TUNING IN TO LOCAL LABOR MARKETS - FINDINGS FROM THE SECTORAL EMPLOYMENT IMPACT STUDY
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
Tuning In to Local Labor Markets Findings from The Sectoral Employment Impact Study
Tuning In to Local Labor Markets
Findings from
The Sectoral Employment
Impact Study
Sheila Maguire
Joshua Freely
Carol Clymer
Maureen Conway*
and
Deena Schwartz
* Maureen Conway is the Director of The Workforce Strategies Initiative at the
Aspen Institute.Public/Private Ventures is a national leader in Board of Directors Research Advisory
creating and strengthening programs that improve
Committee
lives in low-income communities. We do this in
three ways: Matthew T. McGuire, Chair Jacquelynne S. Eccles, Chair
Principal University of Michigan
innovation Origami Capital Partners, LLC Robert Granger
We work with leaders in the field to identify promising existing programs Yvonne Chan William T. Grant Foundation
or develop new ones. Principal Robinson Hollister
Vaughn Learning Center Swarthmore College
research
John J. DiIulio, Jr. Reed Larson
We rigorously evaluate these programs to determine what is effective and Frederic Fox Leadership Professor University of Illinois
what is not. of Politics, Religion and Civil
Jean E. Rhodes
Society
action University of Massachusetts,
University of Pennsylvania
Boston
We reproduce model programs in new locations, provide technical Robert J. LaLonde
Thomas Weisner
assistance where needed and inform policymakers and practitioners Professor
UCLA
about what works. The University of Chicago
John A. Mayer, Jr.
P/PV is a 501(c)(3) nonprofit, nonpartisan organization with offices in Retired, Chief Financial Officer
Philadelphia, New York City and Oakland. For more information, please J. P. Morgan & Co.
visit www.ppv.org. Anne Hodges Morgan
Consultant to Foundations
Siobhan Nicolau
President
Hispanic Policy Development
Project
Marion Pines
Senior Fellow
Institute for Policy Studies
Johns Hopkins University
Clayton S. Rose
Senior Lecturer
Harvard Business School
Sudhir Venkatesh
William B. Ransford Professor of
Sociology
Columbia University
William Julius Wilson
Lewis P. and Linda L. Geyser
University Professor
Harvard University
© 2010 Public/Private VenturesTuning In to Local Labor Markets: Findings From the Sectoral Employment Impact Study
Dedication
This report is dedicated to Eric Parker, who
died suddenly in August 2007. A visionary,
leader, mentor and friend to many in the
workforce development field, Eric founded
the Wisconsin Regional Training Partnership
and devoted much of his career to improving
the lives of low-wage workers and the competi-
tiveness of local employers.
Acknowledgments
Many people worked to make this report possible. Many current and former P/PV colleagues and consul-
Thanks go first and foremost to the three organizations tants made valuable contributions to the study, includ-
that are the subject of the study—Jewish Vocational ing Mark Elliott, Anne Roder, Mae Watson Grote, Rohit
Service–Boston (JVS–Boston), Per Scholas and Wisconsin Reddy, Cathy Danh, Danielle Farrie, Wendy McClanahan,
Regional Training Partnership (WRTP). Participating Jean Grossman, Katie Plat, Chelsea Farley, Laura
in a random assignment evaluation is a challenging and Johnson, Gary Walker and Nadya K. Shmavonian.
risky undertaking, and the study would not have been
possible without the leadership of Barbara Rosenbaum We are deeply appreciative of P/PV Board Member
and Jerry Rubin at JVS–Boston; Plinio Ayala at Per William Julius Wilson’s thoughtful contributions and his
Scholas; and Eric Parker and Earl Buford at WRTP, as support and leadership in hosting a meeting to discuss an
well as the ongoing efforts of dozens of staff members. earlier draft of the report. We thank the attendees of that
We are also grateful to the study participants for sharing meeting for their valuable input about the study’s impli-
their experiences during interviews and in focus groups. cations. We especially thank Lawrence Katz and Robert
LaLonde for their review and comments on earlier drafts.
Over the past two decades, Jack Litzenberg of the Charles
Stewart Mott Foundation has done much to advance the Finally, our thanks to The Institute for Survey Research
practice and study of sectoral strategies. His decision to at Temple University in Philadelphia, which administered
invest in this random assignment study, and his patience the participant surveys; Penelope Malish, who designed
and understanding as it unfolded, is another signal of his the publication; and Lauren Kelley, Jason Warshof, Clare
thoughtful leadership of the field. Thanks also must go O’Shea and Sylvia Foley, who edited the report.
the Mott Foundation’s communications department, and
Duane Elling in particular, for their support of the project.Tuning In to Local Labor Markets: Findings From the Sectoral Employment Impact Study
Contents
Executive Summary................................................................................................................ i
Chapter I: Introduction........................................................................................................... 1
Chapter II: Study Design........................................................................................................ 5
Study Participants.................................................................................................................................. 6
Chapter III: Overall Effects and Key Findings..................................................................... 9
Key Findings........................................................................................................................................ 12
Chapter IV: Program-Specific Findings.............................................................................. 17
Wisconsin Regional Training Partnership.............................................................................................. 18
JVS–Boston......................................................................................................................................... 29
Per Scholas......................................................................................................................................... 37
Chapter V: Programmatic Approaches.............................................................................. 47
Common Elements.............................................................................................................................. 48
Common Challenges........................................................................................................................... 51
Chapter VI: Conclusions and Implications for Further Research.................................... 53
Implications for Further Research......................................................................................................... 55
Concluding Thoughts........................................................................................................................... 57
Endnotes............................................................................................................................... 59
Appendices........................................................................................................................... 61
Appendix A: Selection of the Study Sites............................................................................................. 62
Appendix B: Sample Selection, Randomization and the Follow-Up Sample......................................... 63
Appendix C: Study Methodology......................................................................................................... 70
Appendix D: Employment Outcomes for Selected Subgroups............................................................. 72
Appendix E: Supplementary Tables, WRTP.......................................................................................... 75
Appendix F: The Question of Displacement......................................................................................... 76
Appendix G: Regression Tables for the Overall Sample........................................................................ 77Tuning In to Local Labor Markets: Findings From the Sectoral Employment Impact Study
Tables
Executive Summary Table 1: Baseline Characteristics of the Follow-Up Sample................................................iii
Table 1: Baseline Characteristics of the Follow-Up Sample............................................................................7
Table 2: Training Cohorts, by Site................................................................................................................10
Table 3: Employment Outcomes, Total Sample............................................................................................11
Table 4: Earnings Impacts, Selected Subgroups, All Sites...........................................................................15
Table 5: Baseline Characteristics of the Follow-Up Sample, WRTP..............................................................19
Table 6: Employment Outcomes, WRTP......................................................................................................21
Table 7: Likelihood of Working a Job Offering Benefits, WRTP....................................................................23
Table 8: Likelihood of Working a Unionized Job, WRTP...............................................................................23
Table 9: Employment Outcomes by Industry Sector, WRTP.........................................................................24
Table 10 Likelihood of Receiving a Certification, WRTP................................................................................25
Table 11: Employment Outcomes, Selected Subgroups, WRTP....................................................................27
Table 12: Baseline Characteristics of the Follow-Up Sample, JVS—Boston...................................................30
Table 13: Employment Outcomes, JVS—Boston...........................................................................................33
Table 14: Likelihood of Working a Job Offering Benefits, JVS—Boston.........................................................34
Table 15: Employment Outcomes, Selected Subgroups, JVS—Boston.........................................................35
Table 16: Baseline Characteristics of the Follow-Up Sample, Per Scholas.....................................................39
Table 17: Employment Outcomes, Per Scholas.............................................................................................42
Table 18: Employment Outcomes, Selected Subgroups, Per Scholas...........................................................44
Table 19: Likelihood of Receiving A+ Certification, Per Scholas.....................................................................46
Table 20: Likelihood of Working a Job Offering Benefits, Per Scholas............................................................46
Appendix Table 1: Characteristics of Study Participants at Baseline................................................................64
Appendix Table 2: Regression of Treatment on Selected Baseline Characteristics...........................................65
Appendix Table 3: Characteristics of Study Participants at Follow-Up.............................................................67
Appendix Table 4: Regression of Treatment on Selected Baseline Characteristics at Follow-Up......................68
Appendix Table 5: Employment Outcomes, Selected Subgroups, All Sites......................................................72
Appendix Table 6: Likelihood of Working a Job That Offers Medical Insurance, WRTP....................................75
Appendix Table 7: Likelihood of Working a Job Paying $15 an Hour or More,
WRTP Construction-Track Participants...........................................................................................75
Appendix Table 8: Analysis of Percentage Gains in Earnings and Hours..........................................................76
Appendix Table 9: Regression Tables for the Overall Sample...........................................................................77Tuning In to Local Labor Markets: Findings From the Sectoral Employment Impact Study Contents Figures Figure 1: Total Earnings by Month, Total Sample...........................................................................................12 Figure 2: Hours Worked by Month, Total Sample..........................................................................................12 Figure 3: Likelihood of Employment by Month, Total Sample........................................................................13 Figure 4: Likelihood of Working a Job Paying at Least $11 an Hour by Month, Total Sample.......................13 Figure 5: Likelihood of Working a Job Paying at Least $13 an Hour by Month, Total Sample.......................14 Figure 6: Likelihood of Working a Job Offering Benefits, Total Sample..........................................................14 Figure 7: Total Earnings by Month, WRTP.....................................................................................................22 Figure 8: Likelihood of Working a Job Paying at Least $11 an Hour by Month, WRTP.................................22 Figure 9: Likelihood of Working a Job Paying at Least $13 an Hour by Month, WRTP.................................22 Figure 10: Total Earnings by Month, JVS–Boston............................................................................................32 Figure 11: Likelihood of Employment by Month, JVS–Boston.........................................................................32 Figure 12: Likelihood of Working a Job Paying at Least $11 an Hour by Month, JVS–Boston........................32 Figure 13: Total Earnings by Month, Per Scholas............................................................................................40 Figure 14: Likelihood of Employment by Month, Per Scholas..........................................................................40 Figure 15: Likelihood of Working a Job Paying at Least $11 an Hour by Month, Per Scholas.........................40 Figure 16: Likelihood of Working a Job Paying at Least $13 an Hour by Month, Per Scholas.........................40
Executive Summary
ii Tuning In to Local Labor Markets: Findings From the Sectoral Employment Impact Study
F
Three organizations were selected:
• The Wisconsin Regional Training Partnership
(WRTP) is an association of employers and
unions that seeks to retain and attract high-wage
or American workers, having a high school jobs in Milwaukee and create career opportuni-
or general equivalency diploma (GED)—which ties for low-income and unemployed community
once represented a means of entrance to the residents. WRTP develops training programs
middle class—is no longer adequate for finding (generally lasting between two and eight weeks)
steady employment. In fact, three quarters of low- in response to specific employers’ requests or to
wage workers1 have these qualifications but lack clearly identified labor market needs. Its short-
the relevant occupational skills and connections to term preemployment training programs in the
employers needed to launch a career. At the same construction, manufacturing and healthcare sec-
time, in some regions of the country there are per- tors were included in the study.
sistent skills gaps clustered in particular industries, • Jewish Vocational Service–Boston (JVS–Boston)
such as manufacturing and healthcare.2 Many is a community-based nonprofit that has pro-
of these jobs are expected to grow3 and require vided workforce development services for more
specific technical skills that can be gained only than 70 years, including operating one of three
through focused training that is closely linked to One-Stop Career Centers (One-Stops) funded
the needs of local businesses. by the Workforce Investment Act (WIA) in the
Boston area. The organization aims to serve a
Over the past two decades, an innovative approach diverse range of Boston’s disadvantaged popula-
to workforce development known as sectoral employ- tions, including refugees, immigrants and welfare
ment has emerged, resulting in the creation of recipients. Its training programs in medical bill-
industry-specific training programs that prepare ing and accounting were included in the study.
unemployed and underskilled workers for skilled
• Per Scholas is a social venture in New York City
positions and connect them with employers seek-
that combines a training program with efforts to
ing to fill such vacancies. Based on earlier outcomes
refurbish and recycle “end of life” computers and
studies pointing to the promise of this strategy,
distribute them to low-income people through
Public/Private Ventures (P/PV) set out to conduct
partnerships with nonprofits, schools and com-
a random assignment evaluation to assess whether
munity colleges. Per Scholas’ computer techni-
sector-focused programs could in fact increase the
cian training program—which prepares partici-
earnings of low-income, disadvantaged workers and
pants for jobs in the repair and maintenance of
job seekers.
personal computers, printers and copiers, as well
as the installation and troubleshooting of com-
The Study puter networks—was included in the study.
In 2003, with funding from the Charles Stewart
Mott Foundation, P/PV launched the Sectoral P/PV used an experimental research design to bring
Employment Impact Study. We did not seek orga- as much rigor as possible to the following question:
nizations that followed a common model to par- Do mature sector-focused programs result in signifi-
ticipate in the study, as sectoral programs employ cant labor market gains for low-income, disadvan-
various approaches depending on the organiza- taged workers and job seekers? More specifically, we
tion leading the effort and local employers’ needs. strived to determine whether such programs raise
Instead, we sought mature programs that seemed the earnings of program participants and whether
to be well implemented, since it takes time for an participants were more likely to find employment
organization to both understand employers’ needs and work more consistently. We also wanted to
and craft appropriate responses. explore whether program participants obtained
higher-quality jobs. For example, were participants
more likely to earn higher wages? Did participants
find jobs with better access to benefits? Further,Executive Summary iii
we set out to explore whether specific groups of
people, such as welfare recipients or young adults, Executive Summary Table 1
benefit from participation. We also sought to under- Baseline Characteristics of the
stand the programmatic, contextual and individual Follow-Up Sample
factors that contribute to these outcomes.
Total
To answer these questions, the three sites recruited N 1,014
1,286 people for the study over a two-year period, all Gender
of whom had been through their program’s appli- Male 47%
cation process and met its eligibility criteria. Half Female 53%
of these applicants were selected at random to par- Race/Ethnicity and
Foreign-Born Status
ticipate in the program (the treatment group); the
African American 60%
remaining half (the control group) could not receive
Latino 21%
services from the study sites for the next 24 months,
White 12%
but they were free to attend other employment pro-
Other 6%
grams or seek access to other services. Baseline and
Foreign Born 23%
follow-up surveys were conducted with members of
Age
both groups, eliciting information about their educa-
18 to 24 28%
tion and work histories as well as their employment
18 to 26a 37%
experiences during the two-year study period. The
25 to 54 70%
follow-up survey sample included 1,014 respondents,
55 and Older 2%
reflecting a 79 percent response rate.
Average Age 32.2
Education
In addition to collecting data about individuals,
More Than a High School Diploma 18%
we also conducted annual site visits to each of the
High School Diploma 53%
three organizations to interview staff, participants
GED 22%
and others involved with the programs. The goal
Less Than a High School Diploma 7%
of this qualitative research was to document the
Other Characteristics
structure and content of the programs and to better
Married 18%
understand key practices as well as challenges the
Ever on Welfare 37%
organizations faced.
On Welfare at Baseline 23%
Has Access to a Vehicle 45%
Study Participants Average Number of Children in
1.2
Household
Participants in the study were screened through their Moved in Last Two Years 43%
respective programs to ensure they had the basic Completed Other Training Before
25%
academic skills to read and understand instructional Baseline
material; entrance requirements ranged from sixth Homeless in Year Prior to Baseline 7%
to tenth grade reading and/or math levels. In the Ever Convicted of a Crime 22%
year prior to the study, participants had been in Formerly Incarcerated 17%
and out of the labor market. Only 10 percent had Employment History at Baseline
worked full-time for the entire year, and the aver- Average Months Employed Year Prior
6.8
to Baseline
age participant had worked full-time for three and
Employed (Part-Time or Full-Time) at
a half months. Thirty-four percent were working 34%
Baseline
at the time they enrolled in the study. On average, Worked Full-Time All 12 Months Prior
10%
each had worked (for at least one hour) in seven to Baseline
months of the year prior to the baseline survey, earn- Average Months Working Full-Time
3.5
Year Prior to Baseline
ing $9,872. Nearly 40 percent had received public
Total Earnings Year Prior to Baseline $9,872
assistance at some time,4 including the 23 percent
a Since definitions of “youth” and “young adults” vary among practitioners,
of participants who were on welfare at the time of
researchers and funders, we analyzed the data according to two group-
enrollment.5 (See Executive Summary Table 1.) ings: ages 18 to 24 and ages 18 to 26.iv Tuning In to Local Labor Markets: Findings From the Sectoral Employment Impact Study
Analysis participants also worked significantly more hours—
about 245 more than controls over the 24-month
In evaluating the programs’ impacts, we looked at
study period and 250 more than controls in the
a number of key employment outcomes: total earn-
second year. Employment rates hovered around
ings, the likelihood of finding employment, number
70 percent for program participants in the second
of hours worked, the likelihood of working a job that
year, compared with about 60 percent for controls.
paid an hourly wage of at least $11 and at least $13,
and the likelihood of working a job that offers ben- 3. Program participants were significantly more
efits. Because the outcomes seen during the first 12 likely to work in jobs with higher wages.
months include time spent in training, internships
and the initial job search, we present both the effects Over the full study period, program participants
seen during the full 24-month study period and worked two more months than control group
those observed during the second year of the study members in jobs that paid at least $11 an hour, and
(i.e., months 13 through 24, when participants were 1.5 more months in the second year alone. The
fully available to participate in the labor market). likelihood of ever working a job that paid at least
$11 an hour was 14 percentage points higher for
program participants (59 percent) than controls
Key Findings
(45 percent) over the entire study period and 13
percentage points higher (55 percent for program
1. Participants in sector-focused programs participants and 42 percent for controls) in the sec-
earned significantly more than control group ond year. A similar pattern emerges when we look
members, with most of the earnings gains at the likelihood of working a job that paid at least
occurring in the second year. $13 an hour. Over the entire study period, program
participants worked about a month more in these
Participants in sector-focused training earned 18 jobs and their likelihood of ever working a job at
percent—about $4,500—more than controls over this wage level was eight percentage points higher
the 24-month study period. Not surprisingly, given than it was for controls.
that program participants were in training dur-
ing the first year, most of the increase in earnings 4. Program participants were significantly more
was seen during the second year. During months likely to work in jobs that offered benefits.
13 through 24, participants earned 29 percent
more than controls on average, or $337 more per During the full study period, program partici-
month—about $4,000 more overall. pants spent an average of 11 months working in
jobs that offered benefits (e.g., health insurance,
2. Participants in sector-focused programs were paid vacation, paid sick leave, tuition reimburse-
significantly more likely to work and, in the ment)—about a month and a half longer than
second year, worked more consistently than controls. In the second year, program participants
control group members. spent about seven months working jobs that offered
benefits—1.4 more months than controls. By the
Part of program participants’ earnings gains can be beginning of the second year, and continuing
attributed to the fact that participants were more through the end of the study period, the likelihood
likely to find work and worked more consistently. that program participants were working in jobs that
Over the 24-month study period, program partici- offered benefits was between 50 and 60 percent, as
pants were significantly more likely to be employed, compared with controls, whose likelihood ranged
working on average 1.3 more months than controls. between 40 and 50 percent over the same period.
During the second year, program participants were
significantly more likely than controls to work all 12
months (52 percent versus 41 percent)—an indica-
tion that sector-focused training programs helped
participants find steadier employment. ProgramExecutive Summary v
5. For each subgroup analyzed, program partici- JVS–Boston
pants had significant earnings gains as com-
JVS–Boston’s strategy was to provide participants
pared to their counterpart controls.
with job-specific occupational skills through an
intensive five-and-a-half-month training program
The three organizations in the study serve quite dis-
(the longest in the study) and to supplement this
tinct target populations; therefore, the subgroups
training with a high level of support. JVS–Boston
we examined (men, women, African Americans,
offered substantial support during and after the
Latinos, immigrants, people who were formerly
program. It was able to guide participants into
incarcerated, welfare recipients and young adults)
employment opportunities thanks to its knowledge
were not evenly distributed among the three sites.
of the healthcare sector. JVS–Boston’s results reflect
All subgroups, however, had significant earnings
this approach: Program participants saw 21 percent
gains; the timing of these gains and the programs’
earnings gains over the two-year period and a 35
effects on other employment outcomes (such as
percent earnings gain in the second year alone,
likelihood of being employed, working in jobs with
largely as a result of their being more likely to find
higher wages, etc.) varied among groups. It is likely
employment than their control group counterparts.
that some of these differences are due to differ-
They also worked more hours and were more likely
ences in the approaches at the three sites. It is also
to earn at least $11 an hour. Young adult program
worth noting that not all subgroups had earnings
participants did particularly well, perhaps reflect-
gains at each site.
ing the high level of support provided by program
staff; these younger participants earned almost 50
percent more than young adult controls. African
Program-Specific Findings
American participants and participants who had
Wisconsin Regional Training Partnership ever received welfare also saw earnings gains,
entirely due to working more months and more
The effects we see at WRTP reflect its overall strategy hours. We did not see any significant effects for
of providing short-term, job-specific training and then foreign-born program participants, who were older,
helping guide disadvantaged workers into higher- disproportionately male and more educated than
quality jobs than they might have been able to access the overall sample.
without its assistance. Overall, program participants
earned significantly more, even though they found
Per Scholas
employment at rates similar to their control counter-
parts. They were significantly more likely to work in Per Scholas’ strategy of providing its participants
higher-wage jobs, to secure union jobs and to work in with skills, preparing them to obtain a recognized
jobs that offered benefits. They were also more likely industry certification and offering internships
to obtain certifications in both the healthcare and and work experience is reflected in the program’s
construction tracks. Earnings gains varied across sec- effects. Not surprisingly, given the length of Per
tors: Construction participants saw the highest gains, Scholas’ training and the internship that often fol-
followed by healthcare; participants in manufacturing lows, program participants mainly saw effects in the
did not achieve higher earnings than control group second year. Program participants had significantly
members, which is not surprising given the region’s higher earnings and were significantly more likely
downturn in manufacturing. to work—and work in jobs with higher wages—
than their control counterparts. Program partici-
WRTP’s strategy also had different effects on earn- pants also earned the A+ certification at higher
ings for different types of workers: Both African rates, which may be a critical part of the value
American and women participants earned signifi- contributed by Per Scholas. Latino, immigrant,
cantly more than their counterpart controls, largely and formerly incarcerated program participants
as a result of higher wages. Formerly incarcerated earned significantly more than their control group
program participants also saw earnings gains, which counterparts; immigrant and formerly incarcerated
were attributed to working more hours than con- program participants fared particularly well. Young
trols as well as earning higher wages. For young adults between ages 18 and 24 did not earn signifi-
adult participants and welfare recipients, there were cantly more than their control group counterparts,
no significant earnings gains.vi Tuning In to Local Labor Markets: Findings From the Sectoral Employment Impact Study
though this was possibly due to small sample size. its job developers built strong relationships with
When the range is broadened to 18 to 26, program major employers.
participants did have significantly higher earnings.
3. Job readiness, basic skills and hands-on techni-
cal skills training offered through the lens of a
Common Programmatic Elements specific occupation or sector.
Each organization in the study employed a unique
strategy and crafted its program to respond to Effective adult education is essential to the success
local circumstances. Through site visits, focus of sector-focused training programs. Rather than
groups and interviews, we identified common ele- offering job readiness, basic skills and technical
ments shared by the three programs. While all the skills training separately, WRTP, JVS–Boston and Per
programs focused to some degree on each of these Scholas all addressed these needs together, through
elements, they were implemented differently at the lens of their targeted sectors.
each organization and, in some cases, were stron-
4. Recruitment, screening and intake processes
ger at one than another.
that result in a good match between the appli-
1. Strong organizational capacity—with the cant, the program and the target occupation.
ability to adapt.
Each organization established a screening process
Workforce organizations operate at the nexus that helped identify candidates who had both the
between disadvantaged workers, local employ- ability to benefit from its program and the potential
ers and the public and private agencies that have to be successful in the targeted occupation. This
resources to invest. Each organization in the study process began with outreach and recruitment efforts,
had capacities—resources, staffing, relationships, both of which were integral to each organization’s
institutional memory—that enabled it to under- operation and required considerable staff resources.
stand the specific needs of employers, target appro- The programs’ ability to so carefully target partici-
priate candidates and devise an intervention using pants who were an appropriate match for the target
public and private funding sources. While the sub- occupation (in terms of interest, ability and qualifica-
sequent programmatic elements we discuss are criti- tions) is a critical piece of their success.
cal, each organization’s ability to understand and
5. Individualized services to support training
deal with change—sometimes referred to as adap-
completion and success on the job.
tive capacity or the ability to ask, listen, reflect and
adapt—underlies its success.
For disadvantaged job seekers and workers, help
2. A strong link to local employers that results in with childcare or transportation or a referral for
an understanding of the target occupation and housing or legal services can be critical to staying
connections to jobs. in training or keeping a job. All three organizations
had mechanisms in place to deal with these needs,
An effective sectoral strategy rests on linking to the though delivery of the services varied.
workforce needs of local employers. Organizations
in the study forged this link in various ways. As
Conclusions
an association of employers and unions, WRTP
was able to work collaboratively with individual Mature, nonprofit-led sector-focused programs can
employers, sets of employers and union represen- increase the earnings of disadvantaged populations.
tatives. JVS–Boston’s links to the healthcare sector
were built through its long history of placing peo- This study provides compelling evidence that
ple in jobs with Boston-area employers, as well as nonprofit-led sector-focused training programs can
through the incumbent worker training6 it offered increase the earnings of a range of disadvantaged
to several major healthcare providers. Per Scholas populations. Results of the study also demonstrate
connected to the IT sector through its role as a that this approach can provide disadvantaged
recycling center for “end of life” computers, and people with access to industry-relevant skills andExecutive Summary vii
steady employment. While there has been sig- the three organizations studied. At WRTP, African
nificant growth in both the number of programs Americans, women and formerly incarcerated par-
that target specific industry sectors and the range ticipants experienced significant earnings gains. At
of institutions that operate or sponsor them, it is JVS–Boston, the program showed impacts for young
important to note that the programs in this study adults, African Americans, women and those who
are representative of mature, nonprofit-led sector- had been on welfare. At Per Scholas, immigrants,
focused programs and not all efforts that often fall men, Latinos, formerly incarcerated individuals and
under the umbrella of sectoral training. It is also young adults (18-26) had significant earnings gains.
important to recognize that the programs in this
study were more than simply job training programs. Nonprofit organizations can play a critical role in
Each organization had strong connections to local delivering workforce services. The three programs in
employers and identified specific job opportuni- this study demonstrated an adaptability that allowed
ties for which they trained program participants. them to connect disadvantaged job seekers to employ-
Each organization targeted people who would be ers using a mix of strategies and a range of public and
a good match for the occupation and the training, private funding sources.
provided essential supports and offered skills train- While the three programs in the study did not fol-
ing through the lens of a specific sector. This study low a common model, we found that their ability
points to the promise of programs that combine to combine key elements—good understanding of
these elements. and connection to industry needs, careful screen-
ing to identify appropriate clients, a sector-focused
Variation in approaches can be effective but results in
approach to training and individualized support
different effects on earnings.
services—seemed to contribute to success. The
The programs in this study varied in length, popula- organizations’ ability to keep pace with changes in
tions served and target industry/occupation. Each the local economy, funding agencies or partners
offered a mix of services with differing emphasis was also a key ingredient.
on making connections between participants and
employers, providing supportive services, and
training in occupationally relevant skills. The Implications for Further Research
longer-term training programs, JVS–Boston and These findings suggest the need for additional
Per Scholas, placed a stronger emphasis on skills, research about the effectiveness of sector programs
whereas WRTP emphasized connecting participants for disadvantaged people. Below we outline poten-
to jobs through its networks of unions and employ- tial avenues for further exploration:
ers. These strategies influenced earnings: WRTP’s
participants showed early earnings gains that were Can this approach be scaled?
largely a result of higher wages, while participants
at Per Scholas and JVS–Boston had earnings gains The organizations in the study served small num-
that came later and were a result of participants’ bers of program participants. Scaling up—either for
increased likelihood of finding a job and working these organizations or by other organizations adopt-
more consistently and/or at higher wages. ing this approach—presents some unique chal-
lenges, as sector programs are by their very nature
Mature, nonprofit-led sector-focused programs can be flexible—relying on clearly identified employer
effective with a range of disadvantaged workers and demand as well as available funding (either pub-
job seekers. lic or private) to provide services. More rigorous
research could tell us with greater certainty which
The three programs in the study served a range of the common elements we identified are indeed
of un- and underemployed people, including men essential, if there are other features we missed and
and women, African Americans, Latinos, immi- which combinations of elements are most effective
grants, people who were formerly incarcerated, in various situations. Additional studies could also
welfare recipients and young adults. We saw positive inform the increasing number of organizations
impacts on earnings for all subgroups, though there that are developing sectoral programs and increase
were differing impacts for various groups across the likelihood that their approach could replicateviii Tuning In to Local Labor Markets: Findings From the Sectoral Employment Impact Study
the impacts seen in this study. Research aimed at Closing Thoughts
understanding the costs of these programs is also
Sector-focused programs aim to connect disad-
important in considering how they can be scaled.
vantaged job seekers and low-skilled workers to
employment opportunities, addressing unmet
What about sector programs led by other types of
hiring needs of local employers and improving
institutions?
participants’ prospects in the labor market. As
While our findings show the promise of sectoral we emerge from the Great Recession, which has
programs run by experienced nonprofit organiza- disproportionately affected disadvantaged work-
tions that demonstrate the ability to adapt and ers, these strategies and the organizations that
respond to local circumstances, research is needed implement them may represent a key element in
about the effectiveness of sectoral efforts under- America’s economic recovery—for its workers
taken by other types of institutions, such as commu- and its employers.
nity colleges, Workforce Investment Boards, state
agencies and employer associations.
What about the role of industry certifications? Executive Summary Endnotes
1. Low-wage workers are defined as those who are paid a wage
Both Per Scholas and WRTP offered training
such that, even with full-time, full-year employment, their
that prepared participants to obtain industry- annual earnings fall below the poverty line for a family of four.
recognized certifications—a strategy that may See Loprest, Pamela, Gregory Acs, Caroline Ratcliffe and Katie
have played a major role in participants’ earnings Vinopal. 2009. ASPE Research Brief: Who Are Low-Wage Workers?
Washington DC: US Department of Health and Human
gains. Further research is needed to understand Services, Office of Human Services Policy, Office of the Assistant
how industry certifications affect earnings and Secretary for Planning and Evaluation.
wage gains and the role workforce organizations
2. A 2009 survey conducted by Manpower, Inc., found that 19 per-
can play in helping disadvantaged workers and job cent of United States employers reported having trouble finding
seekers gain access to jobs once they have attained skilled workers to fill vacancies. See Manpower, Inc. 2009. 2009
an industry-recognized certification. Further analy- Talent Shortage Survey Results. Manpower, Inc. For a discussion of
the challenges facing manufacturers looking for skilled workers,
sis using data from this study is forthcoming.
see Jusko, Jill. “The Training Imperative.” Industry Week, March
17, 2010. For a discussion of the shortage of healthcare work-
What strategies are effective for various groups of ers in California, see Lauer, George. “Shortage of Allied Health
job seekers? Care Workers Strains California Clinics.” California Healthline,
January 27, 2009.
Given their flexible design, sector-focused train-
3. Holzer, Harry J. and Robert I. Lerman. 2007. America’s Forgotten
ing programs both targeted and were effective for Middle Skill Jobs: Education and Training Requirements for the Next
many disadvantaged populations. More needs to be Decade and Beyond. Washington, DC: The Workforce Alliance.
understood about what blends of services are most
4. Repeated use of welfare is common. An analysis by the Urban
effective for different groups.
Institute found that 21.9 percent of those who leave welfare
return within two years. For more information, see Loprest,
What about impacts over time? Pamela. 2002. Who Returns to Welfare? Washington, DC: Urban
Institute.
While this study’s 24-month span allowed us to
examine the immediate impact of each strategy, 5. None of the programs in the study included welfare recipients
who had been mandated to attend the training.
longer-term studies would be valuable. Such studies
would allow us to see whether earnings gains grow 6. Incumbent worker training refers to training for currently
or diminish over time, and may cast a different light employed workers.
on the effectiveness of each approach.Introduction
Chapter I2 Tuning In to Local Labor Markets: Findings From the Sectoral Employment Impact Study
F
(e.g., formerly incarcerated individuals, welfare
recipients and people with only a high school
education or less). We did not seek organizations
that followed a common model to participate in
the study, as sectoral programs employ various
or American workers, having a high school approaches depending on the organization lead-
or general equivalency diploma (GED)—which ing the effort and local employers’ needs. Instead,
once represented a means of entrance to the mid- we sought mature programs that seemed to be well
dle class—is no longer adequate for finding steady, implemented, since it takes time for an organiza-
quality employment. In fact, three quarters of low- tion to both understand employers’ needs and craft
wage workers1 have these qualifications but lack appropriate responses.
the relevant occupational skills and connections to
employers needed to launch a career. At the same Through nominations from leaders in the work-
time, in some regions of the country there are per- force development field, P/PV identified 25 orga-
sistent skills gaps clustered in particular industries, nizations that targeted an occupation or cluster of
such as manufacturing and healthcare.2 Many of occupations, that aimed to place participants in jobs
these jobs are expected to grow3 and require spe- paying $8 or more per hour, that served more than
cific technical skills that can be gained only through 100 participants annually and that had been operat-
focused training that is closely linked to the needs ing their programs for at least three years.
of local businesses.
Three organizations were selected to participate in
Over the past two decades, an innovative approach the study (see Appendix A for more details about
to workforce development known as sectoral the selection process):
employment has emerged, resulting in the creation
of industry-specific training programs that prepare • The Wisconsin Regional Training Partnership
unemployed and underskilled workers for skilled (WRTP) is an association of employers and
positions and connect them with employers seeking unions that seeks to retain and attract high-wage
to fill such vacancies. Beginning in the early 1990s, jobs in Milwaukee and create career opportuni-
with support from private foundations, several non- ties for low-income and unemployed community
profit community-based organizations developed residents. WRTP develops training programs
strategies aimed at improving the prospects of (generally lasting between two and eight weeks)
low-income workers by meeting the needs of local in response to specific employers’ requests or to
businesses. In 1998, to explore the potential of such clearly identified labor market needs. Its short-
strategies, Public/Private Ventures (P/PV) launched term preemployment training programs in the
the nine-site, three-year Sectoral Employment construction, manufacturing and healthcare
Initiative, with support from the Charles Stewart sectors were included in the study. Study partici-
Mott Foundation. An evaluation of this initiative pants were primarily African American and were
showed that after two years, participants in pro- about evenly divided between men and women;
grams that offered sectoral training had higher about 40 percent had been incarcerated.
hourly wages, increased annual incomes and better- • Jewish Vocational Service–Boston (JVS–Boston)
quality jobs compared to the year prior to their is a community-based nonprofit that has pro-
enrollment. These findings were echoed in a similar vided workforce development services for more
outcomes study conducted by the Aspen Institute.4 than 70 years, including operating one of three
While these findings were encouraging, more rigor- One-Stop Career Centers (One-Stops) funded
ous research was needed. by the Workforce Investment Act (WIA) in the
Boston area. The organization aims to serve a
With continued funding from the Mott Foundation, diverse range of Boston’s disadvantaged popula-
P/PV set out in 2003 to conduct a random assign- tions, including refugees, immigrants and wel-
ment evaluation to assess whether sector-focused fare recipients. Its training programs in medical
programs could increase the earnings of low- billing and accounting were included in the
income, disadvantaged workers and job seekers study. Each training program was provided overIntroduction 3
Program Components
WRTP JVS–Boston Per Scholas
Sector Focus • Manufacturing, construction • Clerical and medical office • Information technology
and healthcare occupations • Staff identify employers’
• Employers, including orga- • Employers serve on advisory needs and develop relation-
nized labor, are members of committees; staff develop ships through a social venture
WRTP, serving on committees one-on-one relationships with to recycle and refurbish com-
to identify and address needs employers and use an account puters; employers participate
of member businesses, market management system to iden- in job fairs and mock job inter-
services and advise about the tify, address and monitor their views and advise about the
training curriculum. needs. curriculum.
Enrollment • Sixth- to tenth-grade reading • High school diploma or GED • High school diploma or GED
Requirements level, depending on sector • Sixth- to eighth-grade reading • Tenth-grade level (both read-
• Interview to determine career and/or math level, depending ing and math)
goal and participation chal- on sector • Interview to determine career
lenges • Interview to determine career goal and participation challenges
• Driver’s license (no more than goal and participation chal-
five violation points) for con- lenges
struction sector • Staff team agreement of
• Negative drug screen for acceptance
healthcare sector
Preemployment • Training length varies: 2 to 8 • Training length varies: 20 to • Training is 15 weeks, 500
Training and weeks, 40 to 160 hours 22 weeks, 20 to 25 hours per hours
Certifications • Certifications for nursing week • A+ certification
assistants, medical assistants • Certificate of completion for
and construction the training
Employability • “Essential skills” related to • Four- to six-week internship • Internship
Activities timeliness, attendance, strate- • Job readiness training (e.g., • “Life skills” training related to
gies for dealing with childcare, writing resumes and cover let- goal setting, communication,
workplace issues and operat- ters, job interviewing) interviewing for a job and time
ing within the industry culture management
integrated into technical training • Employability workshops
Supports • Case management • Case management • Career mentoring
• Childcare and transportation • Childcare and transportation • Counseling
for those receiving TANF assistance • Job placement
• Job placement • Job placement • Postemployment retention
• Postemployment retention • Postemployment retention services
• Remedial education as services • Assistance with work attire
needed • ESL/basic skills tutoring as
• Assistance to get a driver’s needed
license • Tax preparation assistance4 Tuning In to Local Labor Markets: Findings From the Sectoral Employment Impact Study
20 to 22 weeks for 20 to 25 hours per week.
JVS–Boston engaged its target industry by form-
ing employer advisory committees and building
individual relationships with local businesses.
JVS–Boston study participants were primarily
women and included a large number of young
adults and current or former welfare recipients.
• Per Scholas is a social venture in New York City
that combines a training program with efforts to
refurbish and recycle “end of life” computers and
distribute them to low-income people through
partnerships with nonprofits, schools and com-
munity colleges. Per Scholas’ computer techni-
cian training program—which prepares partici-
pants for jobs in the repair and maintenance
of personal computers, printers and copiers, as
well as the installation and troubleshooting of
computer networks—was included in the study.
The training program consists of 500 hours over
a 15-week period and is aligned closely with the
industry-recognized A+ certification—which
demonstrates computer technician competency.
Program participants also take part in intern-
ships, during which time they work in the Per
Scholas recycling and refurbishing center or with
local employers. At Per Scholas, study partici-
pants were primarily male, and a sizeable propor-
tion was foreign-born.
Since the early 1990s, and indeed since this study
was launched, the number and types of organiza-
tions pursuing sectoral employment strategies
have grown. Today, community colleges, workforce
investment boards, labor-management partner-
ships, business associations and other agencies
have adopted this approach, and many sectoral
programs receive support from federal, state
and local government sources.5 This report pres-
ents the findings of the first rigorous random
assignment study of three nonprofit-led sector-
focused efforts: an employer/union association,
a social venture and a human service organiza-
tion. Chapter II of this report outlines the study’s
design and methodology; Chapter III describes the
findings across all three programs; and Chapter IV
analyzes the strategies and findings for each site
individually. Chapter V presents a discussion of the
common programmatic elements, as well as com-
mon challenges. Finally, Chapter VI summarizes
our conclusions and outlines implications for
further research.Study Design
Chapter II6 Tuning In to Local Labor Markets: Findings From the Sectoral Employment Impact Study
response rate (75 percent for the control group
and 82 percent for the treatment group).7 The
programs’ effects were measured by comparing the
used an experimental progress made by members of the treatment group
research design to bring as much rigor as possible with that made by members of the control group.
to the following question: Do mature sector-focused Because assignment to these groups was random,
programs result in significant labor market gains for any differences found between treatments (here-
low-income, disadvantaged workers and job seekers? after referred to as program participants8) and
More specifically, we strived to determine whether controls can be attributed to participation in the
such programs raise the earnings of program par- sector-focused training programs.
ticipants and whether participants were more likely
to find employment and work more consistently. In addition to collecting data about individuals, we
We also wanted to explore whether program par- conducted regular site visits to each of the three
ticipants obtained higher-quality jobs. For example, organizations. The goal of this qualitative research
were participants more likely to earn higher wages? was to identify key practices as well as challenges
Did participants find jobs with better access to ben- the organizations faced. Once a year, P/PV inter-
efits? Further, we set out to explore whether specific viewed both frontline staff (such as job developers,
groups of people, such as welfare recipients or young case managers and career specialists) and supervi-
adults, benefit from participation. We also sought to sors and senior management. Focus groups were
understand the programmatic, contextual and indi- also held annually with participants, and on occa-
vidual factors that contribute to these outcomes. sion interviews were conducted with employers
and board members of the participating organiza-
To answer these questions, 1,286 people were tions. Although the study design did not include
recruited for the study from across the three pro- the collection of detailed information on program
grams over a two-year period, all of whom had intensity or the costs associated with program
been through their program’s application process implementation, the qualitative component of our
and met its eligibility criteria.6 Baseline data were research did enable us to document the structure
gathered from eligible applicants through a phone and content of the programs.
survey about their education and work histories,
additional sources of income, living situations and
experiences with other employment programs. Study Participants
Then, half of the participants were selected at ran- Table 1 details the baseline characteristics of the
dom to participate in the program (the treatment study’s entire follow-up sample. The study par-
group); the remaining half (the control group) ticipants shared many characteristics across all
could not receive services from the study sites for three sites, though there was some variation in the
the next 24 months, but they were free to attend demographics from site to site. These differences
other employment programs or seek access to other will be explored in the site-specific sections of this
services. No significant differences existed between report. For the baseline survey, we used a number
the treatment and control groups at the time of the of statistical techniques to determine the success of
baseline survey (see Appendix B). random assignment (see Appendix B for details)
and concluded that there were no measurable dif-
Members of both groups were surveyed by phone ferences between program participants and the
between the 24th and 30th months after the base- control group—overall or at each particular site.
line survey was conducted. During the follow-up At the time of the follow-up, we found differences
survey, participants were asked to provide detailed between program participants and controls in 3
information about every job they had worked dur- of 31 characteristics; further analysis using linear
ing the study period, including earnings, months regression suggested that the random nature of the
worked and weekly hours, and whether partici- baseline sample was maintained at the follow-up
pants were offered and had taken advantage of (see Appendix B for more detail about the baseline
benefits. The follow-up survey sample included and follow-up samples).
1,014 respondents, reflecting a 79 percentYou can also read
