The welfare reforms of the late 1990s, along with a strong economy and an expansion of work supports for low earners such as the Earned Income Tax Credit, helped reduce welfare rolls and raise employment rates among low-income single mothers. Not only did employment rates rise for these women, their rates of job retention are currently quite high as well. But most current or former welfare recipients earn low wages—usually in the range of $7 to $8 per hour. Most of these workers do not receive health and other benefits on the job. Nor do they move up the job ladder very much over time. Thus, most former welfare recipients continue to be poor or near poor, even after entering the labor market, and their prospects for escaping poverty or near-poverty in the foreseeable future seem low.
The problem of low wages for unskilled workers extends well beyond the current or former welfare population. The real wages of less-educated males have either stagnated or declined (depending on how we adjust for inflation) over the past 30 years, and the earnings of less-educated men and women have fallen well behind their more educated counterparts. The gains associated with education and training programs for disadvantaged workers have generally been quite modest—at least partly because the investments in such training per person are modest as well.
This policy brief presents new evidence on the determinants of earnings advances for low earners. The results are based on a new source of longitudinal data for very large populations of workers and their employers. This brief also considers what these results imply for a refined job advancement strategy for welfare recipients and other low earners. The evidence strongly suggests that an effective job advancement strategy would stress the placement of low earners into jobs with higher-wage employers and would integrate targeted training with private sector advancement opportunities. Policymakers should consider designing programs to encourage the creation of higher-wage jobs. Private labor market intermediaries—including for-profit temporary help (or “temp”) agencies as well as other nonprofits—could play important roles in these efforts.
A New Study of Low Earners Over Time
Over the past few years, the U.S. Census Bureau has been developing a new source of data for the study of labor markets. The Longitudinal Employer-Household Dynamics (LEHD) data begin with the universe of unemployment insurance records dating back to the early 1990s for participating states (representing over 70 percent of the U.S. workforce). These data are then merged with various household and employer surveys administered by the Census Bureau. The result of this effort is an enormous datafile that tracks virtually all workers in these states over many years, and provides extensive information on their employment histories and the characteristics of the employers for whom they have worked.
Along with Fredrik Andersson and Julia Lane, my coauthors on the book Moving Up or Moving On, I have analyzed the earnings levels and growth rates of prime-age workers in several states over a period of nine years. We defined low earners as those who earned less than $12,000 per year (in 2000 dollars) for three consecutive years in the period 1993-95; we then followed them over the next six years (1996-2001) and evaluated their rates of earnings growth. Among the questions we explored were the following:
How frequently, and to what extent, do low earners transition out of this status? Which groups of low earners do so more frequently than others?
Are rates of advance for low earners heavily dependent on the characteristics of their employers? If so, what kinds of employers in which economic sectors provide the greatest advancement opportunities?
Are low earners more likely to advance by staying with their initial employers over time and gaining work experience and seniority (i.e., job retention), or by moving to other employers who provide better opportunities (job mobility)?
Are returns to any kind of early work experience (as emphasized in work-first approaches) very high, or is it better to wait and gain employment with a higher-wage employer (as welfare recipients have been encouraged to do in Portland, Oregon)? And what role is played in this process by labor market intermediaries—specifically temp agencies—that provide job placements and sometimes training as well?
Rates of Advance: How Much and For Whom?
Among all prime-age workers who consistently earn less than $12,000 a year between 1993 and 1995, significant earnings improvements are observed in the six subsequent years. In fact, over two-thirds of these workers earn above $12,000, and over half earn above $15,000, in one or more of those years.
At the same time, only about 27 percent of these initially low earners consistently earn above $15,000 by the end of this period—which would be needed (along with the Earned Income Tax Credit) to lift the earnings of single parents above the poverty line for a family of four. Earnings advances for women appear to be smaller than those of men, and advances for minorities and the foreign-born (especially among men) generally lag behind those observed for native-born whites. Transition rates out of low earnings are also lower among high school dropouts and others with poor skills, as Helen Connolly and Peter Gottschalk of Boston College noted in their recent work.
The Role of Sector and Employer Characteristics
The earnings gains of these workers vary with the industry, size, and turnover rates of the employer. For example, advancement rates are considerably higher among workers who end up in high-wage sectors such as construction, manufacturing, transportation/utilities, and wholesale trade than among those in retail trade or the services sector. Those who work in larger firms and/or those with lower turnover rates also experience greater earnings gains over time.
Of course, there is considerable variation in wage levels across firms within these broad industrial categories. Within manufacturing, firms producing durable goods such as cars and appliances pay significantly more than those producing nondurables such as textiles and apparel; in retail trade, supermarkets and department stores generally pay considerably better than restaurants and other small retail outlets; and, within the service sector, the financial sector and health services generally pay more than personal services and the hotel and entertainment sectors. In general, low-earning men are much more likely to transition out of low earnings by work in construction, manufacturing, transportation and wholesale trade; while most women (and presumably welfare recipients) who advance do so within the service sectors.
Even within more narrowly defined sectors of the economy, and within particular states, there is wide variation in wages that firms pay comparably skilled workers. These firm-level earnings premiums are highly correlated with the rates of earnings increases experienced by our initially low earners. In fact, just 30 percent of those who completely escape low earnings by 1999-2001 still work in the bottom quartile of firms (in terms of earnings premiums), as opposed to 70 percent of those who remain low earners.
Of course, these findings beg an important question: why do some firms choose to pay higher wages while others do not? As Eileen Appelbaum of Rutgers University and her colleagues point out in their recent volume Low-Wage America: How Employers are Reshaping Opportunity in the Workplace, some employers such as Wal-Mart respond to competitive pressures and new technologies by reducing worker compensation and replacing semi-skilled labor with low-skilled labor; but others respond by increasing training opportunities and providing career ladders so that unskilled workers can become more productive over time and reduce their turnover. Employers generally base their decisions on local labor market conditions, corporate culture and values, pressure from unions, assistance from outside organizations, and the information about alternatives available to them.
Job Mobility vs. Retention among Low Earners
Our data clearly indicate that, on average, earnings growth is higher among those who change jobs than among those who stay with the same employer over time. Roughly three-fourths of those who manage to fully transition out of our low earnings categories do so by changing employers rather than staying with the same one.
On the other hand, these results do not imply that all job mobility is productive, or that returns to work experience and tenure with a given employer are irrelevant. For instance, it is well known that voluntary job changes produce more positive earnings gains than those that are involuntary (or driven by child care or health care problems); and moving from one job to another is preferable to moving from a job into unemployment. Clearly, the gains we observe for job changers occur not because they change jobs per se, but because they have opportunities to improve their earnings with better employers.
But, even among those who change jobs, returns to experience and tenure with an employer are still important. To economists, the latter reflect on-the-job training and advancement opportunities within particular firms. Generally, we find that those who move to new jobs experience earnings growth there at least as high as that observed in their earlier jobs. This growth is in addition to the initial wage gains they enjoy upon making these moves, usually to higher-wage firms. Thus, the greatest earnings gains among initially low earners are experienced by those who change jobs fairly early, and then accumulate work experience and wage growth with their newer and higher-wage employers.
Early Experience: Work First, the Portland Model, or Temp Agencies
Does a worker’s early employment experiences while they remain low earners matter? We have analyzed three characteristics of work experience at earlier employers—accumulating work experience with any employer (i.e., work first); gaining employment at a higher-wage firm right away; or working initially for a temp agency—and how these affect the earnings ultimately attained by low earners who change jobs later on.
The work-first approach suggests that any work experience generates job readiness skills and signals readiness to other potential employers. In contrast, some welfare-to-work service providers (e.g., those in the Portland site of the National Evaluation of Welfare-to-Work Strategies, or NEWWS) have urged recipients to be more selective and choose better initial jobs. Temp agencies often provide low earners with a period of consistent low-wage employment, but the agencies might then place them with good employers subsequent to this initial period.
Our results show some support for each of these three options. For instance, we find some clear returns to the early accumulation of job tenure with any firm, including low-wage employers. On the other hand, these returns are also quite modest. In fact, work with low-wage employers generally raises subsequent earnings by well under 1 percent per quarter worked. In contrast, early work with a high-wage employer has somewhat more positive effects on earnings beyond that period with a different employer. In fact, a 20 percent rise in the wage premium paid by the early employer generally leads to an increase of 4 percent to 8 percent in earnings with a later employer.
Finally, we find that those who work for temp agencies initially have lower earnings than others, but 6 percent to 10 percent have higher earnings with their subsequent employer even after controlling for various personal characteristics and work histories. Furthermore, the higher subsequent earnings are completely accounted for by the higher earnings premiums on the jobs that they ultimately attain. The data strongly suggest that temp agencies improve subsequent job placement opportunities for low earners—perhaps by helping them gain early work experience, or by improving their access to higher-wage employers.
Implications for Policy
Our results indicate that low earners are most likely to advance in the labor market when they have access to higher-wage employers who also provide on-the-job training and/or career ladders. These findings are broadly consistent with the previous literature on advancement for low earners or those with little experience. For example, Rucker Johnson and Mary Corcoran of the University of Michigan have shown that job quality strongly affects earning prospects for welfare recipients. Similarly, Robert Topel and Michael Ward, at the University of Chicago and Unicon Corporation in Los Angeles respectively, have shown that job mobility contributes importantly to earnings growth among young workers. Helen Connolly and Peter Gottschalk of Boston College and Katherine Newman of Harvard also provide evidence of large earnings gains for some very low-wage workers in the 1990s.
Our results imply that job advancement depends not only on the skills that welfare recipients have, but also on their access to higher-wage employers. Unfortunately, low-income and especially minority workers living in poor neighborhoods often have limited access to firms in their local labor markets due to limited transportation options and limited information about (or contacts in) that market.
Efforts to improve the access of welfare recipients to better employers could thus have important payoffs in terms of worker outcomes over time. This conclusion is consistent with the positive results observed in the Portland site of the NEWWS evaluation, in which welfare recipients were encouraged to look for higher-paying jobs with some benefits. Federal legislation can play a role here. The Senate Finance Committee version of the welfare reauthorization bill, for example, replaces the existing credit for caseload reduction with an employment credit for those leaving cash assistance—and placements in jobs with higher earnings receive a larger credit.
Temp agencies seem to be an effective vehicle for improving access to good jobs. In some states they have served as many as 20 percent of welfare recipients entering the workforce, and even higher percentages among minorities. Some critics of these agencies claim that they simply skim the best qualified workers within the population of low-income workers. But, even if this is accurate to some extent, the temps generate access to employers that low-wage workers otherwise would have difficulty reaching on their own.
Many temp agencies provide limited training in computer skills for their clients; some also give their clients chances to obtain new job placements if the first one or two are not successful. Some preliminary work by David Autor of M.I.T. and Susan Houseman of the Upjohn Institute, using an experimental approach to evaluate the impact of temp agencies on welfare recipients in Michigan, suggest that these results might indeed be positive. Other kinds of private (nonprofit) labor market intermediaries might be able to play similar roles.
Besides temp agencies, local one-stop offices and other service providers need to have better data on the kinds of jobs available to less-skilled workers in their local labor markets. In fact, such data can now be accessed through the LEHD program. At their website (http//:lehd.dsd.census.gov), data on recent employment growth and earning levels are available at the detailed industry level for most counties in the United States. These data could enable local providers to improve the quality of their job placements for low earners quite substantially.
Of course, the skills that welfare recipients and other low earners bring to the labor market still matter enormously for their employment and advancement prospects. That is why job placement and training efforts must be better integrated in order for both to be successful. Some of this integration already occurs for welfare recipients and less-skilled workers more broadly. Besides temps, many other labor market intermediaries—such as the Center for Employment and Training (CET) in San Jose and WireNet in Cleveland—work closely with employers to ensure that the workers they refer are adequately skilled for the jobs they will fill.
In fact, the success of CET and other models like it has often been attributed to their close contacts with local employers. Sectoral training strategies, in which growing sectors of particular local labor markets are identified and targeted for participation, are one version of this approach. Training providers and other intermediaries gain significant knowledge about the exact skills needed in the particular sectors they target, and often have close ties with employers in those sectors. Some well-known examples of the sectoral approach—such as Focus: Hope in Detroit (which trains disadvantaged workers for jobs in the auto industry) and QUEST in San Antonio (which builds links between community colleges and employers in health care, financial services, and other key sectors)—have been widely praised in the Aspen Institute’s recent report, Grow Faster Together, or Grow Slowly Apart, among others.
Our research also suggests that successful advancement strategies for welfare recipients and other low-income workers might involve not one particular job, but planning for a sequence of appropriate jobs along with training for each. Most of the training can be provided on the job by employers, as long as the candidates demonstrate the right basic general skills. The emphasis of work-first approaches on early job placement does not appear inappropriate, in light of this need for demonstrated job readiness. But since the returns to these first jobs might diminish fairly rapidly, helping poor workers move on to better jobs after a brief time period would be a sensible strategy. Karin Martinson of the Urban Institute and Julie Strawn of the Center for Law and Social Policy in Washington, D.C. have written about career planning for low-income workers, and the role that private intermediaries could play in that process.
Can public policy also encourage the creation of more higher-wage jobs so we do not simply displace workers who currently have these jobs with others whom we help? A few options are available. Moderate increases in the minimum wage can help lift the earnings levels of some low earners without seriously reducing overall employment levels. Grants or tax credits are already being used in some states to support higher wages through on-the-job (or “incumbent worker”) training; these efforts could be expanded by the federal government.
The potential role of private intermediaries is also important. Intermediaries not only train workers and place them with employers; they also can work with employers to build career ladders and improve their human resources policies. Cooperative Home Care Associates in the Bronx, for example, builds careers for long-term care workers employed in hospitals and nursing homes. Additional examples are provided by the Chicago Manufacturing Institute and the Wisconsin Regional Training Partnership, both of which work closely with employers to build career ladders for their clients.
Of course, many of the ideas presented in this policy brief need to be implemented at the local level and rigorously evaluated. Policymakers also need to assess whether they would be cost-effective, and whether they could be brought to scale. In the meantime, these job advancement strategies provide us with potential models for advancement that heavily rely on the private sector, and therefore might be more successful than many of the training programs that have been undertaken in the past.