Tyler Cowen discusses a large study of panel data on wage changes, which suggests nominal wages are less sticky than many have assumed.
There are some issues with the data. The U.S. sample is self-reported, and some of the wage changes do not reflect hourly rates. But let’s assume that the broad conclusions are accurate. Does this follow?
This paper does not show nominal wages to be fully flexible, nor does it show that observed nominal wage changes were “enough” to re-equilibrate labor markets. Still, this paper should serve as a useful corrective to excess reliance on the sticky nominal wage hypothesis. Nominal wage stickiness is a matter of degree and perhaps we need to turn the dial back a bit on this one.
I disagree with this claim, mostly because the stickiness that matters is not at the individual level, but at the aggregate level. Indeed it would be theoretically possible for individual wages to be highly flexible and aggregate wages to remain very sticky.
During a recession like 2008-09, NGDP growth fell 9% below trend, while nominal wage growth was little changed. W/NGDP soared. Using my musical chairs model, we’d expect about 9% fewer hours worked.
I can already anticipate your objection: ”Yes, but that doesn’t prove causation.” True. For instance, if the Fed had kept NGDP growing at 5%, instead of falling by 4%, perhaps nominal wages would have risen by 13%, instead of 4%. In that case the ratio of W/NGDP still would have soared 9% higher, unemployment would have skyrocketed, but there would have been no NGDP shock. So maybe sticky wages are not the problem.
Yes, that sort of counterfactual would disprove my musical chairs model. And I don’t doubt that something like that has occurred somewhere. Indeed something like that occurred in the U.S. during July through September 1933, due to FDR’s (NIRA) policy of raising nominal wages by 20%. But I don’t believe this sort of thing occurs very often, especially in a large diversified market economy such as the U.S. I’ll believe it when I see it. Until then I’ll continue to assume that had the Fed kept NGDP rising at 5%, nominal wage growth would not have soared to 13%, and unemployment would have risen much less sharply.
There are academic models that show even a small amount of nominal wage and price stickiness at the individual level, can lead to surprisingly large aggregate stickiness. And the study cited by Tyler shows unambiguously that nominal wage stickiness does occur at the individual level. The spike on the wage gain graph at “0% nominal wage increases” is the smoking gun. It’s true that that spike is modest in size, but it doesn’t have to be large, as most workers who got non-zero pay increases also tend to earn non-equilibrium wages.
Tyler also makes this claim:
Note also that this paper need not discriminate against neo-Keynesian and monetarist theories, though it will point our attention toward “zero marginal revenue product” versions of the argument, in which case the flexibility of nominal wages simply doesn’t help much.
I completely agree that the flexibility of individual nominal wages doesn’t help all that much, although a comparison of Hong Kong and Spain suggests that it does help some. Sticky wages are a background assumption, like gravity. When a bridge collapses, you don’t look for ways to make gravity less important in that locale (by shifting the Earth’s mass to other regions) you build a stronger bridge. Wage stickiness will always be with us, the solution is more stable NGDP growth.
If you want to show that wage stickiness is the not the main problem, then find those examples where workers are granted 13% pay raises when NGDP growth is running at 5%. They should be out there. After all, my critics claim that the huge spike in W/NGDP in 2009 was not caused by falling NGDP -- i.e. that W/NGDP would still have shot up if NGDP growth had been steady.
Until then, I’ll continue to believe that the musical chairs model featuring sticky aggregate wages (not necessarily sticky individual wages) is the best way to explain cycles in unemployment.