Opinion pieces and speeches by EPI staff and associates.
TESTIMONY BEFORE THE HOUSE EDUCATION AND THE WORKFORCE COMMITTTEE ON APRIL 27, 1999.
Minimum Wages and Poverty
by Jared Bernstein
Today’s hearing questions whether increases in the minimum wage are a useful anti-poverty tool. In particular, the question is whether raising the minimum wage to $6.15 by September of 2000 will help those workers whose family incomes place them at or near the poverty line. As I will argue, the evidence unequivocally supports the view that increases in the minimum wage, by increasing the earnings of low-income workers without diminishing their employment opportunities, have historically helped to lower poverty rates. What’s more, this evidence is stronger in the 1990s, with large numbers of poor persons moving into the labor market. The facts clearly support the contention that if Congress wants to “make work pay,” the current proposal to increase the minimum wage makes sense.
The following testimony elaborates these four points:
- The historical record is quite clear on this point: increases in the minimum wage have consistently been shown to have their intended effect: they raise the wages of the lowest wage workers, many of whom reside in low-income families, and they do so without lowering their employment opportunities. In this regard, the policy solidly reinforces the “make work pay” platform that Congress and the administration have recently emphasized;
- Economists have evaluated the impact of minimum wage increases practically since the inception of the wage floor in the 1930′s. At this point, I think it is fair to say that the debate over the purported job-loss effect is a debate over whether this effect is slightly below zero, or at zero. While this debate may be an important one among econometricians, from the perspective of policy makers who are looking for ways to help the working poor, it is a distinction without a difference. Even if you were to accept the most negative findings from this research, you simply could not avoid the conclusion that the benefits of the increase far outweigh the costs.
- There is no better example of these points than the most recent increase, which took place in 1996-97. Despite the same dire predictions we hear today, we have solid evidence that the last minimum wage increase lifted the earnings of low-wage workers without diminishing their employment prospects. This last point is key: the employment rates of low-wage workers have rarely been higher than they are today and their unemployment rates have rarely been lower.
- Finally, from the perspective of the working poor, the minimum wage is a useful anti-poverty tool. It cannot and should not, however, be viewed as a sole solution against poverty. This is primarily due to the fact that many employable poor persons have only marginal attachments to the labor market. As the literature evaluating minimum wages and poverty reveals (as well as some new findings I will present regarding the last increase), raising the minimum wage is associated with small decreases in the poverty rate, but the poor need other income supports, such as the EITC and food stamps.
Figure 1 shows the inflation-adjusted minimum wage, from 1955-2000 (projected). The figure shows the sharp decline in the minimum wage over the 1980s, when Congress failed to adjust the wage floor for nine years. Even with the recent increases in the 1990s, the inflation-adjusted minimum is 19% lower today than in 1979.
The dotted line at the end of the figure shows the impact of the new proposal. Using inflation projections by the Congressional Budget Office, in the absence of the proposed increase, the minimum wage would fall to $4.90 (1998 dollars) by the year 2000. The proposed increase would restore the wage floor to slightly above its 1983 level; it would remain 13% below its 1979 peak.
Who Benefits When the Minimum Wage is Raised?
Since increases in the minimum wage have been shown to generate wage gains, an obvious question is: for whom? The data in this section reveal that the current proposal will primarily benefit low-wage workers who are disproportionately adult, female, and minority. Most of the workers reside in low-income families.
Table 1, panel 1, shows that 11.8 million workers (10.1% of the workforce) would receive an increase in their hourly wage rate if the minimum wage were raised to $6.15. Since these workers earn, on average, $5.65 per hour, the average hourly increase would be $0.50.
The rest of the table examines the demographic characteristics of workers by wage levels. The first column looks at those directly affected by the increase in the minimum wage, i.e., workers whose earnings are between their state’s current minimum and $6.15. Most of these workers, 59.2%, are female, and about 72% are adults (age 20 or older). Close to half (48.2%) work full-time, and another third work between 20 and 34 hours.
Comparing the first column to the last (All Workers) shows the extent to which different types of workers are over- or under-represented in the affected range. Note, for example, that the minimum-wage workforce is disproportionately minority. While 11.6% of the overall workforce is African-American, 15.1% in the affected range are in this minority; similarly, 10.6% of the total workforce is Hispanic, compared to 17.4% of the minimum wage workforce. Minimum wage workers are concentrated in the retail trade industry and under-represented in the manufacturing sector. They are also the least likely group of workers to be represented by unions.
Various studies have found that, due to so-called “spillover effects,” the group earning above the minimum (perhaps as much a dollar above) also benefits from the increase. Workers in this group, shown in the second column, are more likely to be older (87% are adults) and to work more hours (69% full-time) than those in the directly affected range.
The benefits of the increase flow mostly to workers with family incomes well below the average. Data from the first stage of the last increase reveal that 35% of the increase in wages went to working families in the bottom 20% of the income distribution (average income: $15,728). Note that the share of national income held by these families in 1997 was only 5%. Combining the first two income fifths (average income, $24,137), we find that 58% of the benefits went to these low-income working families, as compared to 16.4% of national income. Of course, these data imply that some of the benefits from the increase flow to families with above-average incomes (26.7% of the gains go to families in the top 40%). Thus, the policy is well, but not perfectly, targeted.
What About Job Losses?
Despite the compelling evidence of the benefits to poor and near-poor workers, opponents of increases in the minimum continue to raise the same objection: the increase will lead to job loss. This claim is based on the simple textbook notion that is the price of a good is competitively set by the free market, any diversion from that price will lead to lead to an inefficient outcome. In this case, the prediction from the simple model is that the increase will price low-wage workers out of a job.
There are a priori reasons to be skeptical of this simplistic view. Low-wage workers are no
t “goods,” and the low-wage labor market is far from the competitive laboratory that exists in Econ 101 textbooks. But even if this were not the case, when we construct economic policy, we need to empirically test economic theories.
The theory that “small” increases in the minimum wage lead to job losses has been repeatedly tested and repeatedly found lacking (“small” in this case refers to increases which simply serve to replace some portion of the real value of the minimum wage which was eroded by inflation; see Figure 1). The estimates of the so-called “negative employment elasticity” have essentially hovered about zero, leading objective observers, including many economists, to be skeptical of the simple model’s predictions. The state of economists’ understanding of the issue was recently summarized by Nobel laureate Robert Solow, who noted that “the main thing about this research is that the evidence of job loss is weak. And the fact that the evidence is weak suggests that the impact on jobs is small.”
Further proof of this contention comes from a recent survey of labor economists from leading research institutions in the U.S. The median estimate of the disemployment effect among these economists was -0.1%, meaning a 10% increase in the minimum wage would be associated with a 1% decline in the employment rates of teenagers, a group disproportionately employed at or near the minimum. Research that has looked for disemployment effects among older workers has generally come up empty handed. If we were to apply this parameter to the current proposal, the conclusion would be that over 99% of the affected workforce would benefit from a wage increase.
But this practice, while it is sometimes followed in the literature, is an example of what economist John Schmitt calls “applied prejudice.” It simply assigns a negative elasticity, which, by definition, yields a negative result. The accepted practice among empirical economists is to investigate the actual impact of minimum wage increases.
Bernstein and Schmitt (1998) investigate the employment effects of the 1996-97 increase. They ran four separate tests of the employment impact of the increase and found:
- None of the tests show systematic job loss;
- The effect on employment is generally economically small and statistically insignificant; any impact is almost as likely to be positive as negative;
- Our update of the traditional method for estimating the minimum wage’s employment effects (time-series models) reveals a statistically insignificant result that is less than half the magnitude of even the lowest of earlier estimates.
Since that research was completed in early 1998, the labor market conditions facing low-wage labor market have continued to improve, contrary to the predictions of those who opposed the increase. Month after month in the current labor market, low-wage and minority workers post historic gains. In recent months, unemployment rates of African-Americans, Hispanics, and 16-24 year olds (who are likely to be low-earners) all hit 30-year lows. Figure 2 shows the changes in unemployment rates for various groups of disadvantaged workers from the month before the last increase to the most recent month for which data are available (March, 1999; these data are seasonally adjusted). In every case, unemployment rates have declined.
Similarly, the employment rate of young (16-25 year old) African-American women with a high-school degree increased by 6.6 percentage points, 1995-98, and their unemployment rate fell by 4.2 percentage points. These gains are much larger than the average, economy-wide changes in these labor market indicators.
This last group (young, minority females, with high-school degrees) is particularly relevant to the issue of the minimum wage and poverty. Due to welfare reform, many single mothers are required to leave the welfare rolls for the job market. Has the minimum wage increase been a complement or an impediment to their progress?
Due to both the change in welfare policy and the growing economy, welfare caseloads have fallen precipitously since 1996, when the minimum wage was increased. But, in the context of the job loss critique, the relevant question is whether the 1996-97 increase made it tougher for the women to enter the labor market. In fact, the employment rates of single mothers, after stagnating for many years, rose steeply from 62% in 1995 to 69% by 1998, directly over the period when the minimum was increased. Of course, not all single mothers are on welfare, but among those receiving welfare, employment rates grew from 40% in 1995 to 49% in 1997, by far the highest level on record.
Of course, one could always argue that these gains would have been larger in the absence of the minimum wage increase. Such claims are unknowable, and in this context should be discounted. Note, for example, that the seven and nine percentage-point increases in the employment rates of single mothers just cited were six and nine times, respectively, greater than the increase in the overall rate of employment growth over this period. The fact that these low-wage labor market indicators have far outpaced the overall changes in the labor market should convince objective observers that the 1996-97 increase has not hurt the employment prospects of low-wage workers.
Will the Increase Lower Poverty?
In this final section, I address the evidence regarding minimum wage increases and the poverty rate. The key points here are:
- In reviewing literature for this hearing, I found that the vast majority of analyses found small poverty-reducing effects, i.e., increases in the minimum wage were associated with small declines in the poverty rate. In fact, a review of articles on this topic published over the last 10 years in peer reviewed journals revealed no case in which the increase was associated with significant increases in the poverty rate.
- The channel through which this occurs is by delivering more earnings to the working poor. All of the above research finds this to be the case.
- Neumark et al’s recent work on this topic has been raised to cast doubt on this widely held conclusion. In fact, their findings also support the general literature. As these authors note in a recent paper, “one might reasonably interpret our results as indicating that minimum wages have little effect on poverty rates.” As regards the impact of minimum wage increases on the incomes of poor families, these authors write: “we also find that the minimum wages tend to boost the incomes of poor families that remain below the poverty line.” Finally, they note that, as regards the question of whether minimum wages lead to an increase in the family poverty rate, they write: “this estimated net effect is insignificant.”
- The reason the anti-poverty effects of the minimum wage have historically been found to be small is due to the marginal labor market participation of many poor persons. However, as the above employment rate trends reveal, we should expect these effects to grow as more poor and near-poor persons increase their hours of work. This fact also underscores the importance of complementary income policies targeted at low-income workers such as the Earned Income Tax Credit and food stamps.
- When considering the impact of minimum wage increases, it is a mistake to focus solely on those below the poverty line. Many families move in and out of poverty, and
near-poor families should also be considered as potential beneficiaries of the policy. Also, minimum wage increases have been found to increase the incomes of poor families without necessarily lifting them above the poverty line (thus, lowering the poverty gap).
- Finally, there has been only one paper of which I am aware that includes the impact of the most recent increase on poverty rates. This newly published study finds a small poverty-reducing effect of minimum wage increases. My tabulations of the impact of the most recent increase on poverty show 1) a net increase in the share of persons moving out of poverty, 1996-97, and, among those persons who remained in poor families in both 1996 and 1997, an nominal increase in annual earnings of 13% (over this period, the federal nominal minimum wage was raised by 12%).
An objective reading of the literature on minimum wages and poverty leaves one with two salient impressions. First, the minimum wage is a useful, though limited, anti-poverty tool, and second, from the perspective of economic policy analysis, it is a big mistake to limit this analysis solely to the poor, i.e., to ignore the near-poor.
As noted above, the empirical economics literature is widely agreed on the contention that minimum wage increases have small anti-poverty effects. In a 1989 paper, Burkhauser and Finegan find that an increase in the minimum wage would reduce the poverty rate of low-wage workers by 21%. Mincy (1990) also finds an anti-poverty effect of a similar magnitude. However, both of these studies point to the fact that many of the benefits from the increase flow to families above the poverty line, a point I return too shortly. Card and Krueger (1995) also examine this issue, finding some evidence that between 1989 and 1991, states where the minimum wage increase had a greater impact had slightly larger declines in their poverty rates. Their results, however, like most in the literature are statistically fragile. These authors, both of whom are renowned empirical economists, argue that the small magnitude of these effects “points to a modest poverty-reducing effect of the minimum wage” (307).
The most recently published evidence of the minimum wage as an anti-poverty tool comes from the work of Addison and Blackburn (1999). Their work is particularly germane to the question of a new increase because it contains data through 1996. The authors write that their “results provide evidence that increases in minimum wages in the 1990s have served to reduce poverty,” (407, italics in original). One particularly interesting finding from this paper is that in their analysis, minimum wage increases go from having no statistically identifiable effect on poverty in the 1980s to having an anti-poverty effect (specifically for teenagers and high-school dropouts) in the 1990s. These effects are driven by increases in both earnings and weeks worked.
This finding has two implications: first, it supports the speculation, noted above, that as welfare reform pushes poor persons into the labor market, minimum wage increases will be increasing useful as a poverty reduction tool. Second, it lends support to the policy currently under consideration. Raising the minimum wage should help to make work pay for poor and near poor persons entering the labor market.
The evidence strongly supports the contention that minimum wage increases reduce poverty. The effect, however, is small. Why is this the case?
There are two related reasons. First, while the benefits of the increase go disproportionately to low-income families, not all of these families have incomes below the poverty line. It is important to recall that our poverty definition is a fixed income line; thus, a family whose income puts them one dollar above this line is not considered poor in our national accounting system. However, few would deny that these families also need to be supported by a national wage floor. In this regard, the poverty criterion is a somewhat artificial one in this aspect of the debate.
Second, many poor families are only marginally attached to the labor market. In 1997, the most recent year for which we have data on poor families, the average hours of work for poor families (pooled across the family, and excluding one-person families) was 1,042; for families with at least one worker, the average was 1,503. Even with the proposed increase currently under consideration, a single-parent family of three that depended wholly on the minimum would be poor.
As noted throughout this testimony, the work effort of poor families has been growing quite swiftly in recent years, in part due to welfare reform (for all poor families, the increase in average annual hours, 1996-97, was 6.3%). For this reason, it is reasonable to expect the minimum wage will play an increasingly important anti-poverty role as time progresses (the findings in Addison and Blackburn (1999) support this). Nevertheless, if our intention is to help working poor families reach, if not surpass, the poverty line, then these families clearly need other income supports. Three obvious candidates are the EITC, food stamps, and subsidized child care.
Figure 3 shows two stacked bars representing the income of a single mother with two children, with earnings from the minimum wage before and after the increase (minus payroll taxes), the EITC and food stamps. In both cases, I assign the average hours of work, pooled across the family, for poor single mothers with children: 1,164 in 1997. In neither case, does the family reach the poverty line of about $13,000. However, the EITC and food stamps just about double the after-tax earnings of this poor, single parent.
Note also that some of the extra earnings from the increase are lost due to the reduction in food stamps, which are phased out as earnings increase. This has the disadvantage that low-income families on food stamps do not receive the full benefit of the increase in the minimum wage. From the perspective of public spending however, the minimum wage increase clearly reduces public expenditures for the poor. Thus, the composition of family income shifts such that the share of income from earnings increases relative to the share from unearned sources. Given that this Congress has continuously stressed the importance of reducing dependence on unearned income, the minimum wage increase should be viewed favorable in this regard.
By itself, the increase in the minimum wage will not end working poverty. But the evidence presented in this testimony strongly points to the conclusion that the increase will help lift the living standards of working poor families, thus helping to reconnect their economic fortunes to those of the rest of America’s workforce.
[ POSTED TO VIEWPOINTS ON MAY 20 ]
Jared Bernstein is an economist at the Economic Policy Institute. He specializes in labor markets and wage inequality and is a co-author of The State of Working America 1998-99.