[ Main Page of Report | Contents of Report ]
FSA pushed administrators to operate more complex welfare-to-work programs, involve a larger share of the caseload (including the most disadvantaged), and provide more intensive and expanded services -- changes that were likely to increase the costs of the programs relative to those of programs operated in the early to mid-1980s. The legislators' hope was that the higher upfront costs would bring larger long-run benefits, but it was unclear whether the programs' benefits would surpass their costs or whether the benefit-cost trade-off would be the same for all types of program approaches or for all subgroups of welfare recipients.
Under PRWORA, the relationship between costs and benefits remains a concern. A program's cost per person largely determines how many people it can serve. With fixed resources, each state is seeking to maintain the optimal balance between involving as many people as possible in the programs and delivering appropriate, effective services. The extensive analyses conducted as part of NEWWS help reveal the parameters around these program design decisions.
[Go To Contents]
Different types of five-year costs were estimated for the NEWWS programs. The gross cost per program group member is a comprehensive measure of all the costs associated with providing employment services and related support services to people while they were enrolled in a welfare-to-work program as well as after they left the program and/or the welfare rolls; similarly, the gross cost per control group member is the corresponding estimate for the control group. The net cost of the program is the difference between these two estimates; in other words, the cost for the control group is the benchmark used to determine the size of the additional cost of running the program (on a per-person basis).
The costs of providing job search, education, and training -- whether within welfare-to-work programs or when individuals enrolled in these activities on their own -- as well as the costs of case management and support services contribute to both gross and net costs. For example, if a substantial and similar proportion of program and control group members participate in high-cost activities like vocational training and postsecondary education, it is likely that the gross costs for both groups will be high but that the program's net cost will be relatively low. In contrast, if most program group members participate in low-cost job search activities and few control group members do so, it is likely that the gross costs for both groups will be low but that the program's net cost will be relatively high.
Averaged across all the NEWWS programs, the five-year gross cost per program group member, which included a substantial amount of self-initiated activities that sample members engaged in after leaving welfare, was approximately $7,600. Subtracting the gross cost per control group member yields the average five-year net cost per program group member, or about $3,600. The average net cost per program group member over the first two years, during which time many sample members were still on welfare, was about $2,000. The NEWWS programs' average cost was higher than that of other welfare-to-work programs studied by MDRC and comparable to that of programs such as the Alameda and Los Angeles Greater Avenues for Independence (GAIN) programs operated in the late 1980s and early 1990s, which provided extensive education and training services. It is important to note that the NEWWS programs were not mandatory for the entire welfare caseload (as is nearly always the case now, under TANF rules). Thus, the NEWWS costs are per person for a segment of the caseload -- not averaged across the whole caseload -- and do not include women with infant children, who typically have the highest child care costs.
All the NEWWS programs provided job search, education, and training activities as well as child care and transportation services to support participation in these activities. The programs' operating costs -- which included expenditures for case management services, overhead, and program orientation as well as job search, education, and training program activities -- were considerably higher than their support service costs, which included expenditures for child care, transportation, and other needs such as uniforms, tools, and books. Operating costs accounted for about four-fifths of the five-year gross cost, while support services accounted for about one-fifth. Another way to examine program costs is to look solely at the cost of providing in-program activities -- employment-promoting activities (and support services) that people engaged in while on welfare and in the programs -- as opposed to including the costs of self-initiated activities in which people participated after they had left welfare. Averaged across the programs, the in-program cost was approximately 70 percent of the five-year gross cost.
In NEWWS, welfare departments most often directly provided and funded case management, job search activities, and support services (primarily child care and bus passes) for people in program activities. Welfare departments generally did not pay for services such as basic education or vocational training. These costs, which accounted for the majority of the gross cost of most programs, were covered by schools and community agencies.
In all three sites where a side-by-side comparison of an employment-focused program and an education-focused program was conducted, the education-focused program was from one-third more expensive to nearly twice as expensive as the employment-focused program. This is largely because program group members in education-focused programs were more likely to participate in vocational training (a high-cost activity) and because education and training assignments typically lasted longer than job search. Both the welfare department and other agencies had to pay additional expenses for education-focused programs: Whereas the welfare department spent more on case management, nonwelfare agencies spent more on classroom instruction and related expenses.
Recipients who entered NEWWS with a high school diploma or GED were more likely to participate in more expensive activities such as postsecondary education and vocational training, whereas those lacking these credentials were more likely to participate in less expensive activities such as basic education. Control group members' rates of (self-initiated) participation in postsecondary education and vocational training were somewhat lower than those for program group members, while their rates of participation in basic education were much lower. As a result, gross costs were much higher for those with a high school diploma or GED than for those lacking this credential, while the differences in net costs between the two subgroups were less pronounced.
[Go To Contents]
The programs' five-year benefits were also calculated in NEWWS. Benefits included the increases in earnings and decreases in welfare and food stamp payments discussed earlier, as well as dollar valuations of the programs' estimated effects on Medicaid, job fringe benefits, taxes paid, and the costs of administering transfer programs such as food stamps. All of these effects were considered along with the programs' estimated net costs to ascertain the net gains and losses to government budgets as well as to program group members resulting from the programs. Some limits on the comprehensiveness of the benefit-cost analysis, however, should be recognized. The analysis did not, for example, consider education benefits that are not reflected in earnings or put a dollar value on families' or children's well-being.
One measure of cost-effectiveness is the return to government budgets per net dollar invested, that is, the average gain to government budgets from increased tax revenues and savings in transfer programs and associated administrative costs divided by the net cost of program services. By this metric, government budgets come out ahead if a program produces more than a dollar's worth of additional revenues and savings for every extra dollar spent on services for program group members relative to control group members. About half of the NEWWS programs returned at least as much to government budgets as was invested in them. Four of the programs, however, returned only about $.40 to $.80 per dollar invested. Programs that produced larger welfare savings had a better chance of paying for themselves, and those that generated the largest welfare savings resulted in the largest returns, bringing government budgets as much as $2.83 per dollar invested. The programs that generated smaller welfare savings had a harder time breaking even and more often resulted in losses. In contrast, the programs that led to the largest employment increases did not necessarily produce the largest government returns.
Where savings to government budgets were found in NEWWS, they were driven by welfare savings, leaving the sample members no better off financially than they would have been without the programs. In general, the benefit-cost findings from the welfare recipients' perspective mirror the findings on income: In most of the NEWWS programs, program group members' financial losses from decreased welfare and food stamp payments exceeded their financial gains from increased earnings and the EIC. Taking into account the dollar value of program group members' decreased Medicaid eligibility over the whole five-year follow-up period, program group members' losses -- compared with control group members' -- were even more pronounced.
To have generated a return on the government's investment as large as that generated by the employment-focused programs, the education-focused programs would have had to generate larger welfare savings and/or larger increases in tax revenues. Not surprisingly in view of the earnings and welfare impacts discussed earlier, this was not the case (for the average return to government budgets per net dollar invested for each of the four primary categories of programs examined in NEWWS, see Table 7). On average, the greater investment in education-focused programs resulted in a greater loss from the government budget perspective. Likewise, the additional cost of program services for graduates did not result in a larger return to government budgets; in fact, the return to government budgets was not consistently higher or lower for graduates than for nongraduates for any type of program. Finally, no relationship was found between the NEWWS programs' net costs and earnings impacts: For example, the Portland program had one of the lowest net costs, yet programs with much higher net costs produced much smaller effects on earnings and employment than those found in Portland.
| Employment-focused approach | Education-focused approach | ||
|---|---|---|---|
| Job search first | Varied first activity | Education or training first | |
| High enforcement | High enforcement | High enforcement | Low enforcement |
| $1.58 | $2.83 | $0.87 | $0.84 |
| SOURCE: Hamilton et al., 2001. | |||
Top of Page |
Contents
Main Page of Report | Contents
of Report
Home Pages:
National Evaluation of Welfare-to-Work Strategies
(NEWWS)
Human Services Policy (HSP)
Assistant Secretary for Planning and Evaluation
(ASPE)
U.S. Department of Health and Human Services
(HHS)