Benefit-cost analysis of supported employment in Illinois.

This study analyzed the benefits and costs of supported employment in Illinois during a four-year period (i.e., 1987 through 1990). Benefits and costs were identified and valued from three perspectives: society's, taxpayer's, and supported employees'. During this four-year period, society received a return of $.91 for every $1.00 invested; taxpayers received a return of $.77; and supported employees increased their not earnings by 42%. During the fourth year of the project, the societal return was $1.09; the return to taxpayers was $0.89; and not earnings for supported employees increased by 57%. Changes in the costs and benefits of supported employment participants during this time period indicate that the continuation of this type of analysis is critical to an understanding of the progress of supported employment and to identifying ways of improving this emerging service delivery system. Descriptors: Benefit-cost Analysis, Supported Employment, Integration, Work, Severe Disabilities.

In the 1970s, supported employment was introduced to assist people with severe disabilities to obtain and maintain employment on jobs that would otherwise have been filed by persons without disabilities. The target population for supported employment are individuals who would normally be permanently placed in sheltered workshops, work activity centers, adult day care (hereafter, these three types will be referred to collectively as "alternative placements") or who would receive no adult services assistance (cf. Rusch, 1990).

Supported employment consists of providing on-the-job supports for an extended period of time (sometimes for the duration of employment; in an integrated work setting where employees without disabilities perform similar or related work. Initially, supported employment services were usually provided by job coaches paid by an outside organization. More recently, there has been increasing emphasis placed on having extended supports provided by supervisors, co-workers, and relatives and friends. These supports are usually described as "natural supports" (cf Chadsey-Rusch & Johnson).

Data on which to rigorously assess the effectiveness of supported employment services is being collected in Illinois. Since 1985, the Illionois Supported Employment Project at the University of Illinois has been collecting information on the actual aggregate annual earnings of supported employment participants throughout the State as well as the costs of providing these services.

This paper has three purposes. The first is to identify the benefits and costs of supported employment in Illinois between 1985 and 1990, the first four years that this program operated. The second is to compare how these benefits and costs changed over the four-year period. And the third purpose is to discuss implications of the data for long-term benefits and costs of supported employment.

Background

Initially, many people predicted that the earnings, and therefore the benefit/cost ratio of placing people into supported employment would be far greater than if they were in an alternative placement (Conley, 1973). Several early evaluations of supported employment programs appeared to verify these predictions (Hill, Banks, Handrich, Wehman, Hill, & Shafer, 1987; Noble & Conley, 1987). However, more recent and more rigorous analyses utilizing comprehensive longitudinal data bases available in Illinois and New York (Conley & Noble 1990) indicated less promising results. The results were remarkably consistent in the two states, despite some methodological differences.

In both states, the earnings of participants in supported employment were about twice as high as they would have been had they remained in an alternative placement. Unfortunately, the costs of supported employment appeared to increase by more in both states than the increase in earnings of the participants. In consequence, during the initial years of operating the program, the benefit-cost ratio in both states, on the basis of financial variables only, was less than one. These findings fuel questions regarding the advisability of continuing supported employment. Arguments include the possibility that significant intangible benefits are accruing that die usual benefit-cost ratios do not take into account. More importantly for our purposes, it has been argued that these benefit-cost ratios understate the longer term benefit-cost ratio that is expected from supported employment. Clearly, there are substantial start-up costs (e.g., staff training and contacting employers), that will decline over time. In addition, the earnings of participants should improve as participants gain experience on their jobs and as supported employment providers become more experienced and knowledgeable. Finally, these providers can be expected to become more efficient in their operations over time. Some evidence exists that these trends actually occur. Data from the first three years of supported employment in Illinois show that earnings have increased while costs have decreased (McCaughrin, Rusch, Conley, & Tines, in press).

Method

Participants

This benefit-cost analysis is based on data received from 30 Illinois supported employment programs that were started in Fiscal Year 1986 (July 1, 1986 to June 30, 1987) and continued through Fiscal Year 1990 (July 1, 1989 to June 30, 1990). Each of these programs received supported employment funds from the Department of Mental Health and Developmental Disabilities and/or the Department of Rehabilitation Services. During this four-year period, 729 individuals were enrolled in these 30 programs and provided supported employment services. Three hundred and ninety-four individuals entered initially with 69 (18%) leaving during their first year; 519 participated during the second year with 78 (15%) leaving; 511 participated during the third year with 90 (18%,) leaving; and 492 participated during the fourth year with 76 (15%,) leaving. Reasons for termination included inappropriate social behavior while working and job layoffs due to lack of business (cf. Lagomarcino, 1990). There were no significant demographic differences between terminated workers and those who remained employed.

Table 1 summarizes the demographic characteristics of the participants during the four-year period. The average age of these individuals was 32 years; 426 (58%) were male; and 611 (84%) were white. There were 325 (45%) supported employees with mild mental retardation, 161 (22%) with moderate mental retardation, and 56 (8%) with severe or profound mental retardation. The remaining 187 (26%) had other disabilities, such as cerebral palsy, epilepsy, and multiple disabilities.

Characteristics of 729 Participants

                      FY87   FY88   FY89   FY90   TOTAL

Participants           394    519    511    492    729

Gender
 Male                  244    326    300    283    426
 Female                150    193    203    204    298
 Unreported              0      0      8      5      5

Race
 White                 311    430    429    410    611
 Black                  60     63     47    449     68
 Hispanic               21     17     18     19     30
 Asian                   5      6      6      3      4
 American Indian         0      0      2      1      1
 Unreported              2      1     15     15     15

Average Age             30     31     33     32     32

Primary Disability
 LD & Borderline        45     20     22     24     36
 Mild Retardation      146    252    235    227    325
 Mild Retardation       96    124    122    106    161
 Severe Retardation     30     28     32     30     45
 Prof Retardation        6      9      7      4     11
 Mental Illness                25     29     32     45
 Other                  41     45     46     47     78
 Unreported             30     16     18     22     28

Previous Placement
 Adult Day Care          2      2      5      9     10
 Work Activity          92    104    100     78    123
 Reg Work Program      146    190    197    152    244
 Work Adj Training      57     67     61     56     86
 Skill Training         14     16     10     19     20
 Trans Employment        6      7      4      7     10
 Evaluation             11     12      6     13     15
 Voc Development        12     14      9      0      7
 Community               3     14     14     13     18
 School (SWEP)          44     53     43     27     52
 Comp/Support Empl       0     17     36     55     68
 Other                   0      5      0      5      9
 Unreported              7     18     26     58     67

ISEP Placement
 Individual            193    221    227    273    383
 Clustered             171    251    240    179    286
 Mobile Work Crew       30     46     30     38     56
 Unreported              0      1     14      2      4

(*)Total = FY1990 ISEP Workers + Terminated Worker, FY87, FY88, & FY89.

Prior to enrollment in supported employment, participants were enrolled in 1 of 12 different vocational programs provided by local adult service agencies, schools, and state agencies. Two hundred forty four individuals came from regular work programs in sheltered workshops; 123 came from work activity centers, and 10 from daycare centers. Fifty-two were referred from secondary school programs; 18 were referred from the communities they lived in; and 76 came from "other" or "unreported" placements.

Over half of the supported employees were individually placed (383) enabling them to work independently from one another in different locations, while 286 were placed in enclaves where they worked in the same general area on the same shift performing interdependent work with persons without disabilities. Only 56 were placed in mobile work crews, where groups of individuals worked in several locations determined by contractual arrangements with community businesses.

Data Sources

The following data sources were used in this analysis. Data on the characteristics and earnings of the 729 participants in the sample were obtained from data collected on a monthly basis from all program agencies receiving supported employment funds from the state of Illinois. The information management system allowed the funding agents to ascertain worker characteristics, reasons for separation, type and level of support, and benefit-cost information.

Data on the costs of supported employment incurred by the 30 agencies, the costs of the alternative programs in which the participants would have been placed, and the earnings of non-participants still in alternative programs were obtained from the Department of Mental Health and Developmental Disabilities (DMHDD), the Department of Rehabilitation Services (DORS), and the U. S. Department of Labor (Illinois branch).

Method of Calculating Benefits and Costs

The benefits and costs of supported employment were estimated from three perspectives - that of society as a whole, that of taxpayers as the major funding source, and that of supported employees as the most direct beneficiaries.

Social Perspective. Social benefits were measured by the increase in earnings of supported employees over what they would have earned in an alternative program, and the costs that would have been incurred if placed in an alternative program. A key assumption was that all participants would have been placed in an alternative program if not engaged in supported employment. One justification for this assumption is that almost all supported employment participants were selected out of alternative programs.

Benefits and costs were measured over each of the four years that the 30 programs were in operation as well as for the combined four-year period. To estimate increased earnings, the first step was to calculate the gross earnings of participants. For the 729 participants, earnings equaled $3,916,067 over the four years. The earnings listed in Table 2 were obtained by auditing actual Department of Labor income figures and checking these earnings against income figures reported by the 30 participating programs to either the Department of Mental Health and Developmental Disabilities or the Department of Rehabilitation Services.

[TABULAR DATA OMITTED]

The second step was to estimate the earnings that participants would have had in the absence of the supported employment program. It was assumed that participants would have had the same earnings as participants in the alternative-placement program from which the supported employment participants were selected. Average earnings currently reported by all alternative placement programs were used to estimate the probable alternative placement earnings for those few individuals who entered supported employment directly from school or without a reported previous placement. On this basis, it was assumed that supported employees in the sample would have earned $2,144,361. New increased earnings due to supported employment over the entire four-year period were therefore $1,771,706.

To estimate the savings in the costs of alternative placements, it was assumed that each supported employment participant would have incurred the same average costs as current clients in their previous placement. These costs were estimated from placements to these organizations from the Illinois Department of Rehabilitation Services and the Illinois Department of Mental Health and Developmental Disabilities, and by the local Community Mental Health 708 Boards. In those few cases involving participants who entered supported employment directly from school or without a reported previous placement, the average costs of clients in all alternative placement programs combined were used to estimate what would have been experienced by these participants. Total alternative program savings were estimated at $5,922,604. Table 3 lists costs associated with all alternate programs, as well as costs associated with supported employment (top row).

[TABULAR DATA OMITTED

The costs of supported employment programs were estimated by adding all reported payments by state and local agencies to the 30 providers and the tax savings to employers who made use of Target Jobs Tax Credits (TJTC) program for participants involved in these programs. Program costs for the 30 agencies were calculated on the basis of payments by the Illinois Department of Rehabilitation Services, the Illinois Department of Mental Health and Developmental Disabilities, and by local Community Mental Health 708 Boards. Total costs totaled $8,212,812 for the four-year period.

The TJTC program provided a federal income tax credit to employers of up to 40 percent of the first $6,000 paid to each eligible employee during his/her first year of employment. The taxes forgone by the federal government were $205,636 over the four-year period.

Taxpayer perspective. Taxpayer benefits were measured as the total income and payroll taxes paid by supported employees, reductions in public support, and saving in expenditures for alternative programs, as calculated above. All supported employment costs were borne by taxpayers. Benefits from the taxpayers perspective include taxes withheld ($543,740), reduction in government subsidies ($5,147), and savings in operational expenditures for alternate programs ($5,922,684). Tax payer costs were the same as those from society's perspective since participants did not incur costs.

Supported employee perspective. Total social benefits are divided between participants and taxpayers. The benefits to participants were estimated by subtracting taxes withheld ($543,740) and decreases in income support from public agencies ($5,147) from the estimated increased earnings of supported employees ($1,771,706). The net increase in personal, disposable income for supported employees came to $1,222,819 over the four-year period.

Tax payments were estimated from federal and state income taxes withheld, as well as FICA taxes (Social Security), as reported monthly to the Illinois Supported Employment Project by the 30 agencies. For most of the 729 individuals, this was the first time in their lives that taxes had been withheld; no taxes had been deducted in their previous program placement.

Savings in government subsidies were estimated by summing decreases in Supplemental Security Income (SSI), Social Security Disability Insurance (SSDI), Public Aid (PA) administered by the Illinois Department of Public Aid, Medicaid, and Aid to Dependent Children (ADC). Reductions were calculated for each individual by comparing monthly benefits received prior to entering supported employment with those received while in the program. These reductions may be understated because they are based on the amount participants received before entering supported employment and were not adjusted for cost-of-living increases. In addition, some participants receive government subsidies only after entering supported employment.

Results

After yearly benefits and costs had been aggregated from each perspective, net benefits and benefit/cost ratios were calculated. Net benefits were costs subtracted from benefits, with anything greater than $0.00 being considered economically efficient. Benefit-cost ratios greater than 1.00 indicated that supported employment in Illinois was cost-beneficial.

Benefit-cost Analysis from Society's Perspective

As reflected in Table 4, society received an average return of $0.91 for every $1.00 invested during the entire four-year period. Yearly benefit/cost ratios demonstrated that while costs exceeded benefits during the initial year (0.75), benefits began to increase more rapidly than costs during the second year (0.88) and continued to do so during the third year (0.90). By the fourth year, benefits had exceeded costs (1.09). Total benefits increased from $1,254,487 in the first year to $1,928,281 in the second year, $2,177,679 in the third year, and $2,333,943 in the fourth year. Increasing trends were observed in both the earnings attributed to participants enrolled in supported employment and alternate program savings.

[TABULAR DATA OMITTED]

Yearly costs continue to increase during the first three years ($1,632,229; $2,142,004; $2,367,409 respectively). These costs rose as a result of increased enrollments in the supported employment program. During the first year, costs totaled $1,668,752. By the second year, costs rose to $2,191,094 and $2,420,604 by the third year. In the fourth year, total costs began to show a decrease ($2,137,998) due to a reduction in costs per person. Overall, total costs of the supported employment program were $8,418,448.

Benefit-Cost Analysis

from the Taxpayer's Perspective

As reflected in Table 5, the taxpayer received an average return of $0.77 for every $1.00 invested during this four-year period. Benefit-cost ratios from the taxpayers' perspective increased steadily over the four-year period, however. The benefit-cost ratio rose from 0.66 in the first year to 0.78 by the second year. A slight decrease occurred during the third year (0.75), but increased by the fourth year (0.89). In terms of each of the benefit categories, taxes paid increased yearly as did alternate program savings. Reduction in government subsidies increased during; the first year ($26,138) and the second year ($23,61 1), but by the third year, the government had actually paid out an additional $5,066 in subsidies. A slight reduction ($5,754) occurred during the fourth year.

[TABULAR DATA OMITTED]

Benefit-Cost Analysis

from the Employee's Perspective

In terms of gross earnings, supported employees almost doubled their wages by participating in supported employment compared to earnings received while enrolled in their previous placement The average supported employee earned $1,785 during the first year, $1,793 the second year, $2,150 the third year, and $2,406 by the fourth year. Taxes paid gradually rose from $250 during the first year to $366 by the fourth year. Losses due to reduction in government subsidies were slight and amounted to only $5,147.

Economic efficiency was demonstrated throughout the four-year period from the perspective of the supported employee. She increased net earnings by an average of 42%. Total disposable income increased by 28% during the initial year, 30%, in the second year, 52%, during the third year, and 57% by the last year.

Discussion

An analysis of the benefits and costs of the supported employment program in Illinois, during the period under investigation, indicate that trends in supported employment have enabled it to emerge as an economically viable program in Illinois. An earlier analysis of supported employment in Illinois, projecting from data collected between 1978 and 1980, reported that while costs initially exceeded benefits, benefits would begin to exceed costs by the eight year (Schneider, Rusch, Henderson, & Geske, 1981). On the basis of subsequently collected data from this project, McCaughrin (1988) reported that benefits actually began to exceed costs by the fourth year of the project. Evaluations of supported employment in Virginia indicated that benefits began to exceed costs by the third year (Hill & Wehman, 1983; Hill et al., 1987).

It should be noted, however, that the current analysis may understate the true benefit-cost ratio by failing to take account of several benefit it variables. First, no adjustment was made for fringe benefits that supported employees may have received.

For example, some supported employees may have received medical benefits from employers. As another example, employers are required by law to match employee contributions to Social Security. These employer contributions are paid for out of the productivity of employees and could add over 7 percent to the measure of benefits attributable to supported employment.

Second, no adjustment was made for supported employees who may have terminated their supported employment activity and moved into unsupported competitive work. To the extent that this happens, there is a long term stream of increased earnings that were made possible by supported employment. If a significant number of supported employees make this move, this would dramatically increase benefit-cost ratios.

Third, savings in public income support were surprisingly small, a fact that probably results from the relatively low salaries of most supported employees. For example, the average earnings of supported employees in Illinois in the fourth year were less than half the amount that would cause a person to be declared ineligible for the Social Security Disability Income (SSDI) program. In addition, because of low salaries, some participants may have applied for, and received SSDI and Supplemental Security Income benefits after being accepted into supported employment programs, offsetting any reductions in public support resulting from supported employment.

In summary, from society's perspective, benefits showed an increasing trend throughout the four-year period under study with economic efficiency being demonstrated by the fourth year. From the taxpayer's perspective, while an increasing trend was found, the benefit-cost ratio still is less than one. However, a benefit-cost analysis of long-term workers indicates that supported employment eventually will provide economic benefits to taxpayers greater than cost. Supported employees, themselves, have benefitted financially from supported employment since the first year. Importantly, new workers who began in FY1990 were doing better financially than the 394 participants who entered in FY 1987, the first year of the ISEP benefit-cost analysis, indicating that supported employment programs are achieving their primary goal of substantially increasing the productivity of people with severe disabilities. A more in-depth analysis would reveal the factors contributing to the increasing efficiency of supported employment programs.

An unresolved issue is the extent to which supported employment program costs decrease over time. Currently, supported employment program expense reports from Illinois do not differentiate between the initial startup costs of placement and training from later, less expensive maintenance costs. Early model demonstration programs that began in the late 1970's and early 1980's were able to document that these initial costs greatly exceeded later maintenance costs for supported employees on the job (Rusch, 1986). Supported employees should, after placement and training, be more independent and integrated in their work environment so that less input and supervision are required by support personnel resulting in reduced program expenditures for long-term workers. Johnson and Rusch (1990) verified that supervision was reduced over time for individually-placed supported employees, but not for those in enclaves.

To identify the effectiveness of supported employment over time, we compared the earnings of participants who had enrolled in the 30 supported employment programs in the first year of their operation with workers who had enrolled in these programs during the fourth year. These two groups were designated long-term and short-term participants, respectively. There were 78 long-term participants and 144 short-term participants. The average earnings of long-term participants were $1,700 more than if he or she remained in the alternative placement. In contrast, the average earnings of short-term participants was only a little over $700 more than they would have earned had they remained in the alternate program.

Author Notes

(1.) This study was supported in part by the Illinois Department of Mental Health and Developmental Disabilities, Illinois Department of Rehabilitation Services, and the Governor's Planning Council on Developmental Disabilities of the State of Illinois (Grant number STILMIDOR-GPC-MHD938). The opinions expressed herein do not necessarily reflect the position or policy of the Illinois Department of Mental Health and Developmental Disabilities, the Illinois Department of Rehabilitation Services, or the Governor's Planning Council on Developmental Disabilities of the state of Illinois.

References

Conley, R.W. (1973). The economics of mental retardation. Baltimore: The Johns Hopkins University press. Conley, R.W., & Noble, J.H. (1990). Benefit-cost analysis of supported employment. In F. Rusch (Ed.) Supported employment: models, methods, and issues. (pp. 271-288). Sycamore, IL: Sycamore Publishing. Hill, M., Banks, P.D., Handrich, R., Wehman, P., Hill, J., & Shafer, M. (1987). Benefit-cost analysis of supported competitive employment for persons with mental retardation. Research in Developmental Disabilities, 8, 71-89. Hill, M., & Wehman, P. (1983). Cost-benefit analysis of placing moderately and severely handicapped individuals into competitive employment. Journal of the Association for the Severely Handicapped, 8, 30-32 Johnson, J.R., & Rusch, F.R. (1990). An analysis of the hours of direct training provided by employment specialists to supported employees. American Journal on Mental Retardation, 94, 674-682. Lagomarcino, T.R. (1990). Job separation issues in supported employment. In F. R. Rusch (Ed.) Supported employment: Models, methods, and issues (pp. 301-316). Sycamore, IL: Sycamore Publishing. McCaughrin, W.B. (1988). Longitudinal trends of competitive employment for developmentally disabled adults: A benefit-cost analysis. Unpublished doctoral dissertation, University of Illinois. McCaughrin, W.B., Rusch, F.R., Conley, R.W., & Tines, J. (in press). A benefit-cost analysis of supported employment in Illinois: The first two years. The Journal of the Multihandicapped Person. Noble, J., & Conley, R. W. (1987). Accumulating evidence on the benefits and costs of supported and transitional employment for persons with severe disabilities. Journal of the Association for the Severely Handicapped, 12, 163-174. Rusch, F.R. (1986) Competitive employment issues and strategies. Baltimore. Paul H. Brookes Rusch, F.R. (1990). Supported employment models, methods, and issues. Sycamore, IL: Sycamore Publishing. Rusch, R.R., Chadsey-Rusch, J., & Johnson, J.R. (1991) Supported employment: Emerging opportunities for employment integration. In L. Meyer, C. Peck, & L. Brown (Eds.). Critical issues in the lives of people with severe disabilities. (pp. 145-170). Baltimore. Paul H. Brookes. Schneider, K., Rusch, F.R., Henderson, R., & Geske, T. (1981). Competitive employment for mentally retarded persons; Costs versus benefits. Policy Paper Series: Doc. 8, pp. 63-82. Champaign, IL: Leadership Training Institute, Vocational and Special Education, University of Illinois.