Adapted from remarks by Gordon L. Berlin, President, MDRC, at the Recovery and Reemployment Research Conference, sponsored by the Employment and Training Administration of the U.S. Department of Labor, September 15, 2009, Washington, D.C.

This is an extraordinary moment in American workforce history. Extraordinary because the downward secular trends in key employment indicators, once only dimly perceived, are now deeply etched in every employment chart. Extraordinary because havoc is being wreaked in labor markets by this Great Recession that so many unemployed and underemployed Americans are enduring. And extraordinary because the first major act of the new leadership of the Employment and Training Administration (ETA) was to conceive, sponsor, and organize a conference on research and reauthorization, asking the question: What is the evidence base to guide policy and practice in this new environment? Assistant Secretary Jane Oates and Deputy Assistant Secretary Gerri Fiala are signaling a bold and visible commitment to reestablish the role of evidence in policymaking.

As many of the people in this audience recall, the 1970s witnessed perhaps the largest surge of workforce development-related programming and research ever. However, much of that research was terminated prematurely in the early 1980s before it could be finished and before the benefits of the investments could be realized. But now we have an administration that is committed to using evidence to improve program effectiveness and willing to make the needed investments in new knowledge development. In fact, one of this administration’s mantras is: Implement proven programs; evaluate promising ones. This is a weighty responsibility to fulfill, not just for ETA but also for researchers and program operators alike.

Progress means building on what we know and remedying what we don’t. Given the choices that will be forced on us by massive future budget deficits, nothing could be more important in the long term than investing in a strategic research agenda that can identify effective workforce development strategies — for adult workers, dislocated workers, and disconnected youth. This is especially true now, when we are in the midst of the greatest economic downturn in seven decades, a time when every program dollar needs to be spent wisely and every program will be asked to prove its effectiveness. But a commitment to evidence-based policy and practice poses a daunting challenge: Yes, we have learned a lot over the past 30-plus years that can be better applied today, but there is much that we do not know. We need to learn more.

Now, as Workforce Investment Act (WIA) reauthorization begins to gather some momentum, and ETA begins to shape a revitalized research agenda, is the ideal time to step back and ask three critical questions.

• The first question. What is the historical economic and labor market context in which this attempt to remake employment policy is unfolding? What path has employment policy followed? What has happened to the economy and to labor markets? Which trends are long-term secular trends, and which are a result of the current recession?



• The second question. What are the big-picture lessons from over three decades of research into workforce development and related programs? This could point us in directions relevant to the WIA legislation and current program operations.



• The third question. In what critical areas is the record weak or nonexistent? This could point to new directions for future research.

Question 1: Setting the Context

Regrettably, the U.S. Department of Labor has not played a central role in the affairs of the nation’s workforce for more than 40 years, when the response to the employment problems of the late 1970s was the creation of public jobs programs under the Comprehensive Employment and Training Act (CETA). Now in the midst of the Great Recession, employment problems are once again front and center. Are the policies, programs, and systems that currently exist up to the task? Will the principles now being formulated to guide WIA reauthorization not only serve to remedy what has gone awry but also offer a sufficient response to what lies ahead?

We face two critical challenges in answering these questions: an employment system that has until recently been marginalized, largely because of inadequate funding, and a labor market in the midst of transformational changes. For much of the past 40 years, employment policy was viewed as a backwater. The governing mantra during this period was: Get the economy right; unleash the power of capital markets, and these labor market problems will take care of themselves. As a result, employment policies and programs atrophied. The unemployment insurance system eroded to the point where a majority of laid-off workers did not qualify for benefits. And state UI Trust Funds, especially those in the largest states, were allowed to run dangerously low. The summer jobs programs that were a hallmark of youth employment programming were ended entirely. Moreover, the minimum wage did not keep up with inflation until recent increases and remains far below its original historical benchmark (50 percent of the average wage). And because these recent increases are not tied to inflation, its inflation-adjusted value will likely erode again in future years.

Funding for WIA fell roughly 25 percent between 2000 and 2008 at a time when the economy twice fell into recession. Even with the added stimulus money, WIA funding now is less in real terms than what it was under the Job Training Partnership Act (JTPA), which, in turn, considering all programs, was roughly half of CETA’s funding levels in real dollars. As a result, the number of people who can be served relative to the number in need of services is small, and the average investment per enrollee also tends to be small. Yet today’s workforce is more than 50 percent larger than it was in the 1970s, and the economy as measured by Gross Domestic Product is more than two times larger than it was in the CETA years. One consequence of this — and the WIA legislation itself — has been that the system focused on quick placement into jobs of a relatively more advantaged population needing less intensive investments and services. Recent regulatory shifts earlier this year, as well as funding and programs in the American Recovery and Reinvestment Act (aka the Stimulus Bill) offer a start at ameliorating these problems. But only a start.

Given employment policy’s backwater status, it should come as no surprise that the four perennial themes that have challenged it since the passage of the Manpower Development and Training Act remain unresolved:

• The first theme centers on whether the role of employment policy should be purely structural or also countercyclical. Recall this debate in CETA — public service employment began as a countercyclical effort, and then, after criticism, began to target the structurally unemployed and gradually lost favor and support. With unprecedented federal extensions in unemployment benefits, questions about the role of public employment programs in a recession are now upon us.



• A second theme revolves around whether the system is designed to help employers or employees. This was a central debate in JTPA, and it was revisited under WIA.



• A third theme involves the delivery system and the mix and content of services. What would a human resource investment policy for the State of Maryland — or Michigan or Maine or Massachusetts or Mississippi — look like? No state has ever really had one, yet it is the central issue facing many states today.



• Fourth is the theme relating to the link between the workforce system and other systems — for example, welfare or education. What is the jurisdiction of the Department of Labor versus the Department of Education at the federal and the state level in the revamping of the community college system or in the role of career and technical education in high school?

Unfortunately, we have not done enough to build a solid record of evidence that would help us chart our course. We spent an enormous amount of money in the 1970s tackling the problem of displaced workers but have little in the way of institutional memory to keep us from reinventing the wheel. Similarly, we have spent greatly on youth programming, but when the evidence was discouraging, instead of rolling up our sleeves and investing in new approaches, we turned our back on America’s young people. 

If backwater is the context for workforce policy since the 1970s, over this same 40-year period, the economy and labor markets were undergoing profound changes as a result of globalization, technological advances, the enormous influx of immigrants, and other forces. Inequality widened as wealth accumulated in the hands of the top 10 percent and, overwhelmingly, in the hands of the top tenth of 1 percent. The concepts of a career ladder and upward mobility vanished for many Americans. The labor market became more bifurcated, to the point where we are debating whether middle-level jobs are declining and what can be done to support and grow a next generation of middle-skill jobs that rely on community colleges and sectoral training programs.

These economywide changes have had a long-term negative effect on wages, earnings, incomes, and poverty rates. As a nation, we have not fully internalized just how profound these changes have been. Compare the way the economy distributed its benefits after 1973 with the way it distributed benefits before 1973. In the 30-year period following the end of World War II, wages, earnings, and incomes had grown steadily. It was as though everyone was on an Up escalator. Inequality narrowed. Growing numbers of unionized manufacturing and blue-collar jobs paved a road to upward mobility. Those with a high school diploma or less benefited. So did African-Americans and low-income Americans in general. Poverty fell; marriage rates rose; the typical high school graduate could earn enough to support a family of three or four above the poverty line. America was a land of opportunity, promise, and dreams fulfilled. Each generation of Americans was better off than the last.

Then in the early 1970s, that Up escalator came to a slow, grinding halt. Economic recoveries took an ominous new twist — economic growth seemed to become divorced from labor market growth; jobless recoveries became the norm; job growth would lag economic recovery following recession by 18 months or more; and then even when job growth resumed, earnings growth lagged. Today, the average worker’s inflation-adjusted earnings haven’t budged in 36 years. The average full-time male worker in 1973 earned roughly $41,000; today the average full-time nonsupervisory male worker earns $40,000. According to estimates by Sheldon Danziger, if earnings growth in the past 30 years had tracked economic growth in the same way that it did in the postwar era, the average male full-time worker today could be earning as much as $90,000.

In short, getting the economy right was no longer a guarantee of getting employment right. As a result, a set of long-term secular changes have transformed American labor markets:

• Average earnings have been stagnant or declining for more than 36 years.



• Employment rates have been declining among men and, just since 2000, are down roughly 10 percentage points among 20- to 34-year-olds — an enormous change.



• The youth labor market has all but collapsed. In the mid-1970s, half of all 16- to 19-year-olds worked in the summer months; during summer 2009, only 28 percent of teens worked — the lowest U.S. teen employment rate on record. If you believe that young people need both education and employment experience to make a successful transition to adulthood, the inability of young people to obtain summer and school-year part-time jobs is deeply concerning.



• Immigration — and the flow of people and capital across borders — is as great as at any time in recent history, increasing competition, especially at the low end.



• Increasingly, the labor market is placing a premium on higher education for better-paying jobs; at the same time, nearly half of all jobs in 2016 will be low wage, signaling a bifurcated labor market.

Now add to these long-term secular trends the short- and intermediate-term effects of this Great Recession. A generation of workers under age 30 is missing out on critical employment experiences that would have determined their lifetime age-earnings profile. Can they possibly recover, or are these scars permanent? More than a third of the unemployed have been jobless for more than 27 weeks — a record high. The downturn is also having a disproportionately negative effect on African-American and Hispanic males. “Devastation” is too mild a description for what has happened to the manufacturing sector in some regions, especially the Midwest, where the specter of permanently displaced workers looms to an extent not seen since the 1970s. The collapse of the construction sector is having especially grim implications for immigrant workers. If the economic toll is enormous, the human toll is incalculable. 

There is a mantra sweeping through the business community right now: Reset versus recession. It says that every eight years or so, we have a recession, which requires some adjusting, but that every 30 years, we enter not a recession but a “reset” economy, which requires fundamental rethinking, retooling, and restructuring. By necessity, business is now engaged in that reset process. Jettisoning the old and embracing the new — all in a desperate, determined effort to return to profitability. Given the disconnect between policy and labor market change sketched above, what is the equivalent reset for workforce policy? This is the context in which reauthorization will take place.

Question 2: Lessons from Research

Turning to the second question I posed above — What have we learned from research about strategies that work? — five lessons come to mind.

• Targeting matters. The first lesson has to do with targeting: Who should be served when dollars are scarce? Targeting a somewhat harder-to-serve population, as ETA has proposed, is fully consistent with a wide body of research that shows that programs make the biggest difference with a middle to more disadvantaged group. ETA’s principle calling for simpler eligibility criteria and, therefore, simpler documentation requirements would support a shift to serving a more disadvantaged population, since complicated intake rules requiring more than one visit to the office tend to weed out the more disadvantaged. Serving populations with programs that address barriers to employment can make a bigger difference and add greater value, but it requires an eligibility and application system that allows the more disadvantaged in the front door. The large-scale displacement now under way may require further rethinking.



• Performance measures can mislead. Second, there are big differences between the outcomes by which programs are measured, such as the entered-employment rate or postprogram earnings gains, on the one hand, and the impacts and return on investment that programs achieve, on the other hand. In rigorous study after rigorous study, there is simply no correlation between achieving high outcomes, such as a high entered-employment rate, and running effective programs. And this holds true whether you are looking at short-term or long-term outcomes. The reason for this conundrum is a function of the first lesson, on targeting: High outcome standards can drive programs to serve those who would have done nearly as well without the program. A surefire way to increase General Educational Development (GED) completion rates is to enroll tenth-grade readers; a surefire way to increase employment rates is to enroll clients with a recent work history. But these are people who are very likely to have gotten a GED or to have found employment on their own, and the program’s value added was negligible.
  
  It’s also true, however, that programs need standards to strive for. What’s to be done? A stylized response is to start with a program that has been proved effective and then to develop performance standards that (1) measure program fidelity; (2) document who is served, to ensure that the program is serving people with the same characteristics as those served in the proven program; and (3) collect information on the same outcome measures used to judge the proven program’s success. Using this logic, the regression-adjustment methodology that ETA is readopting makes sense as a first step along this path.



• Extensive assessments are costly. Third, making workforce programs accessible by engaging in “quick and effective forms of assessment” so that people get to the right services quickly is a lesson that supports another of DOL’s reform principles. Full-blown and costly assessments have not been found to be an effective vehicle for determining who gets a particular set of services, be it a work-first approach or more intensive and longer-term services, such as skills training, that match both labor market demand and the client’s interests. Indeed, the concept of “work first” as the starting gate to a sequence of services came out of research on mandatory welfare-to-work programs, not voluntary employment and training programs.



• Career and technical education works. A fourth lesson concerns high school and college youth. Effective school-to-work strategies can have a large, long-term payoff, as we know from the study of Career Academies in nine very diverse high schools across the country. After 12 years of follow-up, the earnings of former Career Academy students were 20 percent greater than an identical group of non-Career Academy students — about the same income boost that is estimated from gaining an associate’s degree. In dollar terms, mostly minority males outearned their control group counterparts by $3,731 a year, or nearly $30,000 over the eight-year post-high school follow-up period. Not surprisingly, employers played an important role in the programs, offering internships, job shadowing, and other career and technical information and education.
  
  But the point is not just that this one model was effective but that career and technical education programs can be a worthwhile investment and that their positive results do not come at the expense of college-going. We do not yet have an equivalently strong knowledge base about what works for disconnected, out-of-school youth, but we are making progress. Equally important, we are building an encouraging body of evidence about how best to improve the academic outcomes of low-income community college students by inculcating a data-driven culture in administration and by employing various forms of performance-based scholarships, learning communities, and — to a lesser extent — student support services. And a full-scale assault, now under way, on the quagmire of developmental education courses promises to find new ways to help students make the transition from remedial courses to the credit-granting courses needed for a college degree.



• Earning supplements help. The fifth lesson comes from more recent research showing that the One-Stop system can be effective in providing financial work supports to low-income and low-wage workers. One reason for this is that One-Stops provide an unstigmatized environment where people can go to find out about and in some cases apply for the Earned Income Tax Credit (EITC), child care subsidies, public health insurance, and food stamps. From other studies, we also know that income supports for working people raise employment rates and, importantly, job retention rates — that is, receipt of income supplements has important and positive labor market impacts. So it seems reasonable to think that part of the business of One-Stops is to be an access point for work supports. With the Bureau of Labor Statistics and the Council of Economic Advisers telling us that nearly half of all jobs in 2016 will pay low wages, making work pay will be key. 

Question 3: Where Is the Evidence on Effective Strategies Thin?

Thinking about system building, model building, and targeting, I highlight several potential areas on which to focus.

• Taking models to scale. Identifying effective programs is only the first step in improving a nationwide system. As researchers and program operators, we haven’t devoted enough energy to figuring out how to replicate and scale up what works and avoid doing what doesn’t work. Effective programs must be tested in multiple locations. The program’s services and management strategies need to be well documented so that others seeking to replicate the effective program — or to require it in legislation — have enough information and guidance to act. Moreover, replication itself has to be well managed. In short, the effectiveness research that MDRC and others do that is scientific and credible can point the way. But it takes effective replication to scale up.



• System-building. Once models are operating at scale, matching programs to needs is the next step. Building a human resources development plan for a state, region, or locality requires developing three overarching maps or pictures of (1) employers’ existing jobs, by occupational category, job vacancies, and the skill sets that new workers need to succeed in those jobs; (2) the content and location of high-quality education and training programs, from community colleges to community-based training programs, that can meet those needs; and (3) the education and skill levels and geographic location of the potential workforce. Like the human anatomy textbook that uses transparencies to overlay the skeletal system, the circulatory system, the nervous system, and so on, the idea is that the overlay of these three detailed maps will help to identify and then to fill the gaps that appear.
  
  Last, to foster and maintain quality over time, a workforce system would want to identify, say, all the respiratory therapy training programs, the A-1 computer technology certification programs, the secretarial, auto repair, and other training programs and then to examine the materials, the topics covered, and the training that faculty in these programs get over time. The goal would be to bring more uniformity and teacher training and support to bear on programs. Surprisingly, examples of this kind are rare, suggesting that the know-how to design human resource investment plans of this kind just does not exist.



• Disconnected youth. No group has been more challenging to help than the growing population of 16- to 24-year-olds who aren’t working or in school. WIA is one of the few systems that focus on this population, and it is one deserving of a major national investment in knowledge development. Promising new information that should provide a platform to build on is coming from an ongoing evaluation of the National Guard Youth ChalleNGe program, from a study under way of conservation and related community service programs supported by the Corporation for National and Community Service, and eventually, from studies of YouthBuild USA.



• Low-income men, noncustodial parents, and returning ex-offenders. Among the challenges for employment and training programs and for society at large are the overlapping labor market barriers confronted by low-income men, many of whom are noncustodial parents and ex-offenders returning to the community and desperate to find jobs. Once they get jobs, they struggle to retain them and to advance in the labor market. Just under half of the labor force works in low-wage jobs; the vast majority will not advance, even over long periods of time. It is critical to make work pay for all workers, not just those with children, on whom the current EITC is focused.
  
  In short, are earning supplements to remain the province of income support policy? Or are they to assume the role of employment policy and be tied to an individual’s earnings and income, rather than family income? Moreover, major questions remain about the likely effect of an individual-based EITC on men’s employment and earnings. Jeffrey Grogger, Chinhui Juhn, Harry J. Holzer, and others attribute much of the early decline in low-income men’s employment to the long-term secular decline in earnings. A research agenda that identifies successful advancement strategies is also needed; it could build on a U.S. Department of Health and Human Services-sponsored effort that helped underscore what doesn’t work as well as what might. Evidence from the Parents’ Fair Share Demonstration conducted more than a decade ago also provides some lessons for next steps for the noncustodial parent population. Similarly, ongoing studies of transitional jobs programs for returning ex-offenders that have found encouraging and lasting impacts on recidivism, despite the rapid decay of initially large employment effects, also provide evidence on which to build a next generation of tests.



• Dislocated workers. Since the days of the Downriver Project — an experimental program for dislocated workers in Wayne County, Michigan — and the 1990s reemployment bonus studies, little rigorous research has been aimed at dislocated workers. Yet displaced workers are going to dominate the employment scene for years to come. Most have soft skills, but they often lack the education and technical skills needed in new “middle-skill” jobs. The key is to determine the role of public service employment in a period when cyclical unemployment is so severe that it could plausibly create a new generation of long-term structurally unemployed as the duration of unemployment reaches levels not seen since the Great Depression.



• Immigrant workers. Immigration is at an all-time high, and immigrant workers now make up a substantial share of the workforce. Work has been the primary source of integration into American society. But many immigrant workers have a difficult time finding upwardly mobile jobs, and some are exploited. Although a new set of institutions has emerged to support immigrant workers, these centers are not well integrated into the employment and training and higher education systems. Yet another area for investigation and additional study concerns the role and prospects of immigrant workers and strategies to help integrate them into mainstream systems and services.

In conclusion, this is an extraordinary moment for employment policy. The life course of nearly one of every five would-be workers (comprising the unemployed, the underemployed, the too discouraged to look, and involuntary part-time workers) will be determined in some measure by what we do here today and in the weeks and months that follow, up to and beyond the reauthorization of the WIA. The challenges are daunting, as demonstrated by the striking disconnect between an employment system previously in decline and a labor market that has been undergoing profound secular changes for four decades. Moreover, while we have learned much, the lessons from existing research pale in comparison to these broader demands. Self-appraisal is the first step. Looking at our history, deciding what can work for today’s problems, and forging new policies and programs that can meet tomorrow’s needs will require a lasting, trusting partnership between researchers, practitioners, and policymakers. Otherwise, the past will indeed be prologue. 

Thank you.