The minimum wage is roaring back.
President Joe Biden may have failed to drive through Congress an increase in the federal wage floor, which remains stuck at $7.25. But last month, Connecticut, Nevada and Oregon raised their minimums alongside Washington D.C. and a handful of other cities. Twenty-one states raised their minimum wages at the turn of the year. In San Francisco, it is $16.99 an hour. In Emeryville, California, it hit $17.68.
These increases may be warranted in a moment when high inflation is eroding the paycheck of American workers. Still, the newfound enthusiasm for the minimum wage risks doing a lot of harm to many of the workers it is intended to help.
Proponents “haven’t measured the long term,” Erik Hurst of the University of Chicago told me. “So they think they have a costless solution.”
Recent research by Hurst and three colleagues concludes that President Biden’s proposal to raise the federal minimum to $15 would end up damaging the livelihood of about 15% of the workers earning less than that — mainly those at the very bottom of the pay scale. Even if they benefited from the pay boost in the short term, many would lose their jobs in the end.
This is bad news for advocates of higher minimum wages, who have come to believe that the argument over the pros and cons of the policy — whether it destroys jobs or not — has been decisively settled in their favor.
Three decades ago, economists David Card from the University of California, Berkeley, and Alan Krueger from Princeton University surveyed fast-food restaurants on both sides of the border between New Jersey, which raised its minimum wage, and Pennsylvania, which didn’t. To the surprise of many, they found no evidence that raising the wage floor cost jobs.
The confounding finding seemed in conflict with a basic tenet of economics — that in a competitive market, raising the price of a good, a service or an input in the production process will reduce demand for the thing. But economists noted that it makes a lot of sense if the labor market is not competitive.
Like a monopolist who has the power to raise prices without losing market share to cheaper rivals, employers facing little competition for workers can pay them less than their contribution to the bottom line without fear that a rival will swoop in to hire them away.
Economists have found evidence of so-called monopsonistic behavior among employers in some industries and markets. This has bolstered the proposition that forcing employers to raise wages will not automatically lead them to drop workers altogether, because they are paying them less than what they are worth to the firm. They can pay them more and still make a profit.Lifting the minimum wage in this kind of market might even create new jobs. More people would be drawn to work by the higher pay. As long as the new wage didn’t rise above the value of their contribution to the firm, the employer would still turn a profit on each additional worker.
There has been sniping around the limits of the proposition that the minimum wage can do no harm. Subsequent research in the style of Card and Krueger has sometimes found increases in the minimum wage reducing employment. Others have found no effect. The magnitude of the impact has usually been small.
Some economists argue against the new theoretical framework. “I think the monopsony train is going out of the station way too fast,” David Neumark, an economist at the University of California, Irvine, who is skeptical of the value of the minimum wage, told me. He notes that companies in big markets where most people live and most workers work are not likely to wield much monopsony power. There are too many competitors also hiring.
But perhaps more importantly, the groundswell of support for the minimum wage as the tool of choice to improve the lot of the working poor fails to account for time: all the empirical studies observe changes in employment over a few years at most. Firms don’t usually overhaul their workforce or retool their production lines that fast.
Whether they have monopsony power over their workers or not, businesses will do their best to keep their costs down, substituting high-cost inputs with lower-cost ones.
Workers can be replaced with robots or with other workers who, either because they have more education or more experience, are more profitable for the firm. Just give them time.
This proposition is grounded in empirical research that is just as strong as that by Card and Krueger. (Indeed, Card is co-author of some of it.) The low-wage workers who are suddenly made expensive by a rising minimum wage will largely be replaced in the end.
Hurst, Patrick Kehoe and Elena Pastorino from Stanford University, and Thomas Winberry from the University of Pennsylvania have put together a model of the labor market that fits both the evidence that raising the minimum wage has little or no impact on jobs in the short term, as well as findings regarding worker substitution over the long run.
They conclude that raising the minimum wage in real terms to $15 an hour — compared with the average wage distribution of 2017 to 2019 — would affect a big chunk of the workforce: 40% of workers without any college education and 10% of workers with college made less than that. The group includes workers earning $14.50 and workers making half that.
The main finding by Hurst and his co-authors is that all these workers would benefit over the first few years, as their paychecks shot up. But over the long term — which in some cases could mean as little as four years — the less-productive workers on the bottom end would be fired, to be replaced by others of higher productivity.
Whether raising the minimum wage is ultimately good or bad for workers, then, needs to be calculated over their entire time in the workforce. This will add up to positive gains for many, but it would ultimately damage those at the bottom — including about everybody earning less than $9.
What’s more, the higher minimum would damage the prospects of future low-wage workers who are not yet in the labor market. They wouldn’t benefit from any wage bump in the short term. But they would likely be ignored by employers looking for workers with a latent value to the firm in the $15 range.
The precise effects are hard to estimate. They rely on estimates from empirical-research papers of the elasticity of substitution between different types of workers, as well as between workers and machines, and these things are difficult to measure. The magnitude of firms’ monopsony power is another parameter that would also affect the calibration.
Yet the broad thrust of the finding is solid: The minimum wage is not a cost-free tool to fix the low-wage economy. This might not be obvious at first, but just wait.
