Presented before the U.S. House Committee on Ways and Means, Subcommittee on Income Security and Family Support, on April 26, 2007.

Good afternoon. My name is Gordon Berlin, and I am President of MDRC, a nonprofit, nonpartisan education and social policy research organization that is dedicated to learning what works to improve policies and programs that affect the poor. Founded in 1974, MDRC evaluates existing programs and tries out new solutions to some of the nation’s most pressing social problems, using rigorous random assignment research designs or near equivalents to assess their impact. I appreciate the opportunity to appear before this Committee today to describe what research tells us about the best ways to alleviate poverty.

I will make four points:

• After declining by half between 1959 and 1972, the poverty rate in the United States has remained stuck between 11 and 15 percent ever since. Why? The prime explanations are rising rates of single parenthood and falling real wages, particularly among men with low levels of education. Of the two, the decline in wages is the more instrumental — that is, falling earnings is a problem we can redress and we have good evidence about what works.



• A compelling body of evidence points to effective solutions — both short term and long term — for alleviating poverty related to low earnings today and the intergenerational transfer of poverty tomorrow. In the short term, enhancing the Earned Income Tax Credit (EITC), especially for single individuals, and indexing the minimum wage to inflation could be an effective strategy for boosting employment and earnings and reducing poverty. In the long term, investments in educational reform — from pre-kindergarten classes to community colleges — should equip the next generation with the skills they need to obtain high-paying jobs.



• These short- and long-term two-generation strategies are interdependent: Providing enhanced work supports to adults to move families out of poverty today has positive effects on young children’s school performance — and provides a strong foundation for long-term efforts to prevent poverty tomorrow through improved educational opportunities for poor children.



• An aggressive strategy to address falling wages would redesign and expand the EITC benefit for individuals, regardless of their parenting or marital status, conditioned on working 30 hours a week and determined on the basis of individual income rather than joint income. Retaining the current EITC for families with children while creating a new EITC for single individuals (including noncustodial parents and second earners in two-parent households) could have wide-ranging positive effects on employment, earnings, income, and poverty — as well as on family well-being. But because the costs of such an initiative would be high, a prudent first step would be a demonstration project with a rigorous research design in three or four cities to determine if the plan’s benefits outweigh its costs.

Falling Wages and Poverty

For more than 40 years, the conventional wisdom has been that the best antipoverty strategy is to help the unemployed get jobs. And while work is a necessary precondition to escaping poverty, getting jobs is not the problem it once was for most segments of the population — as unemployment has remained at the historically low rate of between 4 and 6 percent for the past 10 years. The key problem facing most poor people is that many jobs simply don’t pay enough.

In 1959, when we first began to measure poverty, 22 percent of all Americans lived in households with income below the poverty line. By 1972, the poverty rate had been cut in half, falling to 11 percent nationally. But then the poverty rate stopped declining and ranged between 11 percent and 15 percent, depending on the state of the economy, for the next 30 years (see Figure 1). Why didn’t poverty continue falling?

Falling wages and increasing rates of single-parenting are the two principal explanations, and, as I’ll explain, these phenomena are closely related. Economic changes led to stagnant and declining wages at the bottom of the wage distribution, especially among men with a high school diploma or less, and demographic changes saw a near doubling of the fraction of all families with children headed by a single parent.

Let’s focus on wages and earnings. Between 1947 and 1972, average earnings grew in real terms by 60 percent for nonsupervisory workers. As Frank Levy has described, it was as if the whole nation were on an economic up-escalator. It was this rise in earnings that explains much of the postwar decline in poverty until 1972. But then earnings began to tumble. In fact, by 2004, the average production worker’s weekly earnings had fallen to $528 (in inflation-adjusted dollars), a 15 percent decline (see Figure 1).

Wages and earnings declined initially as a result of the recessions of the 1970s. But this era was also the start of a major restructuring of the economy, in which the premium placed on education began to grow. A new skills bias started to dominate the labor market, creating high-paying jobs that required a college degree or better and lots of low-paying jobs that required no more than a high school diploma and often less. As a result, economic inequality — the gap between the richest and poorest Americans — widened during the 1970s and 1980s as earnings for those with college accelerated, while wages for those at the bottom fell in step with the massive loss of high-paying blue-collar jobs as a result of industrial restructuring. The decline of unions, rising competition from low-skilled newcomers, and the erosion of the minimum wage all exacerbated these trends.

How did Americans cope with this decline in earnings? Two-parent families maintained their standard of living by having fewer children and sending both parents into the workforce. Single-parent families, of course, did not have the option of putting another parent to work. In fact, employment rates among single mothers grew rapidly during the 1980s and 1990s — but, because single parents were more likely to be poorly educated and because they only had one earner, inequality widened.

As earnings fell, other manifestations of poverty worsened: employment declined (particularly among less-educated men), marriage rates fell, and crime rates rose. Of course, these problems are intertwined and reinforcing. For instance, as the wages of men with a high school education or less tumbled, the employment rates of these men also fell, and, in turn, the share who could support a family above the poverty line began to decline — and with it the professed willingness of low-income mothers and fathers to marry. Indeed, among men aged 25-54 with a high school diploma or less in 2003, the earnings of a quarter of whites, a third of blacks, and two-fifths of Hispanics were inadequate to support a family of four above the poverty line. Certainly, the choices that individuals make — whether to have children within marriage or not, whether to take a low-paying job or to become involved in criminal activity — play an important role in determining one’s poverty status. Yet, it is hard to argue that technological change, globalization, and other large macroeconomic forces that have transformed the American and world labor markets haven’t played an independent, causal role in poverty’s persistence.

Men with a high school diploma or less, especially men of color, have been particularly hard hit. Over the same period that wages were falling, employment rates among men were also tumbling, down a startling 20-plus percentage points between 1970 and 2000 for men with a high school education or less and roughly 7 percentage points for those with some college. By contrast, as a result of economic necessity, changing norms, and the rise of service sector jobs, women’s employment rates rose dramatically as more and more women entered the labor market, especially in the 1990s.

Why have men’s employment rates been declining? For some men, as blue-collar jobs evaporated and wages fell, employment became less attractive. The strong economy of the 1990s offers a reverse proof: As wages at the bottom rose, the employment rates of white, black, and Hispanic young men stabilized and began to grow. For example, the employment rates of black men aged 16 to 34 rose between 1992 and 2000, as did the rates for young black men (16 to 24) with a high school diploma or less (see Table 1). But once the boom years were over, the employment rates of black men resumed their downward trend, plunging following the 2001 recession much as they did during the 1991 recession. While the reasons for the dismal position of young black men in the labor market are complex (and include racial discrimination and inadequate basic skills and education, as well as the behavioral changes noted by Larry Mead and others), a key part of the explanation is the interaction among low wages, the rewards of illegal activity, and strict drug laws, which have resulted in as many as 30 percent of all young black men becoming entangled with the criminal justice system at some point. Incarceration appears to have its own independent effect — the label of ex-offender further worsens and taints the future employment prospects for all former prisoners reentering society.

In sum, poverty stopped falling in large part because earnings stopped rising. And while poverty is a complex problem with many causes, it seems clear that the nation must address the problem of low-wage work in order to further reduce poverty — because low-wage work is here to stay. The Bureau of Labor Statistics projects that 46 percent of all jobs in 2014 will be filled by workers with a high school diploma or less. The bulk of these jobs — janitor, food service, retail sales, laborer, child care provider, home health aide — are expected to offer either low or very low pay.

Strategies That Work to Reduce Poverty

So, what to do? There are essentially two types of antipoverty strategies the nation could pursue. The first are short-term interventions, most focused on “making work pay” by supporting low-wage workers with earnings supplements and other kinds of supports, including upgraded training. The second type are long-term, intergenerational strategies — principally investments in preschool through postsecondary education — so that the next generations of young people have the knowledge and skills to fill higher-paying jobs. Importantly, the two strategies reinforce each other; for example, lifting a family above the poverty line with an earnings supplement can increase young children’s school performance — in effect, enhancing the payoff of a high-quality early childhood education program. My focus today is on some of the short-run strategies that fall under the jurisdiction of the Ways and Means Committee.

If low wages are the principal problem we face in reversing poverty, one might reasonably ask: Can government successfully intervene to raise earnings and incomes and reduce poverty? Encouragingly, a reliable body of evidence demonstrates that work-based earnings supplements — including the Earned Income Tax Credit (EITC) — can boost employment and earnings and reduce poverty. For very low-wage workers, hiking the minimum wage would likely have similar effects, so long as the increase was not high enough to result in reduced hiring by employers.

The “Make Work Pay” Experiments. Concerned that low-wage work simply did not pay relative to welfare, the state of Minnesota, the New Hope community group in Milwaukee, and two provinces in Canada began to experiment during the 1980s with new approaches designed to increase the payoff from low-wage work — that is, to make work pay. All three provided work incentives in the form of monthly cash payments to supplement the earnings of low-wage workers. The payments were made only when people worked, and the amount of each month’s cash payment depended on the amount of each month’s earnings.

The results were encouraging. The mostly single mothers who were offered earnings supplements in these three large-scale, rigorous studies were more likely to work, earned more, had more income, and were less likely to be in poverty than control group members who were not offered supplements. At their peak, these employment, earnings, and income gains were large — reaching 12- to 14-percentage-point increases in employment rates, about $200-$300 more per quarter in earnings, and $300-$500 more in quarterly income. The earnings supplements also had a secondary benefit for children. Preschool-age children of participating parents did better academically than like children in the control group, in part because their parents had higher incomes and they were more likely to attend high-quality, center-based child care programs. The largest and most persistent effects on adults were found for African-Americans and for the most disadvantaged participants, particularly high school dropouts without recent work history and with long welfare spells. For these groups, the employment and earnings effects continued through the end of the follow-up period — six years in the Minnesota project — implying that early work experience could provide a lasting leg up in the labor market for more disadvantaged populations. The pattern of results for all participants also suggests that income gains — and thus the poverty reduction effects — could be sustained by an ongoing program of supplements. (The earnings supplements in these demonstration projects ended after three years.)

Rent Incentives for Public Housing Residents. A more recent program that used earnings supplements — in this case, in the form of rent breaks for public housing residents conditioned on work — had large positive earnings effects for many different types of residents, including striking earnings effects for immigrant men, and positive but smaller impacts on residents’ employment rates. Called Jobs-Plus, this ambitious place-based effort changed traditional public housing rules so that tenants’ rents did not rise as quickly or at all when their earnings grew (that is, rents were held flat). In addition to this financial work incentive, Jobs-Plus offered employment-related assistance, on-site case management, and job-related information sharing through resident networks.

The Earned Income Tax Credit. Members of this Committee have used employee subsidies as an integral part of the nation’s strategy for reducing poverty since the EITC was first passed in 1975. Today the EITC, which the Committee substantially expanded in 1986, 1990, and 1993, is available to all low-income workers who file tax returns. It is refundable, which means that its benefits are paid out even when the tax filer does not owe any income taxes. More than 20 million taxpayers take advantage of the EITC each year, at a cost approaching $40 billion, making it by far the largest cash benefit program for the poor.

The EITC’s distinguishing feature is its status as a safety net built around work — only people with earnings can claim the credit. The amount varies by both family type and earnings. Families with two or more children can receive a maximum credit of $4,716; those with one child, $2,853; and single adults with no children, $428. However, because the EITC overwhelmingly benefits single parents supporting children, it largely excludes single adults without children who are poor (and disproportionately male) and it creates disincentives to work and marry for some families. Although recent changes have reduced marriage penalties in the EITC, some do remain, particularly when both spouses in a married-couple family have similar earnings.

Based on a comprehensive review of studies, Steve Holt reports that the EITC reduces family poverty by a tenth, reduces poverty among children by a fourth, and closes the poverty gap by a fifth. Note that the Census Bureau’s official poverty estimate doesn’t count the EITC as income. If it did — and if one also subtracted the cost of work expenses and child care — the poverty rate would likely fall by a couple of percentage points, but the trends described in Figure 1 would remain pretty much the same. On the other hand, if certain recommendations of a National Academy of Sciences Panel on Poverty were adopted, the poverty rate would likely be somewhat higher.

Raising the Minimum Wage. Both experience and empirical evidence suggest that the minimum wage can play a valuable role in raising wages and reducing poverty without severely distorting labor markets. However, as of early 2007, the value of the federal minimum wage had fallen to its lowest level in 50 years. Both President Bush and Congressional leaders have vowed to increase the minimum wage to $7.25, although if it is not indexed to inflation, its value will once again gradually erode over time. Thirty states and the District of Columbia have established minimum wages above the federal level.

However, only one in five minimum-wage workers live in families with below-poverty income. Many are between 16 and 24 years old and do not support families, making the minimum wage a relatively inefficient way to reduce family poverty, although its efficiency improves somewhat if the goal is to help workers up to 200 percent of the poverty line. In addition, the political unpredictability of the minimum wage makes it an unreliable policy lever for supporting low-wage workers. A minimum wage increase to $7.25 an hour could substantially boost wages at the bottom, lifting some families above the poverty line, helping other families both below and just above the poverty line, while reducing the overall public cost of the EITC. While available evidence summarized by David Card and Alan Krueger suggests that a boost to $7.00 an hour or so would not have a noticeable effect on employment rates, economic theory and practical experience suggest that, above some wage level, employers would cut back on hiring. For all of these reasons, a higher minimum wage, in and of itself, would not permanently address the problems of persistent poverty.

Considering the Policy Choices: Expanding the EITC and Boosting the Minimum Wage

Over the next 10 to 20 years, it is hard to imagine reducing poverty without finding a way to make low-wage work pay better. The compelling body of evidence that I’ve just described suggests that expanding the EITC, preferably in combination with a boost in the minimum wage, would be an effective way to supplement low earnings. While there are a number of ways that one could envision such an expansion, I believe that it boils down to essentially two broad policy questions:

1. Is it best to expand the current EITC program, with its emphasis on families with children, or to address the imbalance that has emerged between singles and those with children by expanding the EITC program for individuals?



2. Is it enough to raise the minimum wage without indexing it for inflation?

An across-the-board increase in the existing EITC moves more families above the poverty line and increases the incomes of those just above and below the poverty line, but perpetuates current inequities by doing little to address the companion problems of single parenthood, single men’s and women’s low earnings, or remaining marriage penalties in two-earner families. The second approach has the advantage of reducing marriage penalties, but it shares several of the shortcomings of the first, it may create stronger work disincentives for second earners, and it encourages people to marry for the money, running the risk of promoting any marriage over a healthy marriage. Moreover, it fails to tackle the problem of the low wages of single adults, particularly men.

Expanding the EITC for Singles. The third strategy for expanding the federal EITC — further supplementing the earnings of individual workers without children — may seem counterintuitive at first. But single men and women (as well as second earners in two-parent households) have been mostly ignored by the expansion of the EITC in the last 20 years. Single men (many of whom are noncustodial parents with child support obligations) have been the hardest hit by the losses in the manufacturing sector and the decline in earnings since the early 1970s. This imbalance in the EITC has had the unintended effect of distorting incentives to work, marry, and have children. An increase in the EITC for singles would help counter three decades of wage stagnation and persistent poverty, with likely positive corollary effects on employment and parental child support. In addition, if the expansion were accompanied by two admittedly radical changes — (1) creating a full-time work requirement and (2) basing eligibility on individual rather than joint income — it would help both singles and parents with children.

One way to increase the EITC for singles would be to simply double the current maximum benefit for individuals, which currently stands at $428. But this strategy would provide only a limited boost to individual earnings and thus might not have a big effect on work behavior; it would surely exacerbate marriage penalties if done alone; and, if passed in conjunction with a rise in the minimum wage, would primarily affect part-time workers rather than full-time workers.

A bolder expansion would provide all adult low-wage workers (aged 21-54) who work full time (30 hours a week) a payment approaching that of the current family EITC (for example, a 25 percent subsidy rate to a maximum payment amount of $1,950) but with a crucial twist: payment would be based on an individual’s personal income, not joint or family income, and singles would be eligible for the supplement whether they have children or not and whether they marry or not, as would second earners in a married family receiving the existing family EITC. By conditioning this new benefit on full-time work, by targeting individuals regardless of their family status, by keeping the existing EITC for families with children in place, and by calculating EITC eligibility on the basis of individual income (as Canadians and Europeans do) rather than joint income for tax filing purposes, this earnings-based supplement would restore equity to the American social compact while minimizing the distortion of incentives to work, marry, and bear children. (A fuller explanation of this idea will be published in the September 2007 issue of The Future of Children.)

Adults working less than 30 hours a week, including second earners in two-parent households, would have an incentive to increase their work hours, further boosting income, promoting self-sufficiency, and reducing poverty. And those adults not in the labor force would have added incentive to find a full-time job, which would substantially boost total income. To administer the 30-hour requirement, employers would be required to report monthly or quarterly hours worked on the end-of-year W-2 statements that employees rely upon when filing taxes. Next, by supplementing the earnings of single men in low-wage jobs and increasing their income, this plan would encourage more “on the books” work, while helping men meet their child support obligations. As in current law, single people who are parents and owe child support would have their EITC payment attached to pay their child support obligations. Importantly, some of the largest benefits would accrue to two-parent households when both adults can work full time. Roughly, 21 million low-wage married individuals and another 16 million single individuals would receive an EITC payment under this plan. Such an expansion would not be cheap; depending on how one structured the benefits, the annual cost for a national expansion would range from $4 billion to $33 billion.

Would an earnings supplement like this really increase employment rates, particularly among single men? Honestly, we don’t know. But there is good evidence to suggest that it might. Economists estimate that increasing the hourly wage of a low-income worker by 10 percent would boost employment between 2 and 10 percent. Adding credence to these estimates, the three make-work-pay experiments that I described earlier had similar employment, earnings, and income effects, albeit for a population of mostly single mothers. And, the New Hope program, which also served single men, did achieve modest, statistically significant gains in the number of quarters employed for men overall, as well as for single men, when cumulated over the full eight-year follow-up period — although the small number of men in the study sample (by design) makes these findings suggestive at best. And as noted above, the higher wages that came with the economic boom of the 1990s also led to increases in men’s employment rates.

Indexing the Minimum Wage to Inflation. History makes it clear that the value of a boost in the minimum wage declines over time, as political will must be continually rebuilt to adjust it for inflation. To address this problem, policymakers should consider going beyond just raising the minimum wage to $7.25 an hour and also index it to inflation. While this won’t bring the minimum wage back up to its original value of about half the median hourly wage, it would forestall a quick return to the erosion in value it has seen in the last decade.

Boosting the minimum wage and enhancing the EITC are complementary, not substitute, strategies. Allowing the minimum wage to erode while the EITC is indexed to inflation has the perverse effect of substituting public dollars for private wage increases. An expansion of the EITC would exacerbate this problem unless the minimum wage was also indexed. In short, increasing the minimum wage and indexing it for inflation would provide a floor below which wages could not fall, would make the expansion of the EITC more effective and more affordable, and would prevent an inflation-indexed EITC from substituting for wage increases employers would otherwise have provided.

What Do We Know About Other Strategies for Reducing Poverty?

While an expanded EITC, like the one I have described, would do much to help low-wage workers and their families, we have to acknowledge that it would not be enough to address all the causes of poverty. Given the prominent role of single parenthood in persistent poverty, why not propose an expansion in marital education programs? Given the changes in the labor market, why not propose additional investments in job retention and advancement? Given the problems of the “hard to employ,” why not propose additional programs to tackle the problems of youth and adults with low skills, no work history, or mental health and substance abuse problems? The short answer is that we don’t have good evidence about what would make a difference. Fortunately, research is now underway that, I believe, will provide more reliable information about what does and does not work to address these problems.

For instance, marriage and childbearing behaviors and high rates of single-parenting, while related to economic changes, are also largely the product of social norms. Low-income couples face greater challenges to building and maintaining healthy relationships (for instance, because of the stress of financial difficulties), and their families are consequently less likely to experience stable marriages. While an extensive body of evidence on how to strengthen marriages exists, this research consists primarily of small-scale studies of typically short-term programs for middle-class couples. MDRC is involved in two large-scale, random assignment evaluations of new programs to promote healthy marriages and co-parenting relationships among low-income families, which should provide important answers about the value of these interventions.

Similarly, even if we were to boost the earnings and income of low-wage workers through an expanded EITC, real prosperity for most Americans comes from moving up the career ladder. In the U.S., no government agency is tasked with supporting low-wage workers by connecting them to benefits (like public health care, child care and housing subsidies, and food stamps) or helping them find better jobs. Through three large-scale projects in the U.S. and the United Kingdom, MDRC is learning how challenging it is to develop programs that actually promote career advancement. However, early results suggest that one-stop centers (created by the Workforce Investment Act) might be a good venue for these programs and that particular strategies, like using financial incentives, contracting with community-based groups with strong business connections, and combining work supports with advancement services (including community college-based education), could be promising.

Finally, earnings supplements are not much help to people who have a difficult time finding or keeping a job. Few strategies have been developed that have proven effectiveness in helping the hard-to-employ find stable employment, but rigorous studies of new initiatives are currently underway, including transitional jobs programs for reentering prisoners and long-term welfare recipients, intensive case management of single parents suffering from depression, accelerated health benefits for disability recipients, residential youth development programs for dropouts, and employment programs for substance-abusers.

Conclusion

The most direct way to alleviate poverty is to tackle the legacy of falling wages, particularly for men with less education. Solid and reliable evidence demonstrates that earnings supplements have encouraged work and reduced poverty among unemployed and underemployed single parents. Expanding the use of earnings supplement for single adults would go a long way toward reducing poverty among low-wage workers and their families. A first step would be to modestly expand the current EITC in conjunction with an inflation-indexed boost in the minimum wage, paying special attention to singles by doubling or tripling the current annual maximum EITC benefit for single adults with no children.

In addition, the Committee should also consider a limited test of a more enhanced EITC for singles along the lines that I have described: for all adult workers, aged 21 to 54, regardless of parenting or marital status, and conditioned on working 30 hours a week. One could imagine a multiyear demonstration in three or four cities that would determine the new benefit’s effects on poverty, earnings, work, marriage/cohabitation, and childbearing and that would provide guidance about the feasibility of expanding the policy when the EITC is next reauthorized. While the cost of scaling up an EITC for singles, in which eligibility is based on individual rather than joint income, seems daunting, it may well be that the long-term benefits of such a plan could more than pay for itself in increased work effort, increased child support payments, increases in the number of two-parent households, and decreases in crime and nonmarital childbearing. But we won’t know unless we conduct a comparatively inexpensive test of the idea in a few places that relies on random assignment research designs whenever feasible.

As I noted at the beginning, just addressing the effects of low wages will not be enough. To make a significant and long-lasting difference, we will need to invest both in short-term strategies that boost the well-being of poor families today — as well as in long-term educational strategies that ensure that succeeding generations will have the skills to succeed in the labor market. Children growing up in poverty do worse in school, have earnings that are substantially lower as adults, and are more likely to become teen parents, among other problems. By reducing poverty through work supports for parents, their children will be in the position to take advantage of better educational opportunities, as we learn more about what works in early childhood education, K-12, and postsecondary reform. The best incubator for developing human capital tomorrow is a family that is not living in poverty today.