TEENAGE PARENT DEMONSTRATION

Report on Costs of Education and Training Programs, Chapter IV

CHAPTER IV:
COST ANALYSIS RESULTS

CONTENTS OF CHAPTER

1. Introduction
2. Total Resource Costs and Costs by Program Component
3. Unit Costs
4. Cost Estimates as Input to Benefit/Cost Analysis

INTRODUCTION

To deliver the services described in Chapter II, the three demonstration programs and other agencies that served participants incurred substantial costs.  Our analysis indicates that the value of all resources used by the three programs to serve their participants during fiscal year 1989 ranged from almost one million dollars to just over two million dollars.⁽¹⁾  On average, the full resource costs of the site programs were between $206 and $344 for each month that a sample member spent on AFDC and was thus eligible for program services, and between $498 and $920 for each month that a sample member spent involved in a major activity -- education, training, or a job.  These figures translate into an annual resource cost per participant, estimated for fiscal year 1989, ranging from $1,730 to $3,130.  The full resource cost of serving participants who took part in education, training, or employment ranged from just over $3,000 to about $5,400 during fiscal year 1989.  Since between 25 percent and 50 percent of these costs were assumed by other organizations (providing in-kind services), direct costs to the programs were significantly less.  Annual per person costs paid by federal and state agencies (program-paid costs) were between $917 and $2,399, and between $1,636 and $4,067 for participants involved in major activities.

The considerable range in costs among the three programs can be attributed to differences in program scale, the variety and intensity of services used, overall program structure, and rates of participation in program activities.  In Section A of this chapter we present data on the cost of the site programs, broken down by components or service categories, and discuss the relationship between cost and program characteristics.  In Section B we present the calculated unit costs of the programs.  In Section C we discuss how the results presented in this report will contribute to a future benefit-cost analysis.

TOTAL RESOURCE COSTS AND COSTS BY PROGRAM COMPONENT

The cost of services provided by the Teenage Parent Demonstration programs varied significantly across sites.  As shown in Table IV.1, total program-related costs in fiscal year 1989 were estimated at $764,001 for the Newark Teen Progress site, $1,049,765 for the Camden Teen Progress site, and $1,964,002 for Project Advance in Chicago.  Between 53 percent (Chicago) and 76 percent (Camden) of these resource costs were supported by direct program expenditures.  The Chicago program relied most heavily on in-kind services provided by community, state, or federal agencies.  However, the vast majority of education and training services in all three sites were provided to the programs by other agencies at no cost.

SOURCE: New Jersey Department of Human Services Demonstration Expenditure Reports; Illinois Department of Public Aid Obligation Transaction Reports; interviews with demonstration site managers.

Differences in resource costs arise most obviously from differences in the scale of demonstration operations.  Over the full course of the demonstration, the Chicago site enrolled 1,439 teenage parents in the enhanced-services group, compared to 633 in Camden and 575 in Newark.  During fiscal year 1989, the study year for the cost analysis, 1,138 members of the overall Chicago participant group received AFDC at some time and were thus eligible for demonstration services at some point in the year, compared to 337 in Camden and 288 in Newark.⁽²⁾  Much of the variation in overall cost is due to these program scale differences.  The Chicago site enrolled 2.3 and 2.5 times as many participants throughout the demonstration as did the Camden and Newark sites, and used resources in fiscal year 1989 valued at 1.9 and 2.6 times the resources used by the Camden and Newark sites, respectively.

The three demonstration sites provided a similar range of services, but they varied in their allocation of resources to major program components -- case management, workshops, education, job training, support services, and job placement.  Their allocation of resources depended on several factors -- service accessibility, participant characteristics, program strategy, and program management.  The accessibility of particular services clearly can affect the portion of overall program resources devoted to these services.  For example, stringent entrance requirements for job training will restrict access and constrain the share of program costs spent on training.  Participant characteristics, such as the age distribution of enrollees, can affect the number of participants attending or returning to public school, and thus determine the demand for other education or training resources.  Conscious decisions about program strategy can affect cost patterns as well; a decision to conduct very brief initial workshops, for example, is likely to make workshops a less significant aspect of program costs.  A decision to emphasize intensive counseling efforts to find suitable child care arrangements can result in both higher staff costs related to child care and higher utilization of child care subsidies.  Site program management can also affect the distribution of costs.  Tighter monitoring of participant attendance may lead to greater need for case management or clerical support staff, and may increase the percentage of participants who are active in major activities, both of which may result in higher service costs.

In order to distinguish the contributions of different services to overall cost, we allocated all costs to the major program components identified earlier -- case management, workshops, education, job training, support services, and job placement assistance.  Costs associated with site management and supervision and central management, both essential program components, were kept separate rather than grouped together under one "management" category.

Many program costs could clearly be associated with specific components.  The costs of most outside staff or agency services could be readily allocated because they were engaged for a specific purpose, such as running program workshops or providing job training.  Non-labor costs, such as child care payments, were reported at a level of detail that allowed direct allocation to program components.

However, allocating the remaining costs to the major program components required some judgment.  Most importantly, it was necessary to allocate the cost of program staff among the different components according to their job functions.  Salaries and fringe benefits were not reported by the states separately for individual staff persons but for the site as a whole.  Since staff were recruited at different times and in some instances had multiple responsibilities, we had to identify in which parts of fiscal year 1989 each staff person was employed, how each person's time was spent, and their salaries, in order to decompose total salary cost into the service categories.  We used information provided by site managers on employees' periods of employment and their specific roles, along with salary and fringe benefit rate information included in the States' demonstration proposal budgets, to distribute the cost of each staff person's time across component categories.  General site expenditures, such as those for office supplies, furniture, and indirect costs, were allocated to case management services and site supervision in proportions reflecting the approximate distribution of labor costs.

The distribution of site resource costs by service components shows that there were significant differences in program emphases, even when divergences in program scale and the magnitude of overall program costs are taken into account (Table IV.2 and Figure IV.1).⁽³⁾  Most notably, the Chicago site devoted far more resources to job training, even relative to program size, than did the New Jersey sites; Project Advance enrolled more than twice as many demonstration participants as did the Teen Progress sites, but the estimated value of training resources used by Chicago participants was significantly more than twice as high as those used by participants in Camden and Newark.  On the other hand, Chicago site costs were lower, even in absolute terms, than the New Jersey sites for program workshops, reflecting the use of brief three-day initial workshops in Chicago compared to the extensive sequences of initial workshops in Camden and Newark.  Site program costs for support services were also estimated to be lower in Chicago than in the New Jersey sites; Project Advance used 0.8 and 1.7 times the resources for support services as the Camden and Newark sites, despite operating on a general scale at least twice the size of the Teen Progress programs.  This difference most likely arises from the heavier investments in the New Jersey sites in child care counseling staff, and their policy of consciously promoting use of subsidized child care options, but could be due to potential underestimation of support service payments in Chicago, as noted earlier.  Case management resources in Chicago were about twice those devoted to case management in the New Jersey sites, approximately corresponding to differences in overall number of participants throughout the demonstration.  However, case management costs in Chicago were lower than in the other sites when standardized for the sample members' tenures on AFDC.

SOURCE: New Jersey Department of Human Services Demonstration Expenditure Reports; Illinois Department of Public Aid Obligation Transaction Reports; interviews with site managers; Illinois FY 1989 grant renewal application.

SOURCE: New Jersey Department of Human Services Demonstration Expenditure Reports; Illinois Department of Public Aid Obligation Transaction Reports; and interviews with demonstration program staff and other providers.

UNIT COSTS

To provide a useful basis for projecting the full resource costs of programs similar to the Teenage Parent Demonstration in other jurisdictions, and to provide input to the benefit-cost analysis, it is necessary to standardize costs by constructing unit measures that express costs in relation to the size of the program or the extent of program participation.  We have defined two unit cost measures: (1) the average cost of site services for each person-month of AFDC receipt; and (2) the average monthly cost of site services provided to individuals engaged in major activities (education, training, and employment).⁽⁴⁾  Each of these measures, given additional information on the average duration of AFDC receipt and major activity, can be converted into estimates of average cost per person.

Cost Per Person-Month on AFDC

Expressing resource costs in relation to the size of the teenage parent caseload on AFDC is a relevant standardized cost measure because of the way in which the demonstration defined the population to be served.  The demonstration offered services to teenage parents and required their participation only while they were receiving AFDC.  The extent to which teenage parents remained on AFDC affects the size of the ongoing caseload of the program, and variations in the rate of departure from and return to AFDC can affect the overall level of demand for program services.⁽⁵⁾

To standardize costs in relation to the extent of AFDC receipt among the demonstration caseload, we estimated average resource cost per person-month on AFDC, by dividing total site costs by total person-months on AFDC.  Using data on AFDC receipt obtained from the two states' AFDC payment system files (FAMIS in New Jersey and CIS in Illinois), we determined the number of months of AFDC receipt during fiscal year 1989 for every demonstration participant who had completed Teen Progress or Project Advance intake before the start of the fiscal year or who completed intake sometime during the year.  We then summed the number of months on AFDC during the fiscal year across all participants to obtain total person-months on AFDC.  We could then divide resource cost estimates by total AFDC months to estimate average cost per person-month on AFDC.

Cost per person-month of AFDC receipt in effect answers the question: How much do program services cost for each month of a teenage parent's stay on AFDC? This cost was $344 in Camden, $292 in Newark, and $206 in Chicago (Table IV.3).  Average direct expenditures by the programs for each person-month on AFDC were significantly less, since many services were provided in-kind; program-paid costs per person-month on AFDC were about $261 in Camden, $207 in Newark, and $109 in Chicago.

SOURCE: New Jersey Department of Human Services Demonstration Expenditure Reports; Illinois Department of Public Aid Obligation Transaction Reports; New Jersey and Illinois public assistance eligibility systems (FAMIS and CIS).

^a  These figures assume that 23 cases that were unable to be matched with state AFDC records data for FY 1989 were not receiving AFDC during that year.  This assumption potentially biased upward the cost estimates slightly.  For Camden, the bias could be up to five percent; the maximum impact on the cost estimates of the other sites is less than one percent.

During the study year some teenage parents were just entering the demonstration, some were participating in major activities, and others were sanctioned for noncompliance or deferred for medical reasons.  The cost of serving individuals in different stages of their program participation undoubtedly varies, but these estimates represent average cost for a caseload reflecting all of these stages of participation.

These results can be useful in program planning in other contexts.  Since different jurisdictions are likely to have varying degrees of success in finding and relying on community or other in-kind services, average resource cost per person-month on AFDC is more relevant than the program-paid monthly cost figure for projecting maximum costs in other sites.  The resource cost per person month, in conjunction with estimates of the average expected caseload of AFDC teenage parents, can be used by program planners as a basis for estimating the annual cost of offering a program such as the demonstration sites provided, assuming comparable services, resources, and participation rates in major program activities.

There were clear differences in program emphases among the three demonstration programs, some of which were discussed earlier, and these differences are highlighted when costs for particular service components are expressed as unit costs (Table IV.4).  Project Advance case management costs per person-month of AFDC were just over half of case management costs in the New Jersey Teen Progress sites; this finding is consistent with the observation of higher staff caseloads in Chicago.  Chicago workshop costs of $2 per person-month of AFDC receipt were a small fraction of workshop costs in Camden or Newark ($35 and $28, respectively), which largely reflects the use of extensive initial workshops in New Jersey and brief three-day initial workshops in Chicago.  Although the Chicago site offered a variety of workshops for ongoing participants, these were brief workshops offered on a limited schedule, and did not consume substantial resources.

SOURCE: New Jersey Department of Human Services Demonstration Expenditure Reports; Illinois Department of Public Aid Obligation Transaction Reports; New Jersey and Illinois public assistance eligibility systems (FAMIS and CIS).

NOTE: Detailed information on rates of participation in program services can be found in Gleason et al. (1992).

Unit education expenditures varied in the three sites, but were highest in Camden.  This pattern reflects the fact that Newark and Chicago relied primarily on community college GED classes.  In Camden, heavy use was made of the Urban Youth Corps program, which offered not only small-group remedial classes but also work experience activities, at an estimated average monthly cost of

$667 per person, far above the estimated average monthly cost for community college remedial classes of $250.

Project Advance costs per AFDC person-month for job training were substantially higher than unit training costs in the New Jersey sites -- $70 compared to just over $40 -- primarily because participants at the Chicago site were far more likely to have entered job training.  In Chicago, 33 percent of all teenage parents who entered the program took part in some job training, compared to about 18 percent in the New Jersey Teen Progress sites.

Support service resource costs, on the other hand, were considerably lower in Chicago than in the New Jersey sites.  This difference appears to stem from staffing differences as well as differences in subsidy payment practices.  Whereas both New Jersey sites had a full-time child care counselor, the Chicago site relied on case managers to arrange child care.  In addition, the Chicago site followed a general policy of arranging and paying for child care only when participants stated a need for help.  The New Jersey staff were more likely to suggest subsidized child care alternatives.

Project Advance also had considerably lower resource costs per person-month on AFDC for placing participants in training and jobs.  This difference in unit costs between the Chicago and New Jersey sites is due more to the greater number of total months on AFDC in Chicago, than to higher total costs for training and job placement services.  All three sites had either an on-site job and training counselor (Camden and Newark) or an employment specialist (Chicago) and thus had roughly comparable total costs.  However, the substantially higher caseload in Chicago resulted in lower average unit costs for training and job placement functions, compared to the New Jersey sites.

The resource cost per person-month of AFDC receipt for site management and supervision was also lowest in Chicago, although unit-expenditures on these services in Newark were close to those in Chicago.  Like the divergence in training and job placement costs, this difference reflects simple economies of scale -- the fact that there were at least twice as many participants in the Chicago program (and about three times as many person-months on AFDC) than in Camden and Newark.  Although all three sites incurred the costs of a site manager and case management supervision, the New Jersey demonstration programs were smaller and participants spent less time on AFDC.

Cost Per Person-Month of Major Activity

Average resource cost for each person-month on AFDC reflects a composite of costs for individuals in very different stages of activity, and for individuals who ultimately utilize very different levels of program resources.  Costs are likely to be highest when participants are attending full-time job training or educational classes, or are employed.  During these periods, they are likely to be receiving child care subsidies and are using the resources of education and training providers, in addition to receiving case management services.  Participants who never enter these major activities consume some of these same program resources, of course -- for intake and assessment, for program workshops, for case managers' efforts to guide them into major activities, and to some extent for support service payments while they are involved in these activities.  However, the rate at which participants take part in major program activities is a very significant factor in the overall cost per participant.

We, therefore, estimated the average monthly cost of services to participants taking part in education, training, or employment.  This cost per month of major activity is the sum of two components: (1) the average monthly cost of education, training, and support services for people in such courses or employment; and (2) the average monthly cost of other services such as case management, workshops, placement in jobs and training, other counseling, and site supervision and management, services which are provided to both participants in major activities and those who are not.  To estimate the first component, we computed the total resource costs of education, job training, and support service payments, and divided this total by the number of person-months spent in education, training, or employment in fiscal year 1989.⁽⁶⁾  To estimate the second component, we divided all remaining costs by the total number of person-months on AFDC in fiscal year 1989, because individuals involved in major activities are assumed to use other program resources at approximately the same rate while they are in major activities as during other periods when they are not.

This second unit cost measure answers the question: What is the average monthly cost of serving teenage parents while they are involved in education, training, or employment? This cost estimate allows planners to predict how program costs might change if, due to variation in availability of services or characteristics of client populations, they achieved different levels of participation in these major activities.  The average monthly resource cost for those in major activities was $920 in Camden, $560 in Newark, and $498 in Chicago (Table IV.5).  As shown in Table IV.5, differences across sites are somewhat greater for education/training/support payment costs than for the costs of other program services (case management, workshops, counseling on child care, job and training placement, and general site supervision and management) during periods of major activity.  When only direct program expenditures are considered, differences across sites are even more pronounced.  The average overall costs per month of major activity paid by the programs were $699 in Camden, $398 in Newark, and $264 in Chicago.

SOURCE: New Jersey Department of Human Services Demonstration Expenditure Reports; Illinois Department of Public Aid Obligation Transaction Reports; New Jersey and Illinois public assistance eligibility systems (FAMIS and CIS); demonstration participant tracking systems (Teen Progress Management Information System and Project Advance Case Management System).

^a  Education, training, and employment costs include the costs of education and training courses on-site and off-site, and all support service payments for child care and transportation.  Since participants could receive these payments to attend workshops or meet with program staff outside periods of education, training, or employment, this approach overstates the portion of program costs associated with major activities.

^b  Other costs include only the costs of case management, workshops, training and job placement, and site management and supervision.

Cost Per Person

Costs per person-month of AFDC receipt, and per person-month of major activity, can be converted to average costs per person, based on the average duration of participants' stays on AFDC and in education, training, or job.  Based on the average duration of stays on AFDC in fiscal year 1989, the resource cost per person for one year of the demonstration services averaged $3,130 in Camden, $2,657 in Newark, and $1,730 in Chicago (Table IV.3).  For those individuals who were involved in education, training, or employment at any time during that year, average per-person resource costs were $5,351 in Camden, $3,726 in Newark, and $3,087 in Chicago (Table IV.6).  Annual per person costs to federal and state agencies (program-paid costs) were substantially less, ranging from $917 to $2,379 per year of program eligibility (AFDC receipt) and from $1,636 to $4,067 per year for persons in major activities.

Two factors should be noted when interpreting these estimates of average costs per person.  First, the measures of the lengths of stay on AFDC and in education, training, or employment that went into these calculations are based on observation of durations only within a particular year of the demonstration, rather than a complete observation of all enrollees through an extended follow-up period.  Thus, the maximum length of stay on AFDC or in major activities for each participant is artificially constrained to 12 months and our cost per person figures pertain to one year of program participation only.  The average cost per participant would be higher than the annual per-person costs estimated because many participants stayed on AFDC and used program services for longer than one year.  Second, the variation in estimated unit costs across the programs should not be viewed as evidence of which sites had more cost-effective programs or higher quality services.  The cost results are at this point not accompanied by any estimates of long-term program impacts.  It is possible that higher levels of program resources per person will yield greater impacts on key outcomes.

SOURCE: New Jersey Department of Human Services Demonstration Expenditure Reports; Illinois Department of Public Aid Obligation Transaction Reports; New Jersey and Illinois public assistance eligibility systems (FAMIS and CIS); demonstration participant tracking systems (Teen Progress Management Information System and Project Advance Case Management System).

^a  Education, training, and employment costs include the costs of education and training courses on-site and off-site, and all support service payments for child care and transportation.  Since participants could receive these payments to attend workshops or meet with program staff outside periods of education, training, or employment, this approach overstates the portion of program costs associated with major activities.

^b  Other costs include only the costs of case management, workshops, training and job placement, and site management and supervision.

^c  This row reports average months on AFDC in FY 1989 for individuals who participated in education, training, or employment.  Their durations on AFDC in that year were similar to the durations for the overall sample as reported in Table IV.3.

Comparison with Costs of Other Programs

The costs estimated for the Teenage Parent Demonstration can usefully be compared to estimated costs for other teenage parent programs and employment training programs.  In the early 1980s, Project Redirection, sponsored by the Manpower Demonstration Research Corporation, provided comprehensive educational, health, employability, family planning, parenting, and life management training services to over 800 teenage parents in programs run by community organizations in Boston, New York, Phoenix, and Riverside, California.  Program costs were estimated at $3,893 per service year, or $3,536 per participant (Branch, Riccio, and Quint, 1984).  When adjustments are made for inflation, Project Redirection would thus appear to have cost approximately 60 percent more than the Teenage Parent Demonstration across all three sites.⁽⁷⁾  This difference is particularly striking because, unlike our estimates of the Teenage Parent Demonstration service costs, the estimates for Project Redirection did not include the implicit costs of in-kind services provided by other agencies (for example, the in-kind costs of off-site GED classes or job training).  This substantial difference may be due to the smaller scale of Project Redirection and the likelihood that it's sites incurred high fixed costs relative to caseloads.  The four programs enrolled a total of 807 teenage parents over a 30-month period, compared to the total of 2,647 enrolled in approximately the same length of time by the Camden, Newark, and Chicago sites of the Teenage Parent Demonstration.  One probable indicator of these high fixed costs is the fact that Project Redirection sites reported that half of their program costs were for program management as opposed to other major cost components -- program services, ancillary services, stipends to participants, and payments to community mentors.

The estimated costs of the Teenage Parent Demonstration are less than or roughly consistent with the costs estimated for other demonstration programs offering comprehensive services to adults.  For example, the Minority Female Single Parent (MFSP) Demonstration, sponsored in four community-based organizations by the Rockefeller Foundation, offered GED and ABE classes, job training, child care, and counseling to minority single mothers.  At these four sites, average cost per enrollee ranged from $2,679 to $3,861 in fiscal year 1986, and average cost per participant who entered education or training ranged from $3,909 to $6,147 (Handwerger and Thornton, 1988).  These cost estimates are not only higher, in real terms, than the costs of the Teenage Parent Demonstration programs, but also represent durations of enrollment that are substantially less than in these programs (approximately 6 months for MFSP compared to 9 months for TPD).⁽⁸⁾  Costs for the Teenage Parent Demonstration are also roughly comparable to the costs of the comprehensive work-welfare demonstrations of the 1980s.  Some of those programs offered job training as well as basic education and job search assistance, and thus provided a range of services similar to the Teenage Parent Demonstration, albeit for a differently defined population.  Cost per participant in these programs was estimated at $2,089 (Maxfield, 1990).⁽⁹⁾

COST ESTIMATES AS INPUT TO BENEFIT/COST ANALYSIS

The cost estimates presented here are a partial step towards the information needed for the benefit/cost analysis.  To compare program benefits and costs, we must be able to quantify the incremental cost of the demonstration services provided to members of the program sample as compared to services received by the control group, over a consistently defined follow-up period.  Using demonstration participant tracking data, we can estimate the costs of only those services that the enhanced-services program sample received from or through the demonstration program while they were receiving AFDC.  Program data do not allow us to identify services received from other sources.  For the benefit/cost analysis, therefore, we will supplement estimates of demonstration program costs for the enhanced-services program sample with more comprehensive estimates of service utilization, based on follow-up interviews.  These total service costs will be compared to the estimated cost of services reported by control group members in their follow-up interviews.

[End of Chapter IV]

ENDNOTES

1.  As described in the previous chapter, site resource costs include both program-paid costs and estimates of the costs of in-kind services, but exclude central management costs.  [Back to text]

2.  The numbers of participants who received AFDC in FY 1989 are lower than the overall number of participants in the demonstration programs for two reasons.  First, some participants who had enrolled in the demonstration before FY 1989 were no longer receiving AFDC by July 1988 (the start of the fiscal year).  Second, additional sample members were enrolled in fiscal years 1990 and 1991.  [Back to text]

3.  The allocation of the line-item costs in Table IV.1 to program service components is presented in detail in Appendix A.  [Back to text]

4.  Central management costs, the difference between total program costs and site costs, were excluded from the calculations of unit measures.  [Back to text]

5.  As noted earlier, of course, participants could obtain services similar to some of those provided by the demonstration during periods when they were off AFDC.  The extent of such activity cannot be determined from the demonstration program data, and the cost of those non-demonstration services is not included in this analysis.  Estimates of non-demonstration service costs will be based on follow-up survey data and will be included in the benefit-cost study.  [Back to text]

6.  Support service payments were allocated entirely to major activities, although child care and other transportation payments were made during periods when participants were not in major activities (e.g., to attend program workshops).  [Back to text]

7.  For example, cost per participant-year, in 1993 dollars, was $5,709 for Project Redirection and $2,425 for the Teenage Parent Demonstration (based on the Gross Domestic Product implicit price deflator).  [Back to text]

8.  This difference may be due in part to the fact that the MFSP sites achieved considerably higher rates of participation in education and training than did the Teenage Parent Demonstration sites.  In the year studied for the MFSP cost analysis, sample members were active in education or training for 53 percent of the months they were enrolled.  In the Teenage Parent demonstration in the cost study year, participants were reported active in education, training, or employment for 34 percent of the months in which they were on AFDC and thus eligible for demonstration services.  [Back to text]

9.  These programs included the Virginia, Baltimore, and Maine WIN demonstrations, the San Diego Saturation Work Incentive Model, and the San Jose-based Center for Employment Training (one of the sites also included in the Minority Female Single Parent Demonstration).  [Back to text]

Where to?

[ Top | Table of Contents | Cover Page | Acknowledgments | Endnotes | References for this report | References and Other Reports  ]

[ Teenage Parent Demonstration:  Home Page | Timeline | ASPE Home Page | DHHS Home Page ]