond of loyalty”. He doesn’t present much evidence. He notes Lucas arguing in 2003 that the effects of monetary policy on output are unimportant. He interprets this as an expression of loyalty to Prescott. I was in the audience in 2003. Prescott certainly didn’t come to my mind. My thought was that Lucas has finally discovered the obvious critique of New Keynesian macroeconomics which I heard form Brad DeLong 15 years earlier (to be less modest, which came to my mind immediately when I read on of the founding New Keynesian papers by Greg Mankiw who I immediately accosted (in a library) and heard “that wasn’t my focus in that paper”). But I just don’t see a need to bring Prescott into it. It seems to me that Lucas was arguing that, even if he were to concede that his model was incorrect in the way New Keynesians argue, it’s no big deal.

Romer also cites Sargent politely disagreeing with Lucas. This example seems to me to be even weaker and to amount to the courtesy which is normal for Sargent (but not Lucas or Prescott).

I do recall a striking exception to Sargent’s normal courtesy. I recall a talk about Christopher Sims argument that, sometimes one should deseasonalize data rather than explicitly modelling seasons. Here Sargent showed great respect for Sims (he said he wrote the paper to try to understand what Sims had in mind). Now, first of all, it is not easy to be loyal both to Sims and to Prescott. But second of all, in passing, Sargent mentioned detrending data with a Hodrick Prescott filter and said (quoting from memory) “Ed Prescott thinks you should use growth theaory to study everything except growth.” To me that sounds sarcastic and dismissive (but not wrong headed) and makes me very much doubt that Sargent feels loyalty for Prescott. In any case, Romer presents no evidence.

I don’t think Romer understands how macroeconomics reached its current steady state. I think his outspoken polemic is much too kind to Lucas and Prescott.