In his recent State of the Union address, President Barack Obama advocated for a raise in the minimum wage, from $7.25 an hour to $9.00. The president’s call for federally mandated higher wages was met with a flurry of response from all sides of the political sphere, and debate has raged in its aftermath.
Conventional economic wisdom on the minimum wage, for its part, is unequivocal. Despite the presence of a small dissenting minority, the vast majority of the literature on the subject—to the tune of over 85 percent of peer-reviewed studies—is in agreement that the minimum wage increases unemployment. Those who disagree with this conclusion, however, generally refer to a famous 1993 study by Princeton economists David Card and Alan B. Krueger in order to provide support for increases in the minimum wage.
The Card and Krueger study has received widespread attention, and has been cited in support of minimum wage hikes by President Clinton’s Secretary of Labor Robert Reich, as well as former Massachusetts Sens. John Kerry and Ted Kennedy. The appeal of the study to supporters of the minimum wage lies in its finding that, in a one-year survey of fast food restaurants in New Jersey and Pennsylvania, employment at surveyed chains actually increased in New Jersey following an increase in the minimum wage. A host of subsequent studies have also failed to demonstrate an inverse relationship between the minimum wage and employment, which proponents of the minimum wage sometimes cite as further evidence in its favor.
Upon closer inspection, however, those who build their support for the minimum wage on the findings of Card and Krueger might discover that the foundation of their argument is less than sturdy. The problems with the Card and Krueger study are, unfortunately, too numerous to list here—and include everything from highly ambiguous survey questions to a total failure to adjust data for seasonal variations in fast food industry hiring practices—but a brief survey of the most substantial issues is in order.
For one thing, subsequent independent analyses have found it difficult to reproduce the results of the Card and Krueger study—a death knell in the world of scientific publication—even when they work from lists of restaurants that overlap by at least 25 percent with the list used in the original Card and Krueger study. In 1995, for instance, the conservative Employment Policies Institute compared Card and Krueger’s survey data with actual payroll data for fast food restaurants in the same location and over the same time period. The study found that, far from any increase in employment, the far more reliable payroll data indicated that the minimum wage hike had actually resulted in an estimated 5 percent decrease in employment.
But the EPI report doesn’t end there. The numbers used in the Card and Krueger study, it argues, rarely reflect the payroll data. In fact, in a third of cases, the Card and Krueger study mistakenly lists job losses as job gains and vice versa. A subsequent 1996 report from the same institution thus complained of the Card and Krueger study: “The depth of the flaws is shocking.” “Not only are the Card-Krueger numbers wrong,” the report went on, “they are often catastrophically wrong.” Independent economists David Neumark and William Wascher confirmed the findings of the EPI, concluding, “The payroll data from the New Jersey-Pennsylvania minimum wage experiment are consistent with the prediction of the standard competitive model that minimum wage increases reduce employment of low-wage workers.”
There’s more. The variability of the employment change data in the original Card and Krueger study is, upon further inspection, almost nonsensical. Neumark and Wascher argue that such a degree of variability—which somehow shows, for starters, one fast food restaurant moving from six full-time workers to 29 full-time workers and another moving from 50 full-time workers and 35 part-time workers to just 15 full-time workers and 18 part-time workers, all in the span of just eight short months—“raises serious doubts about the quality of their data.” The EPI concluded that such statistics, along with equally absurd data on product price fluctuation, “defy reasonable explanation.”
Now, all of that being said, one methodologically flawed study does not invalidate the case for an increase in the minimum wage. (It does suggest, however, that those who wish to increase the minimum wage abandon the Card and Krueger study as their favorite piece of evidence.) In fact, the shortcomings of the Card and Krueger study, when viewed in combination with the failure of opponents of the minimum wage to empirically demonstrate the orgy of unemployment that they consistently predict would follow proposals such as the president’s, serve best as evidence of the limitations of scientism and empiricism in economic thought.
When political and economic debates devolve into empirical battles like the minimum wage war, we run the risk of reducing our analyses to what economists Russ Roberts and Don Boudreaux call “push button economics,” or, according to Boudreaux, the view that we can “press this economic variable up … and that economic variable down … and thereby improve society.” We forget that such policies have many hidden effects, many of which even vary between regions and industries. We may be so busy haggling over dubious statistical data that we ignore effects on price fluctuations, business practices, employee benefits, workplace amenities, job exportation and a whole host of other complications brought about by our minimum wage policies. Even these effects represent only our “known unknowns,” or the things that we can plausibly say occur but have a difficult time quantifying. That is to say nothing of the “unknown unknowns,” or the things we haven’t even thought to look for.
Despite the resolute certainty of the minimum wage empiricists, however, these things strike economists like Roberts as severely problematic for those who promote tampering with the economy. “The people who are pushing for a large increase in the minimum wage are playing with people’s lives,” he worries. “I don’t understand their certainty.”
Chris Bassil, Trinity ’12, is currently working in Boston, Mass. His column runs every Wednesday. You can follow Chris on Twitter @HamsterdamEcon.