Number A-33 in Series, "New Federalism: Issues and Options for States"
The nonpartisan Urban Institute publishes studies, reports, and books on timely topics worthy of public consideration. The views expressed are those of the authors and should not be attributed to the Urban Institute, its trustees, or its funders.
Dramatic changes in the nation's welfare system passed into law in 1996 are designed to move increasing numbers of single mothers into the job market. The new rules built into the 1996 welfare reform act, the Personal Responsibility and Work Opportunity Reconciliation Act, increase pressure on recipients to work by imposing a five-year lifetime limit on receiving federal welfare benefits (and permitting states to impose even shorter time limits), penalizing states that have too few recipients in work activities, and requiring recipients to participate in work activities within two years of receiving benefits. Other changes in the nation's income support system encourage work by expanding the opportunity for families to mix work with welfare, noncash assistance, or both.
One concern is whether enough jobs will be available to employ most of the welfare recipients induced to enter the job market. Will welfare recipients find jobs or add to the ranks of the unemployed? Another worry is that a massive inflow of welfare recipients will impose hardships on other workers in low-wage markets. Employers may absorb significant numbers of recipients, but only if market forces lower wages enough to accommodate the added supply of workers, leaving all low-wage workers worse off.
This brief examines how the inflows of recipients into the labor force will affect the balance of jobs and low-skill workers in 20 metropolitan areas by looking at the size of welfare inflows, low-skill employment growth, and unemployment rates in those areas. Low-skill workers are defined as people, whether employed or seeking employment, who have no more than a high school diploma. Low-skill jobs are those requiring less than 12 months of training. The analysis deals with metropolitan areas, because data on them are accessible and because these areas are a more natural labor market boundary than cities.(1) Many metropolitan areas are experiencing large employment growth just outside city limits. To focus more clearly on the availability of jobs, we assume that all jobs within metropolitan areas are accessible.
Optimism Nationally, Uncertainty Locally
National studies generally conclude that the inflow of welfare recipients will not generate a significant strain on the labor market and that the number of jobs will be sufficient for most job-seeking recipients.
(2) So far the economy-wide evidence is promising. One indication of the economy's ability to absorb recipients is the experience of never-married mothers, the group most likely to participate in cash assistance programs for families. Between the first quarter of 1996 and the second quarter of 1998, about 741,000 additional never-married mothers began to look for work, and the economy generated enough jobs to employ all of these new entrants.
(3) The 33 percent jump of never-married mothers in the labor force was matched by an astonishing 40 percent rise in the employment of never-married mothers. This 40 percent job growth dwarfed the 9 percent increase in employment for the economy as a whole. Thus, despite the rapid expansion in the numbers seeking employment, never-married mothers experienced a decline in unemployment rates from 19 percent to 14.8 percent. As of early 1999, from the national perspective, the number of jobs created has matched the increased numbers of welfare recipients entering the workforce.
Despite the encouraging national trends, the inflow of recipients may place a serious burden on low-skill job markets in some metropolitan areas, since welfare recipients are concentrated in selected urban areas. The three largest metropolitan areas in the country—Los Angeles, New York City, and Chicago—accounted for about 20 percent of the nation's caseload in 1997 but only 10 percent of the nation's population. Unless job availability is adequate at the local level, many recipients may face serious unemployment problems and low-skill workers may experience wage reductions as well as increased joblessness.
Are Labor Inflows a Problem or a Benefit to an Economy?
Most people greet the news of large inflows of workers into a job market with pessimism. Viewing the number of jobs as fixed, one expects the added workers will either become unemployed, displace other workers, or both. Employers may be in a position to hire more workers, but at lower wages.
(4)
However, according to an alternative perspective, an increase in the supply of labor can be coupled with an increase in the demand for labor (so that the inflow of workers raises the production potential of the economy). This requires that the inflow be balanced with workers of varying skills and that nonlabor inputs expand with the increased workforce. The result is economic growth rapid enough to expand total employment without reducing wages. Theoretically, then, adding large numbers of welfare recipients to the labor force may improve the country's capacity to produce.
Other Studies of Prospective Labor Market Impacts
Many studies have attempted to address the availability of jobs for low-skill welfare recipients.
(5) Often the supply of and demand for low-skill workers are assessed according to the worker-to-job ratio. While having a summary measure such as the worker-to-job ratio is appealing, it must be interpreted carefully because it masks much of the labor market complexity. Labor supply and demand estimates are typically static measures, yet the labor market is dynamic, making it impossible to accurately measure supply of and demand for labor.
One example of the difficulties inherent in worker-to-job ratios can be seen in a study of six midwestern states.(6) Kleppner and Theodore (1997) estimate a worker-to-job ratio of between 2 to 1 and 4 to 1 and conclude that welfare recipients will have significant difficulty finding jobs. When the researchers exclude welfare recipients, their worker-to-job ratio is 1.6 to 1, suggesting that workers would have difficulty finding employment even without the new job seekers. However, the unemployment rates in these six midwestern states are at record lows. The 1998 annual unemployment rates from the six states ranged from 3.4 percent in Wisconsin to 4.5 percent in Illinois. The seemingly high estimate of the worker-to-job ratio in an environment of declining and extremely low rates of unemployment highlights the need to be careful in interpreting these static measures in the context of the dynamic labor market. Just as we would not expect an unemployment rate of zero, we probably should not expect a worker-to-job ratio of 1 to 1. Keeping these issues in mind, we turn to analyzing different aspects of the labor market.
Welfare Recipients Add to Supply of Low-Skill Workers
At what rate will welfare recipients enter the labor market in response to welfare reform changes? How will these added workers affect the size of the low-skill workforce in various urban areas? The answers to these questions depend on the size of the welfare population relative to the local workforce.
To predict the size of the inflow of welfare recipients into the workforce in the 20 metropolitan areas, we use two estimates. The first, probably a low estimate, is a measure of the number of welfare recipients who must engage in a work-related activity in order to meet the federal work requirement passed in 1996 as part of welfare reform. The second estimate is a flat 15 percent of the current adult recipient caseload in each metropolitan area.
Our federal work requirement estimate is based on the mandate that 30 percent of adult welfare recipients with children over age 1 participate in work activities in 1998, but it reflects the reductions due to state caseload declines between 1995 and 1997.(7) Since caseload declines have been substantial across the country since 1995 (over 20 percent nationally), the actual requirement among our 20 metropolitan areas varies from 20 percent to zero. The fact that many recipients are already working also cuts the number of additional welfare recipients who must find work to meet the federal mandates.(8)
In seven of the metropolitan areas—Atlanta, Baltimore, Boston, Chicago, Detroit, Indianapolis, and Jacksonville—enough recipients are already working to meet the federal requirement without further increases. Across all 20 areas, an additional 3.2 percent of welfare recipients will become job seekers (47,644 welfare recipients).(9) Because work is only one of the activities allowed to count toward the federal requirement, these figures may be overestimates.(10) However, they may be underestimates, if employment among recipients in metropolitan areas is lower than in other parts of the state.
Welfare recipients may enter the labor force for reasons other than the federal work requirement. For this reason, the calculation based on the federal work requirement may be too low. Many areas go beyond federal mandates in requiring job search and work activities and changing benefit rules to allow recipients to keep more of what they earn.
In addition, recipients can now earn more from their employment, given the significant increase in the Earned Income Tax Credit, making work an attractive option.(11) Recognizing that inflows may be higher than our first estimate, we use a second estimate based on a flat 15 percent of the current adult recipient caseload. This assumes that an additional 15 percent of each metropolitan area's initial caseload begins seeking work in the coming year, over and above those already working.
Table 1 shows our estimated annual inflows for both of these measures across the 20 metropolitan areas. The variation in our first measure reflects both differences in the percentage of recipients working while on welfare and differences in the state caseload decline for each metropolitan area. Our second 15 percent estimate reflects only differences in the size of the adult caseload. This measure yields 162,468 additional welfare job seekers per year for the combined 20 metropolitan areas, a number more than three times larger than our first estimate.(12) Only the District of Columbia has a lower number of recipients entering the labor market using the 15 percent estimate, due to the low number of recipients working in the District and relatively modest caseload decline.
The annual increase in welfare recipient job seekers as a percentage of employment requiring only relatively low skills is an indicator of how well an economy can absorb the new job seekers. The second and fourth columns in table 1 show that welfare recipient inflows average roughly between 0.3 percent and 1.3 percent of low-skill employment across the 20 metropolitan areas. The highest inflow estimates are for New York City, where the number of recipients entering the labor market is equivalent to 2.4 or 3.7 percent of all low-skill jobs—depending on the assumption made.
High inflow percentages under our first assumption reflect a number of different factors: a lower percentage of welfare recipients working, a lower level of caseload decline, greater numbers of welfare recipients per population, and a smaller percentage of total employment that requires low skill levels. Our second assumption removes some of these differences and reflects mainly the size of the caseload relative to the population and the percent of employment that requires low skill levels. In the majority of metropolitan areas, welfare inflows to the job market amount to less than 1 percent of low-skill employment.
Growth in Low-Skill Jobs
How many new low-skill jobs become available helps determine whether local labor markets can accommodate the inflow of recipients without substantially worsening job opportunities for all low-skill workers. The two components of job availability are net employment growth (the number of jobs created minus the number of jobs destroyed) and replacement demand (openings that become available when workers in low-skill positions are promoted or leave the labor market). This study uses employment growth as the primary estimate of demand. While jobs also become available through low-skill workers' retirement or promotion, most of these openings in recent years have been offset by nonwelfare additions to the low-skill labor force. If metropolitan labor markets do not vary a great deal in this respect, then net employment growth is a good indicator of the relative position of demand for labor.
(13)
Just as not all job seekers compete for the same jobs, not all jobs are available to welfare recipients and other job seekers with low skill levels. We classify certain jobs as low-skill positions as a first step in estimating the jobs accessible to the vast majority of welfare recipients. Using the Bureau of Labor Statistics (BLS 1998) assessment of occupational training requirements, we include jobs that require short-term training (defined as jobs that can be learned in just a few days or weeks) or moderate-term training (jobs that require 1 to 12 months of on-the-job or informal training at the workplace). For each occupation, we calculated the proportion of jobs held by persons with a high school diploma or less. For example, if 70 percent of jobs in a "short-term training required occupation" are held by persons with no more than a high school diploma, then our analysis assumes that 70 percent of the jobs in this particular occupation are available to welfare recipients and other low-skill job seekers. Since these jobs require little to no prior work experience, most welfare recipients would be able to meet their requirements.
The proportion of employment that requires low levels of skill varies widely across labor markets. The percentages of all jobs meeting the low-skill criteria in various metropolitan labor markets are fairly high, ranging from 21 to 37 percent of all employment.
We measure low-skill employment growth as the difference in the number of people employed in low-skill jobs between 1996 and 1997. Table 2 shows the net low-skill employment growth as well as the rate of growth between 1996 and 1997 per 10,000 persons employed in low-skill jobs in 1997 in each metropolitan area. The rate measured as the growth in the number of people employed in low-skill jobs allows us to compare low-skill employment across metropolitan areas, controlling for the varying size of the labor markets. These estimated rates of low-skill employment growth vary widely across metropolitan areas. The highest growth rates are in Atlanta and Jacksonville, with relatively high growth in Dallas, Los Angeles, and Phoenix as well. Some metropolitan areas, including Baltimore, St. Louis, and the District of Columbia, are expected to experience extremely low or negative low-skill employment growth over this time period.
Competition with Low-Skill Job Seekers
Welfare recipients who are entering the workforce must vie for low-skill jobs with other workers. To understand welfare inflows in this context, one must determine the size of the pool of competing unemployed workers—that is, persons not already working who are actively searching for and available for work. This includes new entrants to the labor market (such as youth leaving school), as well as those who have lost jobs, but not adults who are neither working nor actively looking for a job. Of course, not all unemployed persons compete with welfare recipients for employment. Job seekers have a variety of skills and compete for a wide range of jobs. To identify those who compete with welfare recipients, we must focus on the low-skill segment of the unemployed.
Welfare recipients on average have low levels of education. A study by Zedlewski, using the National Survey of America's Families, finds that 41 percent of welfare recipients in 1997 had less than a high school education. The job seekers most likely to compete with welfare recipients are those with similar characteristics, but it is difficult to draw a line dividing the workforce into groups that do or do not compete with recipients. In this analysis, we assume that nonrecipient job seekers with a similar skill level as recipient job seekers are those whose highest education is a high school diploma (or equivalent) or less and who are 40 years old or less. Ideally, our analysis would also consider individuals' work experience when determining skill level, but our data do not allow us to make this distinction. We use the age cutoff to proxy for those with less labor market experience.
Some highly educated job seekers may choose to apply for jobs that have lower skill requirements; similarly, some less-educated job seekers with a number of years of work experience will seek higher-skill jobs. These two components work in opposite directions and will to some extent offset each other. Under the tight labor market conditions that the United States is currently experiencing, it is likely that a larger number of job seekers with no more than a high school education will search for a higher-skill job than job seekers with more than a high school education will search for a low-skill job. If this is the case, then our approach leads us to overestimate the number of nonrecipient job seekers.
The third column of table 2 shows the low-skill unemployment rate across metropolitan areas, combining data for 1996 and 1997. Not surprisingly, in all 20 metropolitan areas low-skill unemployment rates, averaging 7.1 percent, are substantially higher than total unemployment rates, which average around 4.2 percent. In areas such as Indianapolis, Minneapolis, Phoenix, and San Francisco, low-skill unemployment rates are 5 percent or less, quite low by historical standards. Other areas suffer from extremely high low-skill unemployment rates. The rates in Baltimore and New York City are above 10 percent, more than twice the national average in 1997 (4.9 percent).
How Welfare Inflows Affect Metropolitan Area Unemployment Rates
Armed with area information on welfare inflows into the labor force, expected growth in low-skill jobs, and current low-skill unemployment rates, we can analyze how much strain the entry of welfare recipients is likely to place on various metropolitan area job markets.
Starting with the 1997 low-skill unemployment rate, we project a new unemployment rate after job growth that does not take into account new welfare recipient entrants.(14) We assume the new jobs are only taken up by existing unemployed, low-skill workers so that job growth lowers metropolitan area low-skill unemployment rates. We then consider the addition of new welfare recipient job seekers. Any number of recipients entering the labor market will increase the unemployment rate relative to what it would have been without any new welfare entrants.
As a simplifying assumption, we do not consider other entrants or exits from the low-skill labor market. We expect that each area's workforce of less-educated workers will remain constant, as it has over the last six years. Between 1992 and 1998, the number of less-educated workers (those with no more than a high school education) entering the U.S. workforce was only enough to offset the number of less-educated workers leaving the workforce.
Figure 1 shows the results when we bring together the projected impact of employment growth and welfare recipient entrants.(15) When welfare recipient inflows are taken into account along with job growth, the average unemployment rate fell from 7.1 percent to 6.2 percent. This means that low-skill employment growth generally outpaced welfare recipient inflow. While unemployment rates are greater than they would have been in the absence of this new group of job seekers, in many areas they still fall.
There are differences across metropolitan areas, ranging from actual increases in low-skill unemployment rates to a 3.5 percentage point decline. Areas can be broken into three general categories: those with definite overall declines, those with almost no change, and those with overall increases. Of the 12 metropolitan areas expected to have a lower unemployment rate, the largest declines are in Atlanta, Dallas, Indianapolis, Jacksonville, Phoenix, and San Jose. These areas have relatively high employment growth, relatively low numbers of welfare recipient job seekers for the size of the labor market, or a combination of both.
The group with almost no change includes Boston, Detroit, Philadelphia, and San Diego. These areas are expected to encounter average to high rates of recipient inflows and average rates of employment growth, a combination that leaves job markets unchanged.
The final group of metropolitan areas—Baltimore, New York City, St. Louis, and the District of Columbia—is projected to face increases in low-skill unemployment rates. These job markets will have the hardest time absorbing welfare recipients, limiting the decline in job opportunities and wages for other low-skill workers, or both. The negative impacts occur mostly because of low employment growth. Only New York City confronts the additional problem of an unusually large percentage inflow of welfare recipients into the low-skill job market. In the other three areas, recipient inflows are about average for all metropolitan areas.
These projections give a sense of the overall impact of welfare inflows to these metropolitan area job markets. They do not necessarily reflect where individual welfare recipients will have the hardest time finding a job. The ability of welfare recipients and other low-skill workers to find a job in a particular metropolitan area depends not only on the impact of welfare reforms but also on the initial unemployment rate in that area. Taking account of the combined impact of these factors, we project low-skill unemployment rates of 8 percent or higher after job creation and recipient inflows in six metropolitan areas (Baltimore, Los Angeles, New York, Philadelphia, St. Louis, and the District of Columbia) and 7.5 percent in two (Chicago and Houston). In the other 12 metropolitan areas, we estimate the average low-skill unemployment rate at about 4 percent. Nearly all of this difference in job availability is the result of initial differences in unemployment rates. Even before the welfare changes exerted their impact, unemployment rates averaged 9.1 percent for these eight areas, compared to 5.8 percent for the other 12 metropolitan areas.
Caveats and Concluding Comments
On average, our projections show that the 20 metropolitan areas studied will experience decreases in unemployment, even with the entry of welfare recipients into the labor force, largely because of growth in low-skill employment. However, this story varies across areas. Four of the metropolitan areas—Baltimore, New York City, St. Louis, and the District of Columbia—may experience increases in unemployment rates, assuming a 15 percent estimate for the number of recipients seeking jobs.
We should point to three caveats that qualify our findings. The first is that the analysis makes no adjustments for the effects on wages of the increased supply of and demand for workers. If wages are flexible, the added entry of recipients will do more to lower wages and less to increase unemployment than projected in this study.
The second caveat relates to migration. The inflow of large numbers of welfare recipients into an area may be offset in part by outmigration from the area. It is also possible that employers will migrate into a metropolitan area that has recently experienced an inflow of workers. If recipients add to an area's workforce, employers may relocate to take advantage of more abundant, and possibly less costly, workers, thus creating more jobs in these areas.
The third caveat is that we are assuming all low-skill jobs in a metropolitan area are accessible—ignoring the problem of spatial mismatch between job seekers and jobs. Even in metropolitan areas that appear in this analysis to pose few problems for welfare recipient job seekers, special transportation or other policies to enable job seekers to get to jobs may need to be considered.
The effect of an inflow of welfare recipients on a metropolitan area's labor market depends largely on factors distinct to the area. The size of the inflow, the previous employment rate, and the low-skill unemployment growth all influence an area's ability to absorb new job seekers. Because of this variation, the resources needed to move recipients into the labor market are not the same from area to area. However, federal policy does not currently account for differences across areas. Policymakers at both the federal and state levels need to take these differences into consideration when designing and funding work programs in order to meet the unique needs of recipients in different areas.
Tables and Figures
References
Acs, Gregory, Norma Coe, Keith Watson, and Robert Lerman. 1998. "
Does Work Pay? An Analysis of the Work Incentives under TANF."
Assessing the New Federalism Occasional Paper Number 9. Washington, D.C.: Urban Institute.
Bartik, Timothy. 1998. Washington, D.C.: The Labor Supply Effects of Welfare Reform. Kalamazoo, Mich.: W.E. Upjohn Institute for Employment Research.
Bureau of Labor Statistics. 1998. Occupational Projections and Training Data. Washington, D.C.: U.S. Department of Labor, Bulletin 2501.
Burtless, Gary. 1998. "Can the Labor Market Absorb Three Million Welfare Recipients?" Focus 19 (3). Madison, Wisc.: Institute for Research on Poverty.
Carlson, Virginia, and Nikolas C. Theodore. 1995. Are There Enough Jobs? Welfare Reform and Labor Market Reality. Illinois Job Gap Project. Dekalb, Ill.: Office for Social Policy Research, Northern Illinois University.
Holzer, Harry. 1998. Will Employers Hire Welfare Recipients? Recent Survey Evidence from Michigan. Discussion Paper. East Lansing, Mich.: Institute for Research on Poverty, Michigan State University.
Katz, Bruce, and Kate Carnevale. 1998. The State of Welfare Caseloads in America's Cities. Washington, D.C.: Brookings Institution.
Kleppner, Paul, and Nikolas Theodore. 1997. Work after Welfare: Is the Midwest's Booming Economy Creating Enough Jobs? Midwest Job Gap Project. Dekalb, Ill.: Office for Social Policy Research, Northern Illinois University.
Lawson, Leslie, and Christopher T. King. 1997. The Reality of Welfare-to-Work: Employment Opportunities for Women Affected by Welfare Time Limits in Texas. Austin, Tex.: Lyndon B. Johnson School of Public Affairs, University of Texas.
Leete, Laura, and Neil Bania. 1999. "The Impact of Welfare Reform on Local Labor Markets." Journal of Policy Analysis and Management 18 (1): 50-76.
McMurrer, Daniel P., Isabel V. Sawhill, and Robert I. Lerman. 1997. "Welfare Reform and Opportunity in the Low-Wage Labor Market." Opportunity in America Number 5. Washington, D.C.: Urban Institute.
Zedlewski, Sheila. Forthcoming policy brief, "Work-Related Activities and Limitations of Current TANF Recipients: Findings from the 1997 National Survey of America's Families." Assessing the New Federalism. Washington, D.C.: Urban Institute.
Zedlewski, Sheila R., Pamela A. Holcomb, and Amy-Ellen Duke. 1998. "Cash Assistance in Transition: The Story of 13 States." Assessing the New Federalism Occasional Paper Number 16. Washington, D.C.: Urban Institute.
Notes
1. A metropolitan area roughly includes the city and the surrounding counties. We use the Primary Metropolitan Statistical Area (PMSA) or the Metropolitan Statistical Area (MSA) definition, the smallest identifiable geographic region yielding adequate numbers of observations from the Current Population Survey (CPS) data.
2. For example, Burtless (1998) and McMurrer, Sawhill, and Lerman (1997).
3. These are authors' calculations using unpublished Bureau of Labor Statistics tabulations drawn from the CPS.
4. A critical issue is the sensitivity or elasticity of the demand curve in this market. If demand is highly elastic, then nearly all new workers may be absorbed with little reduction in the wages of existing workers. If demand is inelastic, then the new workers will lead to little new employment but sharply lower wages. There is little information on how this sensitivity differs across urban labor markets.
5. National studies also include Bartik (1998) and Holzer (1998). Labor market-specific regionally based studies include Leete and Bania (1999), Carlson and Theodore (1995), Lawson and King (1997), and Kleppner and Theodore (1997).
6. The six midwestern states are: Illinois, Indiana, Michigan, Minnesota, Ohio, and Wisconsin.
7. This analysis uses state, not metropolitan area, welfare recipient employment rates and caseload declines. While metropolitan area figures likely differ from state figures, our analysis assumes the work requirement is spread evenly across states. (See Katz and Carnevale (1998) for information on caseload decline in cities versus states.)
8. For information on waivers prior to TANF in some of these states, see Zedlewski, Holcomb, and Duke (1998).
9. This approach yields results similar to projections by Bartik (1998).
10. The federal definition of work activities includes public employment, work experience programs, job search and job readiness for up to six weeks, community service, and vocational education for up to 12 months.
11. See Acs, Coe, Watson, and Lerman (1998).
12. Our estimates are substantially lower than some estimates developed by other researchers. Kleppner and Theodore (1997) assume, in one of the two inflow measures presented in the paper, that all nonemployed welfare recipients enter the labor market simultaneously. The assumption is unrealistic and double counts unemployed recipients. Carlson and Theodore (1995) estimate that 50 percent of recipients in Illinois enter the labor market simultaneously.
13. We do not consider here openings caused by turnover, that is, a worker leaving a job in the low-skill sector to take another job in the low-skill sector. Turnover does lead to openings that a welfare recipient job seeker could fill, thus increasing the probability of an individual entrant finding a job, though at the expense of the current worker who would have otherwise filled the job.
14. We assume employment grows at the same rate as the 1996 to 1997 period. Of course, some of the unemployed in 1997 may be welfare recipients, but our welfare inflow measures are over and above the unemployed.
15. These calculations use our 15 percent estimate of welfare recipient entrants.
