Ryan Avent Is Sick of Arnold Kling: Invisible Wage-Push Inflation Vigilantes Watch

RA:

Inflation: Life on the Phillips curve: VIA Modeled Behavior, I see that Arnold Kling has written a post which reads:

Mainstream macro in the 1970s (which a lot of people seem to have gone back to) held that there was a NAIRU, meaning the non-accelerating inflation rate of unemployment. If unemployment was above that, inflation would fall. If it was below that, inflation would increase. So, policy should shoot for the NAIRU.These days, unemployment is 8.3 percent, and inflation is increasing. Just sayin’.

Just sayin'...what, exactly? Don't imply, man, argue! Follow the point through to its conclusion and see if it actually holds together!

Since Mr Kling didn't, I'll do it for him.

The NAIRU, as Mr Kling notes, is the non-accelerating inflation rate of unemployment…. [W]ere the government to try to raise employment above that level, fiscally or monetarily, inflation would accelerate…. Got that? Now, Mr Kling says that according to this theory a rate of unemployment below NAIRU will trigger an increase in inflation. He then observes that with 8.3% unemployment, inflation is increasing. And he deploys the just sayin' line to imply that the economy is therefore below NAIRU—that is, at structural full employment, suggesting that further demand stimulus is undesirable.

He is wrong on multiple levels.

First, it's very difficult to discern a steady increase, to say nothing of acceleration, in the inflation figures…. Inflation is positive and has risen on some measures. On others, however, it has declined. Perhaps most importantly, it is difficult to impossible to discern accelerating growth in wages, which is the most relevant price….

But let's take a step back…. Slow growth and high unemployment indicate a problem… it seems probable that the market-clearing real interest rate is below the actual interest rate…. Since the central bank's interest rate is close to zero, however, we have to assume that the market-clearing real interest rate is negative. To reduce the central bank's real policy rate further would require an increase in inflation. And from this we can generate a simple, testable prediction: if the central bank succeeds in raising inflation expectations, then it will move the policy rate closer to the market-clearing real interest rate, and unemployment should fall. 

How does the real world match up against this simple model? Well, in August, short-term inflation expectations (as measured by 2-year breakevens) dropped below 1%. The real interest rate, as computed by the Cleveland Fed, sat above -1%. And the unemployment rate was 9.1%. In August, the Fed introduced language suggesting that short-term rates would be at exceptionally low levels through 2013. Inflation expectations began rising and were comfortably above 1.5% by early 2012. The real interest rate dropped, falling to -1.75% by December of last year. And unemployment has since fallen by nearly a percentage point.

Based on this, it seems reasonable to conclude that the Fed raised expectations which therefore encouraged more hiring…. Mr Kling appears to have gotten both the diagnosis and the direction of causation wrong. And before we begin asking the central bank to engineer a rise in surplus labour, I'd hope to see substantially more proof that America is reaching the structural limits to increased employment.

One other point…. There is a very useful literature on persistent, large output gaps that describes the actual dynamics. Here, for instance, is an IMF paper….

Disinflation has tended to taper off at very low positive inflation rates, arguably reflecting downward nominal rigidities and well-anchored inflation expectations…. [T]he historical patterns suggest little upside inflation risk in advanced economies facing the prospect of persistent large output gaps.

The data match up very, very well with a story of a large and persistent American output gap. I don't really know why folks are so anxious to see something else in the numbers.