Table of Contents | Previous | Next |
Chapter 3: Costs of Jobs-First GAIN
The cost analysis presented in this chapter estimates how much the Los Angeles County welfare department (DPSS) and other government agencies spent to provide employment-related services to experimental and control group members in the Jobs-First GAIN Evaluation. The estimates include expenditures by DPSS for Jobs-First GAIN activities and support services, as well as funds for Jobs-First GAIN activities that came from other agencies’ operating budgets. Further, the chapter estimates the average costs to outside providers of providing employment-related services that experimental and control group members sought out on their own initiative. This information will be useful to administrators and planners who want to understand the level of the government’s investment in Los Angeles’s Jobs-First GAIN program and compare the program’s effects to those of other types of welfare-to-work strategies. Moreover, the results of this analysis provide context for interpreting Jobs-First GAIN’s impacts and cost-effectiveness, which are reported in later chapters.
The primary goal of the cost analysis is to estimate the average net cost to the government of providing employment and education-related services to members of the experimental group. The net cost is the difference between the average cost per experimental group member and the average cost per control group member of all Jobs-First GAIN and non-Jobs-First GAIN services used during the period studied. These costs include those of services provided from the date of random assignment until September 30, 1998. (Starting on October 1, 1998, all experimental and control group members still mandated to participate were enrolled in CalWORKs.) The period studied lasted from two years to two and a half years, depending on each sample member’s date of random assignment. Chapter 8 presents a benefit-cost analysis that addresses whether the economic gains to the government budget exceeded the additional expenditures required to provide Jobs-First GAIN services to experimental group members. This chapter begins with an overview of the major components of the cost analysis. It then discusses the cost estimates in detail for experimental and control group members.
Key Findings
The estimated total, or gross, cost (in 1998 dollars) of Jobs-First GAIN and non-Jobs-First GAIN activities per AFDC-FG experimental group member (including participants and nonparticipants) was $4,305. DPSS paid about $1,771 per experimental group member (41 percent of the gross cost), primarily for case management and the operation of the program's motivational orientation session and job clubs. Another 15 percent of the cost was paid by schools and other agencies, mostly for basic education and vocational training for experimental group members assigned to these activities by Jobs-First GAIN. The remainder of the gross cost (about 44 percent) represents expenditures paid by other agencies and institutions for non-Jobs-First GAIN activities.
The experimental-control difference in cost, or net cost, of Los Angeles's Jobs-First GAIN program — $1,392 per AFDC-FG experimental group member — was much smaller than its gross cost largely because of control group members' relatively high levels of participation in education and training activities.
The estimated gross and net costs of Jobs-First GAIN for AFDC-U experimental group members averaged $2,485 and $1,170, respectively - less than the corresponding costs for AFDC-FGs.
Jobs-First GAIN's gross and net costs fell well below those of the earlier, education-focused Los Angeles GAIN program and were comparable to those of Work First programs previously evaluated by MDRC.
Major Components of the Cost Analysis
Jobs-First GAIN-Related Cost per AFDC-FG Experimental Group Member
Jobs-First GAIN-Related Expenditures by the Welfare Department (Figure 3.1, Box 1)
Operating costs. DPSS covered the expenditures for the day-to-day operation of the program, including expenditures for case management services, overhead, orientation to the program, job club, and work experience.1 Payments made by DPSS to the County Office of Education for running orientations and job clubs and to outside organizations for providing unpaid work experience positions are included in welfare department costs (Box 1) and allocated across activities (see Table 3.2). Welfare department costs also include the expenditures for case management associated with program referrals to education and training activities that other agencies and institutions funded.2 Expenditures for actually operating these activities appear as non-welfare agency costs.
Unit cost estimates are central to cost analysis. The unit cost of an activity is an estimate of the average cost of providing the activity to one person for a specified unit of time. Typically, MDRC collects agency participation and expenditure information for a “steady-state” period — that is, a one- to 12-month period during one fiscal year that falls within the follow-up period for the evaluation. Costs are then distributed among specific activities. It was beyond the scope of the present evaluation to estimate unit costs in this way. As a proxy, this analysis uses the unit costs of orientation, appraisal, and assessment; job search; basic education; and vocational training or post-secondary education calculated for the earlier Los Angeles GAIN program and of work experience or on-the-job training (OJT) calculated for the Riverside Labor Force Attachment (LFA) program. These unit costs were expressed in 1998 dollars and used to estimate gross costs for both AFDC-FGs and AFDC-Us.3
Table 3.1 (columns 1 and 2) shows the welfare department unit costs of the five activities, and Table 3.2 (top panel, column 1) displays the average gross cost to the welfare department per experimental group member. To obtain the cost per experimental group member shown in Table 3.2 (top panel, column 1), the unit cost of each activity was multiplied by experimental group members’ average number of months of participation in that activity. (In addition, each experimental group member was credited with attending one orientation session, and monthly costs of orientation, appraisals, and assessments were added to her total.) Thus, the differences in gross cost between activities can be explained by differences in unit cost, participation level, and length of stay among those who took part. For example, the estimated welfare department unit cost of job search activities ($611) exceeded the unit cost of orientation, appraisal, and assessment ($572) by a small amount. The cost per experimental group member to the welfare department of job search activities was lower ($460 versus $656), however, because 37 percent of experimental group members participated in job search, whereas all experimental group members attended orientation (see Table 2.1). As Table 3.2 shows, the sum of all activities’ costs per experimental group member yields a total Jobs-First GAIN welfare department operating cost of $1,442 per experimental group member.
-
Support service costs. DPSS paid for child care, transportation, and ancillary services (for example, uniforms, tools, equipment, books, registration, and licensing fees) to help sample members participate in Jobs-First GAIN. Data on individual support service expenditures were collected for all experimental group members from the GAIN Employment Activity and Reporting System (GEARS) automated payment records.
Table 3.2 (top panel, column 1) shows that the average total Jobs-First GAIN child care cost in Los Angeles was $258 per experimental group member.4 About 23 percent of experimental group members received a child care payment during the follow-up period (result not shown). These results fall within the same range as those for Work First programs previously evaluated by MDRC; the Job-First GAIN child care cost is somewhat below that of the earlier GAIN program in Los Angeles.5 Jobs-First GAIN experimental group members received an average of $71 to cover expenses for transportation to activities and to appointments with Jobs-First GAIN case managers and for ancillary expenses.
-
Total Jobs-First GAIN-related costs incurred by the welfare department. Table 3.2 (top panel, column 1) shows that Los Angeles’s welfare department paid a total of $1,771 per experimental group member for Jobs-First GAIN services, including both operating and support service expenditures.
Jobs-First GAIN-Related Expenditures by Non-Welfare Agencies (Figure 3.1, Box 2)
Total Jobs-First GAIN-Related Cost (Figure 3.1, Box 3)
Non-Jobs-First GAIN Cost per AFDC-FG Experimental Group Member
Non-Jobs-First GAIN Expenditures for Transitional Child Care (Figure 3.1, Box 4)
Non-Jobs-First GAIN Expenditures by Non-Welfare Agencies (Figure 3.1, Boxes 5 and 6)
Gross Cost per AFDC-FG Experimental Group Member
Gross Cost per AFDC-FG Control Group Member
Non-Jobs-First GAIN Expenditures by the Welfare Department (Figure 3.1, Box 8)
Non-Jobs-First GAIN Expenditures by Non-Welfare Agencies and Gross Cost (Figure 3.1, Boxes 9 and 10)
Net Cost per AFDC-FG Experimental Group Member (Figure 3.1, Box 11)
Summary of Cost Estimates for AFDC-Us (Figure 3.2, Table 3.3)
Figures 3.1 and 3.2 illustrate the cost components for the experimental and control groups in the AFDC-FG and AFDC-U samples, respectively. Costs were calculated for two categories of activities and services: those provided to meet Jobs-First GAIN requirements or support Jobs-First GAIN participation and those provided to support non-Jobs-First GAIN services and activities. In each category, costs are further broken down into those paid by DPSS and those paid by non-welfare agencies. The figures show that the Jobs-First GAIN-related cost per experimental group member (Box 3) equals the expenditures incurred by DPSS to operate the program (Box 1) plus the expenditures incurred by non-welfare agencies (for example, adult basic education schools, community colleges, and proprietary schools) to provide education and training activities that met Jobs-First GAIN requirements (Box 2). The non-Jobs-First GAIN costs per experimental group member (Box 6) include the costs of services that experimental group members received on their own (Box 5), generally after leaving the program. The Jobs-First GAIN and non-Jobs-First GAIN costs per experimental group member make up the gross cost per experimental group member (Box 7).
As discussed in Chapter 2, about 44 percent of control group members participated in non-Jobs-First GAIN activities on their own initiative. Box 9 displays the cost of these activities per control group member. Control group members were eligible to receive support service payments from DPSS to cover child care, transportation, and other expenses related to their participation in non-Jobs-First GAIN activities (Box 8). The sum of these costs (in Boxes 8 and 9) is the gross cost per control group member (Box 10).
This chapter is organized to move through the boxes in Figures 3.l and 3.2, beginning with Jobs-First GAIN-related expenditures and ending with the program’s net cost per experimental group member (Box 11). The net cost equals the gross cost per experimental group member (Box 7) less the amount that would have been spent in the program’s absence, that is, the gross cost per control group member (Box 10). The tables that follow show the above components broken down by activity and type of support service payment.
This section examines expenditures made by the welfare department and by non-welfare agencies for Jobs-First GAIN-related activities and support services for AFDC-FGs.
Welfare department costs consisted of program operating costs paid by DPSS and the costs of support services that experimental group members received to enable their participation in Jobs-First GAIN.
 [D] |
| Activity | Experimental Group Members | Control Group Members | |||
|---|---|---|---|---|---|
| Welfare Department Unit Cost ($) | Non-Welfare Agency Unit Cost ($) | Non-Welfare Agency Unit Cost ($) | |||
| Average per Participant-Month | Average per Appraisal | Average per ADA | Average per Participant-Month | Average per ADA | |
| Orientation, appraisal, and assessment | 572 | ||||
| Job search | 611 | 152 | |||
| Basic education | 317 | 2,342 | 2,344 | ||
| Vocational training/ post-secondary education | 178 | 3,580 | 3,639 | ||
| Work experience/OJT | 565 | 565 | |||
|
SOURCES: Unit cost estimates for Los Angeles GAIN from Riccio et al., Table 3.1, pp. 72-73, and for Riverside LFA from Hamilton et al., 1997, Table 7.1, p. 168. NOTES: Unit costs from the source tables were converted from 1993 into 1998 dollars. "ADA" is a unit of Average Daily Attendance, a measure used by California community colleges and adult schools that is defined as a block of 525 hours of attendance. (One ADA unit equals the total course time for a full-time student during a normal academic year.) |
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
SOURCES: MDRC calculations from the GAIN Employment Activity and Reporting System (GEARS) and from the Two-Year Client Survey, based on unit cost estimates from Riccio et al., 1994, Table 3.1, pp. 72-73, and from Hamilton et al., 1997, Table 7.1, p. 168. NOTES: Rounding may cause slight discrepancies in calculating sums. N/a = not available or applicable. |
As discussed above, outside providers paid most of the costs of operating Jobs-First GAIN’s basic education and vocational training activities. The Jobs-First GAIN-related cost per experimental group member paid by non-welfare agencies for these activities was calculated by multiplying the unit cost estimate for each activity by the average number of hours that experimental group members participated in it.6 Table 3.2 (top panel, column 2) shows that the estimated non-welfare agency cost of Jobs-First GAIN education and training was $638 per experimental group member. Most of these expenses went toward providing vocational training.
Adding the welfare department and non-welfare department costs of Jobs-First GAIN yields a total Jobs-First GAIN-related cost per experimental group member of $2,409 (Table 3.2, top panel, column 3). As expected, most expenditures for Jobs-First GAIN (about 59 percent, excluding support service payments) went toward operating Jobs-First GAIN’s employment-focused activities — job search and work experience (DPSS did not offer OJT) — and the program’s motivational orientation and appraisal sessions. The costs of providing education and training services were also substantial, even though about three and a half times as many experimental group members participated in job search as participated in education and training (see Table 2.1). Most experimental group members who participated in job search activities attended only one three-week session, however, compared with an average of several months of attendance for participants in education and training activities (see Table 2.4). The longer average stay in education and training activities increased their cost relative to that of job search. In addition, the unit costs of operating education and training activities exceeded the unit cost of job search by a wide margin.
As expected, the total cost per experimental group member of Jobs-First GAIN activities ($2,409 in 1998 dollars) fell well below the cost of operating DPSS’s previous education-focused GAIN program ($6,577 in 1998 dollars). In addition, Jobs-First GAIN’s cost fell within the range of those of other employment-focused programs — specifically, above Riverside LFA’s ($1,349 in 1998 dollars) and below Riverside GAIN’s ($3,257 in 1998 dollars).7
As discussed in Chapter 2, some experimental group members entered education and training activities on their own after leaving Jobs-First GAIN or participated in activities that were not approved by Jobs-First GAIN staff while enrolled in the program. Although these services are not considered to be Jobs-First GAIN-related, they have the potential to increase experimental group members' long-term earnings and reduce their reliance on welfare, just as these services have this potential for the control group. Thus, they should be included in the gross cost estimate before the cost per experimental group member is compared with the cost per control group member.
It was beyond the scope of this evaluation to investigate the receipt of subsidized child care while experimental group members were working or participating in employment-related activities outside Jobs-First GAIN. Evidence from survey responses suggests that very few experimental group members received a child care payment and that such assistance contributed very little to the gross cost of serving experimental group members.8
Table 3.2 (top panel, column 5) shows that outside providers spent an average of $1,896 per experimental group member on activities not approved by Jobs-First GAIN staff that experimental group members attended on their own initiative. Most of the expenditures went to pay for vocational training and post-secondary education.9 Notably, the cost of these outside activities is nearly 80 percent of the average cost of Jobs-First GAIN. Longer stays and higher unit costs for education and training explain why costs for non-Jobs-First GAIN activities were relatively high.
The gross cost per experimental group member (Box 7) was determined by adding the Jobs-First GAIN-related cost (Box 3) to the non-Jobs-First GAIN-related cost per experimental group member (Box 6). This estimate is important because it represents the costs of all services, both Jobs-First GAIN-related and non-Jobs-First GAIN-related, that have the potential to increase experimental group members’ long-term earnings and reduce their use of welfare. This total investment must be compared to the gross cost per control group member in order to determine the government’s net investment per experimental group member and, in the benefit-cost analysis presented in Chapter 8, the net payoff of that investment.
Los Angeles’s gross cost per experimental group member averaged $4,305, somewhat above the average costs per experimental group member found in previous evaluations of Work First programs by MDRC. This average gross cost also exceeded those of the two Riverside programs (GAIN, $2,288, and LFA, $3,813, in 1998 dollars), but came nowhere near the gross cost of the earlier GAIN program in Los Angeles ($7,036 in 1998 dollars).10
The gross cost per control group member represents the average cost to government agencies of providing employment-related services to welfare recipients in the absence of Jobs-First GAIN. The difference between the gross cost per experimental group member and the gross cost per control group member represents the net cost of Jobs-First GAIN. As discussed in Chapter 2, about 44 percent of control group members enrolled in employment-related activities on their own initiative. In addition, control group members were eligible for support services from DPSS for their self-initiated activities, although, as noted below, almost no control group members requested such services. Therefore, expenditures paid by non-welfare agencies account for all but a tiny fraction of the gross cost per control group member.
Control group members were eligible to receive child care assistance for education and training activities that they participated in on their own and could receive work-related transitional and other non-Jobs-First GAIN child care. It appears, however, that very few control group members received such assistance. Table 3.2 (bottom panel, column 4) shows that the welfare department's expenditures for support services averaged only $6 per control group member.
Table 3.2 (bottom panel, column 5) shows that non-welfare agencies' expenditures averaged $2,906 per control group member. Most of these expenditures went toward vocational training and post-secondary education ($2,290) or basic education ($507). Adding non-welfare agency expenditures and welfare department expenditures yields an average gross cost per control group member of $2,912.
The net cost per experimental group member is calculated by subtracting the gross cost per control group member (Box 10) from the gross cost per experimental group member (Box 7). The difference is $1,392 (Box 11). As discussed in the previous chapter, Jobs-First GAIN's overall impact on the use of employment-related services resulted mostly from gains in participation in job search (job club) and, to a much lesser extent, from increased participation in unpaid work experience jobs. These activities, coupled with experimental group members’ attendance at Jobs-First GAIN’s motivational orientation and the program’s modest expenditures for support services, produced the net increase in cost. About equal percentages of experimental and control group members attended classes in basic education, vocational training, or post-secondary education, resulting in little net change in the costs of these activities.
Jobs-First GAIN’s net cost is close to those of several other Work First programs evaluated by MDRC, the costs of which range from about $1,200 to $2,500 (in 1998 dollars). As expected, the net cost of Jobs-First GAIN was much lower than — in fact, was less than one-quarter of — the net cost of the Los Angeles GAIN program ($6,363). The difference between the two Los Angeles programs in cost resulted in large part from DPSS’s switch to lower-cost job search services, but also from the much greater use of employment-related services by members of the control group in the Jobs-First GAIN Evaluation.11
Gross and net cost estimates for AFDC-U experimental group members require much more guesswork. As in the AFDC-FG cost analysis, estimates of the welfare department costs for Jobs-First GAIN activities were calculated from DPSS’s automated program tracking and support service payment records. As discussed in Chapter 2, a smaller percentage of AFDC-U than AFDC-FG experimental group members participated in a Jobs-First GAIN activity. On average, AFDC-Us also spent fewer months in Jobs-First GAIN activities and received fewer dollars in support service payments. As would be expected given AFDC-U experimental group members’ less extensive involvement in Jobs-First GAIN, the total welfare department cost for AFDC-U experimental group members averaged $1,350 (Figure 3.2, Box 1), that is, $421 less than what DPSS spent for AFDC-FG experimental group members.
The Two-Year Client Survey sample included no AFDC-Us. Therefore, no survey data were available for estimating their total hours of participation in Jobs-First GAIN basic education and vocational training activities or the resulting operating costs to outside providers. Moreover, the extent of AFDC-Us’ participation in activities outside Jobs-First GAIN and the cost of these services were not measured.
Survey data for AFDC-FG experimental and control group members were used to make the gross and net cost estimates for AFDC-Us, but were adjusted downward. That is, it was assumed that AFDC-U experimental group members’ lower level of participation in Jobs- First GAIN meant that AFDC-Us in both research groups participated less often in activities outside the program as well.
The factor by which the non-welfare agency cost estimates were adjusted was the following: AFDC-U experimental group members’ average number of months of participation in Jobs-First GAIN activities divided by AFDC-FG experimental group members’ average number of months of participation in Jobs-First GAIN activities. For example, based on DPSS’s automated program tracking records, AFDC-U experimental group members averaged 0.19 month of participation in Jobs-First GAIN basic education classes, whereas AFDC-FG experimental group members averaged 0.33 month. It was estimated (from survey responses) that AFDC-FG experimental group members attended basic education classes outside Jobs-First GAIN for an average of about 38 hours. Using the above formula, the average number of hours of attendance in basic education outside Jobs-First GAIN for AFDC-Us was calculated by multiplying 38 hours by (0.19/0.33), which equals about 22 hours. This average was then multiplied by the non-welfare department unit cost of basic education to obtain an estimated gross cost of basic education outside Jobs-First GAIN for AFDC-U experimental group members of $97. A similar procedure was used to estimate costs for AFDC-U control group members.12
According to the same formula, outside providers spent an average of $277 per AFDC-U experimental group member to provide basic education and vocational training. The combined cost of Jobs-First GAIN activities and support services came to $1,627, about a third less than that for AFDC-FG experimental group members. AFDC-U experimental group members required an additional $858 from outside providers for participation in non-Jobs-First GAIN activities, bringing their average gross cost up to $2,485 per experimental group member. The average gross cost for AFDC-U control group members was $1,315. Therefore, the net cost of Jobs-First GAIN for AFDC-Us was $2,485 minus $1,315, or $1,170.
[D]
|
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
SOURCES: MDRC calculations from the GAIN Employment Activity and Reporting System (GEARS) and from the Two-Year Client Survey, based on unit cost estimates from Riccio et al., 1994, Table 3.1, pp. 72-73, and from Hamilton et al., 1997, Table 7.1, p. 168.
NOTES: Rounding may cause slight discrepancies in calculating sums.
N/a = not available or applicable.
|
| Table of Contents | Previous | Next |