| Connecticut’s Jobs First program is a
statewide welfare reform initiative that began operating in
January 1996. Jobs First was one of the earliest statewide
programs to impose a time limit on welfare receipt: Families
are limited to 21 months of cash assistance unless they receive
an exemption or extension. The program also includes generous
financial work incentives and requires recipients to participate
in employment-related services targeted toward rapid job placement.
Jobs First was initiated under waivers of federal welfare
rules that were granted before the passage of the 1996 federal
welfare law (the Personal Responsibility and Work Opportunity
Reconciliation Act, or PRWORA); how the program fares over
time may provide important lessons on the likely results of
welfare reforms implemented in other parts of the country
in response to the federal law.
This report has been prepared as part of a
large-scale evaluation of Jobs First being conducted by the
Manpower Demonstration Research Corporation (MDRC). The evaluation
is funded under a contract with the Connecticut Department
of Social Services (DSS) — the agency that administers
Jobs First — and with support from the U.S. Department
of Health and Human Services, the Ford Foundation, and the
Smith Richardson Foundation. The study focuses on two welfare
offices — Manchester and New Haven — which together
include more than one-fourth of the state’s welfare caseload.
MDRC is a nonprofit, nonpartisan organization with a quarter
century’s experience designing and evaluating programs
and policies for low-income individuals, families, and communities.
This is the third publication in the Jobs
First evaluation. The earlier reports, completed in 1997 and
1998, examined the implementation of Jobs First during its
first two years of program operations. This report updates
the implementation story, and also includes the first information
about Jobs First’s impacts — that is, the
difference Jobs First makes relative to the outcomes generated
by the welfare system that preceded it. To facilitate this
assessment, between January 1996 and February 1997 several
thousand welfare applicants and recipients (most of them single
mothers) were assigned, at random, to one of two groups: the
Jobs First group, whose members are subject to the
welfare reform policies, and the Aid to Families with Dependent
Children (AFDC) group, whose members are subject to the
prior welfare rules. Because people were assigned to the groups
through a random process, any differences that emerge between
the two groups over time — for example, in employment
rates or average family income — can reliably be attributed
to Jobs First.
The report follows early enrollees in the
two groups for up to two and a half years, slightly beyond
the point when Jobs First group members began reaching the
time limit. The study’s final report, scheduled for 2001,
will follow all members of the groups for up to four years,
and will be accompanied by a separate document describing
the impacts of Jobs First for children.
Highlights
of the Findings
Key findings from this report include the following:
- The main features of Jobs First were
successfully put in place in the research sites, but the
program has not been implemented very intensively. Survey
data show that Jobs First group members heard a more employment-focused
message from the welfare system than did AFDC group members.
In addition, a large majority of Jobs First group members
were aware of the key program features: the time limit and
the financial work incentives. Finally, the Jobs First group
was more likely than the AFDC group to participate in employment-related
activities, particularly activities aimed at rapid job placement.
At the same time, owing to start-up problems and certain
features of the program design (for example, limited contact
between staff and clients), recipients’ participation
in employment-related activities was not closely monitored,
and some aspects of the program message were not strongly
reinforced.
- Most Jobs First group members did not
reach the time limit within two and a half years after enrollment.
Of those who did, about half were granted an extension.
Most of those whose cases were closed at the time limit
were employed. About two-fifths of the Jobs First group
reached the time limit within two and a half years after
enrollment; the others still had months remaining because
they had left welfare or were temporarily exempted from
the time limit. Over half of those who reached the time
limit and attended a time limit review meeting had income
below the welfare payment standard (the maximum grant for
their family size) at that point, and almost all of them
were granted at least one six-month extension. Conversely,
most of those whose cases were closed at the time limit
were working and had income above the payment standard.
Overall, roughly one-fifth of Jobs First group members’
cases were closed because of the time limit within the two-and-a-half-year
follow-up period.
- Jobs First increased employment rates
and earnings throughout the follow-up period; impacts were
particularly large for the least job-ready clients.
Just under 82 percent of Jobs First group members were employed
at some point within two and a half years after enrollment
compared with 74 percent of the AFDC group. In addition,
Jobs First group members’ total earnings were about
11 percent higher. But the averages mask an important pattern:
Jobs First nearly doubled the employment rate for those
facing multiple barriers to employment — a group that
was not targeted for services prior to Jobs First —
and generated almost no increase in employment or earnings
for the most job-ready (although it did increase welfare
receipt for the latter group).
- In the first part of the study period,
Jobs First substantially increased both welfare receipt
and family income; as individuals began to reach the time
limit, the program began to reduce welfare receipt and the
income gains diminished. As expected, the Jobs First
group received more welfare than the AFDC group in the early
part of the study period; the financial incentive allowed
many working families to continue receiving benefits. Since
Jobs First group members also had higher earnings, their
combined income from public assistance and earnings was
substantially higher than that of the AFDC group; they were
also more likely to have savings and to own cars. The pattern
changed abruptly when they began reaching the time limit:
In the last part of the study period, the Jobs First group
received substantially less welfare than the AFDC
group, and the lower welfare benefits began to offset the
Jobs First group’s higher earnings. Thus, in the last
three months of the follow-up period, the two groups had
about the same total income. The Jobs First group, however,
derived a greater share of its income from earnings —
a key goal of the program. Data on participants’ income
brackets during this three-month period suggest that Jobs
First caused some families to be worse off financially and
other families to be better off than they would have been
without the program.
Although these findings are encouraging in
many respects, it is too early to draw any firm conclusions
about how Jobs First will ultimately affect eligible families
or government budgets. The longer-term picture will probably
continue to improve from the budgetary perspective because
the program started to reduce public assistance spending after
families began reaching the time limit. But the future is
more uncertain for participants, given the income trends that
emerged at the end of the follow-up period.
The
Policy Context of Jobs First
Between 1993 and mid 1996, more than 40 states
were granted waivers of federal AFDC rules that enabled them
to implement a variety of measures designed to increase employment
and self-sufficiency among welfare recipients. Although the
1996 federal welfare law made major changes in the structure
and funding of public assistance programs, most of the specific
policies that the law encourages states to adopt were already
being implemented as part of state waiver initiatives. For
example, while the 1996 law restricts states from using federal
funds to assist most families for more than five years (and
allows states to set shorter time limits), more than 30 states
had previously obtained waivers to implement some form of
time limit in at least part of the state. Thus, these states’
experiences provide an early look at the likely results of
the new law.
Jobs First is one of the most important initiatives
undertaken under waivers because it includes both some of
the most stringent and some of the most generous provisions
of any state welfare reform program. As of early 1999, a total
of 17 states had imposed time limits that could result in
cancellation of a family’s entire welfare grant after
less than 60 months of assistance, and only six of these states
had imposed lifetime time limits of less than 60 months.
Connecticut’s 21-month limit is the shortest of these.
In addition, Connecticut was one of the first states to impose
a time limit in relatively large cities such as New Haven,
Hartford, and Bridgeport. Many of the other early time limit
programs were initially implemented as pilot programs in relatively
small counties or regions of states. Because the Jobs First
time limit is short and was imposed relatively early, more
than half of the families that have reached a time limit nationwide
are in Connecticut. (In assessing a state’s time limit
policy, however, it is important to understand the design
and implementation of exemption and extension policies; they
are discussed further below.)
In other respects, Connecticut’s welfare
policies are unusually generous. Jobs First includes a financial
work incentive that is both very liberal and distinctive in
its design: All earned income is disregarded (that
is, not counted) in calculating recipients’ monthly welfare
grants as long as their earnings are below the federal poverty
level. This means that most recipients who find a job can
continue receiving their entire welfare grant. While most
states have enhanced earned income disregards, few, if any,
are as generous as Connecticut’s. Jobs First will provide
important new evidence on earned income disregards and on
the complex interaction between disregards and time limits.
About
Connecticut and the Research Sites
Connecticut is a medium-sized state with high
per capita income, but several very poor urban areas. The
state’s welfare grant levels ($543 for a typical family
of three) are high by national standards but lower than those
in most nearby states. Approximately 60,000 families were
receiving cash assistance statewide when Jobs First began.
The caseload declined modestly until late 1997, when recipients
began reaching the 21-month time limit, and then started falling
quickly. By August 1999, fewer than 30,000 families remained
on welfare.
Jobs First has been implemented in a healthy
economic climate: Connecticut’s unemployment rate was
at about the national average when Jobs First began, but has
since dropped substantially below the national rate, which
has also been declining.
The two Jobs First evaluation research sites
were chosen in part because they represent two quite different
environments. New Haven is the third largest city in the state
and one of the poorest cities in the United States. More than
20 percent of the statewide welfare caseload is served by
the New Haven DSS office. The Manchester office serves a more
suburban area near Hartford, accounting for about 6 percent
of the state caseload.
The
Jobs First Program Model
Jobs First aims to replace welfare with earned income. To
this end, the program replaced Connecticut’s AFDC program
with Temporary Family Assistance (TFA). Table
1 describes the key features of Jobs First, along with
the prior policies, which apply to the AFDC group. The key
features are:
- A time limit. Jobs First limits
families to a cumulative total of 21 months of cash assistance
receipt. Certain families, such as those in which the parent
is incapacitated, are exempt from the time limit. (As long
as the exemption applies, months of benefit receipt do not
count toward the time limit.) In addition, recipients who
reach the time limit may receive (renewable) six-month extensions
of their benefits if they have made a good-faith effort
to find employment but have family income below the welfare
payment standard (the maximum monthly grant for their family
size). Families whose cases are closed but who have income
below the payment standard are referred to the Safety Net,
a program administered by nonprofit organizations that aims
to prevent harm to children in such families.
- An earned income disregard. To
encourage and reward work, all earned income is disregarded
(that is, not counted) in calculating recipients’ cash
grants (and Food Stamp benefits) as long as their earned
income is below the federal poverty level. Recipients become
ineligible for cash assistance if their earnings are at
or above the poverty level. A parent with two children who
was working 40 hours per week at $6.25 per hour would have
$688 more in total monthly income under Jobs First than
under AFDC.
- Mandatory "work first" employment
services. Unless they were exempt, Jobs First group
members were required to begin by looking for a job, either
on their own or through Job Search Skills Training (JSST)
courses that teach job-seeking and job-holding skills. Education
and training were generally restricted to those who were
unable to find a job despite lengthy upfront job search
activities. Recipients who failed to meet these requirements
could be sanctioned. During the first 21 months of assistance,
sanctions involve reducing their welfare grant or closing
their case for three months. The penalties become stricter
after the time limit: A single instance of noncompliance
during an extension may result in permanent discontinuance
of the entire welfare grant (the "one-strike"
policy).
Jobs First policies called for other changes
in traditional welfare rules. For example, the program imposes
a partial "family cap": When a recipient gives birth
to a child who was conceived while she received welfare, her
benefits are increased by about half as much as they would
have been under prior rules. In addition, Jobs First participants
receive two years of transitional Medicaid coverage after
leaving welfare while employed (as opposed to the one year
of coverage provided under prior law).
The
Jobs First Evaluation
The Jobs First evaluation was initially required
as a condition of the federal waivers that allowed Connecticut
to operate the program. Then, in 1997, Connecticut received
enhanced federal funding from the U.S. Department of Health
and Human Services to support continuation of the study. (The
state later received a second federal grant to expand the
study to examine Jobs First impacts for children.)
The study has three major components:
- Implementation analysis. This component
examines how Jobs First operates in the research sites.
It assesses whether Jobs First policies have translated
into concrete changes in the day-to-day operations of the
welfare system and identifies obstacles that have been encountered.
This information is necessary in order to understand the
impact results and may also help DSS identify ways to improve
program performance.
- Impact analysis. This part of the
study provides estimates of the changes that Jobs First
generates in employment rates and earnings, rates and amounts
of welfare receipt, family income, the extent of welfare
dependency, child well-being, and other outcomes, relative
to outcomes under the welfare system that preceded it (as
represented by the AFDC group).
- Benefit-cost analysis. This analysis
uses data from the impact study, along with fiscal data,
to compare the financial benefits and costs generated by
Jobs First for both taxpayers and eligible families.
This report focuses on the implementation
and impact analyses. (Longer-term impact results and results
of the benefit-cost analysis will be presented in the study’s
final report.) It uses computerized administrative records
data provided by the state to measure individuals’ monthly
AFDC/TFA benefits, monthly Food Stamp benefits, and quarterly
earnings in jobs covered by Connecticut’s unemployment
insurance (UI) system. The records data are supplemented by
a survey of just under 800 Jobs First and AFDC group members,
which was conducted about 18 months after each person’s
date of random assignment. Finally, data on the program’s
implementation were obtained by interviewing line staff and
supervisors, observing program activities, and reviewing relevant
documents.
For the most part, the results presented in
this summary are for early study enrollees — 2,140 single
parents who were randomly assigned to the Jobs First and AFDC
groups between January and June 1996. At least 30 months of
follow-up data are available for each of these people, allowing
the study to draw some initial conclusions about what happened
after Jobs First group members began reaching the 21-month
time limit. Only two years of follow-up are available for
the full research sample, which includes people randomly assigned
through early 1997.1 Because
results for the early enrollees are similar to results for
the full sample through the first two years, the early group
probably provides a good estimate of the Jobs First impact
beyond that point. However, because the two sets of results
are not identical, the final results presented in future reports
(which will be based on the full sample) will differ somewhat
from the results presented here.
Readers should bear in mind three key features
of the study design. First, almost all of the results in this
report are drawn from the two research sites, and thus may
not represent the implementation or impacts of Jobs First
in other offices.
Second, unlike some earlier studies of welfare-to-work
programs, this one does not compare Jobs First with a control
group that receives no services. Rather, it compares Jobs
First with the AFDC policies that were in place just before
the program began — policies that already included some
emphasis on employment and self-sufficiency and some employment-related
services for welfare recipients. Thus, the study’s impact
analysis is measuring the effects of Jobs First over and above
what was already achieved by the earlier policy.
Third, although the study design has been
well implemented, it seems likely that the behavior of the
AFDC group has been influenced to some extent by the intense
focus on welfare reform at the state and federal levels over
the past few years. This suggests that the study may not capture
the full impact of Jobs First.2
Jobs
First Implementation in the Research Sites
The report examines the implementation of Jobs First during
roughly its first three and a half years of operation —
from early 1996 to mid 1999 — in order to understand how
Jobs First has differed from AFDC in practice. Key findings
include:
- Jobs First group members heard a more
employment-focused message from welfare staff than did AFDC
group members; in addition, staff successfully informed
recipients about the key program features.
A series of questions on the client survey
examined the messages that respondents heard from the welfare
system and generally found large differences between the groups.
The strongest message focused on quick employment: 67 percent
of Jobs First group members and 44 percent of AFDC group members
said that staff urged them to get a job as quickly as possible.
Fewer Jobs First group members (53 percent) reported hearing
that they should get off welfare quickly than that they should
get a job quickly — staff urged them to take advantage
of the earned income disregard by mixing work and welfare.
Still, more Jobs First than AFDC group members reported hearing
the quick-welfare-exit message (53 percent compared with 29
percent).
Nearly 90 percent of Jobs First group respondents
reported that they were subject to a time limit, and most
knew its length. More than 85 percent said that staff stressed
that they could keep part of their welfare benefits if they
went to work.
Just over 20 percent of AFDC group respondents
reported that they were subject to a time limit. Some of them
were correct — they had moved away from the research
sites and become subject to Jobs First policies — but
many had received erroneous information from the media, staff,
family members, or other sources. This fact means that the
evaluation results probably understate the impact of the Jobs
First time limit on recipients’ behavior, especially
during the period before recipients could have reached the
limit.
- Jobs First group members were more likely
than AFDC group members to participate in employment-related
activities, particularly activities focused on quick job
placement.
Figure 1 shows the
rates of participation in employment-related activities for
Jobs First and AFDC group members in the first 18 months after
each person’s date of random assignment. These data are
drawn from the survey and are self-reported. In addition,
they include both activities arranged by DSS and those not
arranged by DSS (for example, activities people participated
in after they left welfare).
The figure shows that members of both groups
were quite likely to report that they had participated in
at least one employment-related activity. However, as expected,
Jobs First group members had a significantly higher participation
rate: 64 percent versus 49 percent for the AFDC group. This
likely reflects the fact that a smaller proportion of Jobs
First group members were exempt from participation mandates
and that mandates were enforced more vigorously for the Jobs
First group (in practice, AFDC group members were generally
not required to participate in employment-related activities,
as had been true prior to Jobs First).
Consistent with the program model, the overall
difference in participation rates is driven mainly by an increase
in participation in job search activities. With the exception
of a small increase in college attendance (not shown in the
figure), Jobs First did not increase the rate of participation
in education or training activities. Because the job search
activities were fairly brief, these data imply that Jobs First
group members were likely not to have been continuously active
in employment-related activities throughout their time on
welfare. Moreover, despite their higher participation rates,
Jobs First group respondents were only slightly more likely
than AFDC group respondents to agree with the statement "I
received help that improved my long-term chances of getting
and keeping a job" (about half the respondents in each
group agreed a little or agreed a lot).
A more detailed analysis of Jobs First group
members who did not participate in any employment activities
found that most of them either left welfare quickly, were
employed for much of the time they received benefits (employed
recipients were generally given low priority for employment
services), or were exempt for a substantial period. (Overall,
about one-quarter of Jobs First group members were exempt
at some point, generally for less than one year). In other
words, a very small proportion of recipients fell through
the cracks entirely (as discussed earlier, this does not mean
that people participated in activities continuously while
on assistance).
Only about 3 percent of Jobs First group members,
and less than 1 percent of AFDC group members, were sanctioned
for failing to comply with employment-related mandates within
18 months after random assignment. The relatively low sanctioning
rates for the Jobs First group probably reflect the modest
scope of the employment-related requirements (that is, most
recipients werenot required to participate in many activities)
and the fact that participation was not very tightly monitored
(see below). The low sanctioning rate means that relatively
few recipients have been offered an Individual Performance
Contract (IPC), which allows noncompliant recipients to restore
their eligibility for time limit extensions based on good-faith
effort (a high percentage of those who were offered an IPC
succeeded in restoring their eligibility for an extension).
- While the key components of Jobs First
were put in place in Manchester and New Haven, start-up
problems and specific features of the program design prevented
it from being implemented very intensively.
As noted earlier, staff succeeded in informing
most Jobs First group members about the key elements of the
new policy and in referring most recipients to employment
services with a work first focus. In addition, DSS was able
to revise its statewide public assistance computer system
to track recipients’ time limit clocks and implement
the enhanced earned income disregard and other changes in
eligibility rules.
At the same time, Jobs First, like virtually
all new programs, has experienced implementation problems.
For example, the New Haven office in particular has faced
persistent difficulties monitoring recipients’ participation
in employment activities, in large part because there have
not been effective systems in place to obtain attendance reports
from contracted service providers. These problems emerged
early on, when employment services were mostly provided by
private organizations working under contract to DSS, but have
persisted since responsibility for employment services was
shifted in mid 1998 to the Connecticut Department of Labor,
Regional Workforce Development Boards, and their subcontractors.
As this report was being completed, a new case management
system was being phased in statewide, in part to improve participant
monitoring.
Start-up problems were particularly likely
to arise in Jobs First because the program has been implemented
in a challenging environment. The program called for radical
changes in the mission and activities of Connecticut’s
welfare system, but was put in place statewide from its inception,
with little time for advance planning. In addition, a variety
of other major initiatives in the past few years have consumed
the time and energy of the staff and managers responsible
for Jobs First, and the program itself has been revised in
significant ways. Because the evaluation has mostly been conducted
during the start-up period, its results probably represent
a conservative estimate of the model’s potential.
Other implementation issues have been related
to the program design. For example, unlike some other state
welfare reforms, Jobs First was implemented with virtually
no increases in staffing (despite a large increase in the
number of recipients who were expected to move toward self-sufficiency).
To facilitate this approach, the programwas designed so that
staff and recipients do not necessarily interact very frequently.
Thus, while most recipients were initially informed about
the time limit and the enhanced disregard, and strongly urged
to seek work, there was limited contact between recipients
and staff in the subsequent months, and thus relatively few
opportunities to reinforce these messages. Large worker caseloads
also contributed to the monitoring problems described above.
Finally, the key tasks of tracking clients’ activities,
assisting those with serious problems, and transmitting a
clear, consistent program message have all become more challenging
as an increasingly complex organizational structure has developed
to implement the various aspects of Jobs First.
The
Jobs First Time Limit
Although it is only one of many program features, the Jobs First
time limit has been the subject of intense scrutiny. Thus, MDRC
examined its implementation in detail. Key findings include:
- Most Jobs First group members did not
reach the time limit within the follow-up period for this
report.
MDRC examined patterns of TFA receipt during
a two-and-a-half-year follow-up period for Jobs First group
members who were randomly assigned between January and June
1996. This analysis found that about one-fourth of them reached
the time limit 21 months after their random assignment date;
that is, they received TFA continuously and were never exempt.
About 39 percent reached the time limit within 30 months after
enrollment. Thus, about 61 percent of Jobs First group members
still had months remaining on their time limit clock 30 months
after enrollment. Most of these individuals had left welfare;
others were exempt from the time limit for at least part of
the time they received benefits.
As discussed below, many of those who reached
the time limit received an extension and were allowed to continue
receiving benefits. Thus, overall, roughly one-fifth of Jobs
First group members’ cases were closed because of the
time limit within the two-and-a-half-year follow-up period.
- Just over half of those who reached
the time limit were granted an extension at that point;
they had very low income and were deemed to have made a
good-faith effort to find a job. Most of these people were
not receiving welfare 15 months later.
MDRC examined a randomly selected group of
100 cases who reached the time limit by March 1998. Figure
2 shows the outcomes for these cases during the 15 months
after they reached the limit.
Recipients are called in for an "exit
interview" in the 20th month of assistance, in order
to determine whether they will receive an extension or have
their case closed. Figure 2 shows that
55 of the 100 recipients MDRC studied were granted a six-month
extension when they reached the time limit (two other cases
were granted an exemption at that point).3
All 55 were granted an extension because they had income below
the welfare payment standard and were deemed to have made
a good-faith effort to find employment. Interviews with staff
indicated that many of the individuals who were granted an
extension had not been closely monitored during the pre-time
limit period; in accordance with the program rules, however,
they were assumed to have made a good-faith effort because
there was no evidence to the contrary (in general, a good-faith
effort is assumed so long as the recipient was not sanctioned
more than once and did not quit a job without good cause in
the final six months of assistance).
There is no limit on the number of six-month
extensions a family may receive. Nevertheless, 35 of the 55
recipients who were initially granted an extension were no
longer receiving TFA benefits 15 months later. Although recipients
in extensions are subject to the one-strike noncompliance
policy described earlier, only 6 of the 35 cases were closed
because they failed to comply with employment requirements.
About 20 of the 35 left TFA because they found a job (the
others moved out of the state or left welfare for other reasons).
Despite the relatively small number of permanent
case closures, dealing with noncompliance during extensions
(and at earlier points) has emerged as a key implementation
challenge. Staff report that the current good-cause policies,
which guide staff in determining whether recipients have a
valid excuse for quitting a job or missing employment activities,
leave many gray areas. As a result, staff must exercise substantial
discretion on decisions with very high stakes, and many workers
believe that recipients in similar situations may be treated
differently by different workers. At the same time, most staff
were skeptical that more detailed rules would improve the
situation.
In addition to the 55 individuals who were
granted an extension when they reached the time limit (and
the two who were granted exemptions), another 5 people initially
had their cases closed at the time limit but were granted
an extension at some point within the next 15 months (see
below). Thus, overall, 62 of the 100 people who reached the
time limit received TFA benefits at some point in the subsequent
15 months, but only 25 were still receiving benefits 15 months
after reaching the time limit.
- The cases of just under half of those
who reached the time limit were closed at that point. A
large proportion of them were working and had income above
the payment standard, and very few of them returned to TFA
during the subsequent 15 months.
As shown in Figure 2,
the cases of 43 of the 100 recipients that MDRC studied were
closed at the time limit. Thirty-two of the 43 were denied
an extension because they had income above the welfare payment
standard. Many of these people would have become ineligible
for welfare earlier had it not been for the enhanced disregard.
Another 10 had their case closed because they failed to attend
their exit interview; it appears that most of these individuals
were employed at that point, although not necessarily earning
above the payment standard. Nevertheless, their cash assistance
and Food Stamp cases were closed because they did not attend
the interview (their Medicaid coverage continued if they were
reporting earnings to DSS).
Only one of the recipients whose case was
closed had income below the payment standard and was deemed
not to have made a good-faith effort. Thus, in all, 57 of
the 58 individuals who attended their exit interview and had
income below the payment standard were initially granted an
extension or exemption. (As noted earlier, another 6 cases
were closed for noncompliance after initially receiving an
extension.)
Recipients whose cases are closed because
their income is over the payment standard (as well as those
who fail to attend the exit interview) may be granted an extension
later if their income drops below the payment standard and
they have made a good-faith effort to find employment (both
before reaching the time limit and afterward). However, of
the 43 people whose cases were closed at the time limit, 38
never returned to TFA in the subsequent 15 months (not shown
in the figure), and only 3 were receiving TFA 15 months after
reaching the time limit. Another 11 applied for TFA at some
point, but did not start to receive benefits, usually because
they were found to be financially ineligible or because they
did not complete necessary paperwork.
MDRC examined the income of the recipients
whose cases were closed at the time limit, both before and
after the time limit, using administrative records. As expected,
their average combined income from earnings and public assistance
fell dramatically, from $4,435 in the quarter before reaching
the time limit to $2,988 in the quarter after the time limit
(a 33 percent drop). Nevertheless, it is important to point
out that even after the time limit, about two-thirds of the
recipients whose cases were closed had more income than a
typical nonworking family would receive from TFA and Food
Stamps. Moreover, these figures do not include the earned
income credit (EIC), which substantially increases the income
of many low-wage workers; child support payments; or other
income sources.
Interestingly, this analysis also showed that
the employment rate among those who had their case closed
at the time limit dropped substantially, from 87 percent in
the quarter when their case was closed to 77 percent in the
third quarter afterward. Despite this drop, however, only
5 percent received TFA in the latter quarter. Although this
could mean several things (for example, some people may have
moved out of Connecticut), it suggests that those who lose
jobs are not necessarily returning to TFA (the earlier discussion
suggests that others may have applied for benefits, but never
received them).
- Only a small number of people who reached
the time limit have had their cases closed despite having
income below the welfare payment standard; thus, the number
of referrals to the Safety Net program has been relatively
small, but may grow over time.
Recipients whose grants are discontinued
despite having income below the payment standard (because
they are deemed not to have made a good-faith effort to find
employment) are referred to the Safety Net program for further
assistance and are generally not eligible for further extensions
(they can receive assistance again if they become exempt or
encounter circumstances beyond their control that prevent
them from working). The cases of 8 of the 100 recipients studied
by MDRC who reached the time limit were closed under these
circumstances within 15 months after reaching the limit.4
Since only about two-fifths of the Jobs First group members
reached the time limit within the report’s follow-up
period, it suggests that less than 5 percent of the entire
group was referred to Safety Net services.
At the same time, it is important to note
that statewide program records show that the number of referrals
to Safety Net services has grown substantially over time.
This is not surprising; as of August 1999, nearly 25 percent
of the statewide TFA caseload (and 40 percent of those subject
to the time limit) were in an extension, and thus subject
to the one-strike noncompliance policy.
Jobs
First Impacts on Employment, Public Assistance Receipt, and Income
As discussed earlier, most of the impact results presented in
this summary are for early enrollees who were assigned to the
research groups between January and June 1996; at least two
and a half years of follow-up data are available for this group.
The full sample — for whom two years of follow-up are available
— is used primarily for assessing the impact of Jobs First
on subgroups. Key findings from the impact analysis include
the following:
- During the first part of the follow-up
period, before anyone reached the time limit, Jobs First
raised employment rates and earnings and also increased
public assistance receipt; thus, the program substantially
increased family income.
Members of the Jobs First group were more
likely to work than members of the AFDC group. The upper panel
of Figure 3 shows that the employment
impact emerged early in the follow-up period and remained
fairly constant thereafter (that is, the distance between
the two lines remained about the same).
The upper panel of Table
2 shows results for the sixth quarter of the follow-up
period, the last complete quarter before recipients began
reaching the time limit. Fifty-six percent of Jobs First group
members worked at some point in that quarter compared with
46 percent of AFDC group members. The difference — about
10 percentage points — is statistically significant,
which means it is very likely that Jobs First really did raise
employment rates. These overall employment impacts are impressive,
especially considering the high employment rate for the AFDC
group: The lower panel of Table 2 shows that about 74 percent
of AFDC group members worked at some point in the follow-up
period, leaving relatively little room for Jobs First to generate
impacts.
Table 2 also shows that
average earnings were about 16 percent higher for the Jobs
First group in quarter 6. It is important to note that the
earnings figures are overall averages, including both those
who worked in the quarter and those who did not. Employed
Jobs First group members earned $2,655, on average, during
the quarter (not shown).
The middle panel of Figure
3 shows that, as expected, Jobs First increased the number
of those receiving cash assistance (AFDC or TFA) during the
period before anyone reached the time limit. This is attributable
to the enhanced earned income disregard, which allowed many
working Jobs First group members to retain their TFA grant
in months when their income would otherwise have made them
ineligible for assistance. Table 2 shows
that, as a result, Jobs First increased cash assistance payments
by about 18 percent in quarter 6. The enhanced disregard also
generated an increase in Food Stamp payments throughout much
of the early period (though not in quarter 6).
Raising family income was never an explicit
goal of Jobs First. However, because Jobs First group members
had both higher earnings and higher public assistance
payments in the period before anyone reached the time limit,
their average combined income from these sources was substantially
higher than the AFDC group average. Table
2 shows that Jobs First group members had about 15 percent
more income from earnings, cash assistance, and Food Stamps
during quarter 6. This is not a complete measure of family
income because it does not include other income sources, such
as child support and the EIC, does not count income of other
household members, and does not include income that was derived
outside Connecticut.
Data from the survey, which includes a fuller
income measure, paint a very similar picture. Survey data
also indicate that Jobs First group members were more likely
than AFDC group members to own a car and to have at least
some savings 18 months into the follow-up period. Jobs First
group members were also more likely to have health coverage
at that point, primarily because they were more likely to
be covered by Medicaid.
- After families began reaching the time
limit, Jobs First began to reduce welfare receipt and payments;
the income gains diminished, and it appears that some families
were made worse off by the program.
Jobs First group members began
to reach the time limit in the 7th quarter of the follow-up
period. The cases of about 13 percent of the Jobs First group
were closed on reaching the time limit in that quarter. As
noted earlier, by the end of the follow-up period (quarter
10), the cases of about 20 percent of the Jobs First group
were closed because of the time limit.
Figure 3 and Table
2 show that the pattern of Jobs First impacts on public
assistance receipt and payments changed abruptly when members
of the Jobs First group began to reach the time limit. As
noted earlier, in the first 18 months of follow-up (before
anyone reached the time limit), Jobs First group members were
more likely than AFDC group members to receive cash assistance.
Beginning in quarter 8, after some people had reached the
time limit, Jobs First group members were less likely
to receive cash assistance. The middle panel of Table
2 shows that in the last three months of the follow-up
period (quarter 10), the rate of cash assistance receipt was
substantially lower for the Jobs First group (37 percent versus
46 percent for the AFDC group), and Jobs First group members
received 18 percent less cash assistance. The patterns of
Food Stamp impacts were similar, though less dramatic: Jobs
First group members received more Food Stamps in the pre-time
limit period, but this increase disappeared in the later quarters.
As might be expected, the total income results
also look quite different in the period after people began
reaching the time limit. Although Jobs First group members
continued to have slightly higher average combined income
just after people began reaching the time limit, this impact
seems to have disappeared by the end of follow-up. The middle
panel of Table 2 shows that in the last
three months of follow-up Jobs First group members’ higher
earnings were almost completely offset by their lower public
assistance amounts. Thus, total income from earnings, AFDC/TFA,
and Food Stamps was nearly the same for the two groups, although
Jobs First group members derived a larger proportion of their
income from earnings and a smaller proportion from public
assistance.5 It is important
to note, however, that this trend did not erase the earlier
income gains — the lower panel of Table
2 shows that Jobs First group members had an average of
$2,378 more combined income than AFDC group members over the
entire 30-month follow-up period. Moreover, since Jobs First
was primarily designed to replace welfare with earned income,
the income results at the end of the follow-up period are
generally consistent with the program’s objectives.
Table 3 shows that the
overall averages mask an important emerging pattern. After
families startedreaching the time limit, Jobs First began
to increase the number of families in both the high and low
income brackets. In other words, the program appears to be
making some people better off financially and others worse
off. The latter result may be attributable to the fact that
some of those who were denied an extension because their income
was over the payment standard either did not or could not
return to welfare when their income later dropped. It may
also reflect the fact that the cases of a growing number of
recipients are being closed despite having income below the
payment standard (for example, because they did not attend
an exit interview or failed to cooperate with program requirements
during an extension). It will be important to see whether
this trend continues in later quarters.
The Jobs First impact on employment rates
did not change much when families began reaching the time
limit. This is not surprising, because most of those whose
grants were discontinued at the time limit were already working.
Essentially, the program allowed a large number of working
families to retain their welfare grant temporarily and then
discontinued their benefits at the time limit. This is illustrated
in Table 4, which shows patterns of employment
and AFDC/TFA receipt in quarter 6 (before families began reaching
the time limit) and quarter 10. In quarter 6, Jobs First generated
a large increase in the percentage of those combining work
and welfare. It did so by decreasing both the percentage who
received welfare without working and the percentage who worked
without receiving welfare (that is, it let many working families
stay on assistance). In quarter 10, after many working families
had been removed from welfare, Jobs First group members were
more likely than AFDC group members to be working and
off welfare. (The program also increased the percentage who
neither worked nor received welfare; this could reflect a
greater number of families moving out of state or relying
on unmeasured income sources.)
- Jobs First impacts on employment and
earnings were concentrated among individuals facing greater
barriers to employment. For welfare applicants — a
more job-ready group — the main impact of Jobs First
was to increase public assistance receipt.
The study examined Jobs First impacts separately
for individuals who were applying for welfare when they entered
the study and for those who were already receiving benefits
at that point. In general, applicants face less serious barriers
to employment than do recipients.6
Jobs First impacts on employment and earnings
were concentrated among the recipient subgroup. Increases
in employment rates and earnings were quite large for them
and even larger for people facing multiple barriers to employment.
For example, in the pre-time limit period, Jobs First nearly
doubled the average quarterly employment rate among those
who were long-term welfare recipients, had no recent work
history, and had no high school diploma (the average employment
rate was 17 percent for the AFDC group and 32 percent for
the Jobs First group).
In contrast, the program did not increase
employment or earnings among applicants, who were quite likely
to find a job without the program (that is, the employment
rate was high for the AFDC group). The program’s primary
impact for this subgroup was to allow those who would have
worked anyway to continue receiving public assistance, thereby
raising their income.
- There was great variation in the quality
of sample members’ jobs.
The impact results suggest that most of the
people who went to work because of Jobs First (that is, who
would not have worked otherwise) initially obtained fairly
low-wage and/or part-time jobs. However, among all
employed Jobs First group members, job characteristics were
diverse. For example, about half of the employed Jobs First
group members who responded to the survey earned below $7.50
an hour in their current or most recent job, while the other
half earned $7.50 an hour or above. About 60 percent worked
30 hours or more per week. About 45 percent worked in a job
with employer-provided health insurance, but only about 15
percent were enrolled in their employer’s health plan
(most of the other respondents were covered by Medicaid).
Overall, just under 40 percent of employed
Jobs First group members worked 30 hours or more per week
in a job that provided at least some fringe benefits (that
is, employer-provided health insurance, paid sick days, or
paid vacation). At the other extreme, just over 20 percent
were in a part-time jobs that provided no benefits.
An
Unfinished Story
Jobs First is an unusual hybrid. On the one
hand, it has the nation’s shortest time limit and a strong
work first focus. On the other hand, its financial work incentive
may be the most generous of any state’s, and many recipients
are granted an extension when they reach the time limit. The
program sought to encourage and assist recipients to find
a job as quickly as possible, and gave them a temporary income
supplement, in the hope that they would gain work experience
and possibly build assets that would prepare them for longer-term
self-sufficiency.
In some respects, the early results are encouraging.
In the pre-time limit period, Jobs First increased employment
rates and family income, especially for the least job-ready
recipients. The program generated higher public assistance
costs, but only in the short term. In addition, the fact that
Jobs First has increased welfare receipt and spending (relative
to what would have occurred under the old rules) is probably
of less concern during a period when the state’s overall
welfare caseload has been declining dramatically. The risk
for recipients has been minimized so far because most people
who were unable to find a job were granted an extension when
they reached the time limit. Conversely, those whose cases
were closed at the time limit mostly comprised employed recipients
who would have left welfare earlier had it not been for the
enhanced disregard; moreover, these individuals may be eligible
for an extension later if their income drops.
But it is too early to draw final conclusions
about Jobs First impacts on participants or government budgets.
This report’s follow-up period is long enough to see
that the pattern of impacts changed abruptly when Jobs First
group members began to reach the 21-month time limit, but
not long enough to reliably predict what the longer-term picture
will look like.
The results will probably continue to improve
from the government budget perspective. Savings in public
assistance payments will probably continue to accumulate over
time, although it is too soon to tell whether the savings
will eventually outweigh the large upfront costs.
The future is more uncertain for participants
and their families. The income trends at the end of the follow-up
period — particularly the income distribution trends
— suggest that the program may be making some families
better off financially and others worse off. Ironically, although
the enhanced disregard may have contributed to the employment
gains, and certainly helped to raise the income of many working
families, it also caused many Jobs First group members to
reach the time limit more quickly than they otherwise would
have (and more quickly than recipients reached time limits
in other states that have been studied). Although most of
those whose cases were closed can theoretically receive an
extension later, the early evidence shows that many of those
who lost their job did not return to welfare. In addition,
the recipients who received an extension became subject to
a very strict policy that could result in permanent benefit
cancellation after a single instance of noncompliance. Finally,
it is not clear how the families whose cases were closed because
of the time limit will fare if the labor market weakens.
A large-scale survey, currently being administered,
will obtain detailed information on family income, material
hardship, and child well-being for several thousand families
about three years after random assignment.
Implications
of the Findings to Date
The results to date highlight several emerging
challenges for Jobs First. First, many staff believe that
the work first focus may need to be revised as the caseload
is increasingly dominated by recipients facing more serious
barriers to employment. This process has already begun, but
the Department of Labor and the Regional Workforce Development
Boards still must confront the substantial challenge of developing
effective employment models for the hardest to employ. As
discussed earlier, this may require some streamlining of the
complex organizational structure that has emerged to assist
such recipients.
Second, with a large proportion of the caseload
subject to the one-strike noncompliance policy, DSS faces
the challenge of ensuring that good cause criteria are implemented
in a way that is flexible enough to account for individual
circumstances, but consistent enough to ensure that individuals
in similar situations receive similar treatment. This will
be difficult because, by definition, the individuals in extensions
are long-term welfare recipients; the same issues and personal
problems that have prevented them leaving welfare may also
contribute to noncompliance and job losses. Interviews with
staff suggest that the focus will need to be on day-to-day
implementation practices; it is not clear that additional
or more detailed policies will, in themselves, ensure the
desired outcomes.
Third, Connecticut, like most other states
that have experienced dramatic welfare caseload declines,
faces the challenge of promoting employment retention and
career advancement among low-wage workers. This is likely
to involve both strategies to promote wage progression and
supports for families who remain in relatively low-wage jobs.
The longer-term success of Jobs First may depend on whether
former recipients are able to support their children over
the long-term without cash assistance.
Finally, the results for the welfare applicant
subgroup suggest that Connecticut may want to reexamine the
Jobs First design at some point in the future. The employment
and earnings gains measured in this study were mostly driven
by people who were already receiving welfare at the time they
entered the Jobs First program. In the future, after this
initial recipient group has moved off welfare, most people
will enter Jobs First when they are applying for benefits;
many will be new to the welfare system. The study results
indicate that the main impact of Jobs First for applicants
is to provide additional welfare benefits and income to people
who would have worked anyway. In addition, it is quite possible
that the generous disregard will begin to draw some low-income
families to TFA.
Given that the primary goal of Jobs First
has been to replace welfare with earned income, the results
for applicants suggest that the program’s work incentives
probably should be targeted more narrowly on people who would
be least likely to work in the absence of incentives. If the
state is now seeking to supplement the earnings of low-income
working families, it will need to consider whether work incentives
within the welfare system are the most appropriate vehicle
for achieving this goal.
Notes:
1. The full sample is used to
assess the impact of Jobs First on subsets of the eligible population
because larger sample sizes are critical for such analyses.
2. In addition, the study is
not designed to measure whether Jobs First has affected the number of
people who apply for welfare.
3. MDRC classified a case as
initially receiving an extension if the recipient reached the time limit
and then continued receiving benefits in month 22. Some of these recipients
did not attend their exit interview when it was first scheduled, but visited
the office and were granted an extension in time to prevent losing a month
of assistance. Conversely, a case was classified as having been closed
at the time limit if the recipient received benefits for 21 countable
months and then did not receive assistance in month 22.
4. One case was closed for lack
of good-faith effort upon reaching the time limit, six were closed for
noncompliance during an extension, and one was denied an extension for
lack of good-faith effort when reapplying for benefits. In addition, it
is possible that some of the 10 people who failed to attend their exit
interview had income below the payment standard.
5. As noted earlier, these income
results do not include the EIC. Because Jobs First group members had higher
average earnings, they probably benefited more from the EIC than did AFDC
group members. However, it is also likely that Jobs First group members
had higher work-related expenses (for example, for transportation or child
care).
6. Results for subgroups are
based on the full research sample because sample sizes are larger. Thus,
the subgroup results include only the two years of follow-up data that
are available for the full sample.
|
