Advocates of a higher minimum wage rarely appeal to scholarly research to bolster their case. Instead, you’re more likely to hear vague, evasive slogans about “living wages,” greedy employers or the top 1 percent. The reason is simple: The vast majority of empirical, peer-reviewed studies of minimum wage hikes find negative employment effects, particularly among teens and low-skill adults.
When pressed for rigorous evidence that minimum wages help more than they harm, progressives often cite a study conducted in 1994 by economists David Card of the University of California and Alan Krueger of Princeton, who documented an apparent example of a minimum wage increase not only failing to provoke job losses but actually causing a small increase in employment in the fast-food industry.
Card and Krueger’s methodology was appealingly simple. In 1992, New Jersey raised its minimum wage by $0.80 an hour, while neighboring Pennsylvania did not. Card and Krueger capitalized on this “quasi-experiment,” an opportunity to compare employment levels in fast-food restaurants between neighboring counties in the two states before and after the increase took effect.
The results, which indicated the minimum wage hike in New Jersey had not reduced employment relative to Pennsylvania, shook the economic establishment and challenged the traditional view that a government-imposed wage floor discourages businesses from hiring low-skill workers.
For a quarter-century, supporters of higher minimum wages have clung to this finding, presenting it as conclusive evidence that employment concerns around the minimum wage are based on an imperfect theoretical model of the labor market that breaks down when confronted with real-world experience.
But the study’s flaws quickly became apparent. For one, Card and Krueger did not include information on the extent to which the minimum wage law was binding for the fast-food restaurants in their sample. If the employers surveyed in New Jersey were already paying wages above the minimum wage before the increase was implemented, then the lack of disemployment effects is not surprising.
The study also did not contain details about income, unemployment, labor force characteristics, and state regulatory and tax changes in the New Jersey and Pennsylvania counties being compared. Card and Krueger assumed that Pennsylvania would serve as an accurate benchmark against which to judge New Jersey’s response to the minimum wage hike, but existing disparities between the two states might have masked the true effect they sought to capture.
A third problem is that while Card and Krueger conducted their first surveys of restaurants a few weeks before the minimum wage increase took effect, the New Jersey law mandating the increase was enacted years earlier, giving fast-food chains ample time to prepare for the policy change.
Since employers typically don’t like to fire workers, it’s reasonable to assume they began to reduce employment through attrition in anticipation of the minimum wage hike, in order to avoid issuing pink slips the day the law took effect. Card and Krueger’s own data suggests this might have occurred, since New Jersey restaurants tended to have fewer employees than Pennsylvania restaurants shortly before the minimum wage increase was implemented.
In 1995, a pair of economists from the National Bureau for Economic Research (NBER) re-examined Card and Krueger’s results using data from actual payroll records from fast-food restaurants in New Jersey and Pennsylvania. The researchers pointed out that Card and Krueger’s data, collected from phone surveys of restaurant staff, showed significantly more variation before and after New Jersey’s minimum wage hike than the payroll data, raising “serious doubts about the quality” of Card and Krueger’s numbers.
For example, their survey did not precisely define the terms “full-time” and “part-time,” opening the door to inconsistencies in reporting. For example, suppose that a restaurant had 10 part-time employees before the minimum wage increased, and that each of those employees had their hours reduced by half in the ensuing months. Card and Krueger’s data would not have detected this adverse effect, since in both surveys the workers would have been categorized as part-time. The NBER analysis based on the more reliable payroll data found that New Jersey’s minimum wage increase led to a 4.6 percent decline in employment relative to Pennsylvania, in line with conventional estimates.
Even if we accept Card and Krueger’s results at face value, they still warrant several qualifications. First, Card and Krueger only looked at an eleven-month period, from a few weeks before the wage hike took effect to a few months after it was implemented. The short-term employment impact may have been insignificant, but that doesn’t give us any information — one way or the other — about how businesses respond in the medium to long-term.
More recent work on the long-term impact of higher minimum wages, including a groundbreaking 2013 paper by Jonathan Meer and Jeremy West of Texas A&M University, shows that they deter business formation and slow the rate of job growth.
Another caveat: In the study, New Jersey increased its minimum wage from $4.25 an hour to $5.05, a 19 percent hike. That relatively modest increase may indeed have had few employment effects. As the Card and Krueger study documents, fast-food restaurants raised menu prices and may have been able to pass the higher payroll costs to their customers.
The situation businesses face is much different when the minimum wage is increased by 60 percent, as Maine is doing, or by 107 percent, as is advocated by those who want a $15 federal minimum wage. The Card and Krueger study has little to say about the effects of such significant increases with short phase-in periods.
In 2015, Krueger himself warned that “a $15-an-hour national minimum wage would put us in uncharted waters, and risk undesirable and unintended consequences.” Indeed, three-quarters of economists oppose a $15 minimum wage, and 43 percent want to see the federal wage floor eliminated entirely.
Advocates of a higher minimum wage should look elsewhere to support their argument. But they won’t find much.