The Minimum We Can DoBy ARINDRAJIT DUBE
These direct interventions, however, were temporary, and unions have become rare in most parts of the United States — virtually disappearing from the private sector. This leaves minimum wage policies as one of the few institutional levers for setting a wage standard. But while we can set a wage floor using policy, should we? Or should we leave it to the market and deal with any adverse consequences, like poverty and inequality, using other policies, like tax credits and transfers? These longstanding questions take on a particular urgency as wage inequality continues to grow, and as we consider specific proposals to raise the federal minimum wage — currently near a record low — and to index future increases to the cost of living.
The idea of fairness has been at the heart of wage standards since their inception. This is evident in the very name of the legislation that established the minimum wage in 1938, the Fair Labor Standards Act. When Roosevelt sent the bill to Congress, he sent along a message declaring that America should be able to provide its working men and women “a fair day’s pay for a fair day’s work.” And he tapped into a popular sentiment years earlier when he declared, “No business which depends for existence on paying less than living wages to its workers has any right to continue in this country.”
This type of concern for fairness actually runs deep in the human psyche. There is a widespread sense that it is unfair of employers to take advantage of workers who may have little recourse but to work at very low wages. For example, the economists Colin F. Camerer and Ernst Fehr have documented in numerous experimental studies that the preference for fairness in transactions is strong: individuals are often willing to sacrifice their own payoffs to punish those who are seen as acting unfairly, and such punishments activate reward-related neural circuits. People also strongly support banning transactions they see as exploitative of others — even if they think such a ban would entail some economic costs.
Of course, if most minimum wage workers were middle-class teenagers, many of us might shrug off concerns about their wages, since they are taken care of in other ways. But in reality, the low-wage work force has become older and more educated over time. In 1979, among low-wage workers earning no more than $10 an hour (adjusted for inflation), 26 percent were teenagers between 16 and 19, and 25 percent had at least some college experience. By 2011, the teenage composition had fallen to 12 percent, while over 43 percent of low-wage workers had spent at least some time in college. Even among those earning no more than the federal minimum wage of $7.25 in 2011, less than a quarter were teenagers.
Support for increasing the minimum wage stretches across the political spectrum. As Larry M. Bartels, a political scientist at Vanderbilt, shows in his book “Unequal Democracy,” support in surveys for increasing the minimum wage averaged between 60 and 70 percent between 1965 and 1975. As the minimum wage eroded relative to other wages and the cost of living, and inequality soared, Mr. Bartels found that the level of support rose to about 80 percent. He also demonstrates that reminding the respondents about possible negative consequences like job losses or price increases does not substantially diminish their support.
These patterns show up in recent survey data as well, as over three-quarters of Americans, including a solid majority of Republicans, say they support raising the minimum wage to either $9 or $10.10 an hour. It is therefore not a surprise that when they have been given a choice, voters in red and blue states alike have consistently supported, by wide margins, initiatives to raise the minimum wage. In 2004, 71 percent of Florida voters opted to raise and inflation-index the minimum wage, which today stands at $7.79 per hour. That same year, 68 percent of Nevadans voted to raise and index their minimum wage, which is now $8.25 for employees without health benefits. Since 1998, 10 states have put minimum wage increases on the ballot; voters have approved them every time.
But the popularity of minimum wages has not translated into legislative success on the federal level. Interest group pressure — especially from the restaurant lobby — has been one factor. Ironically, the very popularity of minimum wages may also have contributed to the failure to automatically index the minimum wage to inflation: Democratic legislators often prefer to increase the wage themselves since it allows them to win more political points. While 11 states currently index the minimum wage, only one, Vermont, did so legislatively; the rest were through ballot measures.
As a result of legislative inaction, inflation-adjusted minimum wages in the United States have declined in both absolute and relative terms for most of the past four decades. The high-water mark for the minimum wage was 1968, when it stood at $10.60 an hour in today’s dollars, or 55 percent of the median full-time wage. In contrast, the current federal minimum wage is $7.25 an hour, constituting 37 percent of the median full-time wage. In other words, if we want to get the minimum wage back to 55 percent of the median full-time wage, we would need to raise it to $10.78 an hour.
International comparisons also show how out of line our current policy is: the United States has the third lowest minimum wage relative to the median of all Organization for Economic Cooperation and Development countries. This erosion of the minimum wage has been an important contributor to wage inequality, especially for women. While there is some disagreement about exact magnitudes, the evidence suggests that around half of the increase in inequality in the bottom half of the wage distribution since 1979 was a result of falling real minimum wages. And unlike inequality that stems from factors like technological change, this growth in inequality was clearly avoidable. All we had to do to prevent it was index the minimum wage to the cost of living.
The social benefits of minimum wages from reduced inequality have to be weighed against possible costs. When it comes to minimum wages, the primary concern is about jobs. The worry comes from basic supply and demand: When labor is made more costly, employers will hire less of it. It’s a valid concern, but what does the evidence show?
For the type of minimum wage increases we have implemented in the United States, the best evidence shows that the impact on jobs is small, although there is still a debate in the literature. There are estimates that do suggest job losses — most prominently associated with work by the economists David Neumark and William Wascher. Since the early 1990s, they have consistently argued that minimum wage increases lead to substantial job losses for low-wage workers: a 10 percent increase in the minimum wage can be expected to reduce jobs among a group like teenagers by between 1 and 3 percent. The methodology pioneered by Mr. Neumark and Mr. Wascher has a critical problem, however: it does not properly account for differences between high- and low-minimum-wage states. Essentially, they make the unrealistic assumption that low-wage employment trajectories are similar in states as diverse as Texas and Massachusetts.
As my colleagues and I show in our research, the states raising minimum wages have had very different trajectories when it comes to trends in demand conditions and business cycle variability. In fact, low-wage employment was often already falling (or growing more slowly) in the states raising the minimum wage — sometimes years before the actual wage increase. Such divergence in trends between the “treatment” and “control” groups is a telltale sign that the control group is being constructed improperly — a major issue for evaluating policies using nonexperimental evidence, otherwise known as real life.
The good news is that today we have much better tools in our toolbox. A particularly reliable methodology compares adjacent counties that are right across the state border but that experience different minimum wage shocks. Originally performed for a single case study of Pennsylvania and New Jersey by the economists David Card and Alan B. Krueger in 1994 and then again in 2000, this methodology has been substantially refined and expanded.
In my work with T. William Lester and Michael Reich, we use nearly two decades’ worth of data and compare all bordering areas in the United States to show that while higher minimum wages raise earnings of low-wage workers, they do not have a detectable impact on employment. Our estimates — published in 2010 in the Review of Economics and Statistics — suggest that a hypothetical 10 percent increase in the minimum wage affects employment in the restaurant or retail industries, by much less than 1 percent; the change is in fact statistically indistinguishable from zero.
In my most recent work with Sylvia Allegretto, Ben Zipperer and Michael Reich, we confirm these results using four data sets covering over two decades, other low-wage groups like teenagers, and five different statistical techniques, including an increasingly popular method that uses past economic trends to construct a “synthetic” control group. And other researchers have independently reached the same conclusion: minimum wage effects on employment are small.
While the evidence may not convince the most strident of critics, it has shifted views among economists. A panel of 41 leading economists was asked recently by the University of Chicago’s Booth School of Business to weigh in on President Obama’s proposal to increase the minimum wage and automatically index it to inflation. A plurality, 47 percent, supported the policy, and only 11 percent opposed it, while the rest were uncertain or had no opinion. Only a third thought that the raise “would make it noticeably harder for low-skilled workers to find employment.”
But how can minimum wages rise without causing job losses? For starters, if the demand for burgers is not price sensitive, some of the cost increase can be passed on to customers without substantially reducing demand or jobs. Existing research suggests that if you raise the minimum wage by 10 percent, you can expect the price of a $3 burger to rise by a few cents, which is enough to absorb a sizable part of the wage increase.
Going beyond simple supply and demand, economic models are getting better at incorporating frictions caused by the costs of finding jobs and filling vacancies, which turn out to be quite important when analyzing labor markets. There are good jobs and bad jobs at the low end of the labor market, and movements between these lead to vacancies and turnover. If McDonald’s is required to pay a higher wage, fewer of its workers will leave to take other jobs. This means fewer vacancies at McDonald’s, and it means other employers are more likely to fill their job openings from the ranks of the unemployed — both of which can help keep unemployment down. So while higher costs may dissuade some employers from creating new positions, it also helps other employers recruit and retain workers. Moderate increases in the minimum wage, in other words, can reduce vacancies and turnover instead of killing jobs. In a follow-up study using our bordering areas methodology, we provide empirical evidence for this argument: while overall employment in low-wage sectors does not change much following a minimum-wage increase, worker turnover falls sharply as workers stay with their jobs longer.
But even if minimum wage policies reduce inequality and improve the functioning of low-wage labor markets, are there better alternatives when it comes to helping low-income families?
In a forthcoming study commissioned by the Department of Labor, I review the evidence using data from the past two decades and find clear evidence that minimum wage raises have helped lift family incomes at the bottom: a 10 percent increase in the minimum wage reduces poverty by around 2 percent.
The minimum wage can also increase the efficacy of a policy that is sometimes pushed as a substitute: the earned-income tax credit. This encourages more people to seek work, but can push wages down; a minimum wage ameliorates this. Of course, many families under the poverty line simply have no workers, making any work-based policy of limited help. This is why raising and indexing the minimum wage is just a part of the portfolio of policies we need to enact to ensure a decent living standard.
What are actual policy options when it comes to raising the minimum wage? At the federal level, the legislation proposed by Senator Tom Harkin, Democrat of Iowa, and Representative George Miller, Democrat of California, would raise the minimum wage to $10.10 an hour, and index it to future cost of living increases. This is a sensible target that would be likely to put the minimum wage right around 50 percent of the median wage for full-time workers — close to the international standard and our own norm during the 1960s and ’70s. Indexation is critical — it replaces politics with economics as the adjustment mechanism and makes changes predictable. This is why even economists opposed to higher minimum wages support indexation.
Other policies can complement the federal minimum wage in building higher wage standards. City and state minimum wages play an important role in ensuring that places with higher costs of living have similarly higher wage standards. A number of cities have instituted “living wage” ordinances covering public sector workers and private city contractors. The most expansive of these ordinances cover major airports, like in the metropolitan areas of San Francisco, Los Angeles and most recently Seattle. Fast food workers in urban centers are beginning to organize and push for substantially higher voluntary wage standards at major chains. Together with a sensible federal minimum wage, these local initiatives can help rebuild wage standards and reduce inequality in a way that reflects our internal sense of fairness.
Arindrajit Dube is an associate professor of economics at the University of Massachusetts, Amherst, and a research fellow at IZA.