Economists Should Not Forget Supply and Demand
By Pierre Lemieux
A Jule 23 Economist article about how “Generous Unemployment Benefits Are Not Keeping Americans from Work” is not easy to understand. “[T]here is little evidence,” it claims, “to suggest that the extra $600 a week [from the federal government] is slowing down the labour-market recovery.” One would think that paying more in benefits to non-working workers would, ceteris paribus, decrease their number. Consider the last two paragraphs of the article:
Another signal that employers were struggling to fill positions would be soaring wages. Workers might hold bosses hostage with the threat of settling for benefits instead. Upon first inspection, this seems to be true. Average hourly earnings in the second quarter of 2020 increased by about 7% from a year ago, according to Goldman Sachs. However this is largely because low-paid workers have lost jobs in disproportionate numbers, dragging average wages upwards.
All of this implies that the main factor behind the high unemployment rate is lack of jobs, not an unwillingness to work. Economists normally fight like cats in a bag. But an astonishing 0% of those surveyed by researchers at the University of Chicago disagreed with the idea that “employment growth is currently constrained more by firms’ lack of interest in hiring than people’s willingness to work at prevailing wages.”
Excluding the misleading last sentence, the first paragraph tells us that, in the second quarter, employment decreased and wages increased, which is confirmed by the chart below (shaded area marks the current recession).
Consider Q2. How could employment decrease while wages increased (all compared to the same quarter one year ago)? Assuming a negatively sloped demand for labor and a positively sloped supply of labor, the only way this combination can obtain is if supply has decreased, that is, the supply curve has shifted to the left. The demand curve may also have shifted, up or down (most likely down in the current outlook), but we know that the dominant factor must have been a decrease in supply.
This is easy to see with a standard supply and demand analysis. The chart below pictures the labor market with demand curve D and supply curve S. Employment and the wage rate were respectively L1 and W1 before Q2. For L1 to decrease to L2 and W1 to increase to W2, you need to have a decreased supply, at S’, that more than compensates for any decrease (or increase) in the demand for labor. Playing with S and D should persuade you of that. (You may want to refer to my post “A Frequent Confusion and the Yo-Yo Model” if you need a more detailed explanation of supply-demand analysis—albeit in a product market instead of a labor market.)
What seems to have happened is the opposite of what The Economist tells us. The dominant factor behind the lower employment is that workers reduced their supply of labor, presumably because, with the $600 federal unemployment supplement, many were paid more non-working than working—and no doubt also because many were too scared of the pandemic to go back to work. The higher unemployment rate (not shown on the chart and again compared to Q2-2019) probably results from the workers who, in the states where it does not compromise their unemployment benefits, chose not to go back to work and not to look intensively for a new job. There is certainly a “firms’ lack of interest in hiring” (compared to what it was before) but the dominant factor is “unwillingness to work.” If that were not the case, wages would not have increased.
One way to question this conclusion, besides making strange assumptions about the supply or demand curves, would be to suggest that percent-change calculations with respect to the preceding quarter (as opposed to the same quarter a year before, as was done in the first chart) would be more representative of short-run changes in supply and demand. The following chart shows that doing this does not change the general picture and thus our conclusion for Q2. (The picture for Q1, with a very tiny increase in employment of 0.09%, suggests that the overall demand for labor rose enough to compensate the unemployment effect of the reduction in labor supply. This is likely an artifact of Q1 being very economically heterogeneous, from a slow start in January and February to a catastrophic drop in March.)
This little analysis of the respective effects of labor supply and labor demand is also confirmed by the monthly evolution of employment and wages (see chart below). In the second quarter and the end of the first, employment and wages moved inversely, implying that the supply of labor played the main role, not the demand for labor.
If all that is correct, the burning question is why none of the 43 economists polled by the Chicago Booth School of Business disagreed with the statement that that “employment growth is currently constrained more by firms’ lack of interest in hiring than people’s willingness to work at prevailing wages.” Perhaps these economists were secretly playing with supply and demand curves in their heads, as the comments of Caroline Hoxby and Darrel Duffie suggest. But they should all come out of the closet.