John Schmitt
The American Prospect, August 28, 2000

With unemployment at a 30-year low, opponents of current proposals to raise the minimum wage by a dollar to $6.15 an hour will be hard-pressed to argue such a move will cost low-wage workers their jobs. But what about that other stock argument that a higher minimum will reduce training for low-wage workers?

New research by MIT economists Daron Acemoglu and Jörn-Steffen Pischke casts doubt on that claim. The researchers noticed that training for low-wage workers fell substantially during the 1980s--a period when the inflation-adjusted value of the minimum wage plummeted. Why didn't the wage decline allow employers to invest more in training?

Since other factors could have been at play, Acemoglu and Pischke designed a test. They followed a sample of low-wage workers over the period 1987-1992, a time when the federal government and some states increased minimum wages. The economists then compared training outcomes across states with different minimum wage levels.

They found "almost no evidence of reduction in training in response to minimum wages." In fact, when Acemoglu and Pischke focused on the lowest-wage workers, the data indicated that, if anything, training was slightly higher where the minimum wage was also higher (though the results were not statistically significant).

Acemoglu and Pischke offer a reasonable explanation. Though economics textbooks may speak of labor markets as "perfectly competitive," in the actual world markets are far from perfect. Employers, for example, often face trouble filling vacancies, motivating their workers, and retaining their best workers even when they pay higher wages. So raising the minimum wage can lead employers to invest more in the skills of their employees. It sometimes can be more profitable for employers to train workers up to their new wage level than to look for new workers.