Northwestern Journal of Law and Social Policy

I.    Introduction

¶ 1         Despite the political rhetoric of "making work pay," poverty continues to be widespread among full-time working adults and their children in the United States. In 2005, some 3.7 million households, with or without children, included a full-time worker who earned too little to lift family income above the poverty line.¹ Many more American children—5.2 million—lived in families in which an adult worked full-time but the family was still poor.² More than three million children with full-time working parents also lacked health insurance.³

¶ 2         For many working poor, particularly single mothers, lack of time can be as problematic as lack of money. Parents at all income levels find it difficult to balance family and work demands, but the challenges are all the more daunting for parents employed in low-wage jobs with few, if any, benefits. Many work irregular hours or take on a second or even third job, yet they find themselves taking home a paycheck that fails to provide the resources needed to juggle the responsibilities of earning a living and give their children the love and supervision they need.⁴

¶ 3         Our Paper makes the case for a national program offering the kind of work supports that were part of the New Hope program, a policy experiment that operated between 1995 and 1998 in Milwaukee, Wisconsin.⁵ New Hope was created by a coalition of community activists and business leaders who believed that work should be the best route out of poverty. It provided a set of work supports for full-time workers—both parents and nonparents, men, and women—that would lift them out of poverty as well as provide essential benefits in the form of health insurance and child care subsidies for people who needed them.⁶

¶ 4         By stipulating that participants must document full-time (thirty hours per week) work to qualify for benefits, New Hope was a social contract rather than a welfare program.⁷ It was designed to increase the incentives and reduce the barriers to full-time work. For parents burdened with extensive overtime work or second jobs, it provided an opportunity to cut back to a more manageable level of full-time work hours without a proportionate drop in income. If a participant was unable to find a job owing, say, to lack of work experience or a criminal record, the program provided opportunities for temporary community-service jobs that paid the minimum wage but still entitled that person to program benefits.

¶ 5         In terms of values, New Hope hit the trifecta: i) its full-time work focus resonated with the business community, the broader public, and participants themselves; ii) it "made work pay," and iii) its social-contract nature was at once respectful and demanding of participants.

¶ 6         The heart of the policy case for New Hope rests on the program's experimentally proven impact on family well-being and children's achievement.⁸ Among all of the people offered the chance to participate in the New Hope program—including single men—work increased and poverty rates fell. Children in New Hope families performed better in school, were more cooperative and independent, and had fewer behavioral problems and loftier schooling expectations than children in the control group. Because boys have a higher risk of school failure and behavior problems than do girls, it is noteworthy that New Hope was especially successful in improving their school performance and behavior. Even a very conservative estimate of the taxpayer dollars saved by the improved behavior of boys amounted to more than the cost of the program.

¶ 7         A New Hope-style program is best conceived as a way of making good on a "make work pay" promise to American workers and their families rather than as a general solution for the myriad problems of all low-income families. Participation in New Hope was voluntary, and only about half of the people who won the right to enter the program actually took up benefits in any given month of its duration.⁹ Reasons for nonparticipation varied widely, with some people unable or unwilling to meet the thirty-hour work threshold to qualify for benefits, others so successful that they no longer qualified for the program's sliding-scale subsidies, and still others not wanting to be bothered with the program's periodic certification requirements. ¹⁰

¶ 8         New Hope was a small, experimental program run in a state with a culture of welfare reform and work-focused initiatives,¹¹ which raises issues about replication.¹² Fortunately, there is additional experimental evidence to support the success of New Hope-type programs. During the period of New Hope's operation, Minnesota and two Canadian provinces also tested employment policies that shared some of New Hope's key features, particularly earnings supplements.¹³ Minnesota's program emphasized training welfare caseworkers to support work efforts rather than simply to process assistance claims.¹⁴ Both Minnesota and Canada evaluated their programs using random assignment and both produced strikingly similar impacts—more work, less poverty, and higher child achievement.¹⁵

¶ 9         All in all, New Hope accomplished its goal of increasing full-time work, lifting workers out of poverty, and providing families with the tools to help balance work and family needs. With its positive effects on children's achievement and behavior, a scaled-up New Hope program would help to break the cycle of poverty for a sizeable number of American families—black and white, urban and rural—in the next generation.

II.    The New Hope Model

¶ 10         The New Hope model was that of a social contract.¹⁶ Low-income (family income below 150% of the poverty line) participants committed themselves to full-time work, and New Hope in turn promised them a package of work supports ensuring that they would not be poor and their families would be able to afford health insurance and licensed child care. New Hope was open to all low-income adults, men and women, regardless of family circumstances.

¶ 11         New Hope offered a "cafeteria" of benefits from which participants could pick and choose—a feature that allowed families with diverse needs and circumstances to tailor the program to their individual situations.¹⁷ Specifically, conditional on participants providing pay stubs or other proof of full-time (thirty or more hours) work per week, New Hope provided three sliding-scale benefits: i) an earnings supplement raising income above the poverty line; ii) subsidized child care; and iii) subsidized health insurance. Individuals unable to provide pay stubs or other proof of full-time work were eligible for: iv) a temporary community-service job.¹⁸ More details on each of these benefits are provided in Section V.¹⁹

¶ 12         New Hope was designed to make work pay while providing the tools for low-income adults to sustain a reasonable work-family balance. Its services were available in a single office to facilitate the time-consuming and confusing process of dealing with multiple agencies.²⁰ Although many participants in the original New Hope program were only interested in benefits, all had access to help from a caseworker who provided information about jobs, educational opportunities, child care, and other community resources in an atmosphere of respect. While the New Hope demonstration was limited to a period of three years, a national New Hope program, as with the Earned Income Tax Credit (EITC), would provide a permanent set of supports available to low-wage, full-time workers as long as they met eligibility criteria.

III.    New Hope's Evaluation

¶ 13         The case for a New Hope-style package of work supports rests on evidence from a rigorous random assignment evaluation. One-half of the individuals who applied for the program were randomly selected to participate in New Hope for three years (totaling 678 participants).²¹ The other 679 formed a comparison group that was ineligible for New Hope benefits.²² All individuals who volunteered for the chance to get into New Hope lived in Milwaukee's two poorest zip codes—a northside neighborhood with many African Americans and a southside neighborhood home to many Hispanic families.²³ Because they volunteered for the program, New Hope participants and their control-group counterparts were more work-ready than average in their communities; indeed about one-third were already working full-time when they volunteered for the program.²⁴ Both the New Hope and comparison groups continued to be eligible for all other federal, state, and local programs (and to be subject to the rules of those programs) during a period of rapidly changing welfare and poverty policies in Wisconsin and across the nation.²⁵ Both groups enjoyed the fruits of Milwaukee's strong economy in the mid-1990s, and both could claim the increasingly generous federal and state EITC that supplemented the earnings of low-income workers.²⁶

¶ 14         Evaluators monitored the implementation of the program and tracked patterns of employment, earnings, and receipt of food stamps and cash assistance through administrative records. Two, five, and eight years after participants entered the program, surveys of participants, older children, and children's teachers provided additional information about job histories, family changes, economic circumstances, mental health, and child well-being.²⁷ We concentrate on impacts found after two years of program operation.²⁸

¶ 15         For a close-up view of how the program was affecting families, part of the evaluation team conducted in-depth interviews during the three years of the families' participation. These interviews were conducted through periodic family visits to a random subset of forty-four parents and their children who had been selected at random from both New Hope families and families in the comparison group.²⁹

IV.    Program Take-Up and Impacts

¶ 16         More than one quarter of New Hope participants were male, half were black, and one-quarter were Hispanic. The vast majority (eighty-five percent) of New Hope participants reported at least some full-time employment prior to random assignment; however, a majority (fifty-five percent) was not employed at the point of random assignment and nearly one-third of participants had no earnings in the previous year.³⁰ The average earnings of sample members in the year prior to random assignment were $5515 (adjusted for inflation to 2005 dollars).³¹ On average, study participants were employed in just over half the quarters of the year prior to random assignment.³² Just under two-thirds were receiving government aid in the form of cash assistance, food stamps, or Medicaid when they entered the study.³³

¶ 17         Among people enrolling in New Hope, almost all (eighty-eight percent) took up benefits for at least some portion of their three-year eligibility periods.³⁴ Most spoke enthusiastically about both their interactions with program staff and the importance of the benefits they received from the program to themselves and their families. But despite the nearly universal taking-up of at least some program services, relatively few participants took advantage of all of the benefits all of the time. In any given month, only about half of participants received program benefits.

¶ 18         The survey and qualitative interviews showed that people used New Hope and other available programs in the context of their existing resources, everyday routines, and family demands.³⁵ The New Hope offer made a big difference for some people, but it was not a good fit for everyone. Some parents refused to entrust their children to the care of someone other than a family member. Many parents worked evenings and weekends, when few child care centers or licensed home settings were available. In such cases the child care subsidy was of no use. Other potential participants had personal difficulties that prevented them from working regularly. The vagaries of low-wage jobs, in particular irregular and unpredictable work hours that fell short of the thirty-hour weekly work requirement, made it difficult for some people to take advantage of New Hope's benefits. A lesson is that no matter how well intentioned and otherwise well-designed a policy is, it must fit the context of people's lives if it is to make a difference. Adults who applied for New Hope were already trying to make ends meet, to sustain a family routine, and to provide some moral direction for their children and their lives. In order to be effective, New Hope had to contribute to these family routines and goals. The evaluation suggests that it did so for most families, but certainly not for all.

¶ 19         Averaged across the three years of program operation, New Hope increased employment relative to controls by about five percent (about $500) per year (Table 1).³⁶ New Hope attracted both nonworkers and full-time workers, so employment impacts varied accordingly. For adults who were working less than full-time when they signed up for the program, New Hope boosted annual payroll-based employment by seven percentage points and annual earnings by about $1000 (an increase of thirteen percent over the earnings of controls). New Hope's community-service jobs played an important role in these impacts. For those already working full-time when they enrolled, the program generally sustained employment, although fewer New Hope participants than controls worked more than fifty hours per week, which suggests that some full-time workers used program benefits to finance reductions in overtime or multiple jobs.

¶ 20         Using the information gathered when people first signed up for the program, evaluators formed an index of obstacles to finding work that New Hope supports might help to reduce, such as a spotty employment history, no high school degree, an arrest record, and the presence of many children or very young children in the family.⁴² About one-quarter of the total participant pool had several such problems, making it less likely that a package of economic benefits would be sufficient to boost their employment.⁴³ At the other end of the spectrum were workers with none of these employment barriers, who appeared able to sustain employment on their own without help from New Hope. Program impacts on both of these groups were generally quite small.⁴⁴

¶ 21         A third group facing only one or two significant but not insurmountable barriers to full-time employment should have been able to profit most from New Hope's package of benefits.⁴⁵ And indeed they did. The impacts on employment and earnings for this group—who collectively made up nearly one-half of the entire set of applicants—were both large and enduring. During its three years of operation, New Hope boosted employment by nine percentage points and annual earnings by $2500—nearly thirty percent higher relative to controls. For participants facing one barrier, employment and earnings impacts continued up to five years after the program ended.⁴⁶

¶ 22         These remarkably long-lasting results show the important niche that a program such as New Hope can fill. For the nearly one-half of those attracted to New Hope's voluntary work supports who were already close to being able to sustain higher levels of employment and earnings, New Hope appeared to arrive at a crucial time in their work careers. In the absence of program supports, they would probably have increased both work and earnings modestly. With New Hope, these work and earnings increases were much larger and longer lasting.

¶ 23         New Hope included men because its designers believed that all adults deserved the opportunity to escape poverty through work.⁴⁷ Although many of them were fathers, they did not fit any single demographic profile. Most were in their twenties and thirties and had never married, but many were living with partners or other family members. Their work efforts were often handicapped by an assortment of employment barriers: more than one-third had been arrested; only one-half had a high school diploma or GED; and only one-third had access to a car. During the three years of New Hope, the employment rate of New Hope men was an average of eight percentage points higher than that of control-group men.

¶ 24         In contrast, the employment and earnings of women who were not living with children were not greatly affected by New Hope.⁴⁸ There were no significant program impacts on either the employment rates or the earnings of these women during the three program years, and there appeared even to be some signs of negative impacts on their earnings a few years after the program ended. The evaluation offered no ready explanation for these gender differences.

¶ 25         Based on administrative sources of earnings and transfer income, poverty rates were dramatically lower for New Hope than for control families—seventeen percentage points lower in the first year, twelve points lower in the second and third years, and eight points lower in the two years after New Hope benefits stopped.⁴⁹ Although New Hope failed to eliminate poverty among all of its families, it was clearly more successful in lifting families out of poverty than was the collection of programs available in Wisconsin in the mid-1990s.

¶ 26         New Hope's greatest success was in increasing school achievement and positive behavior among children, especially boys.⁵⁰ Expressed in SAT-type standard deviation units, teachers ranked the average child in a New Hope family twenty-five points higher in achievement than the average child in a comparison family (Table 2). The impact on boys' achievement (+33 points) was considerably larger than the results for girls (+12 points). Teachers also rated the boys in New Hope families much more favorably than boys in control families on "positive social behavior"— obeying rules in school, being admired and well-liked by other students, and being self-reliant. They reported fewer disciplinary problems and less frequent behavior problems—less arguing, disturbing others, social withdrawal, or sadness.

¶ 27         In the qualitative substudy, parents appeared to worry more about their boys than girls, especially when they reached early adolescence, and there was experimental evidence that New Hope's child care supports were more likely to be used for boys than for their sisters.⁵¹ Mothers often said that their boys were vulnerable, and they use any resources they have to counteract negative influences. As one mother said, "It's different for girls. For boys, it's dangerous . . . [Gangs are] full of older men who want these young ones to do their dirty work. And they'll buy them things and give them money."⁵² New Hope boys were more likely than girls to be in organized after-school programs where they received help with homework and had opportunities for recreation. Impacts for girls were often small and insignificant; indeed, teachers rated New Hope girls as somewhat more disobedient and aggressive. In part, this may have been due to the fact that boys had much more room to gain from an intervention than did girls.

V.    Some Details on Key Elements of the New Hope Program

¶ 28         Were New Hope to be implemented nationally, what would its program components look like?

¶ 29         The New Hope earnings supplement had two components: a wage supplement, directly aimed at increasing the wage rate of low-wage workers, and a child allowance, which supplements the income of families with one or more children.⁵⁶ The combination of these two components produced an earnings supplement that maximized work incentives for individual workers while at the same time effectively supported families of different sizes.

¶ 30         Compared to the earnings supplement, the administration of the New Hope health insurance benefit was relatively straightforward. As long as a family maintained its eligibility for New Hope benefits, all family members automatically qualified for enrollment in a comprehensive health insurance plan if they did not have access to such a plan through their employer or through a public program such as Medicaid or State Children's Health Insurance Program (SCHIP).⁵⁷ Participants paid modest co-payments on a sliding scale that ranged from approximately $10-$150 a month, depending on family size and income. If participants had access to employer-provided health insurance with co-payments larger than those offered by New Hope, they did not have to switch plans; New Hope instead subsidized the employer co-payments, which was almost always less expensive and more convenient than providing health insurance directly. Participants lost their eligibility when they stopped working at least thirty hours a week, although eligibility was extended for a month when participants lost one job and were looking for another.

¶ 31         New Hope's child care subsidy was the most expensive component of the experimental program, although it would cost considerably less now since most states have used Temporary Assistance for Needy Families (TANF) and other funds to expand dramatically their child care assistance programs.⁵⁸ Rules for child care subsidies were fairly simple: as long as a parent met the work requirement, New Hope either assisted families to apply for and utilize existing state child care subsidies or relied on its own funds to subsidize care in any state-licensed or county-certified center or family child care home for all children younger than age thirteen. ⁵⁹ Parents chose their own child care setting and were responsible for paying their share directly to the provider (in co-payments). Participants did not lose subsidies altogether until their incomes reached the greater of either 200% of the poverty level or $30,000 annually—slightly higher than the current income eligibility cap in most state child care assistance programs—although parents' co-payments increased with their income. Parents qualifying for the child care subsidy qualified their children for the rest of the school semester.

¶ 32         Similar to the health insurance benefit, New Hope participants had access to subsidized child care as long as they continued to work at least 30 hours a week and their co-payments were calculated on a sliding scale, with all benefits fading out when family earnings reached 200% of the poverty level. Co-payments ranged from a low of $50 per child per month when earnings were equivalent to full-time employment (or less) at minimum wage and increased to the full child care cost as earnings approached $28,000 a year or twice the poverty level, whichever was less.

¶ 33         Community service jobs (CSJs) were available to participants who were unemployed or whose jobs offered only part-time hours, but only after they spent eight weeks searching for a regular job.⁶⁰ CSJs were not guaranteed; participants had to interview successfully for them and could be fired for cause. They lasted no more than six months, with the option of a second job of the same duration. CSJs were specifically designed not to substitute for private-sector employment; they paid minimum wage and offered no sick days, holiday compensation, or vacation pay. About one-third of the participants in the original program used a CSJ at some point during the three year program.

¶ 34         CSJs were particularly useful to participants who faced behavioral or legal impediments against employment, such as a failed drug test, a poor credit record, or a criminal record. For these participants, the CSJs offered an opportunity to build a resume and accumulate work experience, which could then lead to better employment opportunities down the line. The operational feasibility of this component depends on the context in which the program is implemented. There are areas in the United States (e.g., migrant farming regions in California) where unemployment is so high and full-time full-year jobs are so scarce that a large proportion of potentially eligible New Hope participants would have to rely on CSJs to accumulate thirty hours of employment every week.⁶¹ Implementation of the program in those areas might be too expensive to be feasible.

¶ 35         Administering the CSJ component required close collaboration between the New Hope agency and local employers. This is one area in which One-Stop workforce agencies⁶² may have a particular advantage as they implement a New Hope program. These agencies tend to have good relationships with local employers and often have experience developing similar work experience programs for their other clients.

¶ 36         A distinguishing characteristic of the New Hope program was its close integration with the existing social service infrastructure in Milwaukee. The program's earnings supplements were designed to complement the existing EITC, which, in Wisconsin, included both a federal and a state component.⁶³ Similarly, the health insurance subsidy was designed around an HMO offer that was similar to that available to Wisconsin welfare recipients at the time.⁶⁴ This made for seamless transfers from one system to the other when welfare recipients enrolled in New Hope and left welfare for full-time work.

¶ 37         In scaling up the New Hope model, we foresee similar levels of integration between the New Hope programs operated by the fifty states and the existing benefits and services available in these states. The availability of such benefits has increased significantly since New Hope was first implemented, not just in Wisconsin, but in most states throughout the United States.⁶⁵ Although many of these state efforts are not fully funded and take-up rates are sometimes low, a well-integrated New Hope program could build on these existing services, increase their take-up by participants, and provide a cafeteria-style benefits package like New Hope at a lower cost than in the mid-1990s implementation of the experimental program.⁶⁶

¶ 38         To accomplish integration of New Hope with existing services and to maximize utilization of existing services and funding streams, it is critical that New Hope be implemented with professionalism and community support. Research on the New Hope intervention and subsequent comparisons with Wisconsin Works (Wisconsin's welfare program, sometimes called W-2) underscored the perceived superiority of New Hope for helping participants.⁶⁷

¶ 39         We propose that the program be administered by existing state and local agencies and not-for-profit institutions. Depending on the specific infrastructure in each state, New Hope services could be delivered through the workforce development system, a health maintenance system, county public assistance agencies, or a network of not-for-profit agencies. In a large state, the program could use a combination of these delivery systems tailored to the specific infrastructure of local areas.

¶ 40         One delivery system that has particular promise as a home for a national New Hope program is the One-Stop system funded with the Workforce Investment Act (WIA).⁶⁸ At the time when New Hope was studied in Wisconsin, such a system was not yet fully developed, but after passage of WIA in 1998, One-Stop centers have become a key feature of the workforce investment system.⁶⁹

¶ 41         With its comprehensive package of benefits and case management, the cost of operating our proposed New Hope program would be far from trivial—a little more than $3300 per participant per year in taxpayer costs.⁷⁰ New Hope's collection of benefits would transfer a considerable amount of resources to participants.⁷¹ First, earnings supplements would average $284 per year per participant, given the earnings supplement take-up rate from the New Hope demonstration.⁷² Second, participants would receive health insurance subsidies on the order of $875 per year, based on actual New Hope costs.⁷³ Third, participants would receive about $470 in child care subsidies, on average; families with young children obviously would reap the bulk of the child care subsidies.

¶ 42         The administrative costs of running a New Hope program are also substantial, amounting to a little more than $1700 per year per participant. These costs include case management, setting up and monitoring CSJs, and general office expenses.

¶ 43         What about the benefits that might counterbalance these costs? A key benefit of New Hope is its boost to the economic productivity of adult participants. Using earnings impacts as a measure of productivity, Table 1 shows that the annual earnings advantage of New Hope participants relative to control groups amounted to $497, which is our best guess of the productivity gains associated with a national New Hope policy. Although the earnings impacts were twice as high for the participants who were not working at the start of New Hope and more than $2500 per year for the subset of participants with some (but not too many) employment barriers, the overall productivity estimate of $497 is most appropriate for our program-wide estimate of costs and benefits.

¶ 44         With adjustments for taxes and transfer payments, the estimated costs and benefits sum to a $2088 annual gain for participants and a $3308 net cost to taxpayers.⁷⁴ However, taxpayers also benefit from New Hope's positive impacts on child achievement and behavior, particularly among boys. These impacts are not easily quantified, but are clearly important, since achievement has been linked to significant earnings gains,⁷⁵ and behavioral improvement that reduces crime can generate very large taxpayer benefits.⁷⁶

¶ 45         Two years after their families enrolled in New Hope, children in New Hope families were judged by their teachers to have considerably higher achievement than children in the control group. The size of the impact—one-quarter of standard deviation—is nearly twice as large as the estimated impact on achievement of an experiment in Tennessee that reduced class sizes, called Tennessee STAR.⁷⁷

¶ 46         Relying on methods for converting achievement gains into the monetary value of the future earnings gains they likely occasion, we estimate an annual value of the future earnings gains to be $1295, $259 of which are paid as taxes and the remaining $1036 utilized by the New Hope participants themselves.⁷⁸

¶ 47         Given the costs of crime for both individuals and the government, the behavior improvement in boys has the potential to generate large benefits. Teachers reported that boys in the New Hope program exhibited more positive behavior, had fewer behavior and disciplinary problems, and were more compliant and less hyperactive in classrooms than control-group boys. Since impacts for girls were significantly less than for boys, we concentrate our benefit calculations on boys.

¶ 48         Almost all of the boys benefiting from New Hope were either black or Hispanic, and all lived in low-income families and high-poverty neighborhoods. Clearly, the boys fit a high-risk profile for school dropout, crime, drug abuse, and adult unemployment.

¶ 49         Based on published estimates of the dollar value to taxpayers of saving a youth from serious crime and drug abuse and from becoming a high school dropout, we calculate the taxpayer value of saving each high-risk New Hope child to be $2.4 million, with crime victim and criminal justice costs accounting for the bulk of this total.⁷⁹ Spread over eighteen years of childhood, the annual value of this sum amounts to about $135,000. We estimate that about one in sixteen New Hope boys would have to transition out of the high-risk category to pay for the entire taxpayer cost of the New Hope program, and fewer than one in thirty-one such transitions are required to pay for the entire societal costs of the program. It does not strike us as implausible that New Hope's large impacts on boys' behavior alone would be more than sufficient to cover all of the taxpayer's costs.

VI.    Conclusion

¶ 50         The New Hope program is a promising policy for helping low-income, full-time workers. Its evaluation proved that it reduced poverty, increased employment, put children in better care situations, and improved their school achievement and behavior. It enabled parents to establish more sustainable family routines and to accomplish their family goals. New Hope's broad vision of helping all low-income adults proved successful in boosting work among single men, at least some of whom were parents with obligations to support their children.

¶ 51         New Hope does not replace the need for government to provide for individuals who are too burdened by physical problems, mental health, or drug abuse problems to hold a job, nor does it eliminate the need for other support programs such as food stamps and health insurance for the unemployed. It does not begin to address larger structural issues such as job discrimination. But it does offer some tools for helping adults find full-time work as well as work supports for families and single adults whose full-time employment leaves them struggling with too little money, inadequate child care, and no health insurance. New Hope provides substance to the idea: "If you work you should not be poor." Its experimental evaluation proves that fulfilling this promise can improve family life and boost children's well-being.

ENDNOTES