Over the past 25 years, policymakers have come to acknowledge the link between lack of child support and the pressing problem of child poverty for a broad range of American families. With over 20 million children under age 18 now living with only one parent or neither parent, there is an urgency to develop more effective methods for obtaining support from noncustodial parents. Much of the public concern about child support has focused on the noncustodial parents (usually fathers) of children receiving welfare, a group for whom earnings and support payments tend to be low. Interest in these families has also been heightened by recent changes in federally funded public assistance, which are gradually leading states to impose various time limits on aid. Since poor families will have to rely even more on nongovernment sources of income in the future, their stake in successful child support enforcement (CSE) has dramatically increased.

The noncustodial parents of children receiving welfare have largely been left out of the reform debate and programmatic initiatives, except as targets of increasing CSE efforts. Unfortunately for poor families, most of the recent CSE reforms have been more effective in increasing collections from noncustodial parents with relatively stable jobs and residence; many of the fathers of children receiving welfare do not fall within this group.

The Parents’ Fair Share (PFS) Demonstration tests a new approach: in exchange for current and future cooperation with the child support system, a partnership of local organizations offered fathers services designed to help them (1) find more stable and better-paying jobs, (2) pay child support on a consistent basis, and (3) assume a fuller and more responsible parental role. Among the key services were peer support (focused on issues of responsible parenting), employment and training services, and an offer of voluntary mediation between the custodial and noncustodial parents. During the period in which parents participated in PFS services, the child support system gave them some "breathing room" and an incentive to invest in themselves by temporarily lowering their current obligation to pay support. CSE staff also closely monitored the status of PFS cases. When a parent found employment, CSE staff were to act quickly to raise the support order to an appropriate level (based on the state’s child support payment guidelines), and if a parent ceased to cooperate with PFS program requirements, CSE staff were to act quickly to enforce the pre-PFS child support obligation. The demonstration is a test of the feasibility of implementing this new "bargain" and its effects on parents, children, and the child support system.

PFS rests on an unusual partnership of funders and program operators, including federal agencies, private foundations, states, localities, and nonprofit community-based organizations. Organized by the Manpower Demonstration Research Corporation, it began in 1992 with a pilot phase to refine the program model and test the feasibility of implementing it at the local level and, despite a variety of implementation challenges, moved into a seven-site demonstration phase in 1994.¹

This report presents findings from the demonstration-phase implementation of the program, characteristics of the parents in the sample, and early impacts on two outcomes of interest (fathers’ earnings and child support payments). These impact findings are only the first chapter in the PFS story because they rely solely on administrative records, cover only a part of the full PFS impact sample, provide only six quarters of follow-up, and do not cover several key goals of the program (for example, helping fathers become more effective and responsible parents). Further, an examination of PFS’s effects on direct payments of support to the custodial parents and underground employment (which are not captured by administrative records) must await analysis of the follow-up surveys of parents. Nevertheless, available information (based on a shorter follow-up for the full sample) suggests that the findings presented here are likely to be similar to those for the full sample and provide a look at several key effects of the program.

Findings in Brief

• Implementing PFS presented management challenges that went far beyond identifying agencies with experience in providing the program’s services and seeking funding to support this effort. At a minimum, the local partnership needed to include the CSE agency and the courts, employment and training service providers, and organizations with the capacity to provide peer support and mediation. At the core of the challenge, the intended partners began with different organizational missions and assumptions about their "clients," funding sources, administrative procedures, standards for rating their performance, and experience dealing with those facing a legal mandate to participate (as opposed to volunteers).

• Some PFS services were easier to put in place than others. In general, peer support, job club, extra case review at CSE offices to identify parents for PFS intake, and the offer of voluntary mediation were implemented across most sites. Implementation of "skill-building" education and training options and a quick follow-up when parents found employment or failed to comply with program requirements were more difficult to sustain over time. Further, because of difficulties in identifying potential PFS referrals from the child support caseload and getting them to appear for review hearings, five of the seven sites did not meet their enrollment targets, and, at times, program operations were hampered by this shortfall.

• The majority of the noncustodial parents referred to PFS were living in poverty, or on the edge of poverty, with a recent history of moving from one low-wage job to another. Thus, the challenge was to help these fathers find better jobs than they would otherwise have found or to secure more stable employment. This report is primarily based on a sample of 2,641 parents who were found to be eligible and appropriate for referral to PFS. Many faced substantial barriers to moving into better jobs in the mainstream labor market: nearly 50 percent lacked a high school diploma, and about 70 percent had been arrested for an offense unrelated to child support.

• Slightly more than two-thirds of the noncustodial parents referred to PFS participated in at least one PFS activity. The average participant was active for five months, with about one-half participating for one to three months and about one-quarter continuing to participate for four to six months. Participation was greatest in peer support and job search workshops. Virtually all those who failed to participate and did not have a long-term "excuse" recognized by the program were referred back to the child support agency for further enforcement.

• Parents subject to the extra outreach and case review involved in PFS intake, prior to any referral to the PFS program, made more payments to the child support agency than those subject to traditional child support enforcement. Among other effects, the extra outreach and case review uncovered previously unreported employment, allowing the child support agency to institute wage withholding. In three sites where a special study of the extra review was conducted, the increase in the proportion of parents paying any child support ranged from 6 to 15 percentage points, and average total child support payments per parent subject to the extra review increased by $160 to $200 over the six quarters of follow-up.

• Separate from the effects of this extra outreach effort, a larger number of parents referred to the PFS services and mandates paid child support than would have paid in the absence of access to the program. Across all seven sites combined, the number of parents who paid support during the follow-up quarters increased by about 4.5 to 7.5 percentage points. However, these impacts on child support were mainly the results of substantial impacts in three of the seven sites. In two of these three sites, the average amount of child support paid per parent over the 18 months of follow-up also increased by a statistically significant amount.

• Unfortunately, these increases in child support came without a corresponding increase in fathers’ employment and earnings. No site produced increases in employment and earnings that were consistent and statistically significant during the 18 months of follow-up for this report.

In sum, PFS did lead to an increase in child support for a group sometimes viewed as unlikely to respond to enforcement efforts, but the search for effective means of increasing employment and earnings of low-income men continues. The final section of this summary offers suggestions for program designers and operators based on these PFS findings.

The Policy Context for PFS

PFS emerged out of three interrelated trends and a very concrete dilemma facing courts and child support administrators. PFS had its origins in welfare reform efforts that gradually shifted the balance of responsibility for supporting poor children away from the public sector and toward parents. One goal of a series of reforms, culminating in the passage of the Personal Responsibility and Work Opportunity Reconciliation Act in 1996, was to help custodial parents to increase their earnings and to create and enforce legal obligations of support from more noncustodial parents, so that poor children would be supported by both parents.

Efforts to improve CSE were a second factor contributing to the development of PFS. Since the mid 1970s, federal lawmakers have imposed requirements on states to strengthen CSE, with the goals of helping low-income families stay off welfare and of recouping aid payments made. These efforts have largely focused on noncustodial parents with known income and assets and have been most effective for these cases. Although the CSE system has continued to evolve since the beginning of PFS, public assistance-related cases remain a frustration to CSE agencies and the courts in many jurisdictions at a time when reforms in welfare (especially time limits on receipt of aid) increase the stake that low-income families have in receiving support.

The deteriorating labor market situation of less-educated men also contributed to the emergence of PFS. Over the past 25 years, the inflation-adjusted earnings of men without a high school diploma have dropped substantially. With much of the focus of program development on the custodial parents of poor children, men on the fringes of the labor market — and especially younger men of color — have rarely been the target of employment program outreach. Further, few successful strategies have been developed for increasing their employment and earnings.

For child support administrators and the courts, the factors discussed above created a serious ongoing problem. When a noncustodial parent with little work history claimed he was unable to pay his child support because of unemployment, it was frequently difficult to determine the truth of his claim. In practice, courts and agency staff were left with two unsatisfactory options: threatening jail in an effort to coerce payment or sending the parent out on his own to look for work. While the first option was appropriate for those able but unwilling to pay, neither option was appropriate for those who were unable to support their children. Further, the agencies and courts often struggled to distinguish the unwilling from the unable.

The PFS Demonstration: Testing a New Option

The PFS Demonstration is a test of a third option: referring a specified group of noncustodial parents to a program of employment and other services where participation is mandatory and would be carefully monitored. Parents eligible for PFS (1) were not living with their children who were receiving or had received AFDC; (2) were behind in their child support payments; and (3) had no reported employment or were underemployed or working in a low-paying temporary job. If a parent who was working had not reported the fact to the child support agency, the participation mandate would uncover it because he could not work and participate in program services simultaneously. For those without employment and assets, PFS provided a way to couple enhanced opportunity with a steady message of parental responsibility.

The demonstration’s three goals presented special challenges:

• Increasing the employment and earnings of low-income noncustodial parents of children receiving welfare: PFS faced a different challenge than programs serving custodial parents (usually women) receiving welfare, many of whom had little formal work history. Impacts in these programs were often achieved by getting more women into jobs or getting women who would have worked into jobs faster. In contrast, the great majority of PFS fathers had worked (though usually spottily and in low-wage jobs). Increases in the proportion working at all would be harder to achieve, so the program’s goals included increased job retention and wage levels, as well as higher overall employment rates.

• Increasing child support payments: Many other studies have shown that the frequency and amount of child support payments are related to noncustodial parents’ income; hence the goal of increased earnings is linked to the goal of greater child support. However, fathers’ attitudes toward their parental responsibilities, the custodial parent, and the child support system (which, under most states’ rules, does not pass payments on to families receiving welfare) also influence the payment of support. PFS sought to affect all of these things. It was also implemented as the CSE system was gradually evolving with the development of new methods to track employment and earnings and changes in rules on adjustments of orders, so that the "enhancements" to child support involved in PFS came on top of a changing base of standard enforcement.

• Supporting and improving parenting behavior: Noncustodial parents can help their children in a variety of ways beyond financial support, and PFS sought to help them become more involved as responsible parents, a personal goal of many of the fathers. But lack of money and at times contentious relations with the custodial parent had hampered many fathers’ efforts to play this role. Supporting the importance of the effort was other research indicating that increased parental involvement may also contribute to greater payment of support, suggesting that the goals of the demonstration are interrelated.

The PFS intake process was an important part of the demonstration. In most cases, noncustodial parents were referred to PFS during court hearings or appointments scheduled by CSE staff in response to the parents’ failure to make court-ordered support payments. Several of the sites put in place new procedures to identify parents who appeared to be eligible for PFS (whose child support cases would typically have low enforcement priority) and scheduled special hearings or appointments to review their reasons for nonpayment. Parents who cited unemployment as the reason for their nonsupport were ordered to attend PFS activities until they found a job and began paying support. In some sites, parents just establishing paternity were also referred to PFS when they had no means to meet child support obligations.

Program services were built around four core components, listed in Table 1. In general, parents began their participation with peer support, which was structured around the Responsible Fatherhood curriculum and run by a trained facilitator. The peer support sessions (which typically met a minimum of two to three times per week for a set number of weeks) covered a wide range of topics including self-evaluation, parental roles and responsibilities, relationships, managing anger and communications, problems on the job, coping with racism, and life skills. Some sites offered other services concurrently with peer support, to allow fathers to participate in employment-related activities because of their need for income. (Most did not receive public assistance, despite their low income levels.) In general, services that focused on helping parents quickly find work (employment-readiness workshops, job search assistance, and job clubs) emerged as the most common employment-related services, though sites also offered basic education, classroom occupational training, and on-the-job training as options.

Enhanced child support enforcement consisted of reductions in existing orders (often originally set when the parent had been employed) during PFS participation, close monitoring of parents’ program and job-seeking activities, and immediate modification of support orders when they found employment or failed to comply with PFS requirements. These changes were consistent with efforts under way at the time to link a parent’s child support obligation with his current income; the PFS sites were asked to pursue this objective vigorously. In the first year of the demonstration, case management also emerged as an important part of PFS, and staff in this role identified barriers to employment, developed service plans to address them, tracked participation in PFS services, and informed the child support agency of noncomplying or employed parents. Parents referred back to the agency were then typically subject to traditional CSE.

Formal mediation services were also offered by PFS or outside agencies, and many staff members also served as informed mediators between parents.

Implementation of PFS relied on two important partnerships: a funding partnership and a site operations partnership. Table 2 lists the nonsite funding partners for the demonstration. PFS was authorized by a provision in the Family Support Act of 1988 that permitted use of federal welfare-employment funds (normally restricted to custodial parents) to fund a demonstration of services for the unemployed noncustodial parents of children receiving welfare. In order to access this special federal funding and expanded child support funding, demonstration sites had to provide a state match of cash or in-kind support. Other government and nongovernment organizations also supported the demonstration. Table 3 shows the lead state agency in each site (which took the lead in organizing the original demonstration proposal), the lead local agency (which coordinated program operations), and the agency that housed crucial core services such as peer support and case management (designated the "program home" in the table). Because PFS required diverse kinds of expertise, sites developed local operating partnerships that included child support agencies, the local Job Training Partnership Act (JTPA) agency, and a mixture of community-based organizations.

This report focuses on (1) the nature of PFS services put in place in the sites and the implementation challenges and lessons that emerged from the demonstration, (2) the PFS intake experience and the characteristics of the parents participating in the demonstration, and (3) the early impact of PFS intake and access to services on employment, earnings, and child support payments as measured through administrative records.

The accompanying reports from the demonstration focus on lessons from the PFS intake process for CSE (Working with Low-Income Cases: Lessons for the Child Support Enforcement System from Parents’ Fair Share) and on the lives and attitudes of a sample of parents participating in the program. Future research will continue the impact analysis in this report with longer follow-up using administrative records and will supplement it with survey data to broaden understanding of the initial impact topics (especially effects on underground employment and support provided directly to the custodial parent or child) and will allow analysis of impacts on family relationships.

The remaining sections of this summary present findings on the challenges involved in the PFS partnerships; PFS intake; the characteristics of parents referred to the program; the implementation of PFS services and mandates; participation in PFS services; the early impacts of PFS on child support payments, employment, and earnings; and suggestions for program design.

Key Challenges in Developing and Sustaining the PFS Partnership

PFS called for a shift from a focus on the short term (when noncustodial parents’ child support payments and potential to pay were minimal) to a longer-term perspective of investing in building the capacity of poor, unemployed noncustodial fathers to assume parental roles. This fundamental change in perspective presented a number of implementation challenges that went far beyond assembling the state and local partnerships called for under the demonstration.

• To successfully implement PFS, the local partners had to change their standard operating procedures in ways that often conflicted with pre-existing agency priorities and were therefore difficult to sustain.

At the core of PFS was a commitment by local child support agencies to focus enforcement attention on cases that they typically treated as low priority: low-income, unemployed fathers. Traditionally, enforcement efforts devoted to these cases were seen as unlikely to yield much in support collections, the primary goal of child support agencies. The demonstration also called on child support agencies to broaden their service mission, which typically was focused on serving the financial interests of custodial parents, children, and taxpayers, to include aiding noncustodial parents who were unable to meet their obligations.

PFS called on employment and training agencies funded under JTPA to work with very disadvantaged men who were ordered to participate by the courts (or, in limited cases, the child support agencies). Most of JTPA’s prior participants had been either volunteers who wanted to devote time to building their skills or mandatory referrals from public assistance programs who had an income source during participation in the program. PFS referrals may have been interested in building their skills, but their lack of income created great pressures for them to find a job quickly. Further, the PFS program model called on JTPA agencies to provide many parents with on-the-job training placements, in which participants were placed in a wage-paying job and received training in an occupational skill while the employer received a wage subsidy to cover the training costs. Unfortunately for PFS, the JTPA system sharply curtailed its offering of on-the-job training just as PFS got under way, and the program was continually frustrated in its efforts to expand this program component. Finally, many PFS participants were difficult to place in jobs because of weak work histories, poor education, criminal records, or drug or alcohol problems, and JTPA agencies — which saw the local business community as a continuing customer — were hesitant to push these fathers on employers for fear of spoiling a long-term relationship vital for their continued success.

Many community-based service organizations involved in PFS had a strong organizational commitment to serving clients like the parents in PFS, but had little experience and some concerns about partnering with the child support agency. The PFS parents often saw the CSE system as stacked against them, with legal powers that made it feel like part of the criminal justice system and a mission to serve the taxpayers (by recouping welfare payments) rather than increasing parental support of poor children (who under federal welfare rules received little of the support payment). Reflecting these concerns, the community-based service agencies often were uneasy about monitoring and enforcing legal obligations and referring noncooperating fathers back to the child support agency for further enforcement actions, including the possibility of jail. Despite this, over the course of the demonstration, community-based organizations assumed increasing responsibility for PFS services in several sites.

All of these roles represented major shifts for the participating agencies. In many sites, tensions between PFS procedures and normal practices and between agencies emerged during the implementation of the program, at times causing lack of coordination in services. Thus, operating the program required a level of sustained attention from program managers as well as a commitment to interagency cooperation that could motivate agencies to work through the variety of issues which emerged.

Identifying and Referring Eligible Parents to PFS

• Although staff in the seven participating sites identified over 5,500 parents who were eligible and appropriate for PFS services, this was substantially less than original targets, and the shortfall had programmatic implications.

At the start of the demonstration phase, sites and MDRC staff developed estimates of demonstration-phase samples, based on the available information about the number and status of welfare-related child support cases with arrearages. Only two of the seven sites (Los Angeles and Grand Rapids) were able to meet or closely approach their enrollment targets.

Meeting enrollment targets was more difficult than it had been in the pilot phase, for at least three reasons. First, targets were set higher because the demonstration relied on a random assignment design to estimate program impacts. This involved randomly assigning one-half of the recruited sample to a control group, which served as a benchmark representing the experiences and behavior of parents without the PFS option. Consequently, sites had to double their efforts to get an equal number of enrollees into program services. Second, the economy improved over the course of the demonstration, so that a higher percentage of those who appeared for a review were employed and, thus, ineligible for PFS. Third, it may be that during the pilot phase some sites had "cleaned out" from the caseload the most easily located PFS-eligible group. Despite these pressures on the intake process, the basis characteristics of the sample, discussed in more detail below, showed little measured change from the pilot.

The problem of lower-than-expected enrollment in PFS services did affect PFS negatively, by altering the services provided, making it more difficult for sites to maintain steady funding streams (since much funding was linked to sample buildup), and drawing management attention away from other implementation issues. Peer support and job club were probably most affected programmatically, since both were designed to serve groups of at least five participants. Some sites allowed parents to join ongoing groups, which lessened their cohesiveness, while others operated groups of fewer than five participants. The focus on the problem of enrollment prevented the kind of forward-thinking management style that could have helped sites meet the variety of other implementation challenges PFS posed.

• Most sites instituted extra outreach and case review practices as part of the demonstration, and this effort produced important information about the status of welfare-related child support cases.

In an effort to meet sample targets, sites used a variety of methods to identify potential PFS referrals. Two sites relied primarily on reviewing the regular court dockets for child support cases likely to meet PFS eligibility rules. However, the remaining sites all employed extra outreach efforts to increase the flow of PFS referrals, including conducting reviews of child support cases on the existing caseload, reviewing other lists (such as new referrals of cases from the welfare agency, listings of noncustodial parents about to exhaust unemployment insurance benefits, and Medicaid-supported births in local hospitals), and streamlining the hearing process to review the status of large numbers of noncustodial parents. These efforts appear to have made the greatest difference in Los Angeles, Dayton, and Grand Rapids.

The process of extra outreach to review child support cases produced new information for many of the PFS sites. Some parents could not be located; others contacted the child support agency or the courts and provided information on previously unreported employment. Still others provided information justifying a change in their current support obligation: they were either living with the children for whom support was owed, in ill-health or disabled and unable to work, or incarcerated and unable to work for pay and provide support. A later section of this summary reviews the impacts of the extra outreach and case review on noncustodial parents’ employment and child support payments.

Characteristics of Parents Referred to PFS

• Although the fathers were diverse in terms of race, age, and living arrangements, overall they were a disadvantaged group, and the majority lived in poverty, or on the edge of poverty, with little access to public assistance.

The parents found appropriate for PFS were overwhelmingly male (98 percent) with an average age of 30. About 80 percent of the overall sample were black or Hispanic, but there was great variation among the sites in the racial/ethnic distribution of parents. In Memphis and Trenton, for example, parents were largely black, while in Los Angeles and Springfield they were largely Hispanic. Slightly over 60 percent of the overall sample had never been married, about 50 percent had lived with their own father when they were age 14, and nearly 70 percent had been arrested on a charge unrelated to child support since age 16. Nearly 50 percent of the sample had no educational credential, and about 80 percent had not participated in any education or training program in the year prior to being referred to PFS.

Most were unemployed at referral to PFS (though 17 percent were admitted into the program because of underemployment or very unstable employment), and their recent employment history revealed a tenuous connection to mainstream jobs and very low recent wages. Administrative records on earnings show that 43 percent of the sample had earned a total of $500 or less in the three quarters prior to their entry into the PFS sample, and only 28 percent had earnings exceeding $3,500 in this nine-month period. Despite these low earnings, only 29 percent of the fathers lived in a household that received Food Stamps, and few were receiving AFDC or other cash assistance. This low level of earnings was reflected in low rates of child support payments; slightly more than 20 percent had paid any child support through the child support agency in the quarter prior to their referral to PFS.

• Within the PFS sample there were fathers for whom finding and keeping a job would be an important advance and others for whom the goal was better-paying and more stable employment.

Within the sample, there was a group with continued connection to the job market, albeit a series of relatively short-term and low-paying jobs. For these fathers, the challenge was to help them build their skills to command better pay or expand their access to the kinds of jobs they would otherwise not obtain. But there was also a group for whom any stable job would be an improvement. They tended to be very disadvantaged fathers, who faced serious barriers to employment.

• Many noncustodial parents initially expressed skepticism about the goals and services of PFS, based on their perception that the child support system was "stacked against them," which program staff had to overcome to engage the parents in the program.

Parents referred to the program often reported that their prior experience with the child support system left them feeling that it was fundamentally unfair. Concerns raised included a sense of inequity that (1) payments made by noncustodial parents were largely used to reimburse taxpayers for public assistance expenses rather than passed on to their children; (2) custodial parents had the option of public assistance, when noncustodial parents could be living in similar poverty but face an obligation to pay support; and (3) enforcement was often erratic, with the system interested in them only when they had a job and child support could be withheld from their wages. With their negative histories with the child support system, these parents were at best cautiously interested in the opportunity PFS offered. At the same time, given the eligibility criteria for referral to the program, most of the parents could use the help of an effective program.

Implementation of PFS Services and Mandates

• Peer support was the most consistently well-run component during the demonstration and generally was viewed as the central PFS activity, providing a focal point for participants.

Peer support was more straightforward to operate than employment and training and enhanced child support, which required substantial cooperation across agencies. A single agency typically operated the component, which was the initial activity for most participants after a PFS orientation. Special skills were needed to facilitate the peer support sessions effectively, and the job was intense; to be effective, staff had to make a serious commitment to the fathers. Facilitators usually had prior experience in a similar role, attended training on the PFS peer support curriculum and facilitation techniques, and followed the curriculum fairly closely. It covered 18 topics, and groups generally met a minimum of two or three times per week for a set number of weeks to cover all the topics. At times, the shortfall in sample buildup led some sites to move new PFS referrals immediately into peer support (an open-entry policy) rather than wait for sufficient numbers to begin a new group, and some groups became much smaller than intended.

Peer support was generally well received by the noncustodial parents, providing them an opportunity to relate to a peer group in constructive ways, discuss troubling personal problems, develop new problem-solving skills, and have access to an advocate (the facilitator) who believed in their potential. Two sites stand out for their approaches to peer support. In Dayton, facilitators developed creative new ways to encourage parents to become involved with their children — for example, having parents do activities with their children and report back to the group and holding special events involving participants’ families. In both Dayton and Jacksonville, staff made an effort to develop specialized peer support groups; most notably, Jacksonville instituted a mental health-oriented group, facilitated by a psychologist, for parents who could benefit from a more therapeutic orientation.

• Most sites relied heavily on job search workshops and job club, running these activities and peer support simultaneously because of parents’ strong desire to find work quickly.

The design of PFS assumed that for the program to have a substantial impact on parents’ employment and earnings, sites would have to offer an array of short-term skills training and on-the-job training to help participants obtain higher-paying or longer-lasting jobs and job clubs that would help people find employment. In practice, there was a conflict between the program’s interest in encouraging noncustodial parents to take the time to invest in skill-building activities and the realization that they could not afford to be out of the labor market for an extended period. In most sites, these pressures led to an emphasis on getting parents into jobs quickly.

Basic job search assistance was usually provided through one-to-two-week job search workshops (which taught skills such as résumé writing, interviewing, and search techniques), as well as continuing job club activities that provided participants with ongoing support from PFS staff and their peers as they searched for work by identifying and following up on leads. Implementation of these group job search workshops went generally as planned in most sites, and continuing follow-up of individual job search efforts improved over the course of the demonstration.

• Two sites that emphasized the goal of getting participants better jobs than they could find on their own made job developers an integral part of their program.

Job developers supplement participants’ efforts to find work by identifying and pursuing job leads on their behalf. Springfield and Grand Rapids relied more on job developers than other sites, with these staff members often involved also in developing on-the-job training opportunities. A fundamental constraint on the broader use of job developers was apparent: they often faced conflicting incentives because they valued maintaining good relationships with employers and saw the disadvantaged PFS participants as "risky" to refer for openings.

• Skill-building services, particularly classroom training and on-the-job training, proved to be the PFS activity most difficult to implement.

From the beginning of the demonstration, institutional barriers and differences in practice and procedure made it difficult to provide the full menu of employment services intended for each site. PFS developed arrangements for program slots with local employment and training agencies, which were funded by PFS site grants, JTPA, or some combination of both.

The most flexibility occurred when PFS funded slots for program participants, allowing agencies to develop new program models for these quite disadvantaged men outside the procedures of existing programs. When funds were provided by JTPA, the service providers had to balance many competing priorities. They faced new federal mandates to target services to those with serious and multiple barriers to employment, as was common among this group of disadvantaged men. But at the same time, the agencies expressed concerns that the low employability of the PFS parents could put relations with employers at risk, the pressures they faced to serve many different groups prevented them from committing a specific number of slots to PFS participants, and changes in JTPA on-the-job training rules made this service less attractive to employers. Because of these difficulties, over the course of the demonstration several sites shifted away from JTPA agencies as lead employment service providers to community-based organizations.

• Three sites (Los Angeles, Grand Rapids, and Springfield) were most successful in putting on-the-job training and classroom training in place, and these sites shared some characteristics.

Active leadership that focused on increasing the number of skill-building activities was important in these three sites. In Los Angeles, the state Employment Development Department played an important part in successfully integrating the local JTPA program and its service providers into PFS, and this site generated higher-than-average participation in classroom training. Beyond this leadership commitment, the attitudes of program staff were also an important factor affecting service offerings. Across the PFS sites, staff varied considerably in their assumptions about the employability of PFS participants, which appeared to affect parents’ enrollment in services and the willingness of job developers to market them to employers. Both Springfield and Grand Rapids contracted with agencies experienced in serving very disadvantaged populations with severe barriers to employment; and probably because the staff of these organizations — in contrast to staff in some other PFS sites — did not see the barriers as insurmountable, these two sites produced the highest number of on-the-job training placements.

• Although formal mediation services were offered in every site, most programs did not aggressively market them. There was not a strong interest among either noncustodial or custodial parents in utilizing formal mediation, and many fathers preferred to rely on the informal efforts of known program staff.

Background research during the planning phase of PFS suggested that there were many issues between the custodial and noncustodial parents which could potentially create barriers to fathers’ playing a fuller parental role and providing greater financial and other support to their children. Mediation was included as an optional service within PFS to support efforts of parents to address and resolve these issues. Each site’s partnership included trained family mediators, and program staff informed fathers of the availability of this service. Formal mediation remained little utilized in most sites throughout the demonstration. Program staff reported that few fathers were seriously interested in pursuing this option and, even when they were interested, the mothers might not be. Further, PFS staff generally were more focused on other aspects of the program and did not place as high a priority on encouraging participation in this component as on the mandatory services such as peer support and employment services. Furthermore, many of the PFS staff (who were often trained mediators) performed an informal mediation function, working with the parents to try to resolve issues. Experience suggests that fathers may have been more willing to take problems to a person they already knew, and these staff members were able to serve as an informal go-between for the parents.

• All sites made the PFS participation mandate real by referring parents who did not comply back to the CSE agency for traditional enforcement measures, but sites used this option in different ways.

Each PFS site developed procedures to track parents’ participation in program services (often relying on the management information system developed for the demonstration), and when parents failed to meet program requirements, staff followed up to determine if there was an acceptable reason for nonparticipation. If none existed or if the enrollees did not respond to their communications, sites referred them back to the child support agency for traditional enforcement measures, which usually amounted to a notice to appear at a hearing on the status of the case. Sites were consistent in taking this action (though some were quicker than others to do so), and virtually all (92 percent) parents who never participated in any PFS activity and lacked a recognized long-term excuse were referred back to the agency for traditional enforcement. For some sites, this referral reflected the end of efforts to encourage participation, in effect a statement that the parents were no longer in PFS. For other sites, referral back to the CSE agency was one of many tools used to secure compliance with program requirements, and fathers could and did return to participate in PFS requirements. Those referred back typically faced traditional enforcement actions such as a hearing before a court or referee.

• Sites in which the child support agency played a leading role in PFS showed flexibility in developing new approaches to monitoring the status of cases and encouraging participation in program services.

Because of the differing perspectives of the local agencies involved in PFS, agencies could choose to focus on their part of the program and not seriously engage in the difficult task of coordinating activities. However, in sites in which the child support agency played a leading role, staff were well positioned to work as a problem-solving team, with the child support agency driving the effort. The child support agencies had the most direct financial interest among the partners in developing an effective means to monitor and enforce participation requirements and increase support payments. In Los Angeles and Grand Rapids, staff worked together to hold regular "case conference" meetings to discuss participation problems of particular parents and develop plans on how to respond. Further, in these two agencies (plus two others as well) specific child support workers were assigned to handle all PFS cases, providing staff members who understood the program and the importance of an immediate response when parents found work or failed to comply with program requirements.

Participation in PFS Services

The follow-up for this report extends for 18 months after a parent was referred to PFS by a court or child support agency.

• On average, 70 percent of those parents referred to PFS participated in at least one PFS activity. However, there was substantial variation by site related to differences in intake methods, service offerings, and the way in which referral back to traditional CSE was used.

Participation rates ranged from 82 percent in Los Angeles to under 60 percent in Dayton and Memphis. Rates appear to be higher when (1) the intake process produced parents who were motivated to participate in the program (especially important in Los Angeles); (2) labor market opportunities for those referred to PFS were weaker (either because of higher unemployment or because participants had many barriers to employment); (3) PFS activities started immediately after a parent was referred to the program; (4) the PFS staff members who monitored and responded to participation problems were colocated with those who conducted peer support, the initial component; and (5) referral back to child support was used as a technique to encourage participation rather than indicating an end to PFS involvement. Participation did not differ substantially among subgroups defined by recent earnings, prior arrest status, ethnicity, or age. The low levels of participation in employment and other services just prior to referral to PFS and the absence of peer support outside PFS suggest that PFS participation rates represent a substantial increase over the level of service that would have occurred in the absence of the program.

• Participation was highest in peer support and job club, though four sites did engage a substantial percentage of parents in skill-building activities.

As is common in many multicomponent programs, participation was highest in the activities at the beginning of a service sequence. Sixty-four percent of parents referred to PFS participated in peer support, and 57 percent were active in a job club or workshop. Peer support was the most commonly used activity in all seven sites combined, while job club was the second most common in six of the seven sites. Nevertheless, clear differences in service emphasis emerged. Overall, only 8 percent of the parents participated in classroom occupational skills training, 12 percent in basic education, and 12 percent in on-the-job training; but in Grand Rapids and Springfield, approximately 25 percent participated in on-the-job training; in Memphis almost 50 percent participated in basic education; and in Los Angeles about 29 percent participated in classroom occupational skills training. Again, there were no substantial differences in participation among the key subgroups listed above.

• Participation in PFS was shorter and less intense than originally intended.

The shift from an emphasis on skill-building activities to immediate job placement resulted in a decline in the expected average length of program participation. Services such as on-the-job training and classroom occupation skills training were expected to last several months, while job-readiness workshops and job clubs were relatively short term. Parents who participated in PFS were active in some service for an average of approximately five months of the 18-month follow-up period, though about one-quarter were active for more than six months during follow-up. Parents attended an average of nine sessions during the months they participated in program services. Site variation on measures of length and intensity of participation were small, but parents in Los Angeles participated for more months (7 months) than parents in other sites, and a higher proportion of them were active in 12 months or more (17 percent).

Early Impacts of PFS on Child Support, Employment, and Earnings

Samples and Approach Used in Calculating Impacts

Two samples were used in the impact analysis presented in this report. The first sample consisted of 6,884 parents randomly drawn from the child support caseload in three sites (Dayton, Grand Rapids, and Memphis) because they appeared to fit within the PFS eligibility rules (linked to children receiving welfare, not currently paying support, and having no known employment). Two-thirds of these cases were randomly assigned to a group subject to an extra outreach and case review process to determine their status and appropriateness for PFS. The remaining one-third of these cases were subject to the standard CSE in the site. This process allowed for a comparison of the effects of the extra outreach and case review involved in PFS intake.

The second sample consisted of noncustodial parents who appeared at a hearing or other review of their child support status and met the eligibility rules for PFS. Half of these parents were randomly assigned to be referred to PFS services and subject to the program’s mandates (labeled the "program group"), and the remaining half were subject to the standard enforcement procedures (labeled the "control group"). In many cases, control group members faced an order to seek work on their own and report employment to the support agency, though in some sites these parents faced the prospect of an order to pay a specified amount of support immediately to avoid serving time in jail. The effects of referral to PFS versus the traditional enforcement faced by the control group can be estimated by comparing the outcomes over time for both groups. The sample analyzed for this report consists of parents randomly assigned through June 1995 for whom six quarters of follow-up data are available; they constitute about half of all parents randomly assigned during the demonstration.

For both samples, employment, earnings, and child support payments were analyzed for each individual in the quarter during which that person was randomly assigned (labeled the "quarter of random assignment") and for six additional quarters of follow-up. Impacts findings are reported by quarter relative to the point of random assignment.

The administrative data used in this analysis come from CSE agencies in the PFS sites and from employer reports of earnings to state unemployment insurance systems. Each data source misses some aspects of parental behavior: the child support data do not include payments made directly by the noncustodial parent to the custodial parent and child, and the unemployment insurance system does not cover about 3 percent of wage and salary workers whose jobs are not included and those doing odd jobs or working in the underground economy. Until survey findings are available later in the PFS research, it is not possible to determine whether the observed impacts on child support payments represent a real increase or a shift of activity to a type covered by the administrative data. For example, increased payments of child support to the CSE agency could represent an increase in total support paid or a shift from informal payments to the custodial parent or child to formal payments to the CSE agency. Later research in PFS will explore these issues.

Impacts on Child Support Payments

• The PFS intake process produced a significant increase in child support payments to the CSE agency even before any referral to PFS services and coverage by its mandates.

Two kinds of information support this conclusion. First, as mentioned above, in three sites (Dayton, Grand Rapids, and Memphis) a special study of PFS intake isolated the effect of the extra outreach and case review involved in PFS intake prior to any referral to PFS. For the three sites combined, the PFS intake process produced statistically significant increases in both the percentage paying support to the child support agency and the average total child support payment, as shown in Table 4. These increases occurred in the quarter in which the PFS intake process began and in each of the six quarters of follow-up. For example, in quarter 0 (when random assignment occurred) 21.8 percent of those subject to extra outreach and case review paid some child support compared with 18.0 percent of standard group members, for an impact of 3.8 percentage points. Similarly, in the quarter of random assignment child support payments averaged $98 per parent subject to extra outreach and case review compared with $83 per standard group member. (In this table and subsequent tables on impacts, average figures for child support payments per parent include zero payments for parents who did not pay any support. The average amount for parents who pay is substantially higher than these average payments per parent.) In part, the increase in child support payments occurred because the extra outreach and case review led parents to inform the child support agency of previously unreported employment.

Table 5, each of the three sites experienced a statistically significant increase in the percentage of parents paying child support (present in every post-random assignment quarter) and in the average total amount of child support paid in quarters 1 through 6 of follow-up. For example, in Dayton in quarter 2 of follow-up, 39.6 percent of the extra outreach and case review group paid some child support compared with 31.1 percent of the standard group, for an impact of 8.5 percentage points. Similarly, Dayton parents in the extra outreach and case review group paid an average of $1,506 in child support over the follow-up period, while those in the standard group paid an average of   $1,307, for an impact of $200.

The behavior of the control group in the analysis of the effects of referral to PFS, conducted in all seven demonstration sites, also supports the impact findings. In the quarters prior to that in which they appeared for a hearing and were randomly assigned, the percentage of the control group paying any child support and the average amounts they paid were low and relatively stable. These parents were not in a temporary downturn in child support, but rather had settled into a pattern of low support that had lasted for several quarters.

In the quarter of random assignment, however, for the full sample and six of the seven sites, there was a noticeable increase for the control group in the percentage paying support and average payments. The previous stability of the payment rate and average payment suggests that this abrupt increase in support was not a gradual return to a longer-term pattern of higher support. Instead, the evidence is consistent with a view that the PFS intake process — in which sites reviewed the status of cases they would otherwise not have been intensively worked — in itself produced an increase in child support payment rates and average amount. Therefore, the increases in payment of child support due to referral to PFS reported below are separate from the increase produced by the PFS intake process.

• Separate from the impact of the PFS intake process, a higher percentage of parents referred to PFS paid child support than was the case for parents in the control group. However, there was no statistically significant increase in the average amount of the support payment.

Table 6 shows the child support impacts of referral to PFS, based on administrative records from the child support agency, separate from the impact of the PFS intake process. This analysis of the effect of services focuses on the impact of referral to PFS services and coverage of the mandates rather than the impact of participation in PFS services. Because PFS was a combination of service offerings and a mandate to participate or pay support, it is not possible to isolate the separate effects of participation in services. Parents who did not participate in services might have nevertheless been affected by the PFS mandate and changed their behavior.

In Table 6, in the quarter of random assignment (when many parents in the program group were participating in PFS), a higher percentage of the control group paid support. By the second quarter of follow-up, a higher percentage of the parents in the program group were paying support, and this impact persisted (and remained statistically significant) throughout the remainder of the six quarters of follow-up. Despite there being an impact on payment rates, the overall rate of payment remained low, never reaching 50 percent of the program group in any quarter.

Referral to PFS did not produce a statistically significant increase in average payments, however. (The average payments presented in the table include zero payments for those parents not paying any support, in most quarters somewhat more than half the group.) During the quarter of random assignment, the control group average payment was somewhat higher, and in succeeding quarters the program group average was usually slightly higher, but the difference was not statistically significant. In quarter 5 (months 13 to 15 of follow-up), for example, parents in the program group averaged $269 in child support, while those in the control group averaged $241.

This pattern of impacts on percentage paying but not on average payments could occur for several reasons. First, as part of the PFS model the parents in the PFS program group, at least initially, faced a lower child support order than parents in the control group. The lowering of child support orders with the referral to PFS was intended to be temporary, but it might have persisted long enough to affect average payments for the program group throughout the follow-up period. The second likely reason for this pattern of impacts has to do with the variability in a measure such as amount paid (which could range from zero to a large total) compared with a measure such as ever paid support (yes or no). This greater variability (or variance, in statistical terminology) makes it less likely that differences in measures like amount paid will be statistically significant.

Although it is correct to think of these impacts of referral to PFS services and coverage by its mandates as separate from the impacts of PFS intake (reported earlier in this summary), it is not correct to simply add the impacts produced at the two stages of the PFS process to get a total impact measure. The samples of parents for whom impacts are calculated at the two stages of the process are very different; for intake impacts, the sample included all fathers potentially eligible for PFS, while the impacts of referral to PFS is based on only those fathers actually found to be eligible. Thus, the two types of impacts should be thought of as separate effects of PFS at distinct stages of the program’s process for those fathers who were involved in the relevant stage (intake or referral to PFS services and coverage by its mandates).

• These positive impacts on percentage paying support of referral to PFS were mainly the result of substantial impacts in three of the seven sites: Dayton, Grand Rapids, and Los Angeles.

The findings presented above mask statistically significant variation among the sites. As shown in Table 7, in Dayton, Grand Rapids, and Los Angeles there were statistically significant and substantial impacts on the percentage of parents paying support, in most quarters ranging from 10 to 15 percentage points. Often, this amounted to a 15 to 50 percent increase in the proportion of parents paying support. In some quarters, there were also statistically significant increases in average payments in these three sites, and for Dayton and Grand Rapids the increase in average support paid for the entire period was also statistically significant. In Dayton the increase in the average payment amounted to 55 percent, while in Grand Rapids it was 20 percent. In the remaining four sites, impacts on child support payments were sporadic and generally not statistically significant.

Impacts on Employment and Earnings and Further Analysis of Child Support

• Across the seven PFS sites combined, referral to PFS produced no statistically significant impacts on parents’ employment rates or earnings.

As shown in Table 8, in no quarter did the employment rate or earnings of the program group exceed those of the control group. For both groups the quarterly employment rate remained at approximately 50 percent throughout the follow-up period. Average quarterly earnings, in contrast, did increase over time for both groups, with the amount approximately doubling between the quarter of random assignment and the final quarter of follow-up. In the sixth quarter of follow-up, earnings averaged approximately $1,400 for each group. This includes zero earnings for the parents (approximately 50 percent) who did not work in that quarter. Average earnings for those who worked during the quarter were, therefore, approximately $2,800 over the quarter, or an average of approximately $930 per month.

• Referral to PFS in two sites (Dayton and Los Angeles) did increase the percentage of parents who worked at some point during the follow-up, but no site produced a statistically significant increase in overall earnings.

At the site level, there are few or no positive impacts on employment and earnings and some negative and statistically significant impacts as well. In two sites (Los Angeles and Dayton), referral to PFS produced an 11 percentage point increase in the proportion of parents who worked at some point during the six quarters of follow-up, and this difference was statistically significant. However, no site produced a positive and statistically significant impact on overall earnings, a key objective of PFS, though Los Angeles and Memphis had positive — though statistically insignificant — impacts on earnings in the later quarters of follow-up.

• Persistent increases in child support payment rates came from parents who were employed in the formal economy. Any increases in support from unemployed parents were small and did not persist over the follow-up period.

Impacts on child support and employment were analyzed together in the three sites in which there were child support impacts. This was done using administrative records on earnings; as with all the employment findings in this report, the analysis captures only jobs in the mainstream economy in which employers report wages to the state. The analysis began by examining the percentage of parents who fell into each of four possible categories: (1) unemployed and not paying support, (2) unemployed and paying support, (3) employed and not paying support, and (4) employed and paying support.

During the initial four quarters of follow-up, referral to PFS in the three sites with child support impacts produced increased rates of payment for parents who were employed and those who were not. Using quarter 3 of follow-up as an example, referral to PFS produced a 15 percentage point impact on payment rates, with about 10 percentage points coming from employed parents and 5 percentage points coming from unemployed parents. By quarter 5 of follow-up, impacts on child support payment rates had disappeared for unemployed parents; thus, all longer-term impacts on child support came from employed parents.

• In general, there were few subgroup differences in child support or employment impacts.

Analysis of impacts on child support payment rates, average amounts paid, employment rates, and earnings also included an examination of key subgroups defined by father’s age, ethnicity, recent prior earnings, and prior arrest. There were no significant differences between subgroups in impacts except that non-black parents and parents with earnings of $2,000 or more in the three quarters prior to random assignment had impacts on average child support payment amounts, while black parents and those with earnings of less than $2,000 did not. For employment and earnings, the only difference between key subgroups in impacts concerned the "prior arrest" subgroup: parents with a prior arrest experienced negative impacts on earnings, and those without an arrest experienced a small and statistically insignificant increase in earnings. To the extent that there is a pattern in these subgroup findings, impacts tended to be better for parents with fewer barriers to employment.

Suggestions for Program Design

The findings in this report are far from being the complete PFS story. They rest on six quarters of follow-up for a sample of about half of the parents referred to the program. However, the data available at this time suggest that these findings on program implementation, participation in services, and impacts on child support and employment will be consistent with findings that will be available when comparable follow-up is completed for all parents referred to the program. More limiting is the scope of the topics covered; since information from the survey of parents is not yet available, this report does not discuss PFS’s success in helping noncustodial parents become more effective and involved parents. Finally, the current findings are based on administrative records and do not capture changes in informal support payments made directly to the custodial parent or child or in employment in the underground economy, topics to be covered later in PFS using follow-up surveys.

Despite these cautions, this summary closes with some suggestions about program design. This demonstration is important because of the increasing interest in programs for poor noncustodial parents, in part prompted by time-limiting welfare aid to custodial parents and children and facilitated by new federal "welfare-to-work" funding that can be used to serve noncustodial parents of children receiving assistance.

The PFS eligibility rules and the process of program intake identified a group of parents (largely men) who were both disadvantaged and somewhat diverse. They tended to be made up of fathers who worked in a series of low-paying jobs and fathers who had a tenuous connection to the mainstream labor market and thus had little earnings. The program participation requirements forced those who were working to report this fact (because they could not work and participate), and those who participated in program services typically had serious barriers to employment. As a consequence, PFS programs probably practiced "reverse creaming" rather than focusing their attention on the easier to serve.

The PFS intake process and the referral to PFS did produce impacts on child support payments, in large part coming from fathers who worked during the follow-up period. This suggests that agencies should not discount the possibility of payments from parents without known employment, the parents who initially appear to fit the PFS profile. Even within the disadvantaged population making up the PFS sample, some will find employment. It is important to recognize that continued changes in the CSE system are gradually increasing the speed with which previously unknown employment is discovered. Most notable is the recent requirement that employers covered by the unemployment insurance system report new hires immediately, providing the CSE system with a more accurate employment database against which to match noncustodial parents. But the PFS intake process also illustrates the limits of using administrative records and the payoff from extra outreach.

• The three sites that produced impacts on child support payments share several characteristics: strong involvement of the child support agency in PFS, a strong peer support program that focused on the importance of supporting children, and — in the case of Dayton — low existing levels of support payments.

Two of the three PFS sites with child support impacts were unique among the seven in one respect: in Los Angeles and Grand Rapids the child support agency was the lead local agency, driving the planning process and the management of the program, developing procedures to involve cases that typically would have been given low enforcement priority, and putting in place regular reviews of noncompliant parents involving child support and PFS staff. In other ways, these two sites are not unique, though it appears that a combination of several factors contributed to their strong child support impacts. In Dayton, over the course of the demonstration the child support agency and PFS staff worked together to dramatically change the PFS outreach and intake process, including targeting cases for whom location information was weak, developing new forms of legal notice for hearings, and (after the period of intake for the sample in this report) conducting home visits just prior to hearings to encourage an appearance. In addition, in each site the peer support staff was strong, though other sites also had good peer support facilitators, and some had higher levels of participation in peer support. Finally, existing payment levels were low in Dayton (so that there was "room" for improvement and, therefore, impacts), and the experience suggests that this — coupled with active involvement of the child support agency in program outreach and intake — was an important contributor to child support impacts. Memphis also had very low existing levels of payments, but the child support agency was not as involved as in Dayton, and there was little effort to reach cases who would be otherwise "unworked." As a result, Memphis produced no impacts on child support payments.

Another goal of PFS was to assist parents in working more steadily and earning more money. The program attempted to do this by providing a range of employment and training services and by focusing part of the peer support curriculum on issues relevant to the work setting. As discussed earlier in this summary, full employment services were often not provided, with assistance in job seeking receiving the greatest emphasis.

• A lack of "fit" between the employment and training services emphasized in the sites and the needs of a substantial portion of the PFS parents, as well as limited job opportunities within their neighborhoods, contributed to the lack of overall impacts on employment and earnings.

Because the PFS sample was largely made up of men who had worked — with varying degrees of regularity — at low-paying jobs, the challenge for the program was to help these parents find better jobs than they otherwise would have found and to keep them. Job search assistance and job club services, the most common employment services in PFS, can be effective in helping more people find jobs but are not well suited to helping people who are already in the labor market raise their wage rates or stabilize their work history. In Los Angeles and Memphis, where there was a hint of a trend toward positive earnings impacts at the end of the follow-up, a much higher than average percentage of PFS parents participated in skill-building activities (basic education or occupational training), which might have been better suited to boost earnings for a group already working to some extent.

Experimentation with new combinations of services seems called for in light of the PFS interim results and interest in serving poor, noncustodial parents. Finding new ways to combine work and skill-building services seems important because these parents need income quickly and also need to develop a plan for wage progression over time. Interestingly, the new welfare-to-work grants administered by the U.S. Department of Labor require sites that wish to offer skill- building services to first get a participant into a job and then provide the education or training. This combination seems likely to coincide with the reality of the lives of poor, noncustodial parents (who have little access to cash assistance) and implies a longer-term service plan and a skill-upgrading approach rather than an effort to help these men quickly get better-paying and more stable jobs.

Job retention services, which also would seem to be an important part of this longer-term service strategy and which would have been more directly related to the problems of the PFS sample than job clubs, were added to PFS services part way through the demonstration but were not delivered with intensity to most parents. Finally, jobs were often scarce within the communities of residence for the PFS parents, and they faced — or at least feared — discrimination in the larger metropolitan labor market, especially with their combination of barriers to employment. This suggests that there may be a need in some communities for a pool of time-limited subsidized community service jobs to help men quickly start earning a pay check and build a work history that will make them more appealing to other employers.

A final lesson from the initial PFS experience is also a pressing challenge for program operators. For a program like PFS to work, there must be a strong local service partnership, in which agencies coming from many different perspectives can find a way to work together for common goals. The difficulty of achieving this common purpose plagued the PFS demonstration and — without substantial and continuous attention — it is likely to be a problem in similar programs. As partnerships are developed, there is a danger of glossing over differences in perspective in order to get to the "real" business of putting service components in place. The PFS experience suggests instead that initial investments in team building which acknowledge and seek to reconcile differences in perspective can be money well spent. In the absence of serious efforts to build a sense of common enterprise, the competing priorities of the partners can undermine the best-laid plans to provide a seamless and comprehensive program for low-income fathers.

Notes:

¹The last site to enter the demonstration was Los Angeles, where the random assignment process began later, in February 1995.

Funders of the Parents' Fair Share Demonstration include: U.S. Department of Health and Human Services, The Pew Charitable Trusts, W. K. Kellogg Foundation, Charles Stewart Mott Foundation, U.S. Department of Agriculture, The Annie E. Casey Foundation, U.S. Department of Labor, Smith Richardson Foundation, Ford Foundation, The McKnight Foundation, and Northwest Area Foundation.