generate highly productive workers who are well compensated (Appelbaum et al., 2003).
Furthermore, the ability of employers to choose the “high road” might be very
constrained by the quality of workers whom they perceive to be available for hiring, in
terms of basic and occupational skills. And employers might be reluctant to invest their
own resources in training workers for a variety of reasons.
When employers do, in fact, create good-paying jobs, the evidence suggests that the
match between these jobs and the skills of workers filling them is growing more
important. Drawing on longitudinal data on both workers and firms for a dozen states
during the 1990s and 2000s, a recent book by Holzer et al. (2011) identifies a rising
correlation between measures of worker and job quality over time (see Table 1). The
fractions of the jobs in the top quintile of quality that are filled by workers in the top
quintile of skills grew in this period, as did the fraction of lower-paying jobs (especially
in the fourth quintile) filled by less-skilled workers.
Furthermore, the locus of the “good jobs” is changing, with many fewer available in
manufacturing and more appearing in the professional and financial services, health care,
construction, and even the high end of retail trade (see Table 2). And the decline in good
job availability in manufacturing is concentrated among the least-skilled workers, whose
employment there declined dramatically; in contrast, employment in manufacturing for
workers in the highest skill quintiles declined only mildly.
Fortunately, the data show that good jobs are not disappearing in general. If anything,
Holzer et al. show that the numbers of jobs in the highest quintile of quality were actually
growing in this period. But most of the high-paying jobs require a strong set of basic
cognitive and/or communication skills. And, while many do not require a four-year
college diploma (outside of the professional and financial services), they generally
require some kind of postsecondary training and certification. In health care, these
positions often include a variety of nursing categories as well as technicians. In
construction, they usually include the skilled crafts (electricians, plumbers, carpenters)
which can be filled through apprenticeships or other training models. In manufacturing,
they often include not only engineers but also machinists, precision welders and other
highly-skilled workers (Holzer, 2010). Similarly, Figure 1 (from Carnevale et al. 2010)
indicates a range of sectors and/or occupations where those with some college or an
Firms are often reluctant to pay for general skills training, since workers might not stay with the firm
providing such training
wage rigidities, imperfect or asymmetric information about
prospective workers, liquidity constraints and other market “failures” might make firms reluctant to make
such investments as well (Becker, 1975; Acemoglu and Pischke, 1999).
Specifically, the percentages of top-quintile jobs filled by top-quintile workers rose from 63.6 to 67.7
percent in just over a decade, while the percentages of second-quintile jobs filled by second-quintile
workers rose from 34.1 to 38.6 percent. In contrast, the percentages of four-quintile jobs filled by workers
in the bottom two quintiles of skills rose from 59.5 to 65.4 percent.
Worker and job quality are measured by worker and firm “fixed effects” respectively, calculated from
matched longitudinal data on workers and their employers. See Holzer et al. for more detail.
Other data from Holzer et al. (Table 2.5 of their book) show that manufacturing employed 27.6 and 24.4
percent of all workers in the top two quintiles of skills respectively in 1992, which declined to 24.8 and
18.4 percent respectively by 2003. But for workers in the bottom two quintiles of skills, these percentages
declined much more dramatically, from 17.7 and 10.1 percent to 8.3 and 4.0 percent respectively.
associate’s degree enjoy relatively high earnings. These include managerial and
professional jobs, those in the “STEM” fields (i.e., science, technology, engineering, and
math), healthcare, sales and office support work, and even blue-collar fields. Elsewhere,
Carnevale et al. show that significant percentages of workers with occupational licenses
or certificates as well as those with associate’s degrees earn more than the median worker
with a bachelor’s degree in key fields.
Despite the value of these skills, certain well-documented problems in our education and
workforce systems mean that too-few workers make investments that would allow them
to fill these good-paying jobs. For example, many students currently attend two-year or
four-year institutions but achieve too little there to improve their labor market outcomes.
Dropout rates are extremely high, especially in community college, where so many youth
and adults – especially from minority or low-income communities – get stuck in remedial
classes from which they never emerge and which are completely separated from the
classes providing the relevant occupational training. As a result, most students there
never earn even an occupational certificate, much less an associate’s degree. Data from
the American Association of Community Colleges indicate that 12.4 million students
attended community college in the fall of 2008, and about 7.4 million for credit; yet
fewer than a million associate degrees or certificates were awarded the previous year.
Bailey et al. (2005) also find that fewer than half of all community college students
complete a degree or certificate after five years, and completion rates are lower among
minorities and those with low incomes.
In Germany and elsewhere in Europe, training that helps workers prepare for good labor
market opportunities is delivered through high-quality career and technical education
(CTE). Such systems have not developed in the U.S. at least partly because of historical
controversies here over “tracking” minority students away from college (Hoffman, 2011;
Symonds et al, 2010). But, at its best, CTE would not deter students from attending
postsecondary institutions, and might indeed be structured to better prepare and
encourage more students to do so.
Indeed, it is often not until after entering the labor market and becoming unemployed that
many disadvantaged workers are provided their first valuable career guidance. Such
guidance is provided quite cost-effectively to workers at over 3000 “One-Stop” offices
around the country, funded through the Department of Labor’s Workforce Investment
Act (WIA), in the form of “core” and “intensive” services plus very limited training
(Besharov and Cottingham, 2011). In contrast, fairly little career guidance is provided to
high school or community college students, especially based on local or state labor
market data (Soares 2010). And our colleges and these workforce institutions are largely
isolated from one another in many states, with little effective interaction on the ground.
Local workforce boards, which disperse funds provided through the Workforce
Investment Act (WIA), do not always effectively represent the best-paying employers
with strong demand in growing industries, and are not always integrated with state and
local economic development efforts.
And, even when college students know that earnings and labor market demand are strong
in certain fields (like nursing or health technology), they often find very limited
instructional capacity in these areas in many colleges; perhaps this reflects a lack of
incentives for institutions to meet labor market demand, because their per-student
subsidies from state governments do not depend on degree or certificate completion rates
or on what kinds of credentials they earn.
As a result, not only do too few workers obtain certificates and degrees, but those
obtained are often not well-matched to labor market demand in key sectors. Under these
circumstances, when employers create high-paying jobs at both the middle and high ends
of the skill spectrum, they often seem to have some difficulty filling them with skilled
workers. Indeed, the job vacancy rate has averaged 2.2-2.3 percent over the past year,
which is relatively high given an unemployment rate of over 9 percent. And even in some
sectors where vacancy rates are not high overall – such as manufacturing – the ratio of
vacancies to new hires is striking, suggesting some difficulty that employers have filling
All of this suggests that programs designed to improve the skills and productivity of US
workers, if they also work carefully with targeted employers and industries, could help
fill vacant jobs that currently exist and perhaps encourage employers to create more such
jobs over time. The programs should thus help reduce unemployment and job vacancies
in the short term while also raising worker earnings in the longer term as well.
B. Potential Solutions: Education and Training for Good Jobs
One path to creating good jobs for disadvantaged workers involves raising their skills and
productivity to make them more attractive to potential employers. A rigorous body of
evidence suggests that certain education and training efforts can be cost-effective ways to
address these issues, even when brought to substantial scale. While the overall evidence
on the cost-effectiveness of job training for disadvantaged workers in WIA and elsewhere
is at least modestly positive (Holzer, 2009; Heinrich et al., 2009; Heinrich and King,
2010), there are some particularly strong examples that demonstrate the effectiveness of
education and training that target good-paying jobs on the demand side of the labor
market and is coordinated with employers there. The best studies have demonstrated
results on these programs using experimental methods from randomized control trials
(RCT), though some fairly persuasive nonexperimental evidence exists as well. RCT
studies are important because they allow researchers to compare the labor market
For evidence on the high variance in rewards to community college degrees and certificates for youth and
dislocated adults workers see Jacobson and Mokher (2009) and Jacobson et al. (2003) respectively.
In manufacturing, a job vacancy rate of 1.2 percent and a new hires rate of 1.4 percent in the most recent
JOLTS data from the Bureau of Labor Statistics (April 2011) implies an existing job vacancy for almost
each new hire, which constitutes a much higher ratio of vacancies to hires than exists in any other broad
industry group in these data. Elsby et al. (2010) show evidence that the job vacancy rate associated with a
given unemployment rates in the U.S., known as the “Beveridge Curve,” has shifted out over the past
several years. Uchitelle (2009) and Fletcher (2011) provide journalistic evidence of small employers who
have had difficulty finding skilled workers, even at the trough of the recent downturn.
outcomes of those who receive training to the outcomes of those who do not to
demonstrate the benefits and costs of each intervention.
The most important recent study is the Sectoral Employment Impact Study (SEIS) of
three major programs in Boston, New York and Wisconsin, conducted by PPV (Maguire
et al., 2010). The evaluation used random assignment methods to test for program
impacts on workers’ subsequent earnings, and it found large impacts on earnings in all
three programs in the second full year after random assignment, generated by 3-6 months
per worker of well-targeted training. Net impacts on earnings were about $4000 per
participants over the 24-month period after random assignment but about $4500 in the
second year, once training was completed. Direct costs of the program were estimated to
be about $6000 per worker.
Assuming the large earnings gains persist into the third
year, the program is clearly cost-effective.
The study’s authors have attributed the programs’ success to the close relationships
between employers, training providers (which are sometimes but not always community
colleges) and the intermediaries who coordinate their efforts. Improved earnings were the
results both of higher employment rates and higher wages for this population. Since the
three programs evaluated are moderately large, the evaluation demonstrates that effective
programs can potentially be brought to scale.
Impacts of similar magnitudes have been estimated recently in a random assignment
study of Year Up, a sectoral training program for out-of-school youth. The program
trained 18-24 year olds from low-income urban neighborhoods for jobs in the IT sector in
New England, New York and elsewhere The year after the program took place, the
treatment group reported earnings that were on average $3461 more (30% higher) than
the control group due to higher hourly wages (Roder and Elliott, 2011).
Several other efforts that provide occupational training plus work experience to students
in key sectors have generated impressive estimated impacts in evaluation studies as well.
Regarding CTE in high school, a random assignment evaluation of the Career Academies
have shown large impacts on the earnings of young men, especially those deemed at risk
of dropping out of school, even eight years after random assignment.
focus on particular sectors of the economy, and combine high-quality general academic
instruction with a more occupational kind, and provide critical work experience in those
sectors to students. And, quite importantly, the academies did not “track” students away
from postsecondary education, as the postsecondary enrollment rates of these students
were no lower than those of students in the control group.
Thus, we now have rigorous experimental evidence on highly cost-effective programs for
in-school youth, out-of-school youth, and disadvantaged adults, all of which provide
education or training that closely target good-paying employers or economic sectors and
This information was provided to me by Sheila Maguire, lead investigator on the SEIS study, in a private
See Kemple (2008). Impacts for young men after eight years were still nearly $2000 per year for each
where outreach to and active engagement of employers is a major part of the training
process. This evidence supports nonexperimental evidence on effective training programs
that suggests, for instance, that apprenticeships have also generated large impacts on
earnings in some evaluation studies, as have various state-level programs providing
incumbent worker training (Hollenbeck, 2008; Lerman, 2010).
It is important to note that all of these relatively successful programs have been in
operation for many years, and have developed strong curricula and links to the business
community that might not be easily replicated in a short time period. Furthermore, they
focus on disadvantaged workers with strong enough basic skills and education credentials
to successfully handle moderately technical training. I believe these successes can be
replicated in other settings over time, but only with appropriate screening of candidates
and careful development of their occupational training curricula and ties to employers.
A few other education or employment programs in community colleges low-income
neighborhoods that have undergone evaluation (with varying degrees of rigor) also
deserve some mention. The strongest evidence, based on RCT research designs, shows
positive effects on educational outcomes from the Opening Doors community college
interventions, which include merit-based financial aid, the structuring of “learning
communities” of students, and certain kinds of mandatory counseling on educational
outcomes as well (Richburg-Hayes 2008). Nonexperimental evidence from a program in
the state of Washington that integrates developmental (or remedial) education with
occupational training, known as the Integrated Basic Education and Skills Training (I-
BEST), shows positive effects on educational outcomes in a study of community colleges
(Jenkins et al 2009). And the Youth Opportunities (YO) program at the U.S. Department
of Labor, which provided grants to 36 low-income communities to develop and
coordinate local educational and employment services for youth, generates some positive
impacts as well on both educational and employment outcomes in these sites (Decision
Information Resources 2008). Thus, the potential to improve educational outcomes at
community colleges and to build systemic efforts to provide employment services have
been demonstrated in a range of studies.
III. The Proposal: Competitive Grants to Foster Education and Workforce Systems
Given the strong recent evidence on the efficacy of job training for disadvantaged
populations that carefully involves employers and considers labor demand, there is
clearly some strong potential to raise the skills of workers. Doing so would allow some
currently low-skilled workers to fill existing jobs and could also help create new
employment opportunities if employers respond to a more productive set of workers by
creating more good-paying jobs for them.
The goal of this proposal is to encourage the creation of more effective education and
workforce systems that include evidence-based training models and are more responsive
to employers who create good jobs. Given current and future budget constraints, any new
public expenditure should mostly be designed to improve the efficiency of resources
Documents you may be interested
Documents you may be interested