Table of Contents | Previous | Next |
Introduction
This report presents interim results for the Minnesota site in the national Employment Retention and Advancement (ERA) project. Minnesota’s ERA program, which operated in Hennepin County, was designed to address the needs and barriers of long-term recipients of the state’s Temporary Assistance for Needy Families (TANF) program. The ERA program went beyond the typical services provided through the TANF program to provide more intensive employment and social services to recipients who had received cash assistance for a relatively long period without finding work.
This introduction provides background on the national ERA project and describes key components of the Minnesota program. It also describes the research design for the evaluation and characteristics of the sample.
Overview of the National ERA Project
In the wake of the 1990s welfare reforms, which made long-term welfare receipt much less feasible for families, policymakers and program operators have struggled to learn what kinds of services and supports are best able to assist long-term recipients, or those still left on the welfare rolls, find and keep jobs, and to help former recipients stay employed and increase their earnings. The ERA initiative was developed to increase knowledge of effective strategies to help both of these groups move toward stable employment and increased earnings.
The ERA project began in 1998, when the Administration for Children and Families (ACF) in the U.S. Department of Health and Human Services (HHS) issued planning grants to 13 states to develop new programs. The following year, HHS selected MDRC to conduct an evaluation of the ERA programs.1 From 2000 to 2003, MDRC and its subcontractor, The Lewin Group, worked closely with the states that had received planning grants, and with several other states, to mount tests of ERA programs. MDRC, The Lewin Group, and Cygnet Associates also provided extensive technical assistance to some of the states and program operators, since most were starting the project from scratch, with no proven models on which to build.
Ultimately, a total of 15 ERA experiments were implemented in eight states, including Minnesota. Almost all the programs target current or former recipients of TANF — the cash welfare program that mainly serves single mothers and their children — but the program models vary considerably. One group of programs targets low-wage workers and focuses on advancement. Another group of programs (including Minnesota’s) targets individuals who are considered “hard to employ” and primarily aims to place them in stable jobs. Finally, a third group of programs has mixed goals and targets a diverse population, including former TANF recipients, TANF applicants, and low-wage workers in particular firms. Some of these programs initiate services before individuals go to work, while others begin them after employment. Appendix Table A.1 describes each of the ERA programs and identifies its goals and target populations.
The evaluation design is similar in most of the sites. Individuals who meet ERA eligibility criteria (which vary from site to site) are assigned, at random, to a program group or a control group. Members of the program group are recruited for the ERA program (and, in some sites, are required to participate in it), whereas members of the control group are not eligible for ERA services. The extent and nature of the services and supports available to the control group vary from site to site, but it is important to note that in most sites the ERA program is not being compared with a “no services” control group. The random assignment process ensures that the two groups were comparable at study entry. Thus any differences in outcomes that emerge between the two groups during the follow-up period can be validly tested for the likelihood that they arose because of the program and not by chance variation.
The Minnesota ERA Program
Minnesota’s TANF program, called the Minnesota Family Investment Program (MFIP), was implemented in 1998.2 It consists of several key features: first, a requirement that all recipients work or participate in employment-related services or face financial penalties (sanctions); second, an earned income disregard that allows recipients to keep more of their benefits when they go to work;3 and third, a lifetime limit of 60 months on the receipt of cash benefits.4
Although many individuals have moved off welfare and into work since MFIP began, program administrators have found that a significant number of recipients have remained on the rolls for relatively long periods without working. To address the needs of this segment of the caseload that appeared most likely to reach their benefits time limit, the state legislature enacted the Local Intervention Grants for Self-Sufficiency in 2000, which distributed TANF funds to several counties to design appropriate services. The program implemented in Hennepin County (which includes the city of Minneapolis) was referred to as the Tier 2 program. In Minnesota, the ERA evaluation is a test of the Tier 2 program in Hennepin County.
MFIP in Hennepin County (Tier 1)
MFIP (or the Tier 1 program) in Hennepin County places a strong emphasis on immediate employment, requiring recipients of cash assistance to work or at least prepare for work if they are able. The program also provides case management, support services, and some retention and advancement assistance to help recipients find and keep jobs. Individuals who are eligible for assistance, as determined by county financial workers, are referred to one of 22 contracted providers — primarily nonprofit organizations — for employment services.5 At these providers, recipients are assigned to a case manager who conducts a basic assessment focused on the individual’s employment and education history. Recipients can then generally spend four weeks searching for jobs, sometimes attending a structured job readiness class. If they do not find work, they are assigned to an unpaid or volunteer position at an employer (where they do not receive wages but continue to receive their MFIP grant). For those who find jobs and remain in Tier 1, case managers offer some postemployment assistance. Failure to participate in required activities results in a sanction that lowers the MFIP grant by 10 percent in the first month and 30 percent thereafter.6Recipients are also eligible, although not required, to participate in education or training activities, such as General Educational Development (GED) or college classes, provided they are working at least 20 hours per week. Caseloads for Tier 1 case managers at the contracted providers vary, but typically range from 75 to 100 per worker.
The Tier 2 Program
The Tier 2 program was designed to build on the services provided through Tier 1, and thus the two programs shared some basic characteristics: mandatory participation in work or work-related activities (with sanctions for those who do not comply); a strong focus on employment; four weeks of job search followed by unpaid or volunteer work for those who do not find paid work; support services, including child care and transportation; and some assistance with retention and advancement for working participants. However, Tier 2 differed from Tier 1 in several key ways (see Table 1).
Services | Tier 2 Program | Tier 1 Program | |
---|---|---|---|
Employment-related services | Strong focus on immediate employment | Yes | Yes |
Four weeks of job search followed by community service for those who did not find jobs | Yes | Yes | |
Emphasis on placement in supported work positions | Yes | Limited | |
Case management services | Level of assessment | In-depth "family assessment" by qualified professional of potential barriers to participation and employment (e.g., health problems, domestic abuse, housing issues, and learning disabilities) as well as of employment history and job skills | Standard assessment conducted by MFIP staff focused on employment and educational history and job skills |
Caseload size | 25-30 | 75-100 | |
Referrals for services to address identified barriers to employment (including health care providers, community support groups, individual therapists or counselors) | Yes | Limited | |
Intensive engagement efforts (including frequent contacts with staff and home visits) | Yes | Limited | |
Support services, including child care and transportation | Yes | Yes | |
Assistance with retention/advancement during the earnings disregard period | Yes | Yes |
-
First, Tier 2 case managers worked with substantially smaller caseloads — 25 to 30 participants per worker — to facilitate greater attention to the unique circumstances of individuals and their families. It was felt that smaller caseloads would also allow the Tier 2 case managers to build better relationships with recipients, and even allow them the option of making home visits.
-
Second, the upfront assessment of individuals entering the Tier 2 program was more in-depth and covered a wide range of issues extending beyond those directly pertaining to employment, such as mental health problems, substance abuse, domestic violence, and housing crises. This assessment also explored the needs of recipients’ children and other family members and served as the basis for making referrals to appropriate social services.
-
Third, the Tier 2 program placed greater emphasis on referring individuals to special services to address potential employment barriers, including mental health or substance abuse treatment programs. Tier 2 also put greater emphasis on placing recipients who could not find work in supported employment positions, where they worked for a wage with participating nonprofit or for-profit employers, but also received supervision, training, and job coaching from program staff.
In short, the goals of the Tier 2 program were to better assess the barriers to employment of a segment of the TANF population in Hennepin County who had received welfare for a long time and had not worked recently, to address those barriers through referrals to a wider range of services, and to provide close monitoring and follow-up.
The Hennepin County Context
Hennepin County, an urban county that primarily comprises the city of Minneapolis, is the largest of Minnesota’s 87 counties, and its more than 1.1 million residents make up nearly a quarter of the state’s population. The county’s population is predominantly white, although there is a large foreign-born population. About 10 percent of the county’s residents are immigrants; Southeast Asians and Somalis make up the two major immigrant groups.7
Welfare caseloads fell by 26 percent in Minnesota between 2000 and 2005, from over 39,000 families to about 29,000 families.8 This rate of decline is much higher than that of the average state and may reflect the fact that caseloads fell in Minnesota somewhat more slowly than the national average during the late 1990s.9 In 2005, about 30 percent of the state caseload were in Hennepin County.10
The local economy of Hennepin County is fairly strong. In March 2005, for example, the unemployment rate in the county was 4.1 percent, compared with 5.0 percent for the state and 5.4 percent for the nation.11 In 1999, the poverty rate in the county was 8 percent.12
The Target Population
The Tier 2 program was targeted to individuals who, according to program records: (1) had been assigned to one of the Tier 1 employment service providers for 12 months or longer; (2) were unemployed and had not worked in the preceding three months; (3) were not participating in an approved education or training program; and (4) were not currently being sanctioned. Individuals who were “exempt” from participating in Tier 1 services because, for example, they had young children, were seriously ill or incapacitated, were caring for an ill or incapacitated family member, had a personal or family crisis, or were a domestic violence victim, were not included in the study.
Table 2 presents selected characteristics of the evaluation sample, calculated using data from MFIP administrative records.13 Most recipients were female and most were also black. Nearly half of the sample did not have a high school diploma at program entry, and most had been receiving TANF for at least two years. Nearly 20 percent of the sample did not speak English as their primary language, and 30 percent were not U.S. citizens. The majority (75 percent) of noncitizens were black, and about one-third were Somali (not shown).
Although low education levels and long stays on welfare can be barriers to work, the targeted sample also faced a range of other problems that may act as important barriers. These data were obtained from a baseline assessment form administered just prior to random assignment and are presented in the bottom panel of Table 3. As discussed below, a change in the random assignment process midway through the sample enrollment period led to a reduction in the number of sample members who filled out this baseline form. As a result, the data in Table 3 are available for 60 percent of the overall sample.
Characteristic | Total | |
---|---|---|
Gender (%) | Female | 93.2 |
Male | 6.8 | |
Race/ethnicity (%) | Hispanic | 2.2 |
Black, non-Hispanic | 67.8 | |
White, non-Hispanic | 16.3 | |
American Indian/Alaskan native | 6.8 | |
Asian | 5.2 | |
Other | 0.3 | |
Mixed race | 1.4 | |
Average age (years | 31.4 | |
High school diploma or higher a | 53.6 | |
Number of children (%) | 0 | 0.8 |
1 | 35.0 | |
2 | 32.6 | |
3 or more | 31.6 | |
Age of youngest child in household (%) | Less than 3 years | 38.4 |
3 to 5 years | 22.1 | |
More than 6 years | 39.5 | |
U.S. Citizenship (%) | 70.2 | |
AFDC/TANF receipt history b (%) | Less than 2 years | 30.4 |
2 to 4 years | 69.6 | |
Primary Language (%) | English | 83.0 |
Spanish | 1.1 | |
Russian | 0.3 | |
Vietnamese | 0.5 | |
Other non-English language | 14.8 | |
Unknown | 0.4 | |
Sample size | 1,692 |
SOURCE: MDRC calculations from administrative records from the State of Minnesota. NOTES: Rounding may cause slight discrepancies in calculating sums and differences. Unless otherwise stated, results are for sample members randomly assigned from January 2002 to April 2003. aThose having 12 or more years of education are considered to have a high school diploma. bThis measure goes back only 9 years before random assignment. |
Characteristic | Total | |
---|---|---|
Currently employed (%) | 14.9 | |
Average hourly wage (among those currently employed) ($) | 10.06 | |
Percentage working full time (32+ hours) (among those currently employed) (%) | 40.0 | |
Employment during the past 3 years (%) | Did not work | 16.9 |
Worked less than 6 months | 18.7 | |
Worked 7 to 12 months | 21.8 | |
Worked 13 to 24 months | 24.1 | |
Worked for more than 2 years | 18.4 | |
Current housing status (%) | Rent, public housing | 12.0 |
Rent, subsidized housing | 38.3 | |
Rent, other | 40.2 | |
Own home or apartment | 4.5 | |
Emergency or temporary housing | 1.8 | |
Other | 3.2 | |
Currently pregnant (%) | 7.4 | |
Body Mass Index (BMI) c (%) | Underweight | 1.6 |
Normal weight | 32.7 | |
Overweight | 31.7 | |
Obese | 34.0 | |
Alcohol dependence g (%) | 2.4 | |
Drug dependence g (%) | 2.2 | |
Severe domestic violence f (%) | 13.9 | |
Ever convicted of a felony (%) | 7.8 | |
Potential employment barriers (%) | Poor child health | 21.7 |
Activity limitation | 12.5 | |
Learning disabled d | 18.2 | |
Major depression g | 29.2 | |
Health problems b | 33.1 | |
Limited English ability | 19.9 | |
Any domestic violence e | 21.6 | |
At least one of the above | 71.6 | |
Sample size | 1,015 |
SOURCE: MDRC calculations from Minnesota's Baseline Assessment Data. NOTES: Rounding may cause slight discrepancies in calculating sums and differences. Unless otherwise stated, results are for sample members randomly assigned from January 2002 to April 2003. a Information is provided only for sample members who completed the Baseline Assessment Data.
b If self-rated health as "fair" or "poor."
c Based on BMI.
d Based on the Learning Needs Screening Tool, Washington State Department of Social and Health Services.
e Based on the Modified Conflict Tactics Scale, "Yes" to any abuse. This measure is calculated only for women who reported being in a relationship at some point during the previous year.
f Based on the Modified Conflict Tactics Scale, "Yes" to physical abuse. This measure is calculated only for women who reported being in a relationship at some point during the previous year, which was roughly 84 percent of the sample.
g Based on the Composite International Diagnostic Interview, Short Form. |
A significant fraction of the sample faced health-related obstacles to finding work and achieving self-sufficiency. Over one-third were obese, and one-third reported health problems — that is, self-reporting that they were in “fair” or “poor” health. In addition, nearly 30 percent of recipients met the diagnostic criteria for major depression in the previous year.14 About one-fifth reported having a child with an illness or disability that made it difficult for the parent to attend work or school. Domestic violence was also assessed among women who had been in a relationship in the preceding year. Over one-fifth of these women reported experiencing some form of abuse in the year before study entry, although rates of physical abuse or severe domestic violence were somewhat lower.15 About 20 percent of the respondents likely had a learning disability,16 and the same percentage reported having “limited English ability.” Finally, many recipients faced multiple barriers to employment. Among those who had at least one barrier, representing 72 percent of the sample, over half had two or more barriers (not shown). (See Box 1 for a look at how these barriers cluster together.)
Box 1A Closer Look at Employment BarriersAmong individuals who had at least one employment barrier, over half had two or more. Among those who had two or more barriers, were some combinations more common than others? The table below presents data for the Tier 1 and Tier 2 groups combined, showing associations among barriers. Each column represents a subgroup that has a given barrier and presents the fraction of that subgroup possessing each of seven barriers. Among people who reported poor child health, for example, 38 percent were also depressed. This rate is somewhat higher than that for the full sample (29 percent, shown in the far-right column), but it suggests that this group experiences depression at a rate similar to that of other groups. This method is used to describe the clustering of barriers. If a subgroup has a rate of barrier incidence that is substantially above average, that rate appears in bold type.
The results show that depression is strongly associated with having poor physical health. Among respondents who had activity limitations, for example, 53 percent also met the criteria for depression, a rate that is substantially higher than the overall average. Among those who were in poor health but had problems that were less severe than activity limitations, 47 percent were depressed. Both health and depression are also correlated, although less strongly, with having a learning disability. Depression also appears to be associated with domestic abuse. |
![]() [D] |
Consistent with the prevalence of employment barriers, the top row of Table 3 shows that over 80 percent of the sample reported that they were not working at the time they entered the study (see Box 2 for a look at how these barriers are associated with work). Since one criterion for inclusion in the sample is no recent work experience, it would be expected that no recipients would be working. Nonetheless, it is typical in welfare-to-work studies to find some employed individuals in programs targeted to nonworkers. This outcome most likely reflects a lag in TANF program data on the employment of cash assistance recipients.
How Hard to Employ Is the Target Population?
When considering the potential effects of the Tier 2 program, it is important to place the target population in context. The Minnesota ERA program is considered to be one that focuses on “hard-to-employ” individuals, and the sample does appear to face a high prevalence of barriers to work. But how hard to employ are they compared with other welfare recipients?
This sample can be put in context by comparing the prevalence of employment barriers and rates of employment among the sample members with those of other groups of recipients.17 On one hand, in terms of barriers, data from other studies suggest that the Minnesota sample is generally no more disadvantaged than broader populations of recipients. For example, a study of women receiving welfare in Michigan in 1997 found that 27 percent met the criteria for depression, 19 percent had poor health, 22 percent had a child with a health problem, and 15 percent had experienced severe domestic violence in the past year.18 These rates are very similar to those found in Table 3 and much higher than rates found in the general population.19 A more recent study of women receiving welfare across six states in 2002 also found similar rates of incidence for each of these barriers.20
On the other hand, the Minnesota ERA sample has lower rates of current employment (15 percent) than the six-state sample of recipients (34 percent) and a nationally representative sample of long-term recipients in 2002 (34 percent).21, 22 Although this difference is most likely due to the sample selection criteria, which focuses on the unemployed, it may also be due to differing levels of education across the samples. Just over half of the Minnesota sample have a high school diploma, compared with about 60 percent of the other two samples. Education level has been found to be a strong predictor of employment status in these other studies as well as for this Minnesota sample (see Box 2). Finally, a notable fraction of the sample (about 22 percent) are immigrants, many from Somalia and other African countries. These African immigrants, many of whom may be refugees, may face important and unique barriers to work, such as limited facility in English and the psychological effects of war trauma and torture, coupled with a reluctance to seek help for fear of deportation.23
In sum, the comparisons suggest that, on average, the Minnesota sample is certainly a disadvantaged group with little recent work experience and potentially important employment barriers. Nonetheless, although “hard to employ” relative to the other ERA sites, which target current recipients, former recipients, and low-wage workers who may never have received welfare, they do not appear to be significantly more disadvantaged than other samples of current recipients.
About the Evaluation
Research Questions
The ERA evaluation focuses on the implementation of the programs and their effects, or impacts. Key questions addressed in this report include the following:
-
Implementation. How did Hennepin County implement and operate the Tier 2 program? What services and messages did the program provide and emphasize? How did case managers spend their time?
-
Participation. As a mandatory program, did the Tier 2 program succeed in engaging a substantial proportion of individuals in services? What types of services did people receive? To what extent did the program increase the receipt of services above “normal” levels, as represented by the control (Tier 1) group’s behavior?
-
Impacts. Within the follow-up period, did the Tier 2 program, compared with the Tier 1 program, increase employment and earnings, provide greater employment stability and wage growth, and lead to better jobs?
The Random Assignment Process
Random assignment began in January 2002 and ended in April 2003. MFIP participants who met the study criteria were identified by the county’s administrative database and were contacted by mail and asked to attend a meeting with their Tier 1 case managers. At this meeting, recipients provided baseline data to an interviewer from the Wilder Research Center (WRC) — including information on their sociodemographic background, family circumstances, and physical and mental health. Recipients were then assigned at random either to enroll in the Tier 2 program or to remain in the Tier 1 program.
Beginning in October 2002, the process of random assignment was streamlined, primarily to expedite sample build-up. Instead of waiting for recipients to come to the Tier 1 office in person, those who met the study criteria based on the county’s administrative records were randomly assigned at that point. Those assigned to remain in the Tier 1 program (the control group) were notified by mail of their research status by their Tier 1 case manager, with whom they would continue to work. Those assigned to the Tier 2 program were notified by mail of their assignment to begin working with a new case manager from the Tier 2 program and were subsequently contacted by Tier 2 staff. WRC then attempted to collect baseline data by telephone for all who were randomly assigned. WRC was not able to obtain this information for all sample members, however. As a result, baseline assessment data are not available for about 40 percent of the full sample.24
The Counterfactual: What Is ERA Being Compared With?
Individuals who were randomly assigned to the control group — who represent the counterfactual for the study — were not informed about the Tier 2 program and were treated as though it did not exist. They were required to continue participating in the Tier 1 program in order to receive cash assistance.
As discussed earlier, Tier 1 and Tier 2 shared many basic features, such as mandatory participation in work-related activities, a strong focus on employment, and supportive services. Both Tier 2 and Tier 1 participants were eligible for services in accordance with the rules of programs offering TANF, food stamps, Medicaid, child care, and transitional child care and Medicaid benefits. Tier 2 went beyond Tier 1 in providing smaller caseloads for case managers, more in-depth assessments, and a stronger emphasis on referral to supported work or other services. Of course, recipients in Tier 1 could seek out these types of services either on their own initiative or through referrals by non-Tier 2 program staff. Both groups could also seek out services offered in the community through Workforce Investment Act One-Stop Centers, technical colleges, adult schools, and other education providers, and employment and training organizations.
Data Sources
The data sources for the analyses presented in the report are described below.
Baseline Data: Administrative Records and Baseline Assessment
Demographic and MFIP assistance data were obtained from administrative records for all sample members at the point of random assignment. Additional information was obtained from the baseline assessment form for 60 percent of the sample, including more demographic background and family information, as well as information on physical and mental health, domestic violence, and arrest history. These baseline data are used to describe the sample and to identify key subgroups for which program effects might differ.
Follow-up Data: Administrative Records
Effects on employment and earnings are estimated using automated unemployment insurance wage records data, and effects on public assistance are estimated using automated TANF and food stamp administrative records. One and a half years of follow-up data are available for all sample members.
Program Implementation Data
MDRC staff interviewed case managers, service providers, and program administrators from both Tier 2 and Tier 1. Information was collected about a range of issues, including: program goals; full-family assessments and referrals for special services; the nature of job search; education and training and support services; retention and advancement services; management philosophies and structure; relationships among organizations involved in the program; and enforcement of the participation mandate. Field research was conducted periodically between 2001 and 2003.
MDRC also drew on data collected from a two-week time study that was administered to Tier 2 staff. The time study, administered confidentially to protect the identity of case managers, collected detailed information on the nature of staff-client interactions and on the topics covered in these interactions. In addition, the study collected information on how case managers typically spend their time each day. The time study was administered from September 18 to October 1, 2003, and was completed by a total of 36 staff.
The ERA 12-Month Survey
Information on Tier 1 and Tier 2 sample members’ participation in program services and their employment, income, and other outcomes was gathered by the ERA 12-Month Survey, which was administered to a subset of the Tier 1 and Tier 2 groups approximately 12 months after random assignment.
Sample Sizes
A total of 1,692 single parents were randomly assigned to either the Tier 1 or the Tier 2 group between January 2002 and April 2003. A subset of the single parents who were randomly assigned between October 2002 and March 2003 also completed the ERA 12-Month Survey. The survey sample consists of 503 single parents, representing 76 percent of those who were selected to be interviewed. Additional analyses suggest that the survey sample is representative of the fuller evaluation sample. See Appendix E for more details.
Roadmap of the Report
This report focuses on the Tier 2 program’s implementation and impacts. The next section further describes the program and its implementation and is followed by a section that presents impacts on participation and service receipt; the concluding section presents impacts on employment, earnings, and public assistance receipt.
Table of Contents | Previous | Next |