What the wage data is telling us
By Scott Sumner
During a normal demand-side recession, wage growth tends to fall sharply, with a modest lag due to nominal wage stickiness. In this recession, however, it’s been hard to interpret the data due to “composition effects”. Put simply, lower paid workers were more likely to get laid off, so average hourly wages actually rose faster than normal last year.
Fortunately, the government has another wage series that compares wage growth within a given job category. This shows average wage growth to have been relatively steady over the past year, which is still a bit surprising. Wage growth normally slows sharply during severe recessions:
Wage growth still might slow sharply, after all, wages respond with a lag. But given the severity of the slump last March and April, I’d expect to see some evidence by now. In the case of 2008, the severe slump didn’t begin until the second half of the year, and then wage growth fell a few months later.
One possibility is that this time the recession reflects a decline in both aggregate demand and aggregate supply, so the overall effect on the price of labor is relatively small, even as quantity falls sharply. (Never reason from a quantity change.)
This fits in with all sorts of other evidence that this recession is unusual. For instance, output and employment bounced back after the April 2020 slump much faster than after previous recessions. Indeed the actual downturn lasted only two months. Here’s the Wall Street Journal, discussing reports that companies are having trouble finding workers:
One shouldn’t put too much weight on anecdotes, but these are corroborated by data. Some 7.4 million jobs were open in February, above the pre-pandemic level. By contrast, job vacancies plummeted by half in 2007-09.
Mainstream economists tend to assume that causation goes from high unemployment to falling wages. In my view, the root cause of most recessions is tight money, which reduces the equilibrium average wage rate faster than the decline in actual wages. Because actual wages fall more slowly than equilibrium wages, the price of labor is too high. This causes mass unemployment. This recession was not caused by tight money, however, and hence there was no necessary reason for high unemployment to be correlated with falling wage growth.
To summarize, when we finally get a recession caused by real factors, not a demand shortfall, we can see how different it is from normal recessions. That should make us even more confident that most normal recessions are basically monetary problems. (During my life, the 1974 recession was the one with the largest “real” component—prior to Covid-19.)
One other point. Although I often advocate NGDP targeting, in some models nominal wage targeting does even better. If money were too tight then wage growth would be falling, and vice versa. In my view, the stability of nominal wage growth suggests that monetary policy has been roughly on target during this period.