Two small points for the supply siders in the macro wars

I've been blogging less on the macro wars.  It is something I follow but really cannot comment on with much authority.  I am happy to see that monetary verses fiscal policy discussions have become somewhat more pointed and more clear (see here, here, here, here, here, here, for example).  This was something I had expected to see a lot more of a lot earlier.  But everything takes longer than you expect, no?

I do agree with the Delong-Krugman front that the evidence overwhelmingly supports Keynesian and New Keynesian views, even if these views can be harder to publish in the academic journals.  Indeed, I think on the evidence front, real business cycle models have looked bad for a very long time, well before the recent crisis and recession.  I recall reading Robert Lucas admitting as much, also long before the recent crisis. It certainly is sad that academic fashion constrains research effort on questions that really matter for policy.  That shouldn't happen.

Clearly, political and ideological leanings affect the intellectual debate, with Republican-leaning types mostly ascribing to the modern macro (aka "supply side") approach and Democrat-leaning types ascribing to the older Keynesian (aka "demand side") approach to explaining business cycles (there are notable exceptions, including Rogoff and Mankiw, who are Keynesian but decidedly conservative politically.)

I'm sympathetic to the demand-side view.  Given this, it seems important to bend over backwards to understand where the other side may be coming from.  So, what substance is there, if any, to the supply-side view?

I would award two key points to supply side crowd.

(1)  Sometimes supply-side factors do matter for business cycles. The best examples are the oil price shocks of the 70s and early 80s.  Those came from real supply shocks or reasonable fears of real supply shocks, which fundamentally reduced aggregate output and increased inflation.  It was a rare example of stagflation and counter-example to the Phillips-curve like behavior wherein unemployment and an inflation are inversely related.  The negative effects of oil price shocks on the macroeconomy were real and very much supply-side pheonomena.

(2) In Keynesian models, very little is said about the cause of business cycles.  It's a product of animal spirits, fear, excess demand for money or safe assets, etc.  Combine one or more of these with (very real and easy-to-explain) sticky prices and you have nasty demand-driven business cycles.  But how do we go about modeling the driver of the shock in the first place?  Keynesian models don't have a good answer to that question.  When you read vague notions of "confidence," this is what it's all about.  If everyone loses confidence, we have a downward shift in demand, followed by recession.  In a Keynesian world, fear itself is the seemingly the consummate cause of our malaise.  Blame Steven Colbert.

On the one hand, some might argue that a good answer to the question in 2 is very much beside the point.  We have high unemployment and near-deflationary price expectations, and the Keynesian model gives clear policy prescriptions, like fiscal stimulus, quantitative easing and inflation targeting. On the other hand, it seems hard to put animal spirits into a model, so the Keynesian story seems incomplete, and this is enough for the supply siders to completely ignore it. So, without a clear formal link to the drivers of confidence, it will be hard to communicate with true supply-side believers, who also happen to control the editorial boards of the macro journals.

What to do?

It seems to me there is plenty of room for middle ground here. It seems to me that a formal model of "confidence" should not be so difficult to construct. All we need is to have expectations about the future be critically sensitive to certain kinds of information.  Real productivity shocks, the traditional source of information in RBC models, clearly are not the kinds of information fundamental to big recessions.  The non-bailout of Lehman Brothers did not cause a "great forgetting" about how to produce things.  What it did was change the market's expectations about whether a lot of contracts would be fulfilled.  It changed beliefs about who really owned what, or who would really own what.  Institutionalists might say that property rights suddenly became more ambiguous.  Thus, the non-bailout of Lehman Brothers made the world suddenly appear a lot riskier.  Fear exploded, for not altogether irrational reasons.  The rest follows from sticky prices.

So, is it really so hard to see how to model the effect of an event, like the non-bailout of Lehman Brothers, could lead to such a change in expectations?  I think not.  If I blur my eyes just a little it seems to me such a model could be constructed from a game of incomplete information.  Individuals do not know the precise payoffs of the game they are playing and an innocuous, non-fundamental event rationally changes beliefs about which game (which payoffs) are really on the table.  We have this kind of machinery in economics already. If such a model is then combined with any reasonable model with nominal price rigidities (of which there are many), we will have New Keynesian model that reconciles (2).

Does such a model exist?  If it does, I'm not aware of it. But me being unaware of it wouldn't be so surprising either.


Popular posts from this blog

Nonlinear Temperature Effects Indicate Severe Damages to U.S. Crop Yields Under Climate Change

Commodity Prices and the Fed