Thursday, September 15, 2016

Macro Mess

It seems to be open season on DSGE macroeconomics again (as it was in 2008).

Paul Krugman says that The State of Macro is Sad, and that OJ Blanchard is too polite to write that he agrees. He also says that Tobin was right all along so the rational expectations revolution was not only unsuccessful but also un-necessary (what a Schama). It isn't surprising that Brad DeLong agrees. They are nowhere near as shrill as Paul Romer.

This will be a long boring unoriginal post commenting on Krugman (comments on Romer here).

First I claim that the state of macro has been sad for longer than Krugman recognised back in August. Then (as often) he praises Friedman while criticizing those who think they are Friedman's followers. He agrees that the Stagflation of the 70s demonstrated that Friedman had made a huge contribution when he presented the natural rate hypothesis in 1968.

The theory of a natural rate of unemployment got a big boost when the Phillips curve turned into clockwise spirals, as predicted, during the stagflation of the 1970s.

I have argued that Friedman was setting up and knocking down a straw man, because old Keynesians didn't use models which ruled out stagflation. I also argue that stagflation does not imply that there is a natural rate. I often direct readers to James Forder

This makes Krugman's post on Tobin interesting to me.

along the way I’ve found myself rereading some writings of actual James Tobin from the time; and it has been a revelation.

Let me focus in particular on Tobin’s 1972 presidential address to the American Economic Association, “Inflation and unemployment” (sorry, I don’t see an ungated version.) I remember how that address was seen among my fellow grad students a few later: it was seen as Tobin’s last stand, a desperate rearguard action in the debate with Milton Friedman over the natural rate hypothesis. And everyone knew that Friedman won that debate, vindicated by stagflation.

Except if you read Tobin again now, he’s the one who looks vindicated.

I love to say "I told him so" but I told him so (same link as above). This means that the one victory of fresh water economists over old Keynesians (which Krugman sometimes concedes happened) didn't actually happen. I also look forward to telling Krugman so again (as I have in the past). The recent post on Tobin isn't the first time he has noted that Tobin was right about the Phillips curve and that the standard intellectual history is based on systematic misprepresentations

So how does the decade of the 1980s end up being perceived as a defeat for Keynesians? To see it that way you have to systematically misrepresent both what happened to the economy and what people like Tobin were saying at the time. In reality, Tobinesque economics looks very good in the light of events.
In all the strange thing is that I have the very strong impression that even the shrill one himself can't resist Ballance. He often feels the need to concede something to the Chicago economics Department and to admit some of the blame belongs to old Keynesians. From time to time, he notes that both sides didn't have a point (at least not an original point). But Friedman's defeat of the stable expectations unaugmented Phillips curve keeps coming back from the dead (in Krugman's terms Krugman's concession to Chicago is a cockroach -- uh oh now I've gone and typed it).


Anonymous said...

It's much bigger than "what is the right model". The point Romer is making is that there are limits to mathematical modelling and working in abstraction full stop, (and you can add to this there are also limits to working purely with quantitative data). If you want to understand why things happen you cannot rely entirely on model and/or quantitative data. You need things like primary documentation - which may be non-quantitative. What is patently obvious to most people - historians, psychologists and many others curiously hasn't been to most economists. Romer's critique of Krugman hits hardest when he says that "to say all models are wrong is not good enough". That is very clearly an attack on the Krugman clique.

Robert said...

The phrase is "all models are false" and it isn't an attack on models. It is a claim that models are false by definition so proof that the assumptions in a model aren't literally true is pointless. This is a point on which Romer, Krugman, Lucas and Prescott agree. All consider models potentially useful.

In any case, Romer's line isn't clearly an attack on Krugman. There is no hint of any reference to any distinguishing feature of Krugman's thought.

"In response to the observation that the shocks are imaginary, a standard defense invokes Milton Friedman’s (1953) methodological assertion from unnamed authority that "the more significant the theory, the more unrealistic the assumptions (p.14)."
More recently, "all models are false" seems to have become the universal hand-wave for dismissing any fact that does not conform to the model that is the current favorite. The noncommittal relationship with the truth revealed by these methodological evasions and the "less than totally convinced ..." dismissal of fact goes so far beyond post-modern irony that it deserves its own label. I suggest "post-real.""

Romer makes it very clear whose click he is goring. "Bob Lucas, Ed Prescott, and Tom Sargent led the development of post-real

He is critiquing the clique of Lucas, Prescott and Sargent. The intersections of that clique and the Krugman clique is the null set.

reason said...

are you aware of this post:

This is a new line of ridicule for the ridicule deserving set.

reason said...

I really admire your patience in dealing with anonymous comments (that may of course be students). But come on - if there is one thing that Krugman is not part of, it is a clique.

Anonymous said...

Romer's most devastating zinger was this:

"Math cannot establish the truth value of a fact. Never has. Never will."

That is something not only Sargent and Lucas would disagree with, but Krugman as well. Krugman has spent a career rubbishing historical and institutional approaches to understanding economic development.

Models have a role. So does mathematics. But there are cases where models are not appropriate. For Krugman arguments need to be centred around a model. Its all about the model. His career was built on putting accumulated and meticulously built up knowledge about trade and development into gadgets - and in the process threw the baby out with the bathwater. And what painful lessons we have had to relearn. He comes up with ludicrous arguments about implicit theorising. Romer is saying that there is a limit to the usefulness of models.

Models cannot serve as substitutes for the factual accounts of events - that is where analysis should be centred. Models may be helpful in certain cases, but they are not always helpful and not always necessary. Their place is in technical appendices. It is better not to have a mathematical model than to have one when it should not be applied and say well all models are false.

For sure Romer is specifically attacking DSGE and Sargent and Lucas and economics of the last 30 years. But he does not have kind things to say about the so-called 'Keynesian economics' before than other than that it was less opaque than what followed.

By the time you get to his conclusion it is clear that is critique is clearly even more profound than attacking RBC and DSGE - and was clearly intended to be so.

Robert said...

Huh ? Krugman is the leafer of a clique.I am A member of the Krugman clique. If you want to join, I can recommend you at our next clique meeting.

Robert said...

Romer uses formal models too. On models vs history,he and Krug are similar.

Also Lucas & Sargent agree in principle that math aline can't tell us anything about universe we happen to inhabit.

I think all 4 say same things at the level of abstraction you consider. But Lucas doesn't really believe them.

Blissex said...

«claim that the state of macro has been sad for longer than Krugman recognised back in August»

The main theoretical properties of a model in Economics probably are:

#1 Is it compatible with getting tenure?
#2 Does it help the authors or adopters get funding and consultancies from big sponsors?

I draw your attention in particular to #1, because a model that is not compatible with getting tenure has no theoretical value, because only those who get tenure (and at prestigious institutions) get much of a voice in the Economics profession.

But #2 is also quite important, because it matters a great deal to the success of a model that it helps the success of the authors and adopters after getting tenure, as in getting funding for endowed chairs or institutes or is popular with those who pay large private consulting/speaking fees.

As to points #1 and #2 the state of macro is not sad: by going with the "right" microfoundations and the "right" DSGE models a lot of Economists have got tenure and several have got large funding and consultancy fees from sponsors who like the "parables" and policy conclusions they portend.

MacroEconomics is what successful tenured macroEconomists do, people like Mankiw or Prescott or Hubbard; if it was something applicable to studying what actual economies do then the theorem of second best and the S-M-D theorem and the Cambridges controversy would have had an impact a long time ago. But studying what actual economies do has such little weight that both economic history and history of economic thought have been shrinking for a long time across many anglo-american universities.

Perhaps it would be better if there was by convention a split, an "inner" macroEconomics where macroEconomists could discuss among themselves, without the attention of the public, models that help studying actual economies regardless of whether they help with tenure or funding, as an intellectual exercise, and on "outer" macroEconomics portending the parables and policy conclusions that help get tenure and funding. But what would be the point of "inner" macroEconomics other than the preservation of some knowledge for its own sake?

Blissex said...

My argument above:

«a model that is not compatible with getting tenure has no theoretical value, because only those who get tenure (and at prestigious institutions) get much of a voice in the Economics profession. [ ... ] MacroEconomics is what successful tenured macroEconomists do, people like Mankiw or Prescott or Hubbard; if it was something applicable to studying what actual economies»

I have realized that a recent paper by N Kocherlakota makes a very similar argument with a different wording:
«Macroeconomists use a body of theory that imposes a number of a priori parametric restrictions on households and businesses. [ ... ] The mistake that the novice made is to think that the macroeconomist would rely on data alone to build up his/her theory or model. The expert knows how to build up theory from a priori restrictions that are accepted by a large number of scholars. (Indeed, in the academe, that’s exactly what it means to be an expert macroeconomist.) Those restrictions are what give the models their empirical content.»