Donate
Independent, objective, nonpartisan research

R 101HHR

Authors

R 101HHR

Tagged with:

Publication PDFs

Database

This is the content currently stored in the post and postmeta tables.

View live version

object(Timber\Post)#3742 (44) { ["ImageClass"]=> string(12) "Timber\Image" ["PostClass"]=> string(11) "Timber\Post" ["TermClass"]=> string(11) "Timber\Term" ["object_type"]=> string(4) "post" ["custom"]=> array(5) { ["_wp_attached_file"]=> string(12) "R_101HHR.pdf" ["wpmf_size"]=> string(6) "290763" ["wpmf_filetype"]=> string(3) "pdf" ["wpmf_order"]=> string(1) "0" ["searchwp_content"]=> string(227756) "Employers and Welfare Recipients: The Effects of Welfare Reform in the Workplace ••• Harry J. Holzer Michael A. Stoll 2001 PUBLIC POLICY INSTITUTE OF CALIFORNIA Library of Congress Cataloging-in-Publication Data Holzer, Harry J., 1957– Employers and welfare recipients : the effects of welfare reform in the workplace / Harry J. Holzer, Michael A. Stoll. p. cm. Includes bibliographical references. ISBN: 1-58213-057-4 1. Welfare recipients—Employment—United States. 2. Employer attitude surveys—United States. I. Stoll, Michael A. II. Title. HV95 .H566 2001 331.5'4'0973—dc21 00-051723 Copyright © 2001 by Public Policy Institute of California All rights reserved San Francisco, CA Short sections of text, not to exceed three paragraphs, may be quoted without written permission provided that full attribution is given to the source and the above copyright notice is included. Research publications reflect the views of the authors and do not necessarily reflect the views of the staff, officers, or Board of Directors of the Public Policy Institute of California. Foreword Welfare studies typically focus on the changing size and characteristics of the caseload. If the caseload shrinks, researchers analyze the reasons for the change—is it the economy or is it program incentives? In a similar vein, a recent PPIC study concluded that California’s caseload is declining less dramatically than the nation’s, and that some case types are actually on the rise. Another PPIC study concluded that California’s welfare recipients have lower skills than other adults in the state and across the nation and that their long-term employability depended heavily on the health of the California economy. This study takes a different approach. Rather than tracing the reasons for caseload increases or decreases, it examines the welfare-towork transition from the employer’s perspective. Are employers willing to hire welfare recipients? If so, for what kinds of jobs and at what levels of compensation? How well do these employers think these workers are doing their jobs? And finally, what might these answers mean for the long-term prospects of welfare recipients moving into the job market? Funded by PPIC and the Joyce Foundation, Harry Holzer and Michael Stoll interviewed 750 employers across four metropolitan areas—Los Angeles, Chicago, Cleveland, and Milwaukee. Although the authors remain cautious about the long-term prospects of welfare recipients moving into the labor market, most of their results are encouraging. The boom economy of the late 1990s increased the job vacancy rate in all four cities, and employers actively hired welfare recipients. Although these workers often faced problems associated with job accessibility, their employers consistently reported that their job performance was as good as or better than other workers doing the same work. Their turnover rates were high but no higher than those of other employees. Less encouraging was the fact that the hiring rates for minorities were lower than those for whites. iii The authors also found that the situation in Southern California differed in some ways from what they observed in the Midwest. For example, Los Angeles had a relatively low hiring rate for welfare recipients. This was in part due to a softer economy in Los Angeles, but it also reflected higher skill requirements for the job openings in California compared to those in the Midwest. The average wage was also higher in Los Angeles than in the other three cities. According to the authors, however, all four cities have something in common—an economic downturn will affect their welfare caseloads. The study confirms that there is no better solution to welfare dependency than a good job, and there is no better road to a good job than a healthy, booming economy. But for many welfare recipients, a steady job is a bridge too far. Child care responsibilities, disabilities, poor training, and inadequate transportation are persistent impediments to steady employment. Publicly supported job training programs or supported work have long been a means to help the most needy cross the bridge to self-sustaining income. The authors revisit these policy options and suggest that they will be needed in even the most vibrant economy. The job experience gained by these welfare recipients is a net addition to human capital, one that is generating personal and public benefits today and will continue to do so in the coming decades. By examining the employment prospects of welfare recipients from a rare but critical perspective, this study helps rounds out our understanding of welfare reform even as it helps us monitor its consequences. David W. Lyon President and CEO Public Policy Institute of California iv Summary The recent reforms of the welfare system in the United States, beginning with state-level experiments approved under federal vouchers in the early 1990s and culminating in the 1996 passage of the Personal Responsibility and Work Opportunity Reconciliation Act, remain controversial. Since then, welfare rolls have declined dramatically and employment rates among recipients (and single females more broadly) have risen substantially, partly because of welfare reform and partly because of other factors, such as economic conditions and other policyinduced changes in net earnings and benefits for the working poor. But many questions remain regarding employer willingness and ability to hire from various groups of recipients, their performance and retention rates once hired, and the wages and benefits they receive from their jobs. To address these questions, we analyze data from a new survey of employers, administered to 750 establishments between July 1998 and June 1999 in each of four metropolitan areas: Los Angeles, Chicago, Cleveland, and Milwaukee. The survey focused on employer attitudes about hiring welfare recipients and their experiences with this population. Research Findings A number of important findings emerge from this analysis. For one thing, the current and recent demand for welfare recipients among employers is quite strong, at least in the aggregate. Overall hiring rates of recipients seem sufficient to absorb the numbers seeking employment in the current period and the near future. On the other hand, the rates at which recipients are hired are very sensitive to the job vacancy rates at establishments. This sensitivity suggests that during an economic downturn, a good deal of employer demand for recipients will diminish. Furthermore, although many establishments are open to hiring welfare recipients, actual hiring is much stronger at establishments that v are easily accessible to recipients, such as those near public transit stops or their neighborhoods. Also, recipients have more access to jobs when establishments are minority-owned or when employers have had contact with local welfare-to-work agencies, regardless of who initiates the contact. Limited access to jobs might help to account for the fact that minority groups in the welfare population appear to be hired less frequently than whites, relative to their representation in the welfare population. On the other hand, high school dropouts also seem to be hired less frequently than high school graduates, perhaps because the cognitive and social skills demanded on the jobs filled by welfare recipients are often not trivial. As for the quality of these jobs, those recently filled by welfare recipients in these metropolitan areas pay about $7.00 per hour on average and generally provide 40 hours of work a week. Employers contribute to the provision of health insurance in approximately twothirds of them. Still, significant fractions of these jobs pay low wages, provide few hours, or provide no health insurance. Furthermore, turnover rates for recently hired recipients are near the national average, and most recipients are rated as being as good or better than other workers in the jobs that they fill. But high turnover or weak job performance is a problem in a significant fraction (i.e., a fourth to a third) of cases. Absenteeism is particularly pervasive, often linked to child care and transportation issues. And the average skill levels and work-readiness of welfare recipients may be declining somewhat over time, as the part of the welfare caseload with more significant employment-related problems is entering the workforce now in greater numbers. Finally, there are some noteworthy differences between hiring patterns among employers in Los Angeles and those in the Midwest, particularly Milwaukee. In terms of turnover, job performance, absenteeism, and the like, the average quality of those recipients hired in Los Angeles seems higher than those in the Midwest and those in Milwaukee seem lowest. However, hiring rates have been relatively high in Milwaukee and low in Los Angeles. Thus, the better performance of recipients hired in Los Angeles seems to come at least partly at the vi expense of relatively more disadvantaged recipients who do not get hired there. The relatively limited hiring of welfare recipients in Los Angeles is partly attributable to labor market characteristics, such as the lower job vacancy rate there and the relatively greater skills demanded among the jobs filled. But it may also be related to the slower and less-aggressive implementation of welfare reform there, especially compared to the efforts undertaken as part of “Wisconsin Works” (W-2) in Milwaukee. Employers are significantly more likely to have had contact with a local welfare-to-work agency in Milwaukee than in Los Angeles or the other two metropolitan areas. For the United States overall, these findings suggest that there have been many positive employment outcomes associated with welfare reform to date, both for welfare recipients and for employers. However, some groups (e.g., minorities or high school dropouts) still appear to have difficulty gaining employment, and the bottom group of those being hired (i.e., the lowest one-fourth or one-third along most dimensions) are experiencing poor earnings or benefits when they are employed. The results we have observed to date might also deteriorate over time, as less-employable recipients enter the workforce or when an economic downturn occurs. Implications for Policy Several implications for policy at the federal, state, and local levels can be drawn from these findings. For one thing, the strong state of the U.S. economy and the relative tightness of the labor market have contributed to the ability of welfare recipients to find jobs. This implies that now is a very good time to invest in the training of recipients and in supportive services, such as child care and transportation. Any work experience gained by welfare recipients in this environment should lead to at least some wage growth and employment stability over time. On the other hand, some portion of this demand will no doubt disappear during the next cyclical downturn. Some recipients who cannot find employment at that time will be eligible for Unemployment Insurance, and others can return to the welfare rolls. For those welfare recipients who will not be eligible for Unemployment Insurance and who vii might run into time limits under Temporary Assistance for Needy Families (TANF), other arrangements will need to be made—such as changes in the Unemployment Insurance program, more community service jobs, or temporary suspensions of TANF time limits and work requirements for high-unemployment states. When the labor market slackens, the limited access of many welfare recipients (especially inner-city minorities) to suburban jobs may be more of a problem than it is now. Improving access through innovative approaches to transportation and job placement might become a more pressing need. For example, local intermediary agencies could become more involved in the job placement process for welfare recipients. Because relatively few employers in any given area have contact with local welfare-to-work agencies, improving their funding as well as their outreach to local employers might be worthwhile goals. Also, the need to improve retention and reduce absenteeism among recipients remains strong. Local workforce boards and agencies need to focus more broadly on retention issues as well as job placements. Given that skill requirements on jobs filled by welfare recipients are generally not trivial, and that many long-term recipients have poor skills and other personal characteristics that limit their employability, we may need to develop other employment options (such as community service jobs or “sheltered workshops”) as the most disadvantaged portion of the welfare caseload approaches time limits. Many other groups (e.g., those with criminal records) face barriers to employment even more severe than those faced by welfare recipients, especially in the attitudes of employers. Thus, efforts to promote more training, early employment opportunities, and income supplements to low-income (often noncustodial) fathers and other disadvantaged groups, such as those with criminal records, are critical. Finally, the implementation of welfare reform in Los Angeles will no doubt proceed and gain ground over the coming months and years. As efforts to place welfare recipients into employment expand there, greater efforts will be required to overcome a variety of barriers in the workplace (especially among immigrants) and to deal with any deterioration in employee performance that might develop. The active inclusion of local viii workforce boards and agencies in this process may be crucial to welfare reform’s success there. ix Contents Foreword ..................................... Summary..................................... Figures ...................................... Tables ....................................... Acknowledgments ............................... iii v xiii xv xvii 1. INTRODUCTION ........................... Recent Literature ............................. A New Study of Employers in Metropolitan Areas ........ Outline of This Report ......................... 2. WORKERS, EMPLOYERS, AND JOBS IN FOUR METROPOLITAN AREAS...................... Population and Worker Characteristics ............... The Employer Survey .......................... Employer Characteristics ........................ Labor Market Tightness and Skill Requirements ......... Employer Attitudes and Hiring Behaviors ............. Conclusion ................................. 3. THE DEMAND FOR WELFARE RECIPIENTS: PROSPECTIVE AND ACTUAL HIRING BY EMPLOYERS ............................... Prospective and Actual Demand for Welfare Recipients .... Variation in Demand by Characteristics of Employers ..... Effects of Employer Attitudes and Hiring Behavior ....... Conclusion ................................. 4. WHAT HAPPENS WHEN WELFARE RECIPIENTS ARE HIRED? .................................. Recently Hired Recipients: Who Are They? ............ What Kinds of Jobs Do They Fill? .................. Wages, Benefits, and Hours Worked ................ Worker, Employer, and Job Characteristics ............ 1 1 4 5 7 7 10 11 13 16 20 23 24 28 34 39 43 44 48 50 54 xi Turnover and Job Performance .................... Variation in Performance by Demographics and Time Period................................. Conclusion ................................. 5. CONCLUSION ............................. Summary of Findings .......................... Policy Implications............................ Preparing for a Downturn ...................... Improving Access to Jobs ...................... Improving Job Performance and Retention ........... Employment for the Hardest-to-Serve .............. Serving Low-Income Fathers .................... Appendix A. Regression Results ............................ B. Survey of Employers ........................... 57 66 74 77 78 80 80 82 82 83 84 85 91 Bibliography .................................. 141 About the Authors ............................... 151 xii Figures 2.1. Evidence of Labor Market Tightness .............. 2.2. Percentage of Employers Who Would Hire Disadvantaged Workers, by Disadvantage .......... 3.1. Prospective and Actual Hiring of Welfare Recipients .... 3.2. Prospective and Actual Hiring of Welfare Recipients in Central Cities and Suburbs .................... 3.3. Prospective and Actual Hiring of Welfare Recipients, by Vacancy Rate of Firm ....................... 4.1. Recently Hired Welfare Recipients and Poor Female- Headed Households, by Race .................. 4.2. Starting Hourly Wages Paid to Recently Hired Welfare Recipients ............................... 4.3. Hours Worked per Week by Recently Hired Welfare Recipients ............................... 4.4. Percentage of Recently Hired Welfare Recipients Still with the Firm ............................ 4.5. Months of Employment of Recently Hired Welfare Recipients ............................... 4.6. Performance of Recently Hired Welfare Recipients Relative to Others in the Same Position ............ 4.7. Percentage of Employers Observing Problems with Recently Hired Welfare Recipients ............... 4.8. Race of Recently Hired Welfare Recipients, by Year Hired.................................. 4.9. Starting Hourly Wages Paid to Recently Hired Welfare Recipients, by Year Hired ..................... 4.10. Performance of Recently Hired Welfare Recipients Relative to Others in the Same Position, by Year Hired.................................. 14 17 24 30 32 45 51 51 58 58 61 62 69 69 70 xiii Tables 2.1. Metropolitan Area Characteristics ................ 2.2. Central City Characteristics ................... 2.3. Characteristics of Establishments and Jobs .......... 2.4. Employer Skill Demands ..................... 2.5. Employer Attitudes and Hiring Behaviors .......... 3.1. Prospective and Actual Hiring of Welfare Recipients .... 3.2. Prospective and Actual Hiring of Welfare Recipients, by Industry Size, Type, and Minority Ownership ........ 3.3. Prospective and Actual Hiring, by Location Within Metropolitan Areas ......................... 3.4. Prospective and Actual Hiring of Welfare Recipients, by Skill Demands ............................ 3.5. Prospective and Actual Hiring of Welfare Recipients, by Employer Attitudes ......................... 3.6. Prospective and Actual Hiring of Welfare Recipients, by Employer Hiring Behavior .................... 4.1. Education and Work Experience of Recently Hired Welfare Recipients ......................... 4.2. Occupations Filled and Tasks Performed Daily by Recently Hired Welfare Recipients ............... 4.3. Starting Wages, Hours Worked, and Health Insurance Coverage for Recently Hired Welfare Recipients ...... 4.4. Characteristics and Wages of Recently Hired Welfare Recipients, by Employer Size and Industry .......... 4.5. Characteristics and Wages of Recently Hired Welfare Recipients, by Employer Location ............... 4.6. Characteristics of Recently Hired Welfare Recipients, by Type of Jobs Filled ......................... 4.7. Employment Duration and Turnover of Recently Hired Welfare Recipients ......................... 4.8. Absenteeism Problems of Recently Hired Welfare Recipients ............................... 8 9 12 16 18 25 29 31 33 35 36 45 48 52 54 55 56 59 62 xv 4.9. Performance and Employment Duration of Recently Hired Recipients, by Retention Status ............. 4.10. Turnover of Recently Hired Welfare Recipients, by Characteristics of Recipients ................... 4.11. Performance of Recently Hired Welfare Recipients, by Characteristics of Recipients ................... 4.12. Worker Characteristics and Job Performance of Recently Hired Recipients, by Month and Year Hired ......... A.1. Regression Results: Determinants of Prospective and Actual Hiring ............................ A.2. Regression Results: Determinants of Wages and Relative Performance ............................. 65 67 67 71 85 87 xvi Acknowledgments Many individuals and institutions contributed to this study in some way, shape, or form. We are very grateful to the Public Policy Institute of California and to the Joyce Foundation for providing funding for the survey. We also benefited greatly from the Urban Institute and the Russell Sage Foundation. In particular, we thank Lisa Kahraman at the Russell Sage Foundation for her valuable research assistance. We also thank Keith Ihlanfeldt, Jens Ludwig, and seminar participants at Harvard University, the Urban Institute, the University of Michigan and the Joint Center for Poverty Research for extremely helpful comments on previous versions of this study. At the Public Policy Institute of California, Michael Teitz provided excellent guidance at every stage of the project, read various drafts, and helped sharpen the focus of the study. Maureen Waller and Margaret O’Brien-Strain, also of the Public Policy Institute of California, Sheldon Danziger of the University of Michigan, and David Autor of MIT reviewed final drafts of this study and offered insightful and detailed comments. We thank all of these individuals for their hard work, which has greatly improved the quality of our report. Although many people contributed to this study, the authors remain solely responsible for its content. xvii 1. Introduction With the passage of the Personal Responsibility and Work Opportunity Reconciliation Act of 1996, along with earlier welfare reforms at the state level, the nation has embarked on a dramatic reshaping of its support system for low-income families. Supporters of these revisions in welfare predicted that the poor would enjoy improvements in living standards and self-sufficiency, with most entering the labor market and having little difficulty finding work (albeit sometimes at low wages). In contrast, many critics predicted dire consequences for adults and their children. In particular, the critics (e.g., Edelman, 1997) were concerned with possible high rates of unemployment, frequent job turnover, and low net wages and benefits among those leaving the welfare rolls, especially if they were forced off by time limits or sanctions. This study assesses some of these predictions by analyzing the outcomes associated with welfare reform in the workplace, as experienced by employers and the welfare recipients whom they have hired in the past few years. The analysis is based on data from a new survey of employers in four large metropolitan areas. Before considering these data and our findings in more detail, we review the evidence in the research literature to date, as well as the most important questions that remain to be answered about the effects of welfare reform in the labor market. Recent Literature Two facts are clear about the experiment with welfare reform to date: (1) the welfare (and Food Stamp) caseloads have declined dramatically since the early 1990s, and (2) employment among recipients has risen sharply as well. A variety of studies have analyzed these developments, using pooled time-series and cross-sectional data across states. Some disagreement remains about exact magnitudes, but virtually all these studies agree that welfare reform has contributed to both of these trends 1 without fully accounting for either. In addition, the strong labor market has certainly contributed to both developments, as have other factors, such as the expansion of the Earned Income Tax Credit and Medicaid programs, and perhaps recent increases in the minimum wage.1 Another set of studies has followed the experiences of recipients who have left the welfare rolls, and especially their labor market earnings. These studies of welfare “leavers” use either administrative or survey data from one or more states and are summarized in Loprest (1999). They generally show that, on the one hand, 70–80 percent of welfare leavers gain some employment within a year after departing the rolls; on the other hand, employment rates at any point in time are generally around 50–60 percent, indicating substantial unemployment as well. The extent to which this unemployment reflects high rates of turnover, lengthy spells between jobs, or labor force withdrawals is less clear.2 Furthermore, those studies indicate that earnings levels among former welfare recipients are certainly rising and that many families are better off than they were on welfare (e.g., Acs et al., 1999; Danziger et al., 2000). But low earnings also leave many families with working parents below the poverty line and still dependent on public assistance to varying extents (e.g., Cancian et al., 1999). This is particularly true for some groups, such as high school dropouts and those who have been sanctioned off the rolls. The sense that serious problems plague particular groups of current and former welfare recipients has been reinforced by evidence from Primus et al. (1999), who find declining real incomes for the bottom decile of female-headed families during the late 1990s. Employment and earnings are limited for some current and former welfare recipients by a variety of “barriers,” including low skills, health and mental health ____________ 1The relevant studies on changes in the caseloads include the Council of Economic Advisers (1997, 1999), Figlio and Ziliak (1999), and Wallace and Blank (1999). Studies of trends in female employment include Meyer and Rosenbaum (1999, 2000), Ellwood (1999), and Burtless (1999). For evidence that the minimum wage might actually raise employment and lower welfare dependency among single mothers, see Turner (1999). 2These rates of unemployment among welfare leavers are similar to those that have been observed in the past—e.g., Bane and Ellwood (1994). Since some fraction of those who leave the rolls do so because of marriage or other such household changes, it is not surprising that some leavers do not enter the labor market at all. 2 concerns, and a variety of other personal problems (Danziger et al., 1999). Thus, the evidence to date indicates rising employment and earnings among welfare recipients in general but serious labor market difficulties for at least some of these recipients. Furthermore, a number of important questions remain unanswered by these studies. They include: 1. What are the employment prospects of the least-skilled and least-experienced welfare recipients? Are employers willing to hire them at all, and to what extent? 2. Do black and Hispanic women on welfare face particular problems gaining employment, and if so, why? 3. To what extent do these employment prospects depend on the business cycle and the tightness of current labor markets in the United States? 4. Once welfare recipients are hired, are they plagued by high job turnover and poor work performance? If so, what are the causes of these difficulties in the workplace? 5. Are the earnings and benefits of welfare recipients sufficient to enable them to achieve financial independence over time? Many of these questions involve the characteristics and behavior of welfare recipients themselves but also the characteristics, attitudes, and behaviors of employers who may or may not hire them. In particular, answers to these questions depend on employers’ views of welfare recipients, especially those with very poor skills or work experience; the extent to which they have hired welfare recipients and expect to do so in the future; the extent to which some employers, especially those in smaller establishments or located far from downtown areas, are accessible to recipients residing in poor inner-city neighborhoods; and how well recipients meet the needs of employers in the jobs that they fill. To fully answer these questions, it is necessary to have some data directly from the demand side of the labor market—i.e., from employers and establishments where recipients are potentially or actually hired. To date, few such data have been available to researchers. In this report, we attempt to answer the questions by presenting and analyzing new data on 3 employers, particularly their experiences with and attitudes toward welfare recipients in the aftermath of welfare reform. A New Study of Employers in Metropolitan Areas To deal with the paucity of demand-side labor market data, a new survey was administered to over 3,000 employers in four major metropolitan areas: Los Angeles, Chicago, Cleveland, and Milwaukee.3 The survey builds on earlier work done on employers and their hiring of less-educated workers (e.g., Holzer, 1996) and earlier versions of this survey administered in Michigan (Holzer, 1999a). The survey focuses on employer willingness to hire welfare recipients, the extent to which they have actually done so, and what their experiences have been with recipients. The characteristics of the jobs filled by welfare recipients, their performance in those jobs, and the personal demographic characteristics of those hired are all documented. Any experiences that employers have had with local welfare-to-work agencies, and their openness to a variety of policy interventions designed to aid recipients, are considered as well. The four metropolitan areas considered here are of particular interest because of their different economic and demographic characteristics and policy environments. In particular, Los Angeles differs considerably from the three Midwestern areas in its size and geographic decentralization, its ethnic and immigrant composition (i.e., its heavily Hispanic and Asian populations), and other factors. The Midwestern areas also differ among themselves, with Chicago being much larger than the other two areas and more ethnically diverse as well. The welfare reforms implemented in these areas have clear similarities as well as differences. All of the state programs espouse a strong “work first” component that involves compulsory employment activities, job search assistance, sanctions for noncompliance, exemptions for certain family responsibilities and health problems, and time limits on ____________ 3We received funding from the Public Policy Institute of California (PPIC) for the data collection in Los Angeles and from the Joyce Foundation in the other metropolitan areas. PPIC’s primary focus on California and the Joyce Foundation’s primary interest in the Midwestern region of the United States account for the particular selection of metropolitan areas for this study. 4 recipiency. But given the size and ethnic diversity of the caseload in Los Angeles, implementation of its plan has been relatively slow and not always aggressive (e.g., Quint et al., 1999; Zellman et al., 1999). In contrast, the Wisconsin Works (or W-2) program is widely regarded as being the most aggressive in the country (Wiseman, 1999; Nightingale and Mikelson, 2000). It replaced the entitlement to income benefits with a four-tiered system of employment that rapidly pushes participants toward unsubsidized employment. Generous subsidized child care and transportation assistance are provided, along with a variety of other services. The drop in the welfare caseload in Wisconsin has also been dramatic, with almost 90 percent of the caseload eliminated between 1992 and 1998 (and about 75 percent in Milwaukee County).4 Using our survey data, we analyze the variation in hiring outcomes for welfare recipients across these four areas. Because establishment characteristics vary across these areas, we also examine the extent to which these characteristics account for differences in employment outcomes of welfare recipients across these areas. Finally, we consider differences in local demographics, geography, and policy, as well as their effects on these outcomes. Outline of This Report The outline of the remainder of this report is as follows. Chapter 2 reviews the demographic characteristics of the population in the four metropolitan areas under study as well as some general characteristics of employers and jobs that should be relevant for the analysis of demand for welfare recipients. The employer survey that generated the data for this study will also be discussed in more detail here. Chapter 3 considers data on the demand for welfare recipients. Two kinds of data are considered: prospective demand, or the self-reported willingness of employers to hire ____________ 4The caseload data are available from the Brookings Institution. Caseload reductions in Los Angeles County, Cook County (Chicago), and Cuyahoga County (Cleveland) from their peaks in the early 1990s have been approximately 20 percent, 29 percent, and 32 percent, respectively. In California, a relatively weak economy in the early 1990s and a continued inflow of immigrants caused the rolls to rise until 1995. Consistent with this, analysis of the Los Angeles Jobs-First Greater Avenues of Independence (GAIN) program has demonstrated significant employment increases for participants but fairly small reductions in the welfare rolls (Freedman et al., 1999). 5 recipients currently or in the future, and actual demand, or the hiring of recipients that has occurred to date. Many of the establishment characteristics associated with different levels of demand are considered here as well. Chapter 4 reviews data on what happens after welfare recipients are hired at these establishments. We focus on the most recently hired welfare recipients and consider a wide range of outcomes—demographics of the worker hired, characteristics of the job filled, wages and benefits earned, hours worked, job retention and duration, and performance ratings and problems. Finally, Chapter 5 summarizes the findings of our study and review potential implications for policy in California and elsewhere. 6 2. Workers, Employers, and Jobs in Four Metropolitan Areas Before analyzing the data on employer attitudes toward and experiences with welfare recipients, we begin by reviewing some of the characteristics of workers and employers in the four metropolitan areas: Los Angeles, Chicago, Cleveland, and Milwaukee. First, we review some data on the demographics of the population and employment outcomes of workers in these areas, derived from standard data sources (i.e., the Census and the Bureau of Labor Statistics). We then describe our new survey in greater detail and present summary data from the survey on the characteristics of employers and jobs located in these areas. In particular, we focus on factors such as the current state of overall labor demand (as measured by the job vacancy rate and difficulties reported by employers while seeking new employees), general demands for skills in these establishments, and various hiring practices. These metropolitan area and establishment characteristics will set the stage for the analysis of demand for welfare recipients and experiences with those hired that we present in subsequent chapters. Population and Worker Characteristics We begin in Tables 2.1 and 2.2 by presenting some general characteristics of the four metropolitan areas and their central cities, respectively. Not surprisingly, we find that: • Los Angeles is the most heavily populated of the metropolitan areas, and (along with Chicago) among the greatest in geographic size. • Los Angeles also has the largest minority population of any of these areas, with roughly two-thirds composed of blacks, Hispanics, and Asians and almost half (44 percent) 7 Table 2.1 Metropolitan Area Characteristics Area, square miles, 1990 Los Angeles 4,060 Chicago Cleveland Milwaukee 5,065 2,708 1,460 Population 1997 (March 1)a 1990 (April 1) Change, 1990–1997 9,145,219 8,863,052 .032 7,773,896 7,410,858 .049 2,225,997 2,202,069 .011 1,451,179 1,432,149 .013 Racial characteristics, 1997b White Black Hispanic Asian .340 .623 .758 .795 .083 .201 .171 .125 .439 .127 .047 .057 .127 .043 .010 .017 Population without high school diploma, 1990c .186 .157 .162 .130 White .482 .578 .708 .752 Black .095 .274 .269 .198 Hispanic .600 .205 .024 .054 Asian .074 .024 .005 .014 Households with poor female heads, 1990c White Black Hispanic Asian .034 .041 .042 .045 .327 .194 .325 .296 .323 .705 .640 .638 .472 .137 .044 .072 .057 .010 .004 .008 Metro area unemployment rate, October 1998b .065 .039 .042 .032 Estimated poverty rate, 1993d .238 .137 .150 .137 SOURCES: U.S. Bureau of the Census for relevant years unless noted otherwise. aMarch 1997 Current Population Survey, U.S. Bureau of the Census. bRacial categories include non-Hispanic white, non-Hispanic black, Hispanic, and non-Hispanic Asian. cRacial categories are not exclusive of Hispanic background. Hence, columns do not add to 100 percent. dU.S. Bureau of Labor Statistics, 1999. 8 Table 2.2 Central City Characteristics Los Angeles Chicago Cleveland Milwaukee Population, 1997 (March 1)a 3,553,638 2,721,547 498,246 590,503 Change, 1990–1997 .020 –.022 –.015 –.060 Population, 1997b White Black Hispanic Asian .329 .371 .497 .625 .102 .363 .326 .239 .461 .217 .165 .108 .105 .049 .003 .013 Estimated poverty rate, 1995 .286 .228 .299 .229 Change in jobs, 1993–1996c Central city Suburbs –.048 .020 .004 .045 –.047 .090 .084 .123 SOURCES: U.S. Bureau of the Census for relevant years unless noted otherwise. aMarch 1997 Current Population Survey, U.S. Bureau of the Census. bRacial categories include non-Hispanic white, non-Hispanic black, Hispanic, and non-Hispanic Asian. cDepartment of Housing and Urban Development (HUD) Special Tabulations of County Business Patterns, U.S. Bureau of the Census. composed of Hispanics alone. The central cities in these areas have much larger concentrations of black and Hispanic populations than do the overall metropolitan areas, with minorities accounting for half or more of central city populations everywhere but Milwaukee. Both employment and population growth in the suburbs exceed those of the central cities in each metropolitan area. • Minorities are somewhat more heavily concentrated among high school dropouts than among the overall populations, and much more heavily concentrated among poor female heads of households in all areas. The generally lower educational attainment of minorities than whites, particularly among Hispanics and especially among immigrants, accounts for their 9 greater representation in the population of high school dropouts (Mare, 1995) and the higher fraction of high school dropouts overall in Los Angeles.1 But the very strong majorities of poor female heads of households who are black in the Midwestern areas also reflect the much higher rates of female headship in this population group (e.g., Ellwood and Crane, 1990). • Unemployment and poverty rates are currently higher in Los Angeles than in the Midwest. As of October 1998, Milwaukee had the lowest unemployment rate of these metropolitan areas and Los Angeles had the highest, reflecting the generally tighter labor markets of the Midwestern region relative to that of California. Higher poverty rates in Los Angeles no doubt reflect a longer-term trend, perhaps because of the influx of poor and less-educated immigrants to this area (e.g., Borjas, 1994). The Employer Survey The focus of this study is the survey that was administered to roughly 750 employers in each of these metropolitan areas between July 1998 and June 1999.2 Employers were drawn from lists compiled by Survey Sampling Inc. (SSI), primarily from telephone directories. The sample was random within strata defined by establishment size, and the size strata were chosen so that the size distribution of establishments would resemble that of the overall workforce.3 In each establishment, we ____________ 1Given that overall population characteristics reflect data from 1997 whereas those for high school dropouts and poor female household heads are for 1990, the relative concentration of minorities in the latter groups might actually understate what actually exists today. 2The areas surveyed roughly correspond to the Primary Metropolitan Statistical Areas in each case. They include Cook, DuPage, and McHenry Counties in Chicago; Cuyahoga, Geauga, Lake, and Medina Counties in Cleveland; Milwaukee, Ozaukee, Washington, and Waukesha Counties in Milwaukee; and Los Angeles County in Los Angeles. 3In other words, large establishments are oversampled so that the percentages of establishments in different size groups would reflect the percentages of workers in those groups. This implies a sample of establishments that is already weighted by size to some extent. The size distributions used here were 20 percent in the 1–19 category, 30 percent in the 20–99 category, and 50 percent in the category of 100 and above. Since response rates in small establishments lag behind those of larger establishments, the actual 10 identified and tried to interview the individual responsible for entry-level hiring. Once this person was identified and contacted within these establishments, response rates averaged about 70 percent. The survey instrument, which can be found in Appendix B, was administered over the phone and took roughly 20 minutes to complete. Questions focused on the general characteristics of the employer, including size, industry, and location; composition of the workforce and of current jobs at the establishment, such as skill requirements and job vacancy rates; the demographics of and methods used to hire the worker in the most recently filled noncollege job; the number of welfare recipients hired in the past year or two and characteristics of the most recently hired welfare recipient and the job that he or she filled; experiences with any welfare-to-work agencies that were contacted; and prospective responses to any solicitations that might have come from such agencies. Employer Characteristics Table 2.3 shows some general characteristics of the sample of employers in each metropolitan area, such as their sizes, industries, and locations. The data indicate roughly similar distributions of employers across size and industry categories in these four areas. The results by size and industry indicate that retail trade and services account for about the majority of establishments in each area, and manufacturing accounts for 20 percent or less. The concentrations of employment in manufacturing are somewhat greater in Milwaukee and Cleveland than in Chicago or Los Angeles.4 Somewhat greater variance across metropolitan areas can be found in the locations of establishments. Several measures are presented here, ________________________________________________________ distributions of establishments across these categories are a bit more skewed toward the larger establishments. 4The formula for the standard error on the mean of a dichotomous variable is the square root of p(1 – p)/n, where p is the mean and n is the sample size. Standard errors on the means presented in this and the following chapters are generally in the range of .010–.020, and standard errors on differences in means across categories are generally .020–.025. Any differences discussed below are those that are at least marginally significant by conventional standards. 11 Table 2.3 Characteristics of Establishments and Jobs Firm size 1–19 20–99 100+ Los Angeles Chicago Cleveland Milwaukee .23 .22 .23 .19 .22 .21 .58 .55 .56 .24 .23 .53 Industry Manufacturing Retail trade Services .14 .16 .21 .22 .22 .22 .41 .38 .36 .20 .21 .39 Location Central city Suburbs 1/4 mile or less from public transit 30 min or less from downtown by public transit Both of above Distance from population on public assistance Distance from population not on public assistance Ratio .35 .65 .73 .58 .45 20.25 21.82 .92 .36 .73 .64 .27 .64 .65 .60 .76 .41 .56 19.67 15.75 21.05 19.19 .90 .79 .59 .41 .63 .71 .53 12.10 15.32 .74 Minority ownership .19 .14 .08 .08 Sample size 711 700 718 751 NOTES: Columns under “Firm size” and “Location (central city and suburbs)” sum to approximately one. Those under “Industry” sum approximately to 0.8, as other industries are omitted. Sample sizes are roughly constant throughout the remaining tables in this chapter. such as (a) the percentages of establishments located in the central city versus the suburbs, (b) the percentages located within one-quarter of a mile of a public transit stop, within a 30-minute ride from downtown using public transit, or both,5 and (c) the average distances in miles of ____________ 5The relationship of these locations to public transit is self-reported by the respondents at these establishments. 12 establishments from the populations receiving and not receiving public assistance, as well as the ratios of these average distances.6 The results also show that the majority of employers in Los Angeles and Chicago are in the suburbs, whereas most in Cleveland and Milwaukee are in the central city.7 In all areas, over two-thirds of employers are located near public transit stops, well over half are located within a 30-minute ride of downtown, and between 40 and 60 percent have both attributes. Interestingly, employers in Los Angeles are more likely to be near public transit stops than are those in other areas, reflecting that city’s expansive bus route systems; but the rides are somewhat lengthier, reflecting the greater geographic size and dispersion of the metro area there. As for distances from populations, employers are generally closer to the populations on public assistance than to those that are not; but on average, employers in Los Angeles are relatively farther away from these populations than are employers in the other metro areas.8 Finally, we find larger percentages of minority-owned establishments in Los Angeles than in the other metro areas. Labor Market Tightness and Skill Requirements One of the most salient features of labor markets in the United States today is the degree of tightness that has been achieved in the late 1990s without generating inflation (e.g., Katz and Krueger, 1999). ____________ 6These distances are weighted averages of the distances from the establishments’ Census tracts to every other Census tract in the metro area, weighted by the percentages of each population group that are in those other Census tracts, according to the 1990 Census of Population. For an illustration of these measures and how they affect the hiring practices of establishments, see Holzer and Ihlanfeldt (1996). 7The central city in Los Angeles is defined in such a way as to exclude the San Fernando Valley and to include East Los Angeles. 8The relatively greater proximity of employers to the low-income population parallels their greater proximity to minority groups (Holzer and Ihlanfeldt, 1996). But employment growth is often much more concentrated in outlying suburban areas that are farther away from the minority populations (Hughes and Sternberg, 1992; Raphael, 1997), thus contributing to “spatial mismatch” problems between the locations of employers and minority populations. 13 Tight labor markets have helped generate significant improvements in the employment rates of many groups, particularly minorities (Freeman and Rodgers, 1999). As noted above, economic conditions have also contributed to declining welfare rolls and to the apparent employability of welfare recipients. To gauge the current tightness of labor markets and how it varies within and among these four metro areas, we present several measures of such tightness from the employer’s perspective in Figure 2.1. The first, presented in Panel A, is the job vacancy rate at the establishment, defined as the percentage of all jobs that are vacant, that the employer is currently trying to fill, and that are available for immediate hire. In Panel B, we also show employer responses to questions about how difficult it currently is to find qualified applicants for their positions and how many employers had to fill jobs in the past two years with individuals who were less qualified than usual. Panel A Percentage of jobs vacant Los Angeles Chicago Cleveland Milwaukee 4.0 4.4 4.8 5.2 5.6 6.0 6.4 6.8 7.2 Percentage of employees who had difficulty finding qualified applicants Panel B Somewhat difficult Very difficult Percentage who hired less-qualified applicant in the past two years 10 15 20 25 30 35 40 45 50 55 60 Figure 2.1—Evidence of Labor Market Tightness 14 The results showed that: • Job vacancy rates currently are very high, ranging from just under 5 percent in Los Angeles to 6 percent in Milwaukee; • Roughly 70–80 percent of employers claim it is somewhat or very difficult to find qualified workers; and • One-third to almost one-half have had to hire less-qualified workers than usual in the past two years. All of these results are consistent with the view that labor markets are currently very tight in the United States. Although the more subjective data on hiring difficulties are harder to interpret without some comparisons over time, the job vacancy rate has traditionally been in the range of 1–3 percent and certainly lower than the unemployment rate at most points in time (Abraham, 1983; Holzer, 1989). Thus, it is quite striking to have conventionally measured mean vacancy rates actually exceed local unemployment rates.9 Also, the labor market of Los Angeles appears to be somewhat looser than those in all of the other metro areas, according to the measures examined here. Another potentially important determinant of employer willingness to hire welfare recipients is the overall level of skills required at their establishments. In the survey, we asked respondents about the percentages of all jobs in their establishments that do not explicitly require any particular level of education or experience at the time of hiring; and, of these, the percentages that also do not require the daily use of reading, writing, or arithmetic. We calculated similar measures of skill demand for the most recently filled noncollege jobs at these establishments. These data appear in Table 2.4. The results show that 30–40 percent of all jobs in these establishments do not require education or experience, and roughly 10 percent or more also do not require reading, ____________ 9It is noteworthy that the median job vacancy rate for this sample is just .02, well below the mean that appears in Figure 2.1. The distribution of vacancy rates reported here is heavily skewed to the right, with 5 percent of all establishments reporting vacancy rates of 25 percent or higher and 1 percent reporting them above 50 percent (although these are primarily small establishments). In contrast, about 37 percent of establishments report no vacancies at all. 15 Table 2.4 Employer Skill Demands Jobs that require No education or experience Also no reading, writing, or arithmetic Newly filled jobs that require No education or experience Also no reading, writing, or arithmetic Los Angeles .30 .11 .25 .14 Chicago Cleveland Milwaukee .40 .39 .40 .09 .10 .17 .26 .28 .14 .12 .26 .12 writing, or arithmetic. The generally lower percentages of jobs that require no skills among newly filled ones than among all jobs seem to reflect a secular growth in the demand for these skills among employers, as has been documented elsewhere (e.g., Katz and Murphy, 1992; Murnane and Levy, 1995). The effects of these trends in skill demands on the employment prospects of welfare recipients, who frequently have much lower-than-average cognitive skills and labor market experience, need to be considered in any analysis of employer demand for these recipients. Employer Attitudes and Hiring Behaviors The tendency of employers to hire welfare recipients in any given area should generally be influenced not only by local demographics and employer characteristics but also by their own attitudes toward different groups of workers and their hiring practices and procedures. Attitudes toward groups can encompass potentially discriminatory racial or gender preferences (e.g., Becker, 1971) as well as attitudes toward disadvantaged groups that are frequently disproportionately minorities.10 The relevant hiring practices can include a variety of recruiting and screening ____________ 10The relationship between attitudes toward race and those toward stigmatized groups has been demonstrated in Holzer and Neumark (2000), who show that firms engaging in Affirmative Action express a greater willingness to hire from the various stigmatized groups that we consider below. 16 practices, as well as a tendency to use various kinds of local labor market institutions or intermediaries when hiring (Holzer, 1996). In Figure 2.2 and Table 2.5, we present summary data on various self-reported employer attitudes toward disadvantaged workers, their use of particular hiring screens, and their contacts with local agencies. The attitudinal data in the figure reflect employer responses to questions about whether they would have been willing to hire various kinds of disadvantaged or stigmatized workers into the noncollege job in their establishment that they most recently filled. The categories of workers included welfare recipients, those who had been unemployed for a year or more, those who listed only short-term or part-time work experience, and those with criminal records. In each case, employers were asked whether they would definitely, probably, probably not, or definitely not hire such an individual into the position that was recently filled; either of the first two answers was counted as an affirmative response. Welfare recipient Criminal record Long-term unemployed Los Angeles Chicago Cleveland Milwaukee Only short-term part-time experience 20 30 40 50 60 70 80 90 100 NOTE: “Would hire” indicates that employers would “definitely hire” or “probably hire” someone with the listed attribute into the noncollege job that they most recently filled in their establishment. Figure 2.2—Percentage of Employers Who Would Hire Disadvantaged Workers, by Disadvantage 17 Table 2.5 Employer Attitudes and Hiring Behaviors Would do Drug test Background check Los Angeles Chicago Cleveland Milwaukee .39 .42 .58 .54 .45 .65 .44 .54 When hiring recipient, did Test (including drug test) Background check .45 .49 .36 .40 .25 .52 .39 .34 Had contact with local agency Of these: Employer initiated Hired referral .17 .30 .58 .20 .37 .54 .22 .39 .48 .26 .39 .66 NOTES: “Would do” indicates how employers would prospectively hire welfare recipients referred by a local agency; “When hiring recipient, did” indicates, for the subsample that had hired one or more recipients in the previous two years, actual hiring behavior; and “Had contact with local agency” refers to contact with a local welfare-to-work agency in the past year. The questions about screening used to generate Table 2.5 were limited to whether employers would (or did) use various kinds of tests or background checks on prospective applicants, particularly drug tests and checks for criminal records. The questions about prospective use of such tests were based on a set of questions (described more fully below) on whether employers would hire welfare recipients; another set was asked of those who had recently hired current or former recipients. Finally, a set of questions was asked on whether employers had had any contact with a local welfare-to-work agency, and, if they had, what were the outcomes of that encounter. Here we consider the frequency of such contact, which side initiated it, and whether the employer hired a referral of the agency. The results of Figure 2.2 and Table 2.5 indicate that: • Virtually all employers express a willingness to hire welfare recipients, whereas only about 40 percent express a willingness to hire those with criminal records; 18 • Drug tests or criminal record checks would (or are) being done by roughly 40–60 percent of employers in the hiring of recipients; and • Roughly one-fifth of all firms have had contact with local welfare-to-work agencies, and half or more of these resulted in a referral being hired by the employer. The relative preferences expressed by employers regarding the various disadvantaged groups parallel those found in an earlier study of employers (Holzer, 1996). Although employers are generally quite open to the hiring of welfare recipients or those needing remedial education or training (presumably so long as they met other hiring criteria and skill requirements), they are far more reluctant to consider those with criminal records, even in a very tight labor market.11 The fear of these workers does not bode well for the employment prospects of less-educated young black men, whose high rates of criminal activity and incarceration (e.g., Freeman, 1992) likely generate a reluctance to hire even those young black men without criminal records.12 Indeed, the employment rates of young black men did not grow at all between the cyclical peaks of the late 1980s and the late 1990s, and their labor force participation rates actually declined during that time period—in stark contrast to the dramatic growth in labor force activity that has been observed among young black females and especially single mothers (Aron et al., 2000).13 The greater tendency of firms to engage in background checks than drug tests may be somewhat surprising, given the ease with which the latter can be done. The frequency of contact with agencies also appears ____________ 11The effect of the tight market can still be found in these data; in comparison, in the looser labor market of the early 1990s, only about a third of employers said they would hire such workers (Holzer, 1996). 12See, for instance, Kirschenman and Neckerman (1991) or Kirschenman (1991). Given the imperfect information that many employers have about whether any particular individual has been engaged in criminal activity, they likely discriminate “statistically” against an entire group with high rates of incarceration, such as young black men. 13Although Freeman and Rodgers (1999) focus primarily on improvements in employment rates for less educated young black men over the business cycle during the 1990s, their data are consistent with these observations for the peak-to-peak period. 19 to be rising a bit over time, judging by some comparisons with earlier observations.14 This rise might reflect the very tight labor markets of recent years and a growing tendency of employers to use temporary agencies and other labor market intermediaries (Katz and Krueger, 1999). Finally, although the variance in these attitudes and behaviors across metropolitan areas is not great and shows no particularly consistent pattern, employers in Milwaukee show somewhat greater tolerance of previous criminal activity among their job applicants and greater willingness to use local agencies in the hiring process. In contrast, employers in Los Angeles are relatively less likely to accept candidates with criminal records and have had less contact with local agencies than employers in any of the other metro areas. Conclusion The usefulness of the preceding data will become more apparent in subsequent chapters, but several striking results emerged from this analysis. For one thing, labor markets appear to be extremely tight; mean vacancy rates may exceed unemployment rates in many locations, and most employers note the difficulties of finding qualified job applicants. Although substantial numbers of jobs are available that do not require particular levels of education or experience, most require some minimal cognitive tasks, such as reading, writing, or arithmetic. Further, the number of jobs not requiring skills appears to be dwindling over time. Although most employers are located near public transit, a large number are not easily accessible by short rides from downtown areas, and their proximity to low-income populations varies greatly. Finally, we also note the much greater willingness of employers to hire welfare recipients than other disadvantaged groups, such as young ____________ 14In Michigan during 1997, just under 17 percent of employers had such contact, although a larger fraction of these contacts resulted in welfare recipients being hired (Holzer, 1999a). Although the differences between these results and those presented in Table 2.5 could simply reflect the differences across areas rather than over time, it is noteworthy that the frequency of use in Michigan in 1997 lagged behind all of the rates observed in the other Midwestern metropolitan areas presented here. See Pavetti et al. (2000) for more evidence on the role of intermediaries in the hiring of welfare recipients. 20 men with criminal records. This fact is consistent with the contrast between the dramatic growth of labor force activity among young women (especially single mothers) in the 1990s and the fairly stagnant employment rates for black young men. Also, there has been a growing but still limited tendency of employers to have contacts with welfare-towork agencies in their local areas, as part of a broader trend toward greater use of intermediaries (including temporary agencies) in the labor market. 21 3. The Demand for Welfare Recipients: Prospective and Actual Hiring by Employers The previous chapter considered the characteristics of employers in Los Angeles and the other metropolitan areas, some of which might affect their willingness to hire welfare recipients. In this chapter, we look at data on the demand for welfare recipients by employers in these areas. We distinguish between prospective and actual (or realized) demand for welfare recipients; the former refers to an expressed willingness to hire welfare recipients and the latter measures whether employers have done so to date. Because prospective demand measures a self-reported attitude rather than actual behavior, its reliability might be subject to greater question than actual demand. On the other hand, it might also be a cleaner measure of employers’ demand for the labor of welfare recipients, as actual demand should reflect a mix of demand-side (i.e., employer) and supply-side (i.e., worker) factors that influence hiring. A comparison of outcomes observed for the two kinds of measures should therefore enable us to sort out the relative importance of demand-side and supplyside factors that might be inhibiting the hiring of recipients in some circumstances.1 We begin below by presenting summary measures of prospective and actual hiring of welfare recipients for each of the four metropolitan areas. We then consider how these measures vary by employer characteristics (i.e., employer size, industry, minority ownership, and location within the metropolitan area); measures of labor market tightness and employer skill requirements; and a variety of hiring attitudes and behaviors, ____________ 1See Holzer (1999a) for evidence on prospective and actual demand for welfare recipients from data collected about a year earlier in Michigan. 23 including the use of local intermediary agencies. We conclude the chapter with some discussion of policy implications. Prospective and Actual Demand for Welfare Recipients Figure 3.1 and Table 3.1 present means of several measures of employer demand for welfare recipients by metropolitan area. These include measures of prospective demand, or willingness to hire recipients either currently or over the next year, and of actual demand, indicating whether the firm has actually hired recipients in the previous year or two. The survey questions about prospective hiring begin with, “Suppose you were contacted by an employment agency that was trying to place welfare recipients. We are specifically talking about recipients who do not have a high school diploma or any recent work experience. Do you currently/over the next year have any open positions that you might consider filling with these welfare recipients?” The questions about actual Prospective Panel A. Employers Percentage who have hired recipients in the past year Actual Percentage who would consider hiring recipients currently 0 4 8 12 16 20 24 28 32 36 40 Panel B. Jobs Prospective Percentage that employers would consider filling currently with recipients Actual Percentage filled by recipients in the past year Los Angeles Chicago Cleveland Milwaukee 0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 Figure 3.1—Prospective and Actual Hiring of Welfare Recipients 24 Table 3.1 Prospective and Actual Hiring of Welfare Recipients Prospective hiring Jobs that employers would consider filling with recipients currently Jobs that employers would consider filling with recipients over the next year Los Angeles .013 .055 Chicago Cleveland .014 .020 .037 .037 Milwaukee .016 .048 Actual hiring Jobs filled by recipients in the past year Jobs filled by recipients in the past two years Sample size .015 .029 711 .019 .039 .034 .049 700 718 .030 .045 751 hiring are based on the question, “To your knowledge, has your business in the past year/two years hired any women who had been on welfare?”2 The measure of actual hiring behavior, of course, depends critically on whether employers really know which of their employees are welfare recipients; but most employers in the survey indicated that they were relatively certain of the welfare recipiency status of at least some of these employees when they indicated that they had hired recipients in the recent past.3 ____________ 2The prospective questions impose a stronger skill deficiency on the welfare recipients who would be hired but allow employers to merely consider hiring them. If any employers answer “yes” to either question, they are asked about the numbers that they have hired or would consider hiring. These numbers (including zeros for those who answer “no”) are then calculated as percentages of all current jobs in the establishment, whether filled or vacant. 3For instance, over half of these employers claimed that they were “definitely sure” that their most recently hired recipient had in fact been on welfare, and roughly 80 percent were “definitely” or “fairly” sure. Employers might have an incentive to find out about current or past recipiency to qualify for various federal or state tax credits for the hiring of recipients, although take-up rates on these credits are often quite low (Holzer, 1999a). More important, welfare recipients may need to inform their employers of their status for local Temporary Assistance for Needy Families (TANF) administrators to be able to confirm that they are complying with work requirements in the new program. 25 Figure 3.1 presents two measures of prospective demand: the percentage of employers who claim that they would consider hiring welfare recipients currently and the percentage of all jobs that could be filled by these recipients. Each measure is presented separately by metro area. The measures of actual demand are the percentage of all employers who claim that they have hired welfare recipients during the past year and the percentage of all jobs filled by them. Table 3.1 presents data only on jobs filled but for a broader range of prospective and actual hiring categories: jobs that employers would consider filling either currently or in the next year, and those that they have actually filled in the past year or two. The data indicate that there is substantial demand for welfare recipients in these metropolitan areas, both prospectively and in terms of actual hiring. We find that: • About a third of employers claim that they would consider hiring recipients currently and would potentially fill 1–2 percent of all current jobs this way. Over half also claim that they would consider doing so over the next year and would fill roughly 4–5 percent of all jobs this way.4 • 30–40 percent of employers also claim that they believe they have hired welfare recipients over the past two years, and 18–24 percent believe that they have done so in the previous year. As percentages of current jobs, these new hires would constitute roughly 3–5 percent in the previous two years and roughly 2–4 percent in the previous year. These latter measures reflect total hiring activity (or the flow of workers over the particular period), whereas the prospective measures capture a level of employment (or the stock of workers) at any point in time. In other words, the quantity of new hires over the past year would ________________________________________________________ Finally, firms are likely to have good information about the recipient status of workers hired through welfare-to-work agencies. 4Questions about hiring over the next year presumably involve expectations not only of the welfare recipients and their characteristics as potential employees but also of the employer’s expectations regarding his/her hiring needs over the next year, based on anticipated turnover and net employment growth. 26 have to be adjusted for turnover, which would generate a stock of employment for welfare recipients of over 1–2 percent of all jobs over the previous year.5 These data suggest that the overall level of employer demand for the labor of welfare recipients is quite substantial, as has been suggested by Burtless (1999) and others. To more fully assess the magnitude of demand, we note that, as of 2000, the welfare rolls had fallen by over two million adults since 1994. Studies of welfare “leavers” (e.g., Loprest, 1999; U.S. GAO, 1999) suggest that 70–80 percent have experienced some employment within a year of leaving the rolls, and roughly 50–60 percent are employed at any point in time. This suggests that, out of an overall labor force of about 140 million workers, welfare recipients have recently added over 1 percent to the labor force, in addition to those recipients who had been employed earlier or who were employed but still on the rolls. Those who remain on the rolls and will still enter the labor force over the next few years could constitute an extra 1 percent as well. In comparison, Table 3.1 implies that 1–2 percent of all jobs have already been filled by recipients at some point in the past two years; and roughly 1.5 percent of all current jobs are potentially available for new hires right away, with significantly more available over the coming year. Thus, overall demand for recipients appears sufficiently large to handle their inflow into the labor market, at least in the aggregate. This assumes, of course, that the prospective demand translates into actual demand if the recipients actually appear as applicants to the firm. Prospective demand might overstate actual demand if employers find welfare recipients relatively deficient in skills and other attributes when they appear and apply for work, especially relative to other applicants; but the great difficulty that employers currently have in finding acceptable applicants (as noted in Chapter 2) renders this concern less serious, at least in the context of the very tight labor markets of the late 1990s. Alternatively, employer responses to questions about hypothetical willingness to hire welfare recipients might reflect some desire to appear ____________ 5This calculation assumes a turnover rate of 40 percent, which transforms the 2–4 percent gross hire rate into a 1.2–2.4 percent rate net of turnover. The turnover rate is based on calculations that appear in Chapter 4. 27 “politically correct.” But it is not clear why this desire would surface in a confidential survey and why it would apply more to welfare recipients than to other kinds of disadvantaged applicants.6 Some variation in the levels of prospective and actual demand is also apparent across metropolitan areas. In particular, we find that employer demand for recipients is relatively low in Los Angeles and higher in Cleveland and Milwaukee. These differences across areas will be further explored in the analyses presented below. Variation in Demand by Characteristics of Employers There are no doubt other dimensions along which the demand for welfare recipients varies. Table 3.2 presents measures of demand by various employer characteristics, such as establishment size, industry, and whether the establishment is minority-owned. We limit ourselves to the variables that measure the percentage of all jobs filled rather than the percentage of employers who would fill them. The results of Table 3.2 show that both actual and prospective demand for welfare recipients is higher in the retail trade sector and in minority-owned businesses than elsewhere, whereas demand is particularly low in manufacturing establishments. The greater inclination of minority-owned businesses to hire welfare recipients parallels their greater tendency to hire minorities, all other factors being equal (Bates, 1998; Raphael et al., 2000). Differences by industry and establishment size might well reflect variables such as skill requirements, which are considered below; evidence of lower skill needs in small establishments appears in Brown et al. (1990), and the recent rise in skill requirements in the manufacturing sector has been linked to technological innovations there (Berman et al., 1994).7 ____________ 6As we noted in Chapter 2, employers were clearly willing in this survey to reveal their reluctance to hire applicants with criminal records. 7Larger firms are more likely to have hired at least one recipient, given their larger volume of hiring activity; but the jobs so filled constitute a smaller percentage of all of their jobs than they do in smaller establishments. 28 Table 3.2 Prospective and Actual Hiring of Welfare Recipients, by Industry Size, Type, and Minority Ownership Prospective hiring Jobs that employers would consider filling with recipients currently Jobs that employers would consider filling with recipients over the next year 1–19 Size 20–99 100+ Industry Manufac- Retail turing Trade Minority Ownership Service Yes No .030 .015 .008 .009 .022 .016 .080 .042 .026 .032 .070 .039 .024 .015 .073 .041 29 Actual hiring Jobs filled by recipients in the past year Jobs filled by recipients in the past two years Sample size .047 .022 .016 .008 .042 .034 .069 .032 .029 .016 .056 .057 648 1,014 1,143 504 601 1,054 .042 .022 .070 .033 341 2,414 Some further evidence of variation in employer demand for recipients by location appears in Figure 3.2 and Table 3.3. We analyze measures of employer demand for welfare recipients by whether employers are in the central city or suburbs (in Figure 3.2); whether they are “near” public transit, defined as being within a quarter mile of a transit stop and also within a 30-minute ride of the downtown area; and whether they are located relatively “near” the public assistance population, defined as establishments in which the ratio of their distance to this population relative to the nonassisted population is below the mean (Table 3.3). The results indicate that willingness to hire welfare recipients does not vary substantially by the employer’s location within the metropolitan area, whereas actual demand varies considerably. More specifically, firms located in suburban areas or in areas that are less accessible to welfare recipients are just as willing to hire them as their counterparts who are more accessible, but they actually hire fewer of them. These data indicate that spatial and transportation factors affect the ability of recipients to access potentially available jobs, as has often been suggested in the literature on “spatial mismatch” (Kain, 1992; Ihlanfeldt Prospective Percentage of jobs that employers would consider filling currently with recipients Central City Suburb Actual Percentage of jobs filled by recipients in the past year 0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 Figure 3.2—Prospective and Actual Hiring of Welfare Recipients in Central Cities and Suburbs 30 Table 3.3 Prospective and Actual Hiring, by Location Within Metropolitan Areas Prospective hiring Jobs that employers would consider filling with recipients currently Jobs that employers would consider filling with recipients over the next year Near Public Transit .015 .047 Not Near Public Transit .019 .048 Near Population on Public Assistance .015 .041 Not Near Population on Public Assistance .016 .046 Actual hiring Jobs filled by recipients in the past year Jobs filled by recipients in the past two years .032 .049 .011 .037 .023 .056 .019 .031 Sample size 1,183 1,344 1,018 1,239 NOTE: “Near public transit” refers to employers located within a quarter mile of a transit stop and a 30-minute ride from downtown; “near population on public assistance” refers to employers whose ratio of distance of public assistance population to nonassistance population is below the mean of the entire sample (0.83). and Sjoquist, 1998). These results mirror earlier ones involving the black population’s (Holzer and Ihlanfeldt, 1996) or the welfare population’s (Blumenberg and Ong, 1998) access to employers, as would be expected for groups that frequently lack automobile transportation or information about the broader metropolitan labor market.8 ____________ 8This interpretation assumes, of course, that the lower rate of hiring of welfare recipients in the suburbs reflects a relatively lower rate at which they apply for jobs there, as opposed to employers’ lower rates of hiring these applicants. What little evidence we have in the survey on the tendency of employers to accept or reject welfare recipient applicants is based on a question of how many recipient referrals they have received from agencies and how many they have hired. The data suggest a somewhat greater tendency of central-city employers to hire one or more of their referrals (.58 versus .46), which is partly accounted for by their relatively greater size. The differences in tendency to hire from the applicant pool, therefore, do not appear large enough to fully account for the observed geographic differences in actual hiring, as has also been noted in the literature on minority employment differences (e.g., Holzer and Ihlanfeldt, 1996). 31 These data also imply that there is some unmet prospective demand for these workers in geographic areas that are not accessible to the welfare population. In periods when most welfare recipients face few constraints on their ability to get job offers nearby when they seek them, the existence of this unmet demand may not matter greatly. But in other times or places where nearby job availability may be insufficient to ensure quick employment for these workers (e.g., when labor markets are not as tight, or where larger numbers of recipients enter the market quickly), improving the access of welfare recipients to suburban employers could be an important part of the relevant policy mix for policymakers and intermediaries. Figure 3.3 and Table 3.4 reveal two more dimensions of employer characteristics that might affect their demand for welfare recipients: labor market tightness and skill demands. The data compare employer demand for welfare recipients when their job vacancy rate is above and below the mean for these metropolitan areas (Figure 3.3), and when the percentage of unskilled jobs in their establishments (defined as jobs that Prospective Percentage of jobs that employers would consider filling currently with recipients Job vacancy rate below 5 percent Job vacancy rate at or above 5 percent Actual Percentage of jobs filled by recipients in the past year 0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 Figure 3.3—Prospective and Actual Hiring of Welfare Recipients, by Vacancy Rate of Firm 32 Table 3.4 Prospective and Actual Hiring of Welfare Recipients, by Skill Demands Prospective hiring Jobs that employers would consider filling with recipients currently Jobs that employers would consider filling with recipients over the next year Proportion of Unskilled Workers in Firm Below Mean Above Mean .014 .023 .032 .066 Actual hiring Jobs filled by recipients in the past year Jobs filled by recipients in the past two years .017 .031 .043 .061 Sample size 1,621 1,143 NOTE: “Unskilled workers” refers to the percentage of jobs in the establishment that require no specific levels of education or experience. require no education or experience) is above or below the mean (Table 3.4). The results in Figure 3.3 indicate that labor market tightness, as measured by the job vacancy rate, is a strong determinant of both prospective and actual demand for recipients. These differences are quite large—in particular, prospective demand for welfare recipients currently is five times as high in establishments with higher vacancy rates than in those with lower ones, and even over longer time periods, demand for recipients (both prospective and actual) is roughly twice as high in the highervacancy establishments. This pattern is consistent with the well-known tendency for the employment of less-skilled and less-experienced employees to be more sensitive to business cycles than that of moreskilled or more-experienced groups (e.g., Freeman and Rodgers, 1999). Not surprisingly, Table 3.4 shows that establishments with large percentages of unskilled workers have higher demand (both prospective and actual) for recipients than do those with lower percentages of these workers. Thus, both overall hiring needs and skill requirements at establishments are important influences on an employer’s willingness to hire welfare recipients. 33 These findings suggest that changes in overall labor market tightness or in the relative demand for unskilled workers over time could have important effects on job availability for welfare recipients. During an economic downturn, for example, the hiring rate for welfare recipients should decline substantially as well.9 Whether this affects their employment rates or income levels may depend on a variety of other factors (such as the severity of the downturn or how much work experience they have accumulated at that time), but the potential effects of a downturn should be a source of concern as we consider policy issues for this population. Effects of Employer Attitudes and Hiring Behavior In Chapter 2 we considered data on a variety of employer attitudes and hiring practices that might potentially affect their willingness to hire welfare recipients. These included (1) their willingness to hire workers with some type of stigma, such as a criminal record or limited work history (at least for the last noncollege job that they had to fill), (2) whether they would conduct drug tests or do background checks, and (3) whether they had contact with a local welfare-to-work agency, and if they did, whether they initiated such contact. Tables 3.5 and 3.6 analyze our four measures of demand for welfare recipients by each of these measures of attitudes or behaviors. We find that: • Both prospective and actual demand are higher among employers who are willing to hire job candidates with criminal records or weak employment histories; • Demand is higher among those who do not conduct drug tests or background checks; and • Actual demand is considerably higher among those who have had contact with local agencies, although prospective demand is not. ____________ 9Mean vacancy rates in recessions have generally been roughly 1–2 percent (Abraham, 1983; Holzer, 1989). Thus, the vast majority of employers would be distributed at the low end of the lower-vacancy category during a downturn. 34 Table 3.5 Prospective and Actual Hiring of Welfare Recipients, by Employer Attitudes Prospective hiring Jobs that employers would consider filling with recipients currently Jobs that employers would consider filling with recipients over the next year Criminal Record Yes No Would Hire Long-Term Unemployed Yes No Short-Term Experience Yes No . .020 .014 .017 .010 .018 .011 .051 .044 .048 .038 .051 .034 Actual hiring Jobs filled by recipients in the past year Jobs filled by recipients in the past two years .030 .026 .029 .022 .028 .021 .050 .039 .044 .034 .044 .031 Sample size 926 1,375 2,133 468 1,836 815 NOTE: “Would hire” indicates that employers would “definitely hire” or “probably hire” someone with the listed attribute into the noncollege job they most recently filled in their establishment. The fact that contact with agencies influences actual but not prospective demand indicates that these agencies might help transform potential into realized demand by bridging gaps in access of welfare recipients to available jobs for them. Interestingly, actual demand for recipients is higher when the agency has contacted the firm, although the basic finding holds up in both cases—when either the agency or the employer initiates contact, employer demand for recipients is still higher than when no contact occurs at all. Two other statistical or interpretive questions can be raised about these results. First, it is possible that the very subjective answers on questions about attitudes might be measured with considerable error, especially since the most recently filled jobs to which they refer vary a great deal. However, if the measurement error is random, it is likely to bias the results toward a finding of no difference across attitudinal 35 Table 3.6 Prospective and Actual Hiring of Welfare Recipients, by Employer Hiring Behavior Prospective hiring Jobs that employers would consider filling with recipients currently Jobs that employers would consider filling with recipients over the next year Drug Test Yes No .012 .019 .040 .048 Would Do: Criminal Check Use Agency Yes No Yes No Firm-Initiated Contact Yes No .013 .020 .016 .016 .021 .014 .041 .050 .041 .047 .042 .042 36 Actual hiring Jobs filled by recipients in the past year Jobs filled by recipients in the past two years .021 .036 Sample size 1,128 NOTE: See Table 2.5 for definition of “Would Do.” .030 .042 1,531 .026 .037 1,489 .027 .041 1,121 .037 .019 .060 .027 580 1,920 .036 .026 201 .038 .063 349 categories, which implies that the true differences are likely to be even greater than those observed here. A second question involves the extent to which we are measuring attributes of employers that really have a causal effect on the demand for welfare recipients, as opposed to merely reflecting that demand and its underlying determinants. For example, employers who conduct drug tests or do background checks may be less inclined to hire welfare recipients. Alternatively, drug tests or background checks may objectively screen out some welfare recipients from employment. In truth, we cannot really distinguish between the two interpretations of these results; but even if the effects are strictly correlative and not causal, it is useful to know the extent to which the hiring of welfare recipients is correlated with these attitudes. Given the large number of employer attributes that appear to be correlated with the demand for welfare recipients, it seems reasonable to investigate these issues through multiple regression analysis.10 In Appendix Table A.1, we present results from estimated regression equations in which the dependent variables are the percentage of jobs filled by welfare recipients, either prospectively or actually. The independent variables in the first specification are only dichotomous variables for the metropolitan statistical area of the employer; the second specification also includes dummies for one-digit industry, establishment size, minority ownership, location within the metro area (measured by distance to the public assistance population and to local transit stops), measures of labor market tightness (i.e., the job vacancy rates as well as whether employers have difficulty finding qualified workers), demand for unskilled workers, and agency contact.11 Comparisons between the first and second specification in each case will indicate the extent to which the regressors added in the second equation can account for differences across metro areas in hiring behavior observed in the first specification. ____________ 10Readers who are less familiar with or interested in these more technical research findings can skip to the Conclusion of this chapter. 11We have omitted the attitudinal and behavioral variables in Table 3.5 from these specifications, given the many questions that exist about their accuracy and exogeneity. 37 All equations are estimated using the tobit specification, given the large concentrations of hiring outcomes at zero.12 The results indicate that the relatively low prospective demand of employers for welfare recipients in Los Angeles is little explained by the regression equations, although their actual demand is partly explained. In particular, the lower job vacancy rate in Los Angeles, the greater distance of employers from the low-income population, and their lessfrequent contact with local welfare-to-work agencies all help to account for the relatively low rate of hiring welfare recipients there. Thus, a looser labor market and the geographic features of Los Angeles account for some of the lower tendency of employers there to hire welfare recipients. But the relatively less-frequent contact of employers with local agencies seems to be part of a broader phenomenon of slower and less-aggressive implementation of welfare reform there that has been noted elsewhere (e.g., Zellman et al., 1999). The enormous size and ethnic diversity of the welfare caseload there seems to partly account for the slower implementation, and certain provisions of the CalWorks program also allow much greater discretion to local officials in how strictly they enforce work rules.13 These features of CalWorks stand in sharp contrast to the aggressive Wisconsin Works (or W-2) program, in which entitlement to cash benefits without work has been eliminated and participation in the four-tiered employment program has been quite strictly enforced (Nightingale and Mikelson, 2000). Furthermore, almost all of the effects of industry characteristics on such demand continue to be observed here, suggesting that many of their effects are independent of one another. In particular, the effects of industry (relative to retail trade), the job vacancy rate, and the demand for unskilled workers continue to be significant determinants of both ____________ 12The tobit specification is most appropriate when there is a latent variable that is potentially negative but is “censored” by the zero limit on its values. In this case, such a latent variable might be net employer willingness to hire new recipients, which could be negative in a situation where employers are discharging or laying off more than they are hiring. 13For instance, the five-year time limit on welfare recipiency in California began in January 1998, later than virtually every other state; when sanctions are imposed, they limit the payments of benefits for adults but not for children in their households; and exemptions from work rules can be given more liberally than in most other states. 38 prospective and actual demand. Other factors, such as relative distance from populations receiving public assistance, contact with agencies, and minority ownership, affect actual demand more strongly than prospective demand. These factors suggest that the access of workers receiving public assistance to employers who might demand their services varies across establishments as well, and that improving workers’ access to a broader range of employers might be a useful goal for policy interventions in some cases. It also seems noteworthy that the effect of establishment size on demand, which appeared to be positive in earlier tables, now appears to be negative. In other words, once we control for industry, job vacancy rates, skill demands, and the like, small establishments are less likely to hire welfare recipients than are other employers. This seems consistent with other evidence (e.g., Holzer, 1998) that small firms hire fewer blacks than do others and are perhaps more discriminatory in their hiring behavior. Conclusion The results indicate that, in the aggregate, the demand for welfare recipients seems relatively high, and sufficiently so to absorb the number of recipients entering the labor market recently and in the near future. But there is also substantial variation across establishments in their demand for welfare recipients. Firms in retail trade, those with high job vacancy rates, and those with generally higher demand for unskilled workers have hired more welfare recipients to date (as percentages of their workforces) and are prospectively more willing to do so currently or in the near future. Those near public transit or populations receiving public assistance, are minority-owned, or have had contact with local agencies have also been more able to translate their prospective demand for recipients into actual hiring, indicating that these factors affect the access of welfare recipients to jobs that potentially exist for them in many places. Lower job vacancy rates and less frequent contact with local agencies account for some, but not all, of the lower hiring of welfare recipients in Los Angeles than in Milwaukee or Cleveland. The slower 39 implementation of welfare reform in Los Angeles than elsewhere likely helps to account for this lower level of activity among employers and agencies there as well. The findings above also imply that location, transportation, information, and contacts regarding potential job openings are key determinants of the access of welfare recipients to these jobs. These factors often limit the access of minorities (especially blacks) to jobs in the labor market more broadly, and they might limit employment opportunities for minorities who have also been welfare recipients. Indeed, data presented in the next chapter suggest that minorities have more difficulty gaining employment than white welfare recipients, and perhaps these factors play some role in their relative disadvantage. Although we will discuss implications for policy more fully in Chapter 5, at least a few are worth noting here. For one thing, the strong effect of the job vacancy rate on demand for recipients suggests that, once an economic downturn occurs, the hiring of welfare recipients might be significantly lower than it is today. For welfare recipients (and other lowincome workers) who do not qualify for Unemployment Insurance or TANF (perhaps because of time limits), some other type of “safety net” must be found—whether it is a special Unemployment Insurance program, suspension of time limits on TANF, or public employment programs. Likewise, in areas or time periods in which overall demand for labor is more slack, the differential access of recipients to potentially available jobs (because of location or transportation, information, etc.) may be more critical determinants of their ability to find jobs quickly or easily. In these cases, improving their access to a wider range of establishments could be an important policy goal as well. Because local welfare-to-work agencies and other intermediaries seem to be relatively successful at placing recipients into jobs, strengthening their role in the job-search process (perhaps through greater outreach of such agencies to the employer community or greater public funding of their efforts) might be an important component of a broader effort to improve access to jobs among current or former welfare recipients. The data in this chapter confirm the findings on self-reported employer attitudes toward welfare recipients that we observed in the 40 previous chapter—namely, that employers are very open to hiring welfare recipients. But we also learned above that employers are much more wary about hiring other groups of applicants who are stigmatized in their eyes, especially those with criminal records. As more and more lowincome men in poor neighborhoods have become incarcerated in recent years and are becoming (or will some day become) “ex-cons,” we need to make greater efforts to integrate them into labor markets and to overcome employer concerns over their skills, work experience, and general reliability. More broadly, the lack of recent labor market progress among less-educated young black men needs to be addressed as well. Beyond merely placing welfare recipients into jobs, we must also be concerned with the quality of these jobs—in terms of wages paid, benefits provided, and hours of work available—as well as the characteristics of recipients who obtain them and their ability to retain their employment and perhaps advance in the workplace. We turn to a fuller consideration of these issues in the next chapter. 41 4. What Happens When Welfare Recipients Are Hired? The previous chapter indicated that the demand for welfare recipients, both prospective and actual, is quite extensive, even though the access of recipients to many of these jobs is limited. This finding also raises a number of important questions about which welfare recipients are being hired, into what kinds of jobs, and with what kinds of observed outcomes. More specifically, are minorities within the welfare population being hired at the same rate as their white counterparts, and are those with little formal education or work experience being hired at the same rate as their more educated and experienced counterparts? Do the jobs that welfare recipients fill require a great deal of cognitive, social, or verbal skill? What wages and benefits do they earn, and with what hours of work? From the employers’ point of view, has their performance been satisfactory? Is job turnover a major problem? And how do these outcomes vary according to the characteristics of the workers hired or when they were hired? This chapter explores many of these issues. For those employers who report that they have hired at least one welfare recipient during the previous two years, we focus on the last welfare recipient hired and on the job that person has filled. We begin by examining the demographics of these workers, in terms of race, educational attainment, and recent work experience. We then consider the characteristics of jobs they have filled, along with wages, benefits, and hours worked. Finally, we consider a variety of measures of worker outcomes, such as turnover rates, particular problems such as absenteeism, and overall ratings of worker performance by the employer. These outcomes will determine the extent to which welfare recipients can obtain the kinds of stable employment and work 43 experience that can lead to earnings growth and perhaps self-sufficiency over time. Recently Hired Recipients: Who Are They? Figure 4.1 and Table 4.1 present data on the characteristics of the most recently hired welfare recipients in these establishments. The vast majority of these recipients were hired in 1998 and early 1999, although a small fraction were hired before then.1 We limit the sample to individuals whom the employers were definitely or fairly sure had been on welfare, either currently or in the past.2 Figure 4.1 presents the race of the most recently hired welfare recipients, along with Census data on the race of poor female heads of households (from Table 2.1). Table 4.1 presents data on whether welfare recipients were high school graduates (not including those with General Equivalence Degrees (GEDs)), and whether or not they had any recent work experience when they were hired. The data are presented separately for Los Angeles and each of the other three metropolitan areas from the Midwest. The results of Figure 4.1 and Table 4.1 indicate the following: • The majority of recently hired welfare recipients are black or Hispanic in each metro area. In Los Angeles, almost threefourths of the recipients were black or Hispanic, whereas over half of the recipients hired in Chicago and Milwaukee were black. Only in Cleveland do whites constitute nearly half of those hired.3 ____________ 1The percentages hired in 1997, 1998, and early 1999 are .15, .59, and .22, respectively. The rest (about .03) of the sample was hired before 1997. 2In roughly 42 percent of the cases, employers believed that these employees were still on welfare. In most of the remaining cases, employers believed that the employee had been on welfare in the recent past. 3Because of the smaller sample sizes of employers with recently hired welfare recipients (about 800 in all and somewhat fewer for variables with missing values), the standard errors on means of dichotomous variables in this chapter are generally .02–.04, and standard errors on differences are roughly .03–.05. 44 Los Angeles Percentage of most recent recipients hired Chicago Cleveland Milwaukee Los Angeles Percentage who are poor female heads Chicago Cleveland Milwaukee White Black Hispanic 0 10 20 30 40 50 60 70 80 SOURCE: 1990 U.S. Census. NOTES: “Most recent recipient hired” refers only to cases where employers are “definitely sure” or “fairly sure” of their employees’ recipiency status. The racial background characteristics of those in the category are mutually exclusive and sum to approximately one. The racial background characteristics of poor female heads are not exclusive of Hispanic background, and thus sum to a bit over 100. Figure 4.1—Recently Hired Welfare Recipients and Poor Female-Headed Households, by Race Table 4.1 Education and Work Experience of Recently Hired Welfare Recipients High school graduate Any recent work experience Sample size Los Angeles .69 .65 148 Chicago .60 .60 161 Cleveland .65 .70 214 Milwaukee .57 .68 225 NOTE: “Hired welfare recipients” refers only to cases where employers are “definitely sure” or “fairly sure” of their employees’ recipiency status. • Roughly two-thirds of the recipients hired have high school diplomas, and roughly two-thirds have some recent work experience. Somewhat surprisingly, the recipients hired in Los 45 Angeles have the highest rates of high school completion, despite the relative concentration of Hispanics and especially immigrants among them and the relatively low educational attainment among Hispanics nationwide (Mare, 1995). Educational attainment and work experience among recipients are relatively high in Cleveland and somewhat lower in Chicago and Milwaukee. • Although the hiring rates of minorities and those with little education or experience are significant, they appear to fall somewhat below the representation of these groups in the poor female household head or welfare populations of these areas.4 Comparing the hiring data by race in Figure 4.1 to those on poor female-headed households, the percentage of hired recipients who are black is lower than the percentage of poor female heads of households who are black in each metropolitan area in 1990 except for Los Angeles, whereas the percentage of hired recipients who are Hispanic in Los Angeles lags behind the comparable percentage of the poor female-headed household population there as well.5 Furthermore, the percentages of those hired who have no high school diplomas are lower than the comparable nationwide figure (43–44 percent) among welfare recipients at a point in time (Polivka, 1998). Although caseload data by race or education were generally not available to us at the county level in most cases, what data we did find for the ____________ 4Given that recent welfare hires in our data are from 1997 to 1999 and that poor female-headed households reflect data from 1990, if minorities have become much more concentrated in the latter category over time, these comparisons are likely to understate the underrepresentation of minority recipient hires relative to their representation of the poor female-headed household population. 5The employer and household data by race are not exactly comparable, since Hispanic ethnicity is not mutually exclusive with racial identifications of white or black in the Census household data (see the note to Table 2.1). However, since Hispanics in the Midwestern and Western region of the United States are vastly more likely to self-identify as white rather than black, the discrepancy between the two data sources could be roughly corrected by deducting the entire overlap (8 percentage points in Los Angeles and much less in the other metro areas) from the white category. This would not change any of the results described above. 46 largest counties generally support the notion of lower employment relative to welfare population among minority recipients.6 Thus, it appears that minorities and high school dropouts are being hired somewhat less frequently than whites or high school graduates within the welfare population. These results are consistent with evidence that the employment rates of those on welfare have recently been somewhat lower among blacks than whites and among high school dropouts (Danziger et al., 1999), and that welfare rolls have declined more slowly for blacks than others and in heavily urban and minority counties than in other areas (Katz and Carnevale, 1998). The data also indicate some variation across areas in the extent to which the least-skilled welfare recipients are being hired. The relatively high rate of high school completion among the recipients hired in Los Angeles, despite the high concentration of Hispanics and immigrants in that population, suggests that employers there have been somewhat less willing to hire the least-educated than those in Milwaukee and other areas. Among these recent hires in Los Angeles, the low level of immigrants, who often have English language as well as skill deficiencies, strongly supports this contention.7 Whether the lower rate of hiring high school dropouts in Los Angeles than elsewhere is due to differences in economic factors or to the policy environment is discussed more fully below. ____________ 6For instance, data for Los Angeles County indicate that Hispanics accounted for 53 percent of the welfare population there in 1997, and data for Cook (Chicago) and Milwaukee Counties indicate that blacks accounted for 74 and 73 percent, respectively, of welfare recipients there. All of these percentages are higher than the percentages of recent hires accounted for by these groups in our survey, although the population data would need to be adjusted for the racial compositions of the much smaller outlying counties for the comparison to be exact. Data on the racial composition of caseloads in the 20 largest counties are available from the Brookings Institution at www.brookings.edu/ es/urban/welfarecaseloads/2000report.htm. 7Immigrants account for just 14 percent of the welfare recipients hired in Los Angeles in our survey data, even though they accounted for roughly 40 percent of participants in the GAIN program in Los Angeles as of 1997 (Flaming et al., 1999). Some of this discrepancy, although probably not all of it, might be accounted for by lower participation rates in our survey of Spanish-speaking and informal employers in Los Angeles; unfortunately, we have no direct data on this issue. 47 What Kinds of Jobs Do They Fill? Table 4.2 presents evidence on the kinds of jobs filled by recently hired welfare recipients, by broad occupational category and by whether each of a set of tasks is performed daily on these jobs. The tasks include reading or writing reports, memos, or lengthy instructions; doing arithmetic (including making change); using a personal computer; speaking directly with customers either in person or on the telephone; filling out forms; or keeping a close watch on gauges, dials, or instruments. The tasks thus encompass a wide range of basic cognitive (or “hard”) skills as well as some social and verbal (or “soft”) ones (e.g., Moss and Tilly, 1996). The results indicate that welfare recipients most frequently fill clerical and service jobs; they fill fewer sales jobs and many fewer blue- Table 4.2 Occupations Filled and Tasks Performed Daily by Recently Hired Welfare Recipients Occupations Sales Clerical Craft Operator Laborer Service Los Angeles .20 .46 — .08 .01 .24 Chicago .15 .35 .04 .08 .06 .31 Cleveland .22 .20 .03 .09 .05 .40 Milwaukee .16 .32 .06 .13 .02 .31 Tasks Reading/writing Arithmetic Computer Serving customers Gauges/instruments Completing forms .62 .50 .58 .50 .52 .36 .83 .70 .27 .19 .67 .57 .56 .51 .40 .68 .28 .56 .52 .58 .41 .62 .28 .57 Sample size 143 115 150 166 NOTES: Columns under “Occupations” sum to approximately one. “Tasks” are those that must be performed on a daily basis. 48 collar jobs.8 The semi-skilled and unskilled blue-collar categories (e.g., operatives and laborers) generally account for about 15 percent of these jobs, although less than 10 percent in Los Angeles. In contrast, clerical jobs account for more of these hires in Los Angeles than elsewhere. Together, sales and clerical jobs account for 40 to 50 percent of all jobs filled recently by welfare recipients in the Midwestern areas and almost two-thirds in Los Angeles. The relatively high representation of clerical and sales positions among these jobs is consistent with the frequent use of various cognitive and social tasks on these jobs that we find in the bottom half of the table. For instance, dealing with customers daily is required in about 70 percent of the jobs in each metro area. Reading or writing fairly lengthy material or completing forms is required in over half of all jobs in each metro area, as is the use of arithmetic. Computers are used daily in about 40 percent of these jobs on average (and in over half in Los Angeles), although gauges and instruments are read less frequently. Comparing these job characteristics to broader samples of jobs filled recently in the survey that require relatively little education helps put them in context. Compared to the entire sample of recently filled jobs that do not require college degrees, fewer jobs filled by welfare recipients are managerial or blue-collar and somewhat more fall into the service category, although the daily tasks required by all such jobs are quite comparable.9 Compared to all noncollege jobs filled by women, those filled by welfare recipients include fewer managerial and clerical jobs, more service jobs, and jobs with somewhat lower task requirements.10 But when compared to jobs filled by women that do not require a high ____________ 8More specific occupations that are frequently filled include receptionist, cashier, food service, cleaning, and child care jobs. 9Among all recently filled jobs not requiring college degrees, the percentages in managerial and blue-collar occupations were .03 and .28, respectively, whereas service jobs accounted for .23. The percentages requiring customer contact, reading/writing, arithmetic, and computer use were .64, .54, .55, and .42, respectively. 10Of all noncollege jobs filled by women, .04 are managerial, .44 are clerical, and .21 are in services. The percentages requiring the tasks listed in the previous note are .78, .60, .59, and .53, respectively. 49 school degree, the task requirements on the jobs filled by recipients are actually a bit higher.11 Overall, the data indicate that the jobs filled by welfare recipients generally require at least some minimal amount of cognitive, social, or verbal skills. Recipients who have difficulty performing these tasks will have greater difficulty finding or keeping employment, as has been demonstrated elsewhere (e.g., Danziger et al., 1999). Given the high percentage of individuals with very poor cognitive skills among longterm welfare recipients (e.g., Burtless, 1995; Pavetti, 1997), and particularly the concentration of immigrants with English language difficulties and other skill problems in California (Johnson and Tafoya, 1999), the fraction of these women who should have difficulty becoming or remaining employed in jobs like these may not be trivial at any point in time.12 Furthermore, these low skills are likely correlated with other characteristics, such as disabilities and physical or emotional problems, that limit the kind of “job readiness” that virtually all employers seek when they hire workers (e.g., Regenstein and Meyer, 1998). These facts likely help to explain the high rates of nonemployment among former welfare recipients at any moment in time that we noted in Chapter 1. Wages, Benefits, and Hours Worked Although some of these jobs may be inaccessible to the least-skilled welfare recipients, the better occupations and frequent use of tasks might be associated with higher earnings and benefits or more regular hours of work. Figures 4.2, 4.3, and Table 4.3 present some evidence on these measures by metropolitan area. In particular, the figures present means and medians of starting hourly wages and hours worked, and the table presents frequency distributions on these measures as well as the ____________ 11The percentages in the tasks listed in the footnote above are now .59, .30, .40, and .23, respectively. The percentage of these jobs that are clerical falls to just .12 of the sample. 12 For instance, Danziger et al. (1999) report that about 20 percent of the welfare recipients whom they surveyed had used fewer than four out of nine job skills in previous jobs that they measured, and that the fraction of this group working at least 20 hours a week was much lower than among those with four or more skills (.34 versus .63). 50 Hourly wage, $ 8.5 8.0 7.5 7.0 6.5 6.0 5.5 5.0 Los Angeles Chicago Mean Median Cleveland Milwaukee Figure 4.2—Starting Hourly Wages Paid to Recently Hired Welfare Recipients 45 Mean Median 40 35 30 25 20 Los Angeles Chicago Cleveland Milwaukee Figure 4.3—Hours Worked per Week by Recently Hired Welfare Recipients Hours worked per week 51 Table 4.3 Starting Wages, Hours Worked, and Health Insurance Coverage for Recently Hired Welfare Recipients Starting hourly wages Mean Distribution $5.50 $6.00 $6.50 $7.00 $7.50 $8.00 Los Angeles Chicago Cleveland Milwaukee $7.83 $7.53 $6.94 $7.39 .01 .17 .32 .28 .43 .37 .54 .49 .57 .60 .63 .68 .16 .36 .49 .62 .75 .82 .11 .23 .38 .46 .58 .72 Hours per week Mean Distribution 15 20 25 30 35 40 Employer contributes to health insurance 34.4 34.8 .03 .08 .14 .16 .20 .19 .32 .24 .39 .32 .96 .95 .59 .72 34.0 .03 .14 .20 .35 .44 .95 .65 35.2 .03 .11 .21 .28 .34 .93 .72 Sample size 149 122 155 175 percentages of jobs in which employers contribute to the provision of health insurance for the workers or their families. The results indicate that: • Median wages are generally in the vicinity of $7.00 per hour, and mean wages are somewhat higher; • The median number of hours worked per week is 40; and • Roughly two-thirds of these jobs provide health insurance. The average starting wages earned by recipients and percentage working full-time are somewhat higher than has been reported elsewhere (e.g., Polivka, 1998; Holzer, 1999b; Loprest, 1999), although these data are consistent with other evidence on Milwaukee (Nightingale and 52 Mikelson, 2000).13 Wages and employer contributions to health insurance are lower on these jobs than on all noncollege jobs filled by women recently but somewhat higher in jobs that do not require high school diplomas.14 The frequency with which 35 to 40 hours are worked per week or where health insurance is provided suggests that most of these jobs are “regular” and provide stable hours.15 The use of the Earned Income Tax Credit and a variety of publicly provided subsidies for health care or child care to supplement these earnings would enhance their adequacy and attractiveness even more (Acs et al., 1999; Danziger et al., 2000), as would consistency of such employment over a significant period of time. Indeed, employment at the starting wages and hours per week that we have observed for the average newly hired welfare recipient would generate annual earnings of $12,000–$14,000 per year, or $16,000– $18,000 after including the Earned Income Tax Credit, if such employment could be sustained for the entire year. The issue of job retention and turnover is therefore critical, and we return to it below.16 On the other hand, the data also indicate significant fractions of jobs that pay lower wages and provide fewer hours of work or no health insurance. In fact, roughly one of every three jobs recently filled by a welfare recipient pays $6.00 per hour or less; roughly one-third provide 30 or fewer hours of work per week; and roughly one-third provide no health insurance. In fact, over half (54 percent) of the jobs filled by welfare recipients fall into one or more of these categories; and jobs that ____________ 13Other studies indicate median starting wages of about $6.00–6.50 for welfare recipients, whereas Polivka finds 45 to 47 percent of current or former welfare recipients working fewer than 35 hours per week. 14Mean and median wages on all noncollege jobs filled by women are $8.73 and $8.00, respectively, whereas those on jobs not requiring high school diplomas and filled by women are $6.87 and $6.00, respectively. The percentages with employer contributions to health insurance are .73 and .59, respectively, in the two categories. 15Employers may “provide” health insurance but still require a substantial contribution from employees. In some such cases, the employees might choose to forgo the insurance because of its high cost to them. 16The sufficiency of any given level of earnings will also vary across metropolitan areas with the cost of living, implying potentially greater material hardship for working recipients in Los Angeles than elsewhere. 53 are weaker along any one dimension are likely to be weaker along the others as well.17 Thus, overall earnings and benefits in a large fraction of these cases may not be high enough on their own to enable families to achieve selfsufficiency, as has been stressed by Edin and Lein (1997). Worker, Employer, and Job Characteristics To what extent do the demographics of the workers and their pay vary by establishment characteristics and occupational categories? Table 4.4 provides data on worker demographics, wages, and benefits by Table 4.4 Characteristics and Wages of Recently Hired Welfare Recipients, by Employer Size and Industry Race White Black Hispanic Asian 1–19 Size 20–99 100+ Industry Manufac- Retail turing Trade Service .35 .35 .24 .31 .39 .23 .41 .47 .53 .42 .43 .55 .20 .14 .17 .21 .14 .16 .01 .01 .01 .02 — .01 High school graduate .61 .61 .64 .54 .59 .64 Any recent work experience .69 .67 .64 .62 .70 .68 Starting hourly wages Mean Median $7.37 $7.10 $7.66 $7.79 $7.00 $6.75 $7.25 $7.50 $5.87 $7.67 $5.75 $7.47 Employer contributes to health insurance .55 .74 .63 .85 Sample size 80 205 295 102 NOTE: Columns under “Race” sum approximately to one. .50 .66 148 260 ____________ 17For instance, among jobs that provide 30 or fewer hours per week, median wages are just $6.00 per hour (as opposed to $7.50 for those with more hours) and 66 percent of jobs provide no health insurance. 54 employer size or industry; Table 4.5 presents these by the location of the establishment; and Table 4.6 provides these data by occupation. The data in Tables 4.4–4.6 indicate that: • Black welfare recipients are more likely to be hired in large establishments, in the service sector, and in establishments that are more easily accessible to minority populations (i.e., those located in central cities or near transit stops/low-income populations), and • Wages and the provision of health insurance are highest in large establishments and lowest in the retail trade sector. Both findings are consistent with other research evidence on where blacks are most likely to be hired (e.g., Stoll et al., 2000; Holzer, 1996, Table 4.5 Characteristics and Wages of Recently Hired Welfare Recipients, by Employer Location Race White Black Hispanic Asian Central City Suburb Near Public Transit Not Near Not Near Near Population Population Public on Public on Public Transit Assistance Assistance .22 .40 .22 .39 .61 .40 .58 .45 .16 .19 .18 .16 .01 .00 .01 .01 .21 .62 .15 .01 .36 .43 .19 .01 High school graduate .59 .68 .61 .64 .57 .67 Any recent work experience .66 .66 .68 .64 .67 .66 Starting hourly wages Mean $7.32 Median $7.00 $7.51 $7.00 $7.33 $7.00 $7.41 $7.00 $7.31 $7.00 Employer contributes to health insurance .67 .68 .67 .68 .67 Sample size 311 264 388 300 317 NOTE: Columns under “Race” sum approximately to one. $7.47 $7.00 .67 390 55 Table 4.6 Characteristics of Recently Hired Welfare Recipients, by Type of Jobs Filled Race White Black Hispanic Asian Clerical Sales Craft Operator Laborer Service .28 .30 .52 .30 .45 .30 .51 .53 .42 .44 .40 .55 .19 .17 .05 .26 .15 .15 .02 — — — — .01 High school graduate .77 .63 .65 .52 .40 .56 Any recent work experience .68 .67 .63 .56 .39 .67 Starting hourly wages Mean Median $8.35 $6.74 $7.97 $7.87 $8.00 $6.00 $8.00 $7.25 $6.58 $6.65 $6.25 $6.50 Employer contributes to health insurance .71 .54 .75 .87 Sample size 178 101 20 55 NOTE: Columns under “Race” sum approximately to one. .86 .60 21 175 1998) and on wage and benefit levels by establishment size and industry (e.g., Brown et al., 1990; Krueger and Summers, 1987). They suggest that there is significant variation across establishments in their accessibility not only to the population receiving public assistance but also to minorities versus whites within that population. Spatial factors, size, and industry all seem to play key roles in determining the availability of jobs to minority welfare recipients and the quality of these jobs. The data in Table 4.5 on welfare recipient demographics and job characteristics also suggest the following: • Blacks are hired most frequently into service jobs and Hispanics are relatively more likely to obtain semi-skilled blue-collar (i.e., operative) jobs; and • Wages are highest in clerical and craft jobs, although health insurance coverage is highest in the blue-collar jobs. 56 Thus, minority welfare recipients have mixed access to the betterpaying jobs, with some limitations because of the skill requirements of these jobs. On the other hand, the provision of health insurance in the blue-collar jobs to which Hispanic welfare recipients seem to have greater access than black recipients (even within metropolitan areas) suggests once again that access to good jobs varies across demographic groups in ways that are not fully explained by their relative skill attainments. Overall, cognitive skill requirements seem to pose greater barriers to Hispanic employment, whereas blacks are particularly underrepresented in the better blue-collar categories. Finally, a lack of a high school diploma and work experience prevents at least some welfare recipients from attaining higher-skilled jobs. Turnover and Job Performance Once welfare recipients are hired into jobs, how well do they succeed? Many authors have expressed concerns about issues of job turnover and performance among these recipients (e.g., Burtless, 1995; McMurrer et al., 1997). Indeed, high turnover has been observed among workers with low educational attainment, weak cognitive skills, and low wages—especially among women (Holzer and Lalonde, 2000). High turnover among welfare recipients could potentially lead to frequent and perhaps lengthy periods of unemployment, which would reduce their annual earnings substantially and weaken their ability to enjoy the wage growth that comes with work experience over time.18 In the next few figures and tables, we examine evidence on turnover and job performance among hired welfare recipients in Los Angeles and elsewhere. Figures 4.4 and 4.5 present data on the percentages of recently hired welfare recipients who are still with their employers and on their employment durations (as measured in months). Table 4.7 ____________ 18Taber and Gladden (2000) provide clear evidence that stagnant wage growth among less-skilled workers often results from their limited work experience over time; for those who manage to gain work experience, wage growth is comparable to the rates enjoyed by more-skilled workers. This limited work experience also reflects lengthy durations of nonemployment as well as the high frequencies of nonemployment that result from worker turnover (e.g., Clark and Summers, 1982; Ballen and Freeman, 1986). 57 Percentage 90 85 80 75 70 65 60 55 50 Los Angeles Chicago Cleveland Milwaukee Figure 4.4—Percentage of Recently Hired Welfare Recipients Still with the Firm 12 Mean 10 Median 8 6 4 2 0 Los Angeles Chicago Cleveland Milwaukee Months of employment Figure 4.5—Months of Employment of Recently Hired Welfare Recipients 58 Table 4.7 Employment Duration and Turnover of Recently Hired Welfare Recipients Los Angeles Still with the firm .80 Employment duration (months) All recently hired recipients Median 8 Distribution 3 .17 4 .24 5 .33 6 .38 7 .49 8 .52 Chicago .77 6 .18 .30 .37 .50 .57 .62 Cleveland .76 7 .22 .31 .39 .44 .50 .62 Milwaukee .66 8 .26 .28 .34 .41 .48 .52 Those still with the firm (incomplete spells) Median Distribution 4 5 6 7 8 86 .24 .30 .33 .37 .38 .50 .49 .57 .52 .62 7 .31 .39 .44 .50 .62 8 .28 .34 .41 .48 .52 Those no longer with the firm (completed spells) Median 76 Distribution 3 .32 .25 4 .42 .38 5 .42 .44 6 .42 .50 5 .37 .37 .53 .53 4 .45 .58 .70 .78 Of those who left the firm, reason Quit .70 .54 Discharge .20 .36 Laid off/job disappeared .10 .08 .65 .28 .06 .61 .33 .06 Sample size 167 177 224 241 NOTE: Columns under “Of those who left the firm, reason” sum to approximately one. 59 presents additional detail on these measures, and on the reason for separation among those who have left their jobs. Figure 4.6 also presents data on their performance relative to other workers in the same positions, as perceived by the employers, who were asked how the welfare recipients compare to the “typical one that you hire into this position.” Employers were also asked whether they had experienced each of a set of problems with the welfare recipients they had hired—such as absenteeism (and its particular causes), attitudes toward work, basic and job-related skills, substance abuse, and relations with coworkers (Figure 4.7). The causes of absenteeism problems are presented as well (Table 4.8). All of these results appear separately by metropolitan area. The results show the following: • On average, roughly three-fourths of the most recently hired welfare recipients are still with the firm. The retention rates are actually highest in Los Angeles (at 80 percent) and lowest in Milwaukee (at 66 percent). • Mean durations of employment to date average about eight months and median durations about seven months. Similar differences across metropolitan areas can be observed in these measures as in those described above. • High turnover and short employment durations can be observed among some subset of the recipients hired. For instance, a fourth to a third of recipients have had employment durations of four months or less.19 ____________ 19Incomplete spells of employment are truncated by the interview, which implies a downward bias in their average length. But they are, on average, longer than completed spells, which implies an upward bias in average length if one looks only at those spells. An unbiased estimate of overall spell length can be obtained by analyzing a set of completed spells that all begin within some time period (e.g., Bane and Ellwood, 1986). However, all spells begun in any time period in this study still contain some incomplete spells. The hazard functions estimated and described below incorporate the information from both kinds of spells while analyzing the determinants of employment duration. The data on incomplete spells (i.e., those in progress at the time of the survey) differ somewhat (and suggest longer periods of employment) than those on completed spells. But both of these measures need to be considered when analyzing turnover rates and employment durations. 60 Percentage 100 80 60 40 20 0 Los Angeles Chicago Cleveland Milwaukee Better Similar Worse Figure 4.6—Performance of Recently Hired Welfare Recipients Relative to Others in the Same Position • Welfare recipients are most frequently considered to be as good as or even better than other employees in similar positions. The percentage who are considered better than the typical employee in this position ranges from 40 percent in Los Angeles to 25 percent in Milwaukee. Roughly half are considered comparable in each metro area, and just 10 to 20 percent are considered worse. • Problems with absenteeism are experienced relatively frequently and are often related to child-care and transportation problems. The percentage of employers reporting problems with absenteeism ranges from about a fourth in Los Angeles to over half in Milwaukee. Child care seems to be the single greatest cause of absenteeism, being reported as a problem by over half of employers in each metro area. But transportation problems are frequently mentioned as well, especially in Los Angeles (by 58 percent of respondents), as are health problems (by about 30 percent on average). 61 Absenteeism Attitude toward work Basic skills Job skills Coworker relations Los Angeles Chicago Cleveland Milwaukee 0 5 10 15 20 25 30 35 40 45 50 55 60 Figure 4.7—Percentage of Employers Observing Problems with Recently Hired Welfare Recipients Table 4.8 Absenteeism Problems of Recently Hired Welfare Recipients Absenteeism Los Angeles .24 Chicago Cleveland Milwaukee .35 .39 .52 Due to Physical health Mental health Child care Transportation Domestic violence .26 .28 .07 .03 .63 .62 .58 .41 .05 .05 .35 .05 .58 .39 .07 .37 .06 .67 .39 .12 Sample size 162 176 223 239 NOTE: Columns do not sum to one, since multiple absenteeism problems could be listed by the employer. • Problems with attitudes toward work and basic skills appear somewhat less frequently. The percentages of employers reporting attitude problems range from 12 percent (in Los Angeles) to 23 percent (in Milwaukee), whereas problems with 62 basic skills or coworkers appear in about 10 to 20 percent of cases. Problems related to substance abuse or job-related skills appear even less frequently in these data. Overall, these data present a fairly positive picture of the performance of welfare recipients who are hired. For instance, average retention rates of almost 75 percent over this time period imply job turnover rates that are not dramatically higher than those of the overall workforce and certainly are lower than those of younger and less-skilled workers on average.20 Job performance among these workers is rated fairly positively relative to other workers, as has been noted elsewhere (e.g., Welfare-to-Work Partnership, 1999). Such data need to be interpreted very carefully. Comparisons with the employees most frequently hired into low-skill and low-wage jobs might well lead us to overstate the success of welfare recipients, as those employees often have their own turnover and performance difficulties (especially if they are young workers). What matters for welfare recipients is not how they compare with other unskilled workers but whether their retention rates and performance are good enough to lead to sufficient earnings levels and growth over time. At least for some of the workers considered here, retention and performance do not yet appear high enough to allay these concerns. Significant percentages of these recipients are having trouble keeping jobs for more than three to four months, and problems with absenteeism are quite frequent. On the other hand, not all job turnover is negative. Voluntary turnover or turnover from one job to another (as opposed to the transition from a job to nonemployment) is often associated with wage growth, particularly among the young (Taber and Gladden, 2000; Holzer and Lalonde, 2000). Because the majority of turnover here appears to be voluntary, the extent to which its effects are positive or negative will depend on whether and how quickly these individuals move ____________ 20Given that roughly eight months elapsed, on average, between the time of hire and the interview date, an average turnover rate of .26 in this period implies a rate of about .39. Anderson and Meyer (1994) report that 40 percent of jobs experience turnover in a year, and with multiple incidents of turnover for some jobs, a turnover rate of over 40 percent is implied. But much of the turnover appears to be heavily concentrated among young and inexperienced workers (Holzer and Lalonde, 2000). 63 to other jobs and also on the extent to which even voluntary turnover is associated with child care or related concerns. The differences in results between Los Angeles and the Midwestern metro areas, particularly Milwaukee, are also quite striking. Along virtually every dimension, outcomes for welfare recipients are most positive in Los Angeles and least positive in Milwaukee. The relatively high wages of workers in Los Angeles is consistent with the somewhat higher skills demanded there as well as the higher cost of living, as we noted above. But they are also consistent with the relatively low rate of hires for welfare recipients there compared to the other areas and the high rate observed in Milwaukee. In other words, employers in Los Angeles have not had to dig as deeply into the pool of welfare recipients there (in terms of worker quality) as in Milwaukee and the other areas because of a looser labor market and less-aggressive implementation of welfare reform. The implied tradeoff between the overall quantity and quality of welfare recipients hired also suggests that employers in Los Angeles may face more difficulties from these workers in the coming months and years, as reforms are implemented more forcefully, and driving a larger fraction of welfare recipients (especially less-skilled immigrants) into the workforce. Of course, questions might well be raised about the reliability of the results presented here, especially since the sample sizes of the duration measures in the tables and figures indicate many missing values on these variables; also, the performance measures are subjective and thus likely to be measured with considerable error. One way to ascertain the usefulness of these data is to see whether the different measures of performance are consistent with one another, which would imply that they contain somewhat more real information relative to random error. Table 4.9 presents evidence on performance measures by retention status with the firm and by reason for departure among those who have left. We also present evidence on median durations of employment by retention status and relative performance category. The results of Table 4.9 show that perceived performance of welfare recipients on the job is strongly related to retention status. Specifically, those who are still with the firm are rated much more positively than those who have left. In fact, over 90 percent of those still with the 64 Table 4.9 Performance and Employment Duration of Recently Hired Recipients, by Retention Status Performance relative to average Much better A little better Similar A little worse Much worse Median employment duration All recently hired recipients Those with Better performance Similar performance Worse performance Still with Not with Firm Firm All Quits Discharges .19 .07 .07 .21 .09 .11 .52 .44 .48 .06 .23 .18 .01 .17 .15 .05 .08 .29 .34 .23 7 54 8 99 7 44 8 43 5 9 4 5 Sample size 593 203 122 60 NOTE: Columns under “Performance relative to average” sum to approximately one. employer are rated similar to or better than typical employees in these jobs, whereas almost 40 percent of those who have left are rated worse. Furthermore, those who have been discharged are rated considerably worse than those who have quit, although even the latter are viewed considerably more negatively than those who stayed.21 Also, among those with completed spells, those with relatively better performance have considerably longer median durations of employment than those with similar or worse performance. Indeed, median durations of employment among those who have left the firm after performance that was similar to or worse than average is just four months. The consistency of performance measures with one another thus gives us some reason to believe that these data are capturing real dimensions of performance, at least to some extent. ____________ 21Of course, the relatively negative views of “quitters” may be reactions to their quits rather than unbiased assessments of their performance while working. 65 There may be other reasons as well to be concerned about measurement error or biases in one direction or another. For instance, the relatively high average retention rates (as well as the percentages of welfare recipients working full-time or receiving health insurance) might reflect the fact that this survey does not capture information on jobs for welfare recipients in the irregular or informal sector of the economy, thus leading to upward biases in all of these measures (or downward biases in turnover). Higher response rates to the survey among larger employers and older employers could have the same effect. Also, the large numbers of missing values on durations of employment could systematically lead to upward bias in retention or performance measures, if the less successful recipients on the job are less well-remembered by the employer (or if they chose to reveal their status as welfare recipients to employers less frequently). Of course, this latter bias could just as easily go in the opposite direction, if shorter-term employees are more recent and therefore more easily remembered. Furthermore, focusing on the last recipient hired might also generate downward biases in retention (or upward biases in turnover), since the most recent hire is more likely to be in a high-turnover job category than the average employee at the establishment. Thus, potential biases certainly exist here, but their net effects on the direction of our estimates as well as their magnitudes are unclear. Variation in Performance by Demographics and Time Period To what extent do observed measures of job performance differ according to the demographic characteristics of the welfare recipients hired? Tables 4.10 and 4.11 present data on job retention and performance measures, respectively, by race, education, and experience of the last welfare recipient hired. The results show at least some evidence of variation in performance by demographic group. For instance, durations of employment among whites are longer than those among minorities, even though blacks and Hispanics are more likely to still be working with their employers. Whites are somewhat more likely to be considered better than the typical 66 Table 4.10 Turnover of Recently Hired Welfare Recipients, by Characteristics of Recipients Still with the firm White .65 Race Black Hispanic .72 .80 High School Graduate Yes No .73 .75 Any Recent Work Experience Yes No .75 .73 Duration of employment (months) Mean Median 9.56 8 8.16 7 8.59 8.41 8.42 8.15 8.26 7 77 7 7 Sample size 173 296 97 442 264 496 252 Table 4.11 Performance of Recently Hired Welfare Recipients, by Characteristics of Recipients Rating of employee relative to others in this position Much better A little better Similar A little worse Much worse White Race Black Hispanic High School Graduate Yes No Any Recent Work Experience Yes No .16 .15 .21 .18 .44 .48 .10 .13 .08 .05 .18 .18 .16 .19 .12 .16 .16 .19 .18 .16 .58 .49 .51 .47 .55 .05 .09 .10 .10 .12 .02 .07 .03 .05 .04 Problems with Absenteeism Attitude Basic skills Job skills Substance abuse Coworker relationships .45 .21 .09 .07 .04 .18 .44 .20 .12 .10 .02 .16 .25 .44 .36 .40 .41 .10 .20 .16 .18 .18 .20 .16 .09 .10 .14 .05 .09 .09 .09 .10 .03 .03 .02 .03 .01 .07 .18 .13 .15 .15 Sample size 173 296 97 442 264 496 252 NOTES: Columns under “Rating of employee” sum to approximately one. Columns under “Problems with” do not sum to one since multiple problems could be listed by the employer. 67 employee than are minorities, although they are also more likely to be rated as much worse. Hispanics are the least likely group to show significant problems related to poor relative performance, absenteeism, and relationships with coworkers, although they suffer from greater problems with basic skills than do other groups. High school graduates are a bit more likely than dropouts to be rated positively and are less likely to have problems with absenteeism, attitudes, and basic skills. Recent work experience has little effect on observed outcomes, except that those without experience have more problems with basic skills. Although we see differences in performance across demographic groups, these differences are often modest. This pattern could reflect measurement error in the demographic variables (which will bias results toward a finding of no differences across groups) or selection effects at the hiring stage that tend to equalize the attributes of those who are actually hired across demographic groups.22 The higher retention rates but shorter employment durations of minorities than whites also suggest that, on average, the former groups were hired more recently than the latter. This might have occurred if, for example, the relatively more skilled or white recipients left the rolls or sought employment most quickly, whereas minorities or those with greater difficulties did so somewhat less quickly. Alternatively, employers might have chosen to hire white recipients first and then minority or lessskilled recipients only as labor markets grew increasingly tight over time and their hiring needs were increasingly unmet (as we saw in the chapters above). We analyze this last issue somewhat more systematically in Figures 4.8 through 4.10 and Table 4.12, which present race, wages, hours of work, and retention/performance measures by the time period in which the individual worker was hired. The figures compare those hired in 1997 with those hired in 1998–1999. In Table 4.12, we cut the data into three time periods: those hired between January and September of 1997, those hired between October 1997 and June 1998, and those hired ____________ 22In other words, the hiring screen eliminates more potential minority or high school dropout candidates with performance problems, so that those who are actually hired are more similar to their white counterparts than those in the general population. 68 Percentage 60 Black 50 Hispanic 40 30 20 10 0 1997 1998–1999 Figure 4.8—Race of Recently Hired Welfare Recipients, by Year Hired 8.0 Mean Median 7.5 7.0 6.5 6.0 5.5 5.0 1997 1998–1999 Figure 4.9—Starting Hourly Wages Paid to Recently Hired Welfare Recipients, by Year Hired Starting hourly wage, $ 69 100 80 Percentage 60 40 20 0 1997 1998–1999 Better Similar Worse Figure 4.10—Performance of Recently Hired Welfare Recipients Relative to Others in the Same Position, by Year Hired since July of 1998. The benefit of the latter cut is that it enables us to isolate the effects of other factors, such as the most recent increase in the federal minimum wage in October 1997, and still have two other periods of relatively similar length to compare to one another. Using either cut of the time period, we find that those hired relatively later in this time period are more likely to be minorities and to have lower median starting wages, shorter durations of employment, and lower performance ratings than those hired relatively earlier. Despite the increase in the minimum wage after October 1997, starting wages still deteriorated for those hired more recently, implying a potentially larger decline in the absence of that mandated wage increase. Also, the more recently hired workers are more likely to still remain with the employer, and their employment durations up to the survey date are shorter. This suggests that more welfare recipients have been hired relatively recently, closer to the interview dates of these firms, 70 Table 4.12 Worker Characteristics and Job Performance of Recently Hired Recipients, by Month and Year Hired Race Black Hispanic Month and Year Hired 1/97–9/97 10/97–6/98 7/98–6/99 .41 .57 .49 .11 .17 .18 Starting hourly wages Mean Median 7.56 7.47 7.36 7.50 7.00 7.00 Hours Median Below 30 40 40 40 .27 .27 .31 Still with the firm .58 .70 .78 Median employment durations All recently hired recipients Those still with the firm Those no longer with the firm 19 22 14 11 13 5 5 5 3 Performance relative to average All recently hired recipients Better .45 .36 .32 Similar .33 .43 .52 Worse .19 .20 .13 Those still with the firm Better .54 .46 .37 Similar .38 .43 .54 Worse .05 .11 .06 Those no longer with the firm Better .32 .14 .13 Similar .26 .45 .45 Worse .39 .40 .38 Sample size 62 132 399 NOTE: Columns under “Performance relative to average” sum to approximately one. which also implies an increasing rate at which recipients are hired over time (perhaps because more of them enter the labor force over time, or labor market tightness grows among employers leading them to hire 71 recipients at a faster rate).23 Since performance tends to be positively correlated with retention, a higher retention rate over time might suggest higher performance as well. But, within the samples of those who remained employed as well as those who did not, the decline in performance in more recent time periods is quite striking.24 The evidence thus suggests that the quality of welfare recipients hired has been falling over time, consistent with the notion that the easiest to hire were employed fairly quickly, whereas those with more problems made up a growing part of this pool of workers over time. For this growing group, outcomes will not be as positive as those observed for the overall sample of recently hired welfare recipients. Before concluding the chapter, we consider some evidence from multiple regression analysis on these topics.25 In Appendix Table A.2, we present results of regression equations in which we explore the determinants of three primary outcome measures presented above: wages, relative performance, and duration of employment.26 Three specifications are presented for each of the three outcome measures. In the first, we begin only with independent variables for the metropolitan areas (relative to the omitted category of Milwaukee); the second specification then adds worker demographics, establishment size ____________ 23In these data, the date of interview is strongly correlated with the date of hiring, so that the most recently hired workers are frequently in the most recently interviewed establishments. But the average time elapsed between date of interview and date of most recent hire is also declining with more recent interviews, implying a greater frequency of new hires on welfare over time. 24An alternative interpretation of these results is that the most recently hired welfare recipients entered the labor market at the same time as those hired earlier but experienced more job turnover and therefore have been hired more recently by their current employers. We have no way of distinguishing between these interpretations with the current data. 25Readers who are less familiar with or interested in these more technical research findings can skip to the Conclusion of this chapter. 26The wage equation uses the log of real hourly wages as the dependent variable, where wages have been deflated by the gross domestic product deflator for Personal Consumption Expenditures. For relative performance, the dependent variable is simply a dichotomous variable equaling one if the individual’s performance is deemed better than the typical employee in that position and zero otherwise. Finally, the duration equation is an estimated hazard function, using the Weibull functional form, which accounts for the truncation of duration values among incomplete spells. 72 and location, occupation, and task performance as independent variables; the third also includes industry dummies. Thus, we can ascertain the extent to which these observable characteristics of employers and their workers can help to account for average differences between Los Angeles and the other metropolitan areas in these outcomes. These regression equations indicate, as we noted above, that employers in Los Angeles pay higher real wages and have workers with higher average performance and longer durations of employment than do those in other metropolitan areas, particularly Milwaukee. Although much of the wage differential favoring Los Angeles can be accounted for by the characteristics of employers and workers, the Los Angeles advantage for the other dependent variables cannot be easily accounted for and might be attributed to the policy differences on implementation of welfare reform that we noted above. The regressions also confirm the effects of many groups of worker and establishment/job characteristics on observed outcomes. Thus, being a high school dropout and working in smaller establishments, in the laborer/service occupations, or in jobs with no computer use are associated with lower wages among welfare recipients.27 Reported relative performance is lower among minorities, higher for those with previous experience, higher in smaller establishments, and lower in establishments located relatively farther from populations on public assistance. Also, employment durations are lower for minorities, lower in smaller establishments, lower for those establishments located farther from populations on public assistance, and lower in the sales, blue-collar, and service occupations than in clerical or white-collar jobs. The results thus confirm many of our findings noted above and indicate that a wide range of personal, employer, and job characteristics influence the success of welfare recipients on the job. Overall, significant variation in these employment outcomes has been observed among welfare recipients, and this variation can be partly, but not fully, explained by these characteristics. ____________ 27The effect of a high school diploma is roughly .10 before including occupational and task dummies. The effect of a high school diploma on turnover is also somewhat reduced by including these dummies. 73 Conclusion The data in this chapter indicate some positive outcomes for welfare recipients who have been recently hired into the workplace. For example, significant numbers of minorities are being hired; many recipients wind up with fairly good jobs, and most work full-time; average wages (at least in these areas) are somewhat higher than previously reported, with a median of $7.00 per hour or more; and turnover as well as relative performance are often as good as or better than that achieved by other groups of low-wage workers. But certain findings also suggest some areas of concern. In particular (1) minorities and high school dropouts are not hired as frequently as might be implied by their representation in the welfare population, suggesting somewhat greater problems for these groups at the hiring stage, (2) skill requirements are not trivial in many jobs filled, indicating potential problems with hiring or retention for the very unskilled, (3) low wages and hours of work, high turnover, and poor performance plague significant percentages of those hired, and (4) absenteeism, especially related to child care and transportation, is a pervasive problem. Furthermore, the quality of recipients hired and their performance on the job have been deteriorating, suggesting that the relatively more employable recipients were hired more quickly and those with somewhat greater problems are now entering the workforce or being hired in greater numbers. Virtually all outcomes among hired recipients in Los Angeles are more positive than those observed in other metropolitan areas, especially Milwaukee. Although this is certainly good news at some level, it likely reflects the lower hiring rates of welfare recipients in Los Angeles that we noted in Chapter 3—especially those welfare recipients whose are less “work-ready”—and the slower implementation of welfare reform there. In contrast, the aggressive strategy of placing welfare recipients into jobs, which is central to the Wisconsin Works program, has resulted in somewhat greater employer difficulties with more marginal workers there. Eventually, as implementation of welfare reform proceeds, Los Angeles employers will need to grapple with particular skill deficiencies 74 plaguing welfare recipients (especially those who are immigrants), and other employment barriers recipients may face. In the next chapter, we discuss at greater length some of the implications of these findings for public policy, but a few of these implications are briefly noted here. In particular: • Transportation and child care problems remain barriers to successful job retention, if not to job placement, in large numbers of cases involving welfare recipients in the workplace. These issues, and performance/retention issues more broadly, deserve greater attention from state and local workforce boards and agencies in all of these locations. • The low wages and benefits earned by significant fractions of welfare recipients imply a continuing need to enhance their returns from work, perhaps through the expansion of the Earned Income Tax Credit and subsidies or credits for the provision of health insurance to workers and their families, at the federal or state levels. • Finally, the lower employment rates observed among recipients who are minorities or high school dropouts, as well as the nontrivial cognitive and social skills required on most of the jobs filled, suggest a need for greater skill development among the most disadvantaged recipients or perhaps more use of public service jobs to enhance their transitions to private sector employment. Both options are considered more extensively in the next chapter. 75 5. Conclusion The recent reforms of the welfare system in the United States, beginning with state-level experiments approved under federal vouchers in the early 1990s and culminating in the 1996 passage of the Personal Responsibility and Work Opportunity Reconciliation Act, remain controversial. Since then, welfare rolls have declined dramatically and employment rates among recipients (and single females more broadly) have risen substantially, partly because of welfare reform and partly because of other factors (such as economic conditions and other policyinduced changes in net earnings and benefits for the working poor). But many questions remain unanswered. For instance, are employers willing to hire even the most unskilled welfare recipients? Do minorities on welfare have fairly easy access to available jobs, including those located in suburban areas, and are employers as willing to hire them as white recipients? What role does the current strength of the economy play in those decisions, and what will happen during an economic downturn? What is the ability of recipients to retain jobs once they are hired, and what determines performance and retention rates? Finally, are the levels of pay, benefits, and hours of work sufficient for families to achieve independence? To address many of these questions, we have analyzed data from a new survey of employers, administered to 750 establishments in each of four metropolitan areas: Los Angeles, Chicago, Cleveland, and Milwaukee. The survey was administered by phone between July 1998 and June 1999 and focused on employers’ attitudes toward the hiring of welfare recipients and any experiences they have had to date with this population. In particular, we asked about the prospective willingness of employers to hire very unskilled recipients, the numbers actually hired in the past year or two, the characteristics and performance of the last such recipient hired, and any contact the employer might have had with a local welfare-to-work agency. 77 Summary of Findings Our most important findings can be summarized as follows: • The current demand for welfare recipients among employers is quite strong, at least in the aggregate. Employers express much greater willingness to hire welfare recipients than members of other stigmatized groups, such as those with criminal records. Hiring rates for welfare recipients in the aggregate seem sufficient to absorb the numbers of recipients seeking employment in the current period and the near future. • The rate at which recipients are hired is quite sensitive to the job vacancy rate across establishments, at a time when the aggregate job vacancy rate is quite high and labor markets are extremely tight. This suggests that a good deal of employer demand for recipients will diminish as well when these vacancy rates decline during an economic downturn. • Although many establishments are prospectively open to hiring welfare recipients, actual hiring is much stronger at establishments that are easily accessible to recipients, such as those near public transit stops or near their neighborhoods. Access to jobs for welfare recipients also seems related to minority ownership and to employer contact with local welfareto-work agencies, regardless of who initiates the contact. • Although many of the welfare recipients being hired are minorities, they appear to be hired somewhat less frequently (relative to their representation in the welfare population) than are whites, perhaps because of their lower access to jobs. High school dropouts are also hired less frequently than high school graduates, consistent with the fact that the cognitive and social tasks demanded on the jobs filled by welfare recipients are generally not trivial. • The median job filled by a recently hired welfare recipient in these metropolitan areas pays about $7.00 per hour and provides 40 hours of work a week. Employers contribute to the provision of health insurance in approximately two-thirds of these jobs. 78 But a significant portion of these jobs pay low wages, provide few hours, or provide no health insurance. • On average, turnover rates for recently hired recipients are near the national average, and most recipients are rated as being as good as or better than other workers in the jobs that they fill. But high turnover and weak job performance are problems in a significant fraction (i.e., a fourth to a third) of cases. Absenteeism is particularly pervasive, often linked to child care and transportation issues. • The average quality of welfare recipients hired also appears to be deteriorating somewhat over time, which implies that the part of the welfare caseload with more significant employment-related problems is entering the workforce now in greater numbers. • Finally, there are some noteworthy differences between hiring patterns among employers in Los Angeles and those in the Midwest, particularly Milwaukee. Those in Los Angeles hire fewer welfare recipients, but the average quality of those hired seems higher there. Thus, the better performance of recipients hired in Los Angeles seems to come at the expense of relatively more disadvantaged recipients who do not get hired there. Differences in employment outcomes across these metropolitan areas partly reflect differences in job vacancy rates or geographic accessibility of establishments, but other differences are not fully accounted for and may well reflect policy differences as well. All in all, these findings indicate many positive employment outcomes for welfare recipients in the workplace—such as strong demand for their work, average wages well above the minimum, significant health care provision, and job performance and retention that is above the average of many other less-skilled groups of workers. However, some groups (e.g., blacks and high school dropouts) still seem to face difficulties becoming employed, which likely contributes to the fact that over 40 percent of former welfare recipients are not employed at any point in time. Furthermore, at least some of those employed (i.e., the lowest one-fourth or one-third along most dimensions) experience a variety of serious difficulties on the job. Results over the coming months 79 or years might also deteriorate, as relatively less-employable recipients enter the workforce or when an economic downturn finally occurs. Before proceeding to policy implications that flow from these results, it is worth restating a few caveats about the data. Specifically, some of our results might be more positive than the real outcomes experienced by welfare recipients in the labor market, as our data likely miss the informal sector of the market and underrepresent smaller or newer establishments. But other potential biases (such as an overrepresentation of high-turnover jobs among the most recently hired) actually go in the other direction, and the net effects of these biases remain unclear. Policy Implications The discussion below describes the implications of these findings for policy at the federal, state, or local level regarding welfare reform in general, and especially for the goal of encouraging the success of welfare recipients in the labor market. Preparing for a Downturn There is no doubt that the strong state of the U.S. economy and the relative tightness of the labor market have helped welfare recipients find jobs. Because employers are very open to hiring them and working with local agencies to promote their success, this is a very good time to invest in the training of recipients and in supportive services (such as child care and transportation). Any work experience gained in this environment should help lead to wage growth and employment stability over time. On the other hand, at least some of this demand will no doubt disappear during the next cyclical downturn. Some recipients who cannot find employment at that time will be eligible for Unemployment Insurance, and many others will return to the welfare rolls (the growth of which will likely be financed out of TANF surpluses that many states are now accumulating). But some will not be eligible for either program— because of time limits in the case of TANF and insufficient prior earnings or other reasons in the case of Unemployment Insurance.1 The ____________ 1For a discussion of the Unemployment Insurance eligibility of welfare recipients see Gustafson and Levine (1998) and Vroman (1998). Most states require minimum levels 80 eligibility for either program among low-income males with little recent labor market experience will also be quite low, thus denying the children of these (often noncustodial) fathers an additional source of support. Just how many will fall into this category is unclear and will depend on factors such as the severity and duration of the downturn. Nevertheless, other arrangements will need to be made to preserve some “safety net” for such individuals. These might include: • A special Unemployment Insurance program for those who do not qualify for the usual program, as well as reforms in the Unemployment Insurance system that might make it easier for low-wage workers with limited experience to qualify; • More community service jobs for welfare recipients during this time period; and • Temporary suspensions of time limits or work requirements under the current TANF system, perhaps linked to state-level unemployment rates.2 This approach and its implementation raise several questions. For instance, exactly who might be eligible for the community service jobs or special Unemployment Insurance funds? Who would administer the program? How would state and federal funding be balanced? At a minimum, there needs to be more discussion of and planning for this period, at both the federal and state levels. Discussions regarding the reauthorization of TANF during the next year or two would be a perfect occasion for addressing these issues. ________________________________________________________ of earnings in a recent four-quarter “base period” to establish eligibility, with the current and most recent quarter excluded from that period. Many others are disqualified if they quit their jobs or were discharged for cause. Finally, those who work in the informal sector will generally not be covered by Unemployment Insurance programs. 2Reforms in Unemployment Insurance that might improve eligibility among lowwage workers include the use of an “alternative base period,” which includes the most recent quarter of earnings in calculations of prior earnings (which is potentially important for those who have entered the labor market recently); and providing some eligibility for part-time workers. A precedent for the development of a special Unemployment Insurance program can be found in the federal Special Unemployment Assistance program, funded by general revenues, that was put in place during the 1974–1975 recession. All of these possibilities are discussed in Vroman (1998). 81 Improving Access to Jobs When the labor market develops more slack, the relatively limited access of many welfare recipients (especially inner-city minorities) to suburban jobs may be more of a problem than it is currently. Improving access through innovative approaches to transportation and job placement might become a more pressing need. In fact, several federal initiatives (such as the “Bridges to Work” and “Access to Jobs” programs run by the U.S. Department of Housing and Urban Development) have already been implemented, as well as several at the state level. However, the scale of such efforts remains limited, and the problems associated with providing public transit access to outlying suburban employers remain substantial. We thus need continuing efforts to improve transportation to jobs by state and local agencies. More generally, the data presented here suggest that local welfare-towork agencies and other intermediaries can help recipients overcome some of the difficulties in gaining access to employers and achieving basic employability. But since relatively few employers in any given area have contact with these agencies, improving funding for their efforts and raising their outreach to local employer communities might be worthwhile goals. Improving Job Performance and Retention The need to improve retention and especially to reduce absenteeism among at least some current or former welfare recipients in the workplace remains strong. Improved access to reliable transportation and child care will certainly help deal with these problems. But local workforce boards and agencies need to focus more broadly on retention issues, rather than simply job placements for recipients. Approaches to job retention often include the provision of support services and counseling to both employers and employees in the workplace. Unfortunately, we have little clear evidence to date on successful models for improving job retention. The results in the Post Employment Services Demonstration (sponsored by the U.S. Department of Health and Human Services) have not been encouraging to date, though other strategies offer somewhat greater promise (Rangarajan, 1998). Continued experimentation with and evaluation of 82 different approaches in this area should remain a high priority for state and local agencies. Employment for the Hardest-to-Serve Given that skill requirements on jobs filled by welfare recipients are generally not trivial, and that many long-term recipients have poor skills and other personal characteristics that limit their employability, we may need to either improve their skills or develop various public sector employment options (at least in the short run) for the most disadvantaged portion of the caseload. The need to improve the skills of welfare recipients is sometimes seen as contrary to the “work first” notion that is embodied in most state-level welfare reform efforts. Indeed, programs that emphasize “work first” have generally proven to be more cost-effective than stand-alone training programs for this population. On the other hand, programs that combine some elements of both have often proven to be the most successful at raising the subsequent earnings of disadvantaged recipients (Michalopoulos et al., 2000). Public sector employment options might be useful for dealing with the most disadvantaged welfare recipients, as well as for time periods and local areas in which job availability is limited (such as during an economic downturn). These options might include various kinds of community service jobs or “sheltered workshops,” either as transition to the private sector or as ends in themselves. Although these approaches have largely fallen out of favor in recent years, Ellwood and Welty (2000) review the evidence on past programs and argue convincingly that such programs can be structured to provide useful public services, limited substitution for private sector employment, and subsequent gains for the individuals employed. Indeed, evidence on the Supported Work program for welfare recipients indicates strong positive postprogram effects on employment retention and earnings for very disadvantaged welfare recipients (Ham and Lalonde, 1996). And, although most states and local areas are not currently using these approaches, both New York City and Wisconsin 83 have developed relatively large-scale programs of public employment as part of their welfare reform efforts.3 Serving Low-Income Fathers Many other groups face barriers in the labor market similar to or even greater than those faced by welfare recipients. Indeed, employer reluctance to hire from some of these other stigmatized groups, such as those with criminal records, is substantially stronger than for welfare recipients, and they are the focus of much less assistance in the labor market today. Thus, efforts to expand various kinds of training and assistance to low-income fathers and other disadvantaged groups are critical (Committee for Economic Development, 2000). In fact, low-income young men and especially noncustodial fathers are now eligible to participate in local welfare-to-work programs as well as the new Youth Opportunity grants to high-poverty neighborhoods across the country (both funded by the U.S. Department of Labor), and strong outreach is needed to ensure their participation in these efforts. Given the very diverse and decentralized nature of these programs, strong evaluation efforts to identify effective interventions are critical as well. In addition, efforts to deal specifically with the employment issues faced by those with criminal records, who are rapidly becoming a much larger part of the population, need to become a higher priority. Finally, we should consider making the Earned Income Tax Credit and other wage/income supplements available to noncustodial fathers and other low-income men or women without children. ____________ 3Ellwood and Welty (2000) emphasize that New York’s Work Experience Program (“WEP”) has been a “workfare” model for welfare recipients, as opposed to the wageearning model used in Wisconsin and elsewhere. Nevertheless, both have the potential of providing needed work experience for recipients as they transition to the private sector labor market. 84 Appendix A Regression Results These tables present results from multiple regression equations that were estimated using the employer survey data. Table A.1 is discussed in Chapter 3 and Table A.2 in Chapter 4. Table A.1 Regression Results: Determinants of Prospective and Actual Hiring Metro area Chicago Cleveland Los Angeles Industry Agriculture Construction Manufacturing Transportation/communications Wholesale trade Finance Services Size 1–19 20–49 50–99 Prospective Demand: Current (1) (2) –.014 (.015) –.003 (.015) –.043** (.015) –.028* (.020) .001 (.017) –.047** (.020) — .027 (.070) — –.057* (.041) — –.111** (.022) — –.049* (.030) — –.077** (.033) — –.108** (.030) — –.062** (.017) — –.060** (.019) — –.047** (.018) — –.014 (.019) Actual Demand: Last Year (1) (2) –.043 (.024) .020 (.023) –.080 (.025) –.023 (.030) .033 (.026) –.050* (.031) — –.217* (.163) — –.133** (.070) — –.106** (.032) — –.178** (.053) — –.174** (.056) — –.042 (.043) — –.007 (.026) — –.076** (.028) — –.032 (.027) — .012 (.028) 85 Table A.1 (continued) Relative distance to public assistance population Distance to transit < .25 miles .25 – 1.0 miles Vacancy rate Difficulty finding qualified workers Percentage of employees unskilled Contact with agency Minority ownership Log L Prospective Demand: Current (1) (2) — .117** (.039) — –.003 (.020) — –.011 (.025) — .380** (.060) — .017* (.013) — –.031 (.376) — .027** (.016) — .009 (.020) 708.60 Actual Demand: Last Year (1) (2) — –.187** (.060) — — — — — — — 408.40 .037 (.032) .039 (.034) .215** (.093) .016 (.019) .030** (.004) .078** (.023) .063** (.029) 994.19 NOTES: Statistical significance: *p < .10, **p < .05, one-tailed test. Equations are estimated by tobit. Dependent variables in both cases are percentages of jobs, rather than employers. Reference groups for categorical variables are Milwaukee (metro area), retail trade (industry), 100 and above (size), and greater than one mile (relative distance to public assistance population). Standard errors are in parentheses. 86 87 Metro area Chicago Cleveland Los Angeles Industry Race Black Hispanic High school graduate General experience Size 1–19 20–49 50–99 Table A.2 Regression Results: Determinants of Wages and Relative Performance ln (Real Wages) (1) (2) (3) Better Performance Than Average (1) (2) (3) Duration (*Weibull Hazard Function) (1) (2) (3) .016 (.036) –.091** (.033) .051* (.033) No .021 (.046) –.065** (.038) .046 (.047) No .008 (.043) –.076** (.036) .016 (.044) Yes .122** (.048) .087** (.045) .173** (.049) No .118* (.080) .022 (.067) .205** (.081) No .133 (.079) .033 (.066) .211 (.080) Yes .457** (.259) .585** (.246) .569** (.245) No 1.034** (.391) .798** (.307) .848** (.348) No 1.012** (.381) .808** (.299) .755** (.336) Yes — .019 .005 — –.130** –.156** — –.523** –.564** (.035) (.033) (.062) (.062) (.301) (.304) — –.010 –.021 — –.124* –.145** — –.165 –.173 (.048) (.045) (.081) (.080) (.381) (.369) — .052** .037* — –.036 –.026 — .223 .187 (.031) (.029) (.055) (.055) (.261) (.259) — .001 .005 — .117** .136** — .010 –.051 (.031) (.029) (.054) (.054) (.244) (.244) — –.081** –.063* — .174** .151** — –.594** –.201 (.042) (.040) (.075) (.076) (.295) (.307) — –.118** –.089** — .084 .097* — –.363 –.105 (.039) (.038) (.070) (.071) (.309) (.301) — –.128** –.093** — .072 .049 — –.050 .280 (.042) (.040) (.073) (.073) (.359) (.394) 88 Relative distance Occupation Sales Clerical Craft Operative Laborer Service Tasks Read/write Arithmetic Computer Customer Table A.2 (continued) ln (Real Wages) (1) (2) (3) — –.029 –.005 (.087) (.082) Better Performance Than Average (1) (2) (3) — –.266** –.277** (.149) (.149) — –.129* .046 — –.065 –.091 (.091) (.091) (.139) (.149) — .055 .063 — –.074 –.028 (.087) (.082) (.133) (.132) — –.040 –.028 — –.109 .002 (.115) (.112) (.180) (.184) — –.077 –.057 — –.172 –.047 (.100) (.100) (.156) (.167) — –.125 .011 — .085 .091 (.120) (.116) (.189) (.196) — –.168** –.092 — –.082 –.087 (.087) (.082) (.133) (.133) — –.000 .010 — .039 .071 (.034) (.032) (.059) (.059) — –.004 .038 — –.044 –.063 (.032) (.032) (.056) (.058) — .063** .042 — .022 .018 (.038) (.036) (.067) (.067) — –.071** –.023 — .018 .006 (.041) (.040) (.071) (.073) Duration (*Weibull Hazard Function) (1) (2) (3) — –1.94** –2.02** (.736) (.724) — –1.194 –1.297* (.857) (.884) — –.652 –.640 (.838) (.823) — –1.57** –1.85** (.936) (.926) — –.661 –1.155 (.945) (.958) — –1.615* –1.75** (.993) (1.000) — –.110 –.063 (.857) (.840) — –.125 .005 (.265) (.259) — .278 .237 (.236) (.237) — .415* .499** (.273) (.277) — –.489 –.442** (.348) (.339) Table A.2 (continued) Forms Gauges R2 Log L ln (Real Wages) (1) (2) (3) — .024 –.015 (.037) (.035) — –.001 .010 (.034) (.032) .035 .205 .312 —— — Better Performance Than Average (1) (2) (3) — .031 .012 (.063) (.064) — .134** .152** (.059) (.059) .017 .102 .149 — —— Duration (*Weibull Hazard Function) (1) (2) (3) — .071 .026 (.289) (.285) — –.080 –.055 (.274) (.271) —— — 292.5 155.9 152.2 NOTES: Statistical significance: *p < .10, **p < .05, one-tailed test. The equation for “Better Performance Than Average” is estimated as a linear probability model. Coefficients in the hazard function can be interpreted as marginal effects of characteristics on the log of the expected duration of employment (in months). Reference groups for categorical variables are Milwaukee (metro area), white (race), 100 and above (size), and technical/managerial. 89 Appendix B Survey of Employers The actual text of the survey that was administered to employers by phone is reproduced below. U1 Before we begin, let me tell you that all the information that you give me will be kept strictly confidential. Should we come to any question that you do not want to answer, just let me know and we will go on to the next question. PART A: GENERAL BUSINESS QUESTIONS A2 (I’d like to start by asking you some general questions about your company.) Is this a for-profit or a nonprofit company? 1 FOR PROFIT 5 NONPROFIT 8 DO NOT KNOW 9 REFUSED A3 Does this company operate at more than one site? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED 91 A4a Is this a minority-owned company? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED A5a (I’d like to start by asking you a couple general questions about your organization.) About how far is your location from the nearest public transit stop? 0–9 TENTHS OF MILE 25 QUARTER OF A MILE 50 HALF OF A MILE 75 THREE QUARTERS OF A MILE 10 1 MILE OR MORE 97 NO PUBLIC TRANSPORTATION[goto A6a] 98 DO NOT KNOW[goto A6a] 99 REFUSED[goto A6a] A5b About how long do you think it would take for a worker to get to your business/organization from the center of the downtown business area if they used public transportation? 0–59 MINUTES 60 ONE HOUR OR MORE 98 DO NOT KNOW 99 REFUSED A6a Thinking about your customer or client base, what percent of your customers or clients are African American (or black)? 0–100 PERCENT AFRICAN AMERICAN 998 DO NOT KNOW 92 999 REFUSED A6b (What percent of your customers or clients are) Asian? 0–100 PERCENT ASIAN 998 DO NOT KNOW 999 REFUSED A6c (What percent of your customers or clients are) Hispanic? 0–100 PERCENT HISPANIC 998 DO NOT KNOW 999 REFUSED PART B: SITE EMPLOYMENT B1 All of the questions from now until the end of the interview refer only to the business (name) located at (address) (in city). How many employees currently work at (business name) (at this location?) 1–9996 EMPLOYEES AT SITE 9998 DO NOT KNOW 9999 REFUSED B2 How many people worked there approximately one year ago (one year ago would be (month) 1997)? 0–9996 EMPLOYEES YEAR AGO 9998 DO NOT KNOW 9999 REFUSED B4 What percentage of your current employees are covered by a collective bargaining agreement (belong to a union)? 93 0–100 PERCENT 998 DO NOT KNOW 999 REFUSED B6 How many of your (current) employees are in jobs that do not require any particular skills, education, previous training, or experience when they are hired? 0 EMPLOYEES[goto B7] 1–9996 EMPLOYEES 9998[goto B7] DO NOT KNOW[goto B7] 9999[goto B7] REFUSED[goto B7] B6b Of these employees, how many perform no significant reading, writing, or arithmetic on the job? 0 NO READING/MATH SKILLS[goto B7] 1–9996 NO READING/MATH SKILLS 9998 DO NOT KNOW/NOT APPLICABLE 9999 REFUSED B6c Of these employees (employees that perform no significant reading, writing, or arithmetic on the job), how many have been hired in the past year? 0 EMPLOYEES 1–9996 NUMBER HIRED PAST YEAR 9998 DO NOT KNOW/NOT APPLICABLE 9999 REFUSED B6d Of these [fill nskl] employees (those that perform no significant reading, writing, or arithmetic), how many are women? 0 NO WOMEN 1–9996 WOMEN 94 9998 DO NOT KNOW 9999 REFUSED B7 Now, thinking about all the different types of positions that you have at (business name), approximately how many vacancies are you currently trying to fill? 0–9996 JOB VACANCIES 9998 DO NOT KNOW 9999 REFUSED all3 During the past year, have you used the Internet when trying to fill job vacancies? 1 YES 5 NO[goto C1] 8 DO NOT KNOW[goto C1] 9 REFUSED[goto C1] all4 Have you listed job openings with America’s Job Bank on the Internet? [r]IWER: FYI: America’s Job Bank is an Internet Resource in which employers post job openings and people looking for jobs can search for job openings that meet their qualifications all over the nation 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED all5 Have you considered job applicants from America’s Talent Bank? 95 [r]IWER: FYI: On America’s Talent Bank, people post their resumes for viewing by prospective employers who are looking for people to fill positions 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED ml13 Have you used “JobNet”? [r]IWER: FYI: JobNet is Wisconsin’s job search system. Both employers and job seekers can use the system to either find workers or work. It contains all the jobs posted with Wisconsin Job Service 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED ml14 Have you participated in a Job Fair sponsored by a state job center or a public agency? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED[goto C1] PART C: LAST WORKER HIRED C1 This next section focuses entirely on the last worker you hired in the past year (that is since (month), 1997) into a position that does not require a college degree. 96 What type of work does this person do, that is, what is their position with your business/organization? 0 SPECIFY POSITION[#specify] 1–9996 SOC CODE 95 HAVE NOT HIRED SINCE [fill cMON] 1997 9998 DO NOT KNOW 9999 REFUSED C2a For this position, how necessary is a high school diploma? Would you say it is absolutely necessary, strongly preferred, mildly preferred, or not at all? 1 ABSOLUTELY NECESSARY 2 STRONGLY PREFERRED 3 MILDLY PREFERRED 4 NOT AT ALL 8 DO NOT KNOW 9 REFUSED C2b (How necessary is) previous experience in this line of work? (Would you say absolutely necessary, strongly preferred, mildly preferred, or not at all)? 1 ABSOLUTELY NECESSARY 2 STRONGLY PREFERRED 3 MILDLY PREFERRED 4 NOT AT ALL 8 DO NOT KNOW 9 REFUSED 97 C2c (How necessary is) some previous training or skill certification? (Would you say absolutely necessary, strongly preferred, mildly preferred, or not at all)? 1 ABSOLUTELY NECESSARY 2 STRONGLY PREFERRED 3 MILDLY PREFERRED 4 NOT AT ALL 8 DO NOT KNOW 9 REFUSED C3a Next, I would like to ask you some questions about the tasks this employee performs on a daily basis. Does this position involve speaking directly with customers in person or over the phone on a daily basis? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED C3b (Does this position involve) reading or writing reports, memos, or lengthy instructions on a daily basis? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED C3c (Does this position involve) doing arithmetic including making change (on a daily basis)? 1 YES 5 NO 98 8 DO NOT KNOW 9 REFUSED C3d (Does this position involve) using a personal computer (on a daily basis)? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED C5a Next, I would like to ask you some general questions about the last person you hired. Is this person male or female? 1 MALE 5 FEMALE 8 DO NOT KNOW 9 REFUSED C5b What is this employee’s racial or ethnic background? (Are they black, Asian, Native American, white, or Hispanic?) 1 AFRICAN-AMERICAN OR BLACK 2 ASIAN OR PACIFIC ISLANDER 3 NATIVE AMERICAN 4 WHITE OR CAUCASIAN 5 HISPANIC 0 OTHER: SPECIFY[#specify] 7 OTHER 8 DO NOT KNOW 9 REFUSED TO ANSWER 99 C5ba Is this employee an immigrant? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED C5d What is the highest level of education they have completed? 1 8TH GRADE OR LESS 2 SOME HIGH SCHOOL 3 GED 4 HIGH SCHOOL GRADUATE 5 SOME COLLEGE 6 ASSOCIATE DEGREE 7 COLLEGE GRADUATE 8 GRADUATE OR PROFESSIONAL DEGREE 9 TECHNICAL OR SKILLS TRAINING 0 OTHER: MISCELLANEOUS[#specify] 98 DO NOT KNOW 99 REFUSED all6 Did you use the Internet while recruiting this person? 1 YES 5 NO[goto C8ca] 8 DO NOT KNOW[goto C8ca] 9 REFUSED[goto C8ca] all7 Did the Internet generate this particular employee? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED[goto C8ca] 100 C6b When this person was first hired, were they hired through a temporary agency? 1 YES[goto C8ca] 5 NO 8 DO NOT KNOW 9 REFUSED[goto C8ca] C7a How was this employee recruited? (Was it through state employment agency, a private employment agency, referral from another current employee, referral from an acquaintance, walk-in from the street, newspaper ad, or a referral from a school, union, or community agency?) 1 STATE EMPLOYMENT AGENCY 2 PRIVATE EMPLOYMENT AGENCY 3 REFERRAL FROM ANOTHER EMPLOYEE 4 REFERRAL FROM AN ACQUAINTANCE/FRIEND 5 WALK-IN FROM THE STREET/HELP WANTED SIGN 6 NEWSPAPER AD 7 SCHOOL REFERRAL 8 UNION REFERRAL 9 COMMUNITY AGENCY 10 JOB LINE 11 PROMOTED FROM WITHIN/ALREADY AN EMPLOYEE 12 TRAINING PROGRAM 0 OTHER: SPECIFY[#specify] 98 DO NOT KNOW 99 REFUSED[goto all6] C8ca Did you have the applicants take any kind of tests? 1 YES 5 NO[goto C10a] 101 8 DO NOT KNOW[goto C10a] 9 REFUSED[goto C10a] C8cb Which kinds of test(s) did you administer? 1 BASIC SKILLS APTITUDE 2 BOOK-KEEPING TEST 3 CIVIL SERVICE EXAM 4 COMPUTER TEST 5 DATA ENTRY 6 DRIVING TEST 7 DRUG TEST 8 ENGLISH TEST 9 LOGIC TEST 10 MATH TEST 11 MECHANICAL SKILLS/APTITUDE 12 READING TEST 13 PSYCHOLOGICAL, PERSONALITY TESTS 14 PHYSICAL EXAM 15 PHYSICAL AGILITY, PHYSICAL APTITUDE 16 SPECIFIC JOB SKILL TEST (receptionist, bar tender, waitress, map reading, field test, company tests) 17 WRITING TEST/SAMPLES 18 TYPING TEST 19 VERBAL TEST 98 DO NOT KNOW 99 REFUSED C10a Of the applications you received for this position, roughly what percentage were from African American males? 0-100 PERCENT AFRICAN AMERICAN MALES 998 DO NOT KNOW 999 REFUSED 102 C10b (Roughly, what percentage were from) African American females? 0–100 PERCENT AFRICAN AMERICAN FEMALES 998 DO NOT KNOW 999 REFUSED C10c (Roughly, what percentage were from) Hispanics? [r]IWER IF NECESSARY USE: Your best estimate is fine. 0–100 PERCENT HISPANICS 998 DO NOT KNOW 999 REFUSED C10d (Roughly, what percentage were from) Asians? [r]IWER IF NECESSARY USE: Your best estimate is fine. 0–100 PERCENT ASIANS 998 DO NOT KNOW 999 REFUSED W1 What was their starting wage or salary for this position? [r]IWER: ENTER THE DOLLARS ONLY AT THIS SCREEN [r]IT IS POSSIBLE TO ENTER UP TO 100,000 AT THIS SCREEN 0–999997 DOLLARS 999998 DO NOT KNOW 999999 REFUSED 103 W1a (What is the starting wage?)[allow 3] [r]IWER: ENTER ONLY THE CENTS AT THIS SCREEN 00–99 CENTS 998 DO NOT KNOW 999 REFUSED W1b (Is this figure hourly, weekly, monthly, or yearly)? [r]IWER: YOU DO NOT HAVE TO ASK THIS QUESTION, JUST RECORD WHAT THE RESPONDENT HAS SAID TO YOU ABOUT THE WAGE STRUCTURE 1 HOURLY 2 WEEKLY 3 MONTHLY 4 YEARLY 5 SOMETHING ELSE[#specify] 6 BI-WEEKLY 7 COMMISSION 0 DAILY 8 DO NOT KNOW 9 REFUSED C12 In addition to their wage, does this employee get tips, commission, or profit sharing? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED C13a Do you provide health insurance to the employee in this position? 104 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED C14a How many hours per week does this person usually work? 0–90 HOURS PER WEEK 98 DO NOT KNOW 99 REFUSED C16a Did this employee receive any formal training after being hired? (By formal training, we mean with an instructor or in a classroom)? 1 YES 5 NO[goto C20] 8 DO NOT KNOW[goto C20] 9 REFUSED[goto C20] C16b Approximately how many hours of formal training did he or she receive? 1–996 HOURS OF TRAINING 998 DO NOT KNOW 999 REFUSED C20 Would you say it is easy, somewhat difficult, or very difficult to find qualified applicants for this type of position at the present time? 1 EASY 3 SOMEWHAT DIFFICULT 5 VERY DIFFICULT 105 8 DO NOT KNOW 9 REFUSED C21a In the past two years, have you hired someone into a non-college degree position who did not meet all of the qualifications you usually require because you really needed the employee? 1 YES 5 NO[goto C22a] 8 DO NOT KNOW[goto C22a] 9 REFUSED[goto C22a] Ca21 How many have you hired in the past two years? 1–9996 HIRES 9998 DO NOT KNOW 9999 REFUSED C22a Next, I am going to list a few different types of applicants. Please tell me if you would accept each type of applicant for a position in your business/organization. A person who is or has been on welfare? (Would you definitely, probably not, or definitely not accept this applicant)? 1 DEFINITELY WOULD 2 PROBABLY WOULD 3 PROBABLY NOT 4 DEFINITELY NOT 8 DO NOT KNOW 9 REFUSED C22b A person who is or has been in a government employment program or had a GED instead of a high school diploma? 106 (Would you definitely, probably not, or definitely not accept this applicant)? 1 DEFINITELY WOULD 2 PROBABLY WOULD 3 PROBABLY NOT 4 DEFINITELY NOT 8 DO NOT KNOW 9 REFUSED C22c A person who has a criminal record? (Would you definitely, probably not, or definitely not accept this applicant)? 1 DEFINITELY WOULD 2 PROBABLY WOULD 3 PROBABLY NOT 4 DEFINITELY NOT 8 DO NOT KNOW 9 REFUSED C22d A person who only lists short term or part time jobs for work experience? (Would you definitely, probably not, or definitely not accept this applicant)? 1 DEFINITELY WOULD 2 PROBABLY WOULD 3 PROBABLY NOT 4 DEFINITELY NOT 8 DO NOT KNOW 9 REFUSED 107 C22e A person who has been unemployed for a year or more? (Would you definitely, probably not, or definitely not accept this applicant)? 1 DEFINITELY WOULD 2 PROBABLY WOULD 3 PROBABLY NOT 4 DEFINITELY NOT 8 DO NOT KNOW 9 REFUSED PART D: WELFARE RECIPIENTS D1 The next set of questions focus on hiring individuals who have received welfare. Does your business/organization have any type of a formal welfare to work program at this workplace? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED Da1 To your knowledge, has your business/organization in the past two years (that is, since (month) 1996) hired any women who had been on welfare? 1 YES 5 NO[goto E1] 8 DO NOT KNOW[goto E1] 9 REFUSED[goto E1] Da2 How many (women who have been on welfare) has your business/organization hired in the past two years? 108 1–500 HIRES PAST 2 YEARS 998 DO NOT KNOW 999 REFUSED Da3 How many (women who have been on welfare) has your business/organization hired in the past year? 0–500 HIRES PAST 1 YEAR 998 DO NOT KNOW 999 REFUSED Da4 I’d like to ask you some questions about the last woman you have hired that had been a welfare recipient. First, how sure are you that she had been a welfare recipient? Would you say you are definitely sure, fairly sure, or not really sure? 1 DEFINITELY SURE 2 FAIRLY SURE 3 NOT REALLY SURE[goto E1] 8 DO NOT KNOW[goto E1] 9 REFUSED[goto E1] Da5 Was she still on welfare, had she just come off of welfare, or had she been off welfare for some time when she was hired? 1 STILL ON WELFARE 2 JUST COME OFF WELFARE 3 BEEN OFF WELFARE FOR SOME TIME 8 DO NOT KNOW 9 REFUSED 109 Da6 Did she have a regular high school diploma? 1 YES[goto Da8] 5 NO 8 DO NOT KNOW 9 REFUSED[goto Da8] Da7 Did she have a GED? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED Da8 Was she recruited through some type of agency or by some other means? 1 AGENCY 5 OTHER MEANS OF RECRUITMENT[goto same] 8 DO NOT KNOW[goto same] 9 REFUSED[goto same] Da9 Did you contact the agency from which you hired this person, or did the agency contact you? 1 CONTACTED THE AGENCY 5 AGENCY CONTACT BUSINESS 8 DO NOT KNOW 9 REFUSED Is this the person that we have already discussed above? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED 110 Da10 What is this employee’s racial or ethnic background? (Are they black, Asian, Native American, white, or Hispanic)? 1 AFRICAN-AMERICAN OR BLACK 2 ASIAN OR PACIFIC ISLANDER 3 NATIVE AMERICAN 4 WHITE OR CAUCASIAN 5 HISPANIC 7 OTHER 8 DO NOT KNOW 9 REFUSED TO ANSWER Da11 Is she an immigrant? 1 YES 5 NO[goto D1c] 8 DO NOT KNOW[goto D1c] 9 REFUSED[goto D1c] Da12 Does this employee speak English fluently? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED D1c In what month and year was this person hired? 1–12 MONTH 98 DO NOT KNOW 99 REFUSED 111 Dc1 In what month and year was this person hired? 90–99 YEAR HIRED[goto sm3] 998 DO NOT KNOW 999 REFUSED D2 Into what position was she hired? 1–9996 SOC CODE 9998 DO NOT KNOW 9999 REFUSED W3 What was the starting wage or salary for her position? 0–999997 DOLLARS 999998 DO NOT KNOW 999999 REFUSED W3a (What is the starting wage?)[allow 3] 00–99 CENTS 998 DO NOT KNOW 998 REFUSED W3b (Is this figure hourly, weekly, monthly, or yearly)? 1 HOURLY 2 WEEKLY 3 MONTHLY 4 YEARLY 5 SOMETHING ELSE [#specify] 6 BI-WEEKLY 7 COMMISSION 8 DO NOT KNOW 9 REFUSED 112 D2a In addition to their wage, does this employee get tips, commission, or profit sharing? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED D4 Do you provide health insurance to her and her family? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED D4a How many hours per week does this person usually work? 0–80 HOURS PER WEEK 98 DO NOT KNOW 99 REFUSED D4b Thinking about this employee’s work schedule, is it at all flexible or are the hours strictly fixed? 1 FLEXIBLE 5 FIXED 8 DO NOT KNOW 9 REFUSED D6a Did she have any recent work experience before you hired her? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED 113 D6b (Did she have) previous experience in this particular type of work? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED D6c Had she received any previous training that was relevant to this position or this line of work? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED D7a Before hiring her, did you do a criminal background check? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED Db7 Did you have the applicants take any kind of tests? 1 YES 5 NO[goto D7b] 8 DO NOT KNOW[goto D7b] 9 REFUSED[goto D7b] Dc7 Which kinds of test(s) did you administer? 1 BASIC SKILLS APTITUDE 2 BOOK-KEEPING TEST 3 CIVIL SERVICE EXAM 4 COMPUTER TEST 114 5 DATA ENTRY 6 DRIVING TEST 7 DRUG TEST 8 ENGLISH TEST 9 LOGIC TEST 10 MATH TEST 11 MECHANICAL SKILLS/APTITUDE TEST/SAMPLES 12 READING TEST 13 PSYCHOLOGICAL, PERSONALITY TESTS 14 PHYSICAL EXAM 15 PHYSICAL AGILITY, PHYSICAL APTITUDE 16 SPECIFIC JOB SKILL TEST (receptionist, bar tender, waitress, map reading, field test, company tests) 17 WRITING 18 TYPING TEST 19 VERBAL TEST 98 DO NOT KNOW 99 REFUSED D7b How much did each of the following factors weigh into your decision to hire her for this position? Her overall appearance, dress, and demeanor? Would you say a lot, some, or none at all? 1 A LOT 2 SOME 3 NONE AT ALL 8 DO NOT KNOW 9 REFUSED D7c Her recommendations and referrals? (Would you say a lot, some, or none at all)? 1 A LOT 2 SOME 115 3 NONE AT ALL 8 DO NOT KNOW 9 REFUSED D7g How many other applicants for this position did you pass over to hire her? 0–500 APPLICANTS 998 DO NOT KNOW/NOT APPLICABLE 999 REFUSED D8a Next, I would like to ask you some questions about the tasks this employee performs on a daily basis. Does this position involve speaking directly with customers in person or over the phone on a daily basis? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED D8b (Does this position involve) reading or writing reports, memos, or lengthy instructions on a daily basis? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED D8c (Does this position involve) doing arithmetic including making change (on a daily basis)? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED 116 D8d (Does this position involve) using a personal computer? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED[goto D8e] de8 Next, I have a couple of questions about other tasks this employee performs on a daily basis. D8e (Does this position involve) filling out forms on a daily basis? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED D8f (Does this position involve) keeping a close watch over gauges, dials, or instruments of any kind (on a daily basis)? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED D9a For this position, did she receive any formal training after being hired? 1 YES 5 NO[goto D9c] 8 DO NOT KNOW[goto D9c] 9 REFUSED[goto D9c] 117 D9b Was any of the training she received after being hired remedial of basic skills? Basic skills include the employee’s ability to read or write, do arithmetic, or speak English. 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED D9c After hiring her, did she pursue any additional education on her own? 1 YES 5 NO[goto D9f] 8 DO NOT KNOW[goto D9f] 9 REFUSED[goto D9f] D9d Did you support her with tuition reimbursements? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED D9e Did you support her with reduced or flexible work hours? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED D9f Do you provide child care assistance to this employee? 1 YES 118 5 NO 8 DO NOT KNOW 9 REFUSED D9g (Do you provide) transportation assistance? 1 YES 5 NO[##md20] 8 DO NOT KNOW 9 REFUSED D10a For the next set of questions, please tell me if you had problems with this employee in each of the following areas. Absenteeism or tardiness? 1 YES 5 NO[goto D10g] 8 DO NOT KNOW[goto D10g] 9 REFUSED[goto D10g] D10b Were her problems with absenteeism or tardiness due to any of the following? Physical health? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED D10c (Were her problems with absenteeism or tardiness due to) mental health or depression? 1 YES 5 NO 119 8 DO NOT KNOW 9 REFUSED D10d (Were her problems with absenteeism or tardiness due to) child care? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED D10e (Were her problems with absenteeism or tardiness due to) transportation? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED D10f (Were her problems with absenteeism or tardiness due to) domestic violence? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED D10g Did you have any problems with this employee’s attitude toward work? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED 120 D10h (Did you have any problems with this employee’s) basic verbal, math, or reading skills? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED D10i (Did you have any problems with this employee’s) other job-related skills? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED D10j (Did you have any problems with this employee’s) substance abuse (alcohol or drug abuse)? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED D10k (Did you have any problems with this employee’s) relationships with co-workers, supervisors, or customers? 1 YES 5 NO 9 REFUSED D13 If this person performs well, what are the chances that she could be promoted? Would you say excellent, good, fair or poor? 121 1 EXCELLENT 2 GOOD 3 FAIR 4 POOR 8 DO NOT KNOW 9 REFUSED D14 Overall, how would you rate the employee relative to the typical one that you hire into this position? Would you say she was much better, a little better, about the same, a little worse, or much worse? 1 MUCH BETTER 2 A LITTLE BETTER 3 ABOUT THE SAME 4 A LITTLE WORSE 5 MUCH WORSE 8 DO NOT KNOW 9 REFUSED D5 Is she still with your firm? 1 YES[goto all8] 5 NO 8 DO NOT KNOW[goto all8] 9 REFUSED [goto all8] D5a In what month and year did she leave your business/ organization? 1–12 MONTH 98 DO NOT KNOW 99 REFUSED D5b (In what year did she leave your business/organization)? 122 19 94–99 YEAR 98 DO NOT KNOW 99 REFUSED D5c Did she voluntarily quit, get discharged, get laid off, or something else? 1 VOLUNTARILY QUIT 2 DISCHARGED 3 LAID OFF 4 POSITION ENDED 0 OTHER: SPECIFY[#specify] 8 DO NOT KNOW 9 REFUSED[goto all8] E1 Have you had any contact during the past year with any agency that is trying to place welfare recipients into jobs? 1 YES 5 NO[goto F1] 8 DO NOT KNOW[goto F1] 9 REFUSED[goto F1] Ea1 Did you contact the agency, or did the agency contact you? 1 CONTACTED THE AGENCY 5 AGENCY CONTACT BUSINESS 8 DO NOT KNOW 9 REFUSED[goto fl2] 1e Next, I have some questions about contacts you may have had with agencies who find work for welfare recipients. E1a Were they a(n)(name of agency) agency or contractor? 1 YES 123 5 NO 8 DO NOT KNOW 9 REFUSED m11 Have you claimed the subsidy for “Trial Jobs”? 1 YES 5 NO[goto skp5] 8 DO NOT KNOW[goto skp5] 9 REFUSED[goto skp5] ml12 For how many employees have you claimed the subsidy at any time? 1–500 EMPLOYEES 998 DO NOT KNOW 999 REFUSED 1000 [goto skp5] ml1 I would like to read you a list of agencies in the Milwaukee area. Please tell me if you have had any contact with the agency during the past year. YW Works? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED ml2 (Have you had any contact with) Opportunities Industrialization Center? 1 YES 5 NO 124 8 DO NOT KNOW 9 REFUSED ml3 (Have you had any contact with) United Migrant Opportunities Services? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED ml4 (Have you had any contact with) Workforce Solutions? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED ml5 (Have you had any contact with) Maximus? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED ml6 (Have you had any contact with) The Milwaukee Private Industry Council? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED ml7 (Have you had any contact with) Waukesha Workforce Development Center? 125 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED ml8 (Have you had any contact with) Washington County Department of Social Services? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED ml9 (Have you had any contact with) Ozaukee County Workforce Development Center? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED[goto E3] E3 Did you agree to consider an application or interview at least one of these referrals? 1 YES 5 NO[goto E3b3] 8 DO NOT KNOW[goto E3b3] 9 REFUSED[goto E3b3] E3a How many applications did you consider or interviews did you conduct? 0–996 APPLICANTS/INTERVIEWS 998 DO NOT KNOW/NOT APPLICABLE 999 REFUSED 126 E3b Did you hire any of their referrals? 1 YES 5 NO[goto E3b3] 8 DO NOT KNOW[goto E3b3] 9 REFUSED[goto E3b3] E3b1 How many (referrals) did you hire? 0–996 REFERRALS 998 DO NOT KNOW 999 REFUSED E3b2 How many (referrals) still work for you? 0–996 REFERRALS 998 DO NOT KNOW 999 REFUSED Ea3 Did the agency provide any of the following assistance to the applicants you hired? General work experience not related to the type of work at your business/organization? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED Eb3 Experience in the particular type of work? (Did the agency provide any assistance to the applicants you hired)? 1 YES 127 5 NO 8 DO NOT KNOW 9 REFUSED Ec3 Training relevant to the type of work your business/organization does? (Did the agency provide any assistance to the applicants you hired)? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED Ed3 Did the agency provide any services to your business/organization after you hired referrals to assist in retaining referrals or dealing with problems that arise with referrals? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED E3b3 Would you consider any future referrals from this agency? 1 YES 3 MAYBE 5 NO 8 DO NOT KNOW 9 REFUSED F1 Suppose you were contacted by an employment agency that was trying to place welfare recipients. We are specifically talking 128 about welfare recipients who do not have a high school diploma or any recent work experience. Do you currently have any open positions that you might consider filling with these welfare recipients? 1 YES 5 NO[goto F1a] 8 DO NOT KNOW[goto F1a] 9 REFUSED[goto F1a] 1F How many of them would you consider employing right away? 1–996 IMMEDIATE HIRES 998 DO NOT KNOW 999 REFUSED F1a Do you think you will have open positions during the next year that you might consider filling with these welfare recipients? 1 YES 5 NO[goto Fa5] 8 DO NOT KNOW[goto Fa5] 9 REFUSED[goto Fa5] 2F How many of them would you possibly employ at any time during the next year? 1–996 HIRES DURING THE NEXT YEAR 998 DO NOT KNOW 999 REFUSED Fa5 Did you know that tax credits for hiring welfare recipients are now available from the federal government? 1 YES 129 5 NO 8 DO NOT KNOW 9 REFUSED Fa51 Did you know that you can receive these credits per employee for up to 2 years? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED ml10 Did you know that you can receive a subsidy from the state of Wisconsin for putting people into “Trial Jobs”? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED[goto Fa6] Fa6 Do you think this will make you any more likely than before to hire poorly educated young women who are single mothers? 1 YES 3 MAYBE 5 NO 8 DO NOT KNOW 9 REFUSED Fa10 If you discover that an employee is eligible, will you or your company actually claim the credit next year? 1 YES 3 MAYBE 5 NO 8 DO NOT KNOW 130 9 REFUSED[goto all1] all1 Have you claimed the Federal Tax Credits for any employee during the past year? 1 YES 5 NO[goto skp5] 8 DO NOT KNOW[goto skp5] 9 REFUSED[goto skp5] all2 For how many employees have you claimed the credits at any time? 1–500 EMPLOYEES 998 DO NOT KNOW 999 REFUSED all8 When hiring this person, did you remind them that they may be eligible for the Earned Income Credit (on their Federal Income Taxes)? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED ml15 (When hiring this person) did you remind them about child care subsidies from the state? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED[goto skp4] 131 skp5 [if F1 ge 5][if F1a ge 5][goto F11a][endif][endif] would not hire right away and would not hire in the next year skip to question F11a. F5 Is there some specific job into which you would most likely hire a worker who was a welfare recipient who did not have a high school diploma or recent work experience? 1000–9996 SOC CODE 1 YES, SPECIFY POSITION(S)[#specify] 5 NO, NO SPECIFIC POSITION[goto F11a] 8 DO NOT KNOW, UNSURE OF POSITION[goto F11a] 9 REFUSED[goto F11a] W5 What is the starting wage or salary for this position? 0–999997 DOLLARS 999999 DO NOT KNOW 999998 REFUSED W5a (What is the starting wage?) 00–99 CENTS 998 DO NOT KNOW 999 REFUSED W5b (Is this figure hourly, weekly, monthly, or yearly)? 1 HOURLY 2 WEEKLY 3 MONTHLY 4 YEARLY 5 SOMETHING ELSE[#specify] 6 BI-WEEKLY 7 COMMISSION 0 DAILY 132 8 DO NOT KNOW 9 REFUSED F6 In addition to their wage, would this employee get tips, commission, or profit sharing? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED F7 Would this position provide health benefits to the employee and the employee’s family? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED F8 How many hours per week would this person usually work? 0–60 HOURS PER WEEK 98 DO NOT KNOW 99 REFUSED F10 If this person performs well, what are the chances that this person could be promoted? Would you say excellent, good, fair or poor? 1 EXCELLENT 2 GOOD 3 FAIR 4 POOR 8 DO NOT KNOW 9 REFUSED 133 F11a Would you be more likely to hire welfare recipients, especially those without a high school diploma or recent work experience, if the agency could provide the following. A 50% tax credit against their wages for at least one year? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED F11b (Would you be more likely to hire welfare recipients, especially those without a high school diploma or recent work experience, if the agency could provide) A 100% tax credit against their wages for at least one year? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED F11c (Would you be more likely to hire welfare recipients, especially those without a high school diploma or recent work experience, if the agency could provide) Applicants with any kind of recent work experience? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED 134 F11d (Would you be more likely to hire welfare recipients, especially those without a high school diploma or recent work experience, if the agency could provide) Work experience in your particular line of work? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED F11e (Would you be more likely to hire welfare recipients, especially those without a high school diploma or recent work experience, if the agency could provide) Training in your particular line of work? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED F11f (Would you be more likely to hire welfare recipients, especially those without a high school diploma or recent work experience, if the agency could provide) Assurances that applicants have basic math or reading skills? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED F11g (Would you be more likely to hire welfare recipients, especially those without a high school diploma or recent work experience, if the agency could provide) 135 Assurances that applicants have good attitudes toward work? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED F11h (Would you be more likely to hire welfare recipients, especially those without a high school diploma or recent work experience, if the agency could provide) (Assurances that the applicant) has no criminal record? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED F11i (Would you be more likely to hire welfare recipients, especially those without a high school diploma or recent work experience, if the agency could provide) (Assurances that the applicant) does not have any substance abuse problems? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED F11j (Would you be more likely to hire welfare recipients, especially those without a high school diploma or recent work experience, if the agency could provide) (Assurances that the applicant) has good dress and appearance? 136 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED F11k (Would you be more likely to hire welfare recipients, especially those without a high school diploma or recent work experience, if the agency could provide) (Assurances that the applicant) has good social skills? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED F12a Before hiring a referral, would you do a drug test? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED F12b (Before hiring a referral), would you check into their background for a possible criminal record? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED F13a Which of the following types of assistance would you be willing to provide, if any, in order to improve a welfare recipient’s ability to hold a job with your business/organization? 137 Transportation? 1 YES 3 MAYBE 5 NO 8 DO NOT KNOW 9 REFUSED F13b Child Care? 1 YES 3 MAYBE 5 NO 8 DO NOT KNOW 9 REFUSED F13c Remedial basic skills training? 1 YES 3 MAYBE 5 NO 8 DO NOT KNOW 9 REFUSED F13d Specific job skills training? 1 YES 3 MAYBE 5 NO 8 DO NOT KNOW 9 REFUSED F13e Tuition reimbursements for school? 1 YES 3 MAYBE 138 5 NO 8 DO NOT KNOW 9 REFUSED F13f Reduced or flexible hours for schooling? 1 YES 3 MAYBE 5 NO 8 DO NOT KNOW 9 REFUSED F14a Would you be more likely to provide some training if you could get a tax credit against at least some of the costs? 1 YES 3 MAYBE 5 NO 8 DO NOT KNOW 9 REFUSED F14b Would you be more likely to provide some training if you could get some technical assistance with the training? 1 YES 3 MAYBE 5 NO 8 DO NOT KNOW 9 REFUSED H1 Would you be more likely to hire someone from an agency trying to place workers such as the ones we just spoke about if they provided services to help you retain the employees or to deal with problems that arose? 1 YES 139 5 NO 8 DO NOT KNOW 9 REFUSED H2 Would you be willing to work with such an agency in developing relevant work experience or training for a client that you might hire afterward? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED G1 Finally, I have two questions about another group of workers. Suppose you were contacted by an employment agency that was trying to place young males with criminal records. Do you currently have any open positions that you might consider filling with this group of workers? [r]IWER: IF THE RESPONDENT SAYS, DEPENDS OR DO NOT KNOW, USE THE FOLLOWING PROBE: “Just assume that they are young males who have committed crimes that have lead to their incarceration” 1 YES 5 NO[goto MOD7] 8 DO NOT KNOW[goto MOD7] 9 REFUSED[goto MOD7] 1G How many of them would you consider employing right away? 1–996 IMMEDIATE HIRES 998 DO NOT KNOW 999 REFUSED 140 Bibliography Abraham, Katherine, “Structural/Frictional v. Demand-Deficient Unemployment,” American Economic Review, Vol. 73, 1983, pp. 708– 724. Acs, Gregory, Norma Coe, Keith Watson, and Robert Lerman, Does Work Pay? An Analysis of the Work Incentives Under TANF, The Urban Institute, Washington, D.C., 1999. Anderson, Patricia, and Bruce Meyer, “The Extent and Consequences of Job Turnover,” Brookings Papers on Economic Activity— Microeconomics, 1994, pp. 177–248. Aron, Laudan, Robert Lerman, and Stephanie Riegg, Youth in 2010: Trends and Policy Implications, The Urban Institute, Washington, D.C., 2000. Ballen, John, and Richard Freeman, “Transitions from Employment to Nonemployment,” in R. Freeman and H. Holzer (eds.), The Black Youth Employment Crisis, University of Chicago Press, Chicago, 1986. Bane, Mary Jo, and David Ellwood, “Slipping Into and Out of Poverty—The Dynamics of Spells,” Journal of Human Resources, Vol. 21, 1986, pp. 2–23. Bane, Mary Jo, and David Ellwood, Welfare Realities, Harvard University Press, Cambridge, 1994. Bates, Timothy, “Job Creation Through Improved Access to Markets for Minority-Owned Businesses,” in W. Leigh and M. Simms (eds.), The Black Worker in the 21st Century: Job Creation Prospects and 141 Strategies, The Joint Center for Political and Economic Studies, Washington, D.C., 1998. Becker, Gary, The Economics of Discrimination, University of Chicago Press, Chicago, 1971. Berman, Eli, John Bound, and Zvi Griliches, “Changes in the Demand for Skilled Labor Within U.S. Manufacturing: Evidence from the Annual Survey of Manufacturers, Quarterly Journal of Economics, Vol. 109, 1994, pp. 367–397. Blumenberg, Evelyn, and Paul Ong, “Accessibility and Welfare Usage Rates: Evidence from Los Angeles,” Journal of Policy Analysis and Management, Vol. 17, No. 4, 1998. Borjas, George, “The Economics of Immigration,” Journal of Economic Literature, Vol. 32, 1994, pp. 1667–1717. Brown, Charles, James Hamilton, and James Medoff, Employers Large and Small, Harvard University Press, Cambridge, 1990. Burtless, Gary, “The Employment Prospects of Welfare Recipients,” in D. Nightingale (ed.), The Work Alternative, The Urban Institute Press, Washington, D.C., 1995. Burtless, Gary, “Can the Labor Market Absorb Three Million Welfare Recipients?” The Low Wage Labor Market: Challenges and Opportunities for Economic Self-Sufficiency, Office of the Assistant Secretary for Policy and Evaluation, U.S. Department of Health and Human Services, Washington, D.C., 1999. Cancian, Maria, Robert Haveman, Thomas Kaplan, Daniel Meyer, and Barbara Wolfe, “Work, Earnings and Well-Being after Welfare,” in S. Danziger (ed.), Economic Conditions and Welfare Reform, W. E. Upjohn Institute for Employment Research, Kalamazoo, MI, 1999. 142 Clark, Kim, and Lawrence Summers, “The Dynamics of Youth Unemployment,” in R. Freeman and D. Wise (eds.), The Youth Labor Market Problem: Its Nature, Causes and Consequences, University of Chicago Press, Chicago, 1982. Committee for Economic Development, Welfare Reform and Beyond: Making Work Work, Washington, D.C., 2000. Council of Economic Advisers, Technical Report: Explaining the Decline in Welfare Receipt, 1993–96, Executive Office of the President, Washington, D.C., 1997. Council of Economic Advisers, Technical Report: Economic Expansion, Welfare Reform, and the Decline of Welfare Caseloads: An Update, Executive Office of the President, Washington, D.C., 1999. County of Los Angeles, Evaluating CalWorks in Los Angeles County, Urban Research Division, Chief Administrative Office, 1999. Danziger, Sandra, Mary Corcoran, Sheldon Danziger, Colleen Heflin, Ariel Kalil, Judith Levine, Daniel Rosen, Kristin Seefeldt, Kristine Siefert, and Richard Tolman, Barriers to the Employment of Welfare Recipients, Poverty Research and Training Center, University of Michigan, 1999. Danziger, Sheldon, Colleen Heflin, and Mary Corcoran, “Does It Pay to Move from Welfare to Work?” Mimeo, University of Michigan, 2000. Edelman, Peter, “The Worst Thing Bill Clinton Has Done,” The Atlantic Monthly, March 1997, pp. 43–58. Edin Kathryn, and Laura Lein, Making Ends Meet: How Single Mothers Survive Welfare and Low-Wage Work, Russell Sage Foundation, New York, 1997. 143 Ellwood, David, “The Impact of the Earned Income Tax Credit and Other Social Policy Changes on Work and Marriage in the United States,” Working Paper, Harvard University, 1999. Ellwood, David, and Jonathan Crane, “Family Change among Black Americans: What Do We Know?” Journal of Economic Perspectives, Vol. 4, 1990, pp. 65–84. Ellwood, David, and Elizabeth Welty, “Public Sector Employment and Mandatory Work: A Policy Whose Time Has Come and Gone and Come Again?” in D. Card and R. Blank (eds.), Finding Jobs: Work and Welfare Reform, Russell Sage Foundation, New York, 2000. Figlio, David, and James Ziliak, “Welfare Reform, the Business Cycle, and the Decline in AFDC Caseloads,” in S. Danziger (ed.), Economic Conditions and Welfare Reform, 1999. Flaming, Daniel, Mark Drayse, and Peter Force, On the Edge: A Progress Report on Welfare to Work in Los Angeles, Economic Roundtable, Los Angeles, 1999. Freedman, Stephen, Marisa Mitchell, and David Navarro, The Los Angeles Jobs-First GAIN Evaluation, New York Manpower Demonstration Research Corporation, 1999. Freeman, Richard, “Crime and the Employment of Disadvantaged Youth,” in G. Peterson and W. Vroman (eds.), Urban Labor Markets and Job Opportunities, The Urban Institute Press, Washington, D.C., 1992. Freeman, Richard, and William Rodgers, “Area Economic Conditions and the Labor Market Outcomes of Young Men in the 1990’s Expansion,” National Bureau of Economic Research Working Paper, 1999. 144 Gustafson, Cindy, and Philip Levine, “Less-Skilled Workers, Welfare Reform, and the Unemployment Insurance System,” NBER Working Paper, 1998. Ham, John, and Robert Lalonde, “The Effect of Sample Selection and Initial Conditions in Duration Models: Evidence from Experimental Data on Training,” Econometrica, Vol. 64, No. 1, 1996. Holzer, Harry, Unemployment, Vacancies and Local Labor Markets, W. E. Upjohn Institute for Employment Research, Kalamazoo, MI, 1989. Holzer, Harry, What Employers Want: Job Prospects for Less-Educated Workers, Russell Sage Foundation, New York, 1996. Holzer, Harry, “Why Do Small Employers Hire Fewer Blacks Than Larger Ones?” Journal of Human Resources, Vol. 33, 1998, pp. 896– 914. Holzer, Harry, “Will Employers Hire Welfare Recipients? New Survey Evidence from Michigan,” Journal of Policy Analysis and Management, 1999a. Holzer, Harry, “Employer Demand for Welfare Recipients and the Business Cycle,” in S. Danziger (ed.), Economic Conditions and Welfare Reform, W. E. Upjohn Institute for Employment Research, Kalamazoo, Michigan, 1999b. Holzer, Harry, and Keith Ihlanfeldt, “Spatial Factors and the Employment of Blacks at the Firm Level,” New England Economic Review, May/June 1996, pp. 65–86. Holzer, Harry, and Robert Lalonde, “Job Change and Job Stability Among Less-Skilled Young Workers,” in R. Blank and D. Card (eds.), Finding Jobs: Work and Welfare Reform, Russell Sage Foundation, New York, 2000. 145 Holzer, Harry, and David Neumark, “What Does Affirmative Action Do?” Industrial and Labor Relations Review, Vol. 53, 2000, pp. 240– 271. Hughes, Mark, and Julie Sternberg, The New Metropolitan Reality: Where the Rubber Meets the Road in Antipoverty Policy, The Urban Institute, Washington, D.C., 1992. Ihlanfeldt, Keith, and David Sjoquist, “The Spatial Mismatch Hypothesis: A Review of Recent Studies and Their Implications for Welfare Reform,” Housing Policy Debate, Vol. 9, 1998, pp. 849–892. Johnson, Hans, and Sonya Tafoya, The Basic Skills of Welfare Recipients: Implications for Welfare Reform, Public Policy Institute of California, San Francisco, 1999. Kain, John, “The Spatial Mismatch Hypothesis Three Decades Later,” Housing Policy Debate, Vol. 3, 1992, pp. 371–460. Katz, Bruce, and Kate Carnevale, The State of Welfare Caseloads in American Cities, Center on Urban and Metropolitan Policy, The Brookings Institution, Washington, D.C., 1998. Katz, Lawrence, and Alan Krueger, “The High-Pressure U.S. Labor Market of the 1990’s,” Brookings Papers on Economic Activity, No. 1, 1999. Katz, Lawrence, and Kevin Murphy, “Changes in Relative Wages, 1963– 87: The Role of Supply and Demand Factors,” Quarterly Journal of Economics, Vol. 107, 1992, pp. 35–78. Kirschenman, Joleen, “Gender Within Race in the Labor Market,” Unpublished Paper, University of Chicago, 1991. 146 Kirschenman, Joleen, and Kathryn Neckerman, “We’d Love to Hire Them But . . . ,” in C. Jencks and P. Peterson (eds.), The Urban Underclass, The Brookings Institution, Washington, D.C., 1991. Krueger, Alan B., and Lawrence Summers, “Reflections on the InterIndustry Wage Structure,” in Kevin Lang and Jonathan Leonard (eds.), Unemployment and the Structure of Labor Markets, Basil Blackwell Press, London, 1987, pp. 17–47. Loprest, Pamela, “Families Who Left Welfare: Who Are They and How Are They Doing?” Assessing the New Federalism Discussion Paper, The Urban Institute, Washington, D.C., 1999. Mare, Robert, “Changes in Educational Attainment and School Enrollment,” in R. Farley (ed.), State of the Union, Vol. 1, Russell Sage Foundation, New York, 1995. McMurrer, Daniel, Isabel Sawhill, and Robert Lerman, “Welfare Reform and Opportunity in the Low-Wage Labor Market,” Opportunity in America, No. 5, The Urban Institute, Washington, D.C., 1997. Meyer, Bruce, and Dan Rosenbaum, “Welfare, the Earned Income Tax Credit, and the Labor Suppy of Single Mothers,” NBER Working Paper, 1999. Meyer, Bruce, and Dan Rosenbaum, “Making Single Mothers Work: Recent Tax and Welfare Policy and Its Effects,” NBER Working Paper, 2000. Michalopoulos, Charles, and Christine Schwartz, with Diana AdamsCiardullo, What Works Best for Whom: Impacts of Twenty Welfare-toWork Programs by Subgroup, prepared for the U.S. Department of Health and Human Services, 2000. Moss, Phil, and Chris Tilly, “Soft Skills and Race,” Working Paper, Russell Sage Foundation, New York, 1996. 147 Murnane, Richard, and Frank Levy, Teaching the New Basic Skills, Free Press, New York, 1995. Nightingale, Demetra, and Kelly Mikelson, An Overview of Research Related to Wisconsin Works (W-2), The Urban Institute, Washington, D.C., 2000. Pavetti, LaDonna, How Much More Can They Work? Setting Realistic Expectations for Welfare Mothers, The Urban Institute, Washington, D.C., 1997. Pavetti, LaDonna, “What Will the States Do When Jobs Are Not Plentiful?” in S. Danziger (ed.), Economic Conditions and Welfare Reform, W. E. Upjohn Institute for Employment Research, Kalamazoo, Michigan, 1999. Pavetti, LaDonna, Michelle Derr, Jacquelyn Anderson, Carole Trippe, and Sidnee Paschal, The Role of Intermediaries in Linking TANF Recipients with Jobs, prepared for the U.S. Department of Health and Human Services, 2000. Polivka, Anne, “Notes on the Possible Effects of Welfare Reform on Labor Market Activities: What Can Be Gleaned from the March CPS,” Unpublished manuscript, Bureau of Labor Statistics, Washington, D.C., 1998. Primus, Wendell, Lynette Rawlings, Kathy Larin, and Kathryn Porter, The Initial Impacts of Welfare Reform on the Incomes of Single-Mother Families, Center on Budget and Policy Priorities, Washington, D.C., 1999. Quint, Janet, Kathryn Edin, Maria Buck, Barbara Fink, Yolanda Padilla, Olis Simmons-Hewitt, and Mary Valmont, Big Cities and Welfare Reform: Early Implementation and Ethnographic Findings from the Project on Devolution and Urban Change, Manpower Demonstration Research Corporation, New York, 1999. 148 Rangarajan, Anu, Keeping Welfare Recipients Employed: A Guide for States Designing Job Retention Services, Mathematica Policy Research Inc., Princeton, NJ, 1998. Raphael, Stephen, “The Spatial Mismatch Hypothesis and Black Youth Joblessness: Evidence from the San Francisco Bay Area,” Journal of Urban Economics, Vol. 43, 1997, pp. 79–111. Raphael, Stephen, Michael Stoll, and Harry Holzer, “Do Suburban Employers Discriminate More Against African-Americans?” Journal of Urban Economics, forthcoming (2000). Regenstein, Marsha, and Jack Meyer, “Job Prospects for Welfare Recipients: Employers Speak Out,” Assessing the New Federalism Occasional Papers, The Urban Institute, Washington, D.C., 1998. Stoll, Michael, Harry Holzer, and Keith Ihlanfeldt, “Within Cities and Suburbs: Neighborhood Composition and Employment Opportunities for Whites and Minorities,” Journal of Policy Analysis and Management, forthcoming (2000). Taber, Christopher, and Tricia Gladden, “Wage Progressions Among Less-Skilled Workers,” in R. Blank and D. Card (eds.), Finding Jobs: Work and Welfare Reform, Russell Sage Foundation, New York, 2000. Turner, Mark, “The Effects of Minimum Wages on Welfare Recipiency,” Final Report, The Urban Institute, Washington, D.C., 1999. U.S. Department of Labor, Futurework: Trends and Challenges for Work in the 21st Century, Office of the Secretary, Washington, D.C., 1999. U.S. General Accounting Office (GAO),Welfare Reform: Information on Former Recipients’ Status, Washington, D.C., 1999. 149 Vroman, Wayne, “Effects of Welfare Reform on Unemployment Insurance,” The New Federalism: Issues and Options for States, Series A, No. 22, The Urban Institute, Washington, D.C., 1998. Wallace, Geoffrey, and Rebecca Blank, “What Goes Up Must Come Down?” in S. Danziger (ed.), Economic Conditions and Welfare Reform, 1999. Welfare-to-Work Partnership, “Member Survey: Promotion and Partnership,” No. 1, Washington, D.C., 1999. Wiseman, Michael, “In the Midst of Reform,” Assessing the New Federalism Discussion Paper, The Urban Institute, Washington, D.C., 1999. Zellman, Gail L., Jacob A. Klerman, Elaine Reardon, Donna Farley, Nicole Humphrey, Tammi Chun, and Paul Steinberg, Welfare Reform in California: State and County Implementation of CalWorks in the First Year, RAND, Santa Monica, California, 1999. 150 About the Authors HARRY J. HOLZER Harry J. Holzer is a professor of Public Policy at Georgetown University and a visiting fellow at the Urban Institute in Washington, D.C. He is a former chief economist of the U.S. Department of Labor. He is also a senior affiliate of the Joint Center for Poverty Research (University of Chicago and Northwestern University), research affiliate of the Institute for Research on Poverty (University of Wisconsin), and a national fellow of the Program on Inequality and Social Policy (Harvard University). He received his Ph.D. in economics from Harvard University in 1983. Dr. Holzer has written extensively on the labor market problems of minorities and the urban poor. His books include The Black Youth Employment Crisis (coedited with Richard B. Freeman, University of Chicago Press, 1986); What Employers Want: Job Prospects for LessEducated Workers (Russell Sage Foundation, 1996); and Detroit Divided (with Reynolds Farley and Sheldon Danziger, Russell Sage Foundation, 2000). MICHAEL A. STOLL Michael A. Stoll is an assistant professor of policy studies in the School of Public Policy and Social Research at UCLA. He is a former visiting scholar at the Russell Sage Foundation, and a research associate at the Center for the Study of Urban Poverty at UCLA. He received his Ph.D. in urban planning from MIT in 1996. Dr. Stoll has written on the interplay of labor markets, race, and geography. In particular, his work explores the explanations of the unemployment problems experienced by low-skilled workers in urban areas, especially blacks. His recent book is titled Race, Space and Youth Labor Markets (Garland Publishing, 1999). 151" } ["___content":protected]=> string(102) "

R 101HHR

" ["_permalink":protected]=> string(122) "https://www.ppic.org/publication/employers-and-welfare-recipients-the-effects-of-welfare-reform-in-the-workplace/r_101hhr/" ["_next":protected]=> array(0) { } ["_prev":protected]=> array(0) { } ["_css_class":protected]=> NULL ["id"]=> int(8059) ["ID"]=> int(8059) ["post_author"]=> string(1) "1" ["post_content"]=> string(0) "" ["post_date"]=> string(19) "2017-05-20 02:34:35" ["post_excerpt"]=> string(0) "" ["post_parent"]=> int(3134) ["post_status"]=> string(7) "inherit" ["post_title"]=> string(8) "R 101HHR" ["post_type"]=> string(10) "attachment" ["slug"]=> string(8) "r_101hhr" ["__type":protected]=> NULL ["_wp_attached_file"]=> string(12) "R_101HHR.pdf" ["wpmf_size"]=> string(6) "290763" ["wpmf_filetype"]=> string(3) "pdf" ["wpmf_order"]=> string(1) "0" ["searchwp_content"]=> string(227756) "Employers and Welfare Recipients: The Effects of Welfare Reform in the Workplace ••• Harry J. Holzer Michael A. Stoll 2001 PUBLIC POLICY INSTITUTE OF CALIFORNIA Library of Congress Cataloging-in-Publication Data Holzer, Harry J., 1957– Employers and welfare recipients : the effects of welfare reform in the workplace / Harry J. Holzer, Michael A. Stoll. p. cm. Includes bibliographical references. ISBN: 1-58213-057-4 1. Welfare recipients—Employment—United States. 2. Employer attitude surveys—United States. I. Stoll, Michael A. II. Title. HV95 .H566 2001 331.5'4'0973—dc21 00-051723 Copyright © 2001 by Public Policy Institute of California All rights reserved San Francisco, CA Short sections of text, not to exceed three paragraphs, may be quoted without written permission provided that full attribution is given to the source and the above copyright notice is included. Research publications reflect the views of the authors and do not necessarily reflect the views of the staff, officers, or Board of Directors of the Public Policy Institute of California. Foreword Welfare studies typically focus on the changing size and characteristics of the caseload. If the caseload shrinks, researchers analyze the reasons for the change—is it the economy or is it program incentives? In a similar vein, a recent PPIC study concluded that California’s caseload is declining less dramatically than the nation’s, and that some case types are actually on the rise. Another PPIC study concluded that California’s welfare recipients have lower skills than other adults in the state and across the nation and that their long-term employability depended heavily on the health of the California economy. This study takes a different approach. Rather than tracing the reasons for caseload increases or decreases, it examines the welfare-towork transition from the employer’s perspective. Are employers willing to hire welfare recipients? If so, for what kinds of jobs and at what levels of compensation? How well do these employers think these workers are doing their jobs? And finally, what might these answers mean for the long-term prospects of welfare recipients moving into the job market? Funded by PPIC and the Joyce Foundation, Harry Holzer and Michael Stoll interviewed 750 employers across four metropolitan areas—Los Angeles, Chicago, Cleveland, and Milwaukee. Although the authors remain cautious about the long-term prospects of welfare recipients moving into the labor market, most of their results are encouraging. The boom economy of the late 1990s increased the job vacancy rate in all four cities, and employers actively hired welfare recipients. Although these workers often faced problems associated with job accessibility, their employers consistently reported that their job performance was as good as or better than other workers doing the same work. Their turnover rates were high but no higher than those of other employees. Less encouraging was the fact that the hiring rates for minorities were lower than those for whites. iii The authors also found that the situation in Southern California differed in some ways from what they observed in the Midwest. For example, Los Angeles had a relatively low hiring rate for welfare recipients. This was in part due to a softer economy in Los Angeles, but it also reflected higher skill requirements for the job openings in California compared to those in the Midwest. The average wage was also higher in Los Angeles than in the other three cities. According to the authors, however, all four cities have something in common—an economic downturn will affect their welfare caseloads. The study confirms that there is no better solution to welfare dependency than a good job, and there is no better road to a good job than a healthy, booming economy. But for many welfare recipients, a steady job is a bridge too far. Child care responsibilities, disabilities, poor training, and inadequate transportation are persistent impediments to steady employment. Publicly supported job training programs or supported work have long been a means to help the most needy cross the bridge to self-sustaining income. The authors revisit these policy options and suggest that they will be needed in even the most vibrant economy. The job experience gained by these welfare recipients is a net addition to human capital, one that is generating personal and public benefits today and will continue to do so in the coming decades. By examining the employment prospects of welfare recipients from a rare but critical perspective, this study helps rounds out our understanding of welfare reform even as it helps us monitor its consequences. David W. Lyon President and CEO Public Policy Institute of California iv Summary The recent reforms of the welfare system in the United States, beginning with state-level experiments approved under federal vouchers in the early 1990s and culminating in the 1996 passage of the Personal Responsibility and Work Opportunity Reconciliation Act, remain controversial. Since then, welfare rolls have declined dramatically and employment rates among recipients (and single females more broadly) have risen substantially, partly because of welfare reform and partly because of other factors, such as economic conditions and other policyinduced changes in net earnings and benefits for the working poor. But many questions remain regarding employer willingness and ability to hire from various groups of recipients, their performance and retention rates once hired, and the wages and benefits they receive from their jobs. To address these questions, we analyze data from a new survey of employers, administered to 750 establishments between July 1998 and June 1999 in each of four metropolitan areas: Los Angeles, Chicago, Cleveland, and Milwaukee. The survey focused on employer attitudes about hiring welfare recipients and their experiences with this population. Research Findings A number of important findings emerge from this analysis. For one thing, the current and recent demand for welfare recipients among employers is quite strong, at least in the aggregate. Overall hiring rates of recipients seem sufficient to absorb the numbers seeking employment in the current period and the near future. On the other hand, the rates at which recipients are hired are very sensitive to the job vacancy rates at establishments. This sensitivity suggests that during an economic downturn, a good deal of employer demand for recipients will diminish. Furthermore, although many establishments are open to hiring welfare recipients, actual hiring is much stronger at establishments that v are easily accessible to recipients, such as those near public transit stops or their neighborhoods. Also, recipients have more access to jobs when establishments are minority-owned or when employers have had contact with local welfare-to-work agencies, regardless of who initiates the contact. Limited access to jobs might help to account for the fact that minority groups in the welfare population appear to be hired less frequently than whites, relative to their representation in the welfare population. On the other hand, high school dropouts also seem to be hired less frequently than high school graduates, perhaps because the cognitive and social skills demanded on the jobs filled by welfare recipients are often not trivial. As for the quality of these jobs, those recently filled by welfare recipients in these metropolitan areas pay about $7.00 per hour on average and generally provide 40 hours of work a week. Employers contribute to the provision of health insurance in approximately twothirds of them. Still, significant fractions of these jobs pay low wages, provide few hours, or provide no health insurance. Furthermore, turnover rates for recently hired recipients are near the national average, and most recipients are rated as being as good or better than other workers in the jobs that they fill. But high turnover or weak job performance is a problem in a significant fraction (i.e., a fourth to a third) of cases. Absenteeism is particularly pervasive, often linked to child care and transportation issues. And the average skill levels and work-readiness of welfare recipients may be declining somewhat over time, as the part of the welfare caseload with more significant employment-related problems is entering the workforce now in greater numbers. Finally, there are some noteworthy differences between hiring patterns among employers in Los Angeles and those in the Midwest, particularly Milwaukee. In terms of turnover, job performance, absenteeism, and the like, the average quality of those recipients hired in Los Angeles seems higher than those in the Midwest and those in Milwaukee seem lowest. However, hiring rates have been relatively high in Milwaukee and low in Los Angeles. Thus, the better performance of recipients hired in Los Angeles seems to come at least partly at the vi expense of relatively more disadvantaged recipients who do not get hired there. The relatively limited hiring of welfare recipients in Los Angeles is partly attributable to labor market characteristics, such as the lower job vacancy rate there and the relatively greater skills demanded among the jobs filled. But it may also be related to the slower and less-aggressive implementation of welfare reform there, especially compared to the efforts undertaken as part of “Wisconsin Works” (W-2) in Milwaukee. Employers are significantly more likely to have had contact with a local welfare-to-work agency in Milwaukee than in Los Angeles or the other two metropolitan areas. For the United States overall, these findings suggest that there have been many positive employment outcomes associated with welfare reform to date, both for welfare recipients and for employers. However, some groups (e.g., minorities or high school dropouts) still appear to have difficulty gaining employment, and the bottom group of those being hired (i.e., the lowest one-fourth or one-third along most dimensions) are experiencing poor earnings or benefits when they are employed. The results we have observed to date might also deteriorate over time, as less-employable recipients enter the workforce or when an economic downturn occurs. Implications for Policy Several implications for policy at the federal, state, and local levels can be drawn from these findings. For one thing, the strong state of the U.S. economy and the relative tightness of the labor market have contributed to the ability of welfare recipients to find jobs. This implies that now is a very good time to invest in the training of recipients and in supportive services, such as child care and transportation. Any work experience gained by welfare recipients in this environment should lead to at least some wage growth and employment stability over time. On the other hand, some portion of this demand will no doubt disappear during the next cyclical downturn. Some recipients who cannot find employment at that time will be eligible for Unemployment Insurance, and others can return to the welfare rolls. For those welfare recipients who will not be eligible for Unemployment Insurance and who vii might run into time limits under Temporary Assistance for Needy Families (TANF), other arrangements will need to be made—such as changes in the Unemployment Insurance program, more community service jobs, or temporary suspensions of TANF time limits and work requirements for high-unemployment states. When the labor market slackens, the limited access of many welfare recipients (especially inner-city minorities) to suburban jobs may be more of a problem than it is now. Improving access through innovative approaches to transportation and job placement might become a more pressing need. For example, local intermediary agencies could become more involved in the job placement process for welfare recipients. Because relatively few employers in any given area have contact with local welfare-to-work agencies, improving their funding as well as their outreach to local employers might be worthwhile goals. Also, the need to improve retention and reduce absenteeism among recipients remains strong. Local workforce boards and agencies need to focus more broadly on retention issues as well as job placements. Given that skill requirements on jobs filled by welfare recipients are generally not trivial, and that many long-term recipients have poor skills and other personal characteristics that limit their employability, we may need to develop other employment options (such as community service jobs or “sheltered workshops”) as the most disadvantaged portion of the welfare caseload approaches time limits. Many other groups (e.g., those with criminal records) face barriers to employment even more severe than those faced by welfare recipients, especially in the attitudes of employers. Thus, efforts to promote more training, early employment opportunities, and income supplements to low-income (often noncustodial) fathers and other disadvantaged groups, such as those with criminal records, are critical. Finally, the implementation of welfare reform in Los Angeles will no doubt proceed and gain ground over the coming months and years. As efforts to place welfare recipients into employment expand there, greater efforts will be required to overcome a variety of barriers in the workplace (especially among immigrants) and to deal with any deterioration in employee performance that might develop. The active inclusion of local viii workforce boards and agencies in this process may be crucial to welfare reform’s success there. ix Contents Foreword ..................................... Summary..................................... Figures ...................................... Tables ....................................... Acknowledgments ............................... iii v xiii xv xvii 1. INTRODUCTION ........................... Recent Literature ............................. A New Study of Employers in Metropolitan Areas ........ Outline of This Report ......................... 2. WORKERS, EMPLOYERS, AND JOBS IN FOUR METROPOLITAN AREAS...................... Population and Worker Characteristics ............... The Employer Survey .......................... Employer Characteristics ........................ Labor Market Tightness and Skill Requirements ......... Employer Attitudes and Hiring Behaviors ............. Conclusion ................................. 3. THE DEMAND FOR WELFARE RECIPIENTS: PROSPECTIVE AND ACTUAL HIRING BY EMPLOYERS ............................... Prospective and Actual Demand for Welfare Recipients .... Variation in Demand by Characteristics of Employers ..... Effects of Employer Attitudes and Hiring Behavior ....... Conclusion ................................. 4. WHAT HAPPENS WHEN WELFARE RECIPIENTS ARE HIRED? .................................. Recently Hired Recipients: Who Are They? ............ What Kinds of Jobs Do They Fill? .................. Wages, Benefits, and Hours Worked ................ Worker, Employer, and Job Characteristics ............ 1 1 4 5 7 7 10 11 13 16 20 23 24 28 34 39 43 44 48 50 54 xi Turnover and Job Performance .................... Variation in Performance by Demographics and Time Period................................. Conclusion ................................. 5. CONCLUSION ............................. Summary of Findings .......................... Policy Implications............................ Preparing for a Downturn ...................... Improving Access to Jobs ...................... Improving Job Performance and Retention ........... Employment for the Hardest-to-Serve .............. Serving Low-Income Fathers .................... Appendix A. Regression Results ............................ B. Survey of Employers ........................... 57 66 74 77 78 80 80 82 82 83 84 85 91 Bibliography .................................. 141 About the Authors ............................... 151 xii Figures 2.1. Evidence of Labor Market Tightness .............. 2.2. Percentage of Employers Who Would Hire Disadvantaged Workers, by Disadvantage .......... 3.1. Prospective and Actual Hiring of Welfare Recipients .... 3.2. Prospective and Actual Hiring of Welfare Recipients in Central Cities and Suburbs .................... 3.3. Prospective and Actual Hiring of Welfare Recipients, by Vacancy Rate of Firm ....................... 4.1. Recently Hired Welfare Recipients and Poor Female- Headed Households, by Race .................. 4.2. Starting Hourly Wages Paid to Recently Hired Welfare Recipients ............................... 4.3. Hours Worked per Week by Recently Hired Welfare Recipients ............................... 4.4. Percentage of Recently Hired Welfare Recipients Still with the Firm ............................ 4.5. Months of Employment of Recently Hired Welfare Recipients ............................... 4.6. Performance of Recently Hired Welfare Recipients Relative to Others in the Same Position ............ 4.7. Percentage of Employers Observing Problems with Recently Hired Welfare Recipients ............... 4.8. Race of Recently Hired Welfare Recipients, by Year Hired.................................. 4.9. Starting Hourly Wages Paid to Recently Hired Welfare Recipients, by Year Hired ..................... 4.10. Performance of Recently Hired Welfare Recipients Relative to Others in the Same Position, by Year Hired.................................. 14 17 24 30 32 45 51 51 58 58 61 62 69 69 70 xiii Tables 2.1. Metropolitan Area Characteristics ................ 2.2. Central City Characteristics ................... 2.3. Characteristics of Establishments and Jobs .......... 2.4. Employer Skill Demands ..................... 2.5. Employer Attitudes and Hiring Behaviors .......... 3.1. Prospective and Actual Hiring of Welfare Recipients .... 3.2. Prospective and Actual Hiring of Welfare Recipients, by Industry Size, Type, and Minority Ownership ........ 3.3. Prospective and Actual Hiring, by Location Within Metropolitan Areas ......................... 3.4. Prospective and Actual Hiring of Welfare Recipients, by Skill Demands ............................ 3.5. Prospective and Actual Hiring of Welfare Recipients, by Employer Attitudes ......................... 3.6. Prospective and Actual Hiring of Welfare Recipients, by Employer Hiring Behavior .................... 4.1. Education and Work Experience of Recently Hired Welfare Recipients ......................... 4.2. Occupations Filled and Tasks Performed Daily by Recently Hired Welfare Recipients ............... 4.3. Starting Wages, Hours Worked, and Health Insurance Coverage for Recently Hired Welfare Recipients ...... 4.4. Characteristics and Wages of Recently Hired Welfare Recipients, by Employer Size and Industry .......... 4.5. Characteristics and Wages of Recently Hired Welfare Recipients, by Employer Location ............... 4.6. Characteristics of Recently Hired Welfare Recipients, by Type of Jobs Filled ......................... 4.7. Employment Duration and Turnover of Recently Hired Welfare Recipients ......................... 4.8. Absenteeism Problems of Recently Hired Welfare Recipients ............................... 8 9 12 16 18 25 29 31 33 35 36 45 48 52 54 55 56 59 62 xv 4.9. Performance and Employment Duration of Recently Hired Recipients, by Retention Status ............. 4.10. Turnover of Recently Hired Welfare Recipients, by Characteristics of Recipients ................... 4.11. Performance of Recently Hired Welfare Recipients, by Characteristics of Recipients ................... 4.12. Worker Characteristics and Job Performance of Recently Hired Recipients, by Month and Year Hired ......... A.1. Regression Results: Determinants of Prospective and Actual Hiring ............................ A.2. Regression Results: Determinants of Wages and Relative Performance ............................. 65 67 67 71 85 87 xvi Acknowledgments Many individuals and institutions contributed to this study in some way, shape, or form. We are very grateful to the Public Policy Institute of California and to the Joyce Foundation for providing funding for the survey. We also benefited greatly from the Urban Institute and the Russell Sage Foundation. In particular, we thank Lisa Kahraman at the Russell Sage Foundation for her valuable research assistance. We also thank Keith Ihlanfeldt, Jens Ludwig, and seminar participants at Harvard University, the Urban Institute, the University of Michigan and the Joint Center for Poverty Research for extremely helpful comments on previous versions of this study. At the Public Policy Institute of California, Michael Teitz provided excellent guidance at every stage of the project, read various drafts, and helped sharpen the focus of the study. Maureen Waller and Margaret O’Brien-Strain, also of the Public Policy Institute of California, Sheldon Danziger of the University of Michigan, and David Autor of MIT reviewed final drafts of this study and offered insightful and detailed comments. We thank all of these individuals for their hard work, which has greatly improved the quality of our report. Although many people contributed to this study, the authors remain solely responsible for its content. xvii 1. Introduction With the passage of the Personal Responsibility and Work Opportunity Reconciliation Act of 1996, along with earlier welfare reforms at the state level, the nation has embarked on a dramatic reshaping of its support system for low-income families. Supporters of these revisions in welfare predicted that the poor would enjoy improvements in living standards and self-sufficiency, with most entering the labor market and having little difficulty finding work (albeit sometimes at low wages). In contrast, many critics predicted dire consequences for adults and their children. In particular, the critics (e.g., Edelman, 1997) were concerned with possible high rates of unemployment, frequent job turnover, and low net wages and benefits among those leaving the welfare rolls, especially if they were forced off by time limits or sanctions. This study assesses some of these predictions by analyzing the outcomes associated with welfare reform in the workplace, as experienced by employers and the welfare recipients whom they have hired in the past few years. The analysis is based on data from a new survey of employers in four large metropolitan areas. Before considering these data and our findings in more detail, we review the evidence in the research literature to date, as well as the most important questions that remain to be answered about the effects of welfare reform in the labor market. Recent Literature Two facts are clear about the experiment with welfare reform to date: (1) the welfare (and Food Stamp) caseloads have declined dramatically since the early 1990s, and (2) employment among recipients has risen sharply as well. A variety of studies have analyzed these developments, using pooled time-series and cross-sectional data across states. Some disagreement remains about exact magnitudes, but virtually all these studies agree that welfare reform has contributed to both of these trends 1 without fully accounting for either. In addition, the strong labor market has certainly contributed to both developments, as have other factors, such as the expansion of the Earned Income Tax Credit and Medicaid programs, and perhaps recent increases in the minimum wage.1 Another set of studies has followed the experiences of recipients who have left the welfare rolls, and especially their labor market earnings. These studies of welfare “leavers” use either administrative or survey data from one or more states and are summarized in Loprest (1999). They generally show that, on the one hand, 70–80 percent of welfare leavers gain some employment within a year after departing the rolls; on the other hand, employment rates at any point in time are generally around 50–60 percent, indicating substantial unemployment as well. The extent to which this unemployment reflects high rates of turnover, lengthy spells between jobs, or labor force withdrawals is less clear.2 Furthermore, those studies indicate that earnings levels among former welfare recipients are certainly rising and that many families are better off than they were on welfare (e.g., Acs et al., 1999; Danziger et al., 2000). But low earnings also leave many families with working parents below the poverty line and still dependent on public assistance to varying extents (e.g., Cancian et al., 1999). This is particularly true for some groups, such as high school dropouts and those who have been sanctioned off the rolls. The sense that serious problems plague particular groups of current and former welfare recipients has been reinforced by evidence from Primus et al. (1999), who find declining real incomes for the bottom decile of female-headed families during the late 1990s. Employment and earnings are limited for some current and former welfare recipients by a variety of “barriers,” including low skills, health and mental health ____________ 1The relevant studies on changes in the caseloads include the Council of Economic Advisers (1997, 1999), Figlio and Ziliak (1999), and Wallace and Blank (1999). Studies of trends in female employment include Meyer and Rosenbaum (1999, 2000), Ellwood (1999), and Burtless (1999). For evidence that the minimum wage might actually raise employment and lower welfare dependency among single mothers, see Turner (1999). 2These rates of unemployment among welfare leavers are similar to those that have been observed in the past—e.g., Bane and Ellwood (1994). Since some fraction of those who leave the rolls do so because of marriage or other such household changes, it is not surprising that some leavers do not enter the labor market at all. 2 concerns, and a variety of other personal problems (Danziger et al., 1999). Thus, the evidence to date indicates rising employment and earnings among welfare recipients in general but serious labor market difficulties for at least some of these recipients. Furthermore, a number of important questions remain unanswered by these studies. They include: 1. What are the employment prospects of the least-skilled and least-experienced welfare recipients? Are employers willing to hire them at all, and to what extent? 2. Do black and Hispanic women on welfare face particular problems gaining employment, and if so, why? 3. To what extent do these employment prospects depend on the business cycle and the tightness of current labor markets in the United States? 4. Once welfare recipients are hired, are they plagued by high job turnover and poor work performance? If so, what are the causes of these difficulties in the workplace? 5. Are the earnings and benefits of welfare recipients sufficient to enable them to achieve financial independence over time? Many of these questions involve the characteristics and behavior of welfare recipients themselves but also the characteristics, attitudes, and behaviors of employers who may or may not hire them. In particular, answers to these questions depend on employers’ views of welfare recipients, especially those with very poor skills or work experience; the extent to which they have hired welfare recipients and expect to do so in the future; the extent to which some employers, especially those in smaller establishments or located far from downtown areas, are accessible to recipients residing in poor inner-city neighborhoods; and how well recipients meet the needs of employers in the jobs that they fill. To fully answer these questions, it is necessary to have some data directly from the demand side of the labor market—i.e., from employers and establishments where recipients are potentially or actually hired. To date, few such data have been available to researchers. In this report, we attempt to answer the questions by presenting and analyzing new data on 3 employers, particularly their experiences with and attitudes toward welfare recipients in the aftermath of welfare reform. A New Study of Employers in Metropolitan Areas To deal with the paucity of demand-side labor market data, a new survey was administered to over 3,000 employers in four major metropolitan areas: Los Angeles, Chicago, Cleveland, and Milwaukee.3 The survey builds on earlier work done on employers and their hiring of less-educated workers (e.g., Holzer, 1996) and earlier versions of this survey administered in Michigan (Holzer, 1999a). The survey focuses on employer willingness to hire welfare recipients, the extent to which they have actually done so, and what their experiences have been with recipients. The characteristics of the jobs filled by welfare recipients, their performance in those jobs, and the personal demographic characteristics of those hired are all documented. Any experiences that employers have had with local welfare-to-work agencies, and their openness to a variety of policy interventions designed to aid recipients, are considered as well. The four metropolitan areas considered here are of particular interest because of their different economic and demographic characteristics and policy environments. In particular, Los Angeles differs considerably from the three Midwestern areas in its size and geographic decentralization, its ethnic and immigrant composition (i.e., its heavily Hispanic and Asian populations), and other factors. The Midwestern areas also differ among themselves, with Chicago being much larger than the other two areas and more ethnically diverse as well. The welfare reforms implemented in these areas have clear similarities as well as differences. All of the state programs espouse a strong “work first” component that involves compulsory employment activities, job search assistance, sanctions for noncompliance, exemptions for certain family responsibilities and health problems, and time limits on ____________ 3We received funding from the Public Policy Institute of California (PPIC) for the data collection in Los Angeles and from the Joyce Foundation in the other metropolitan areas. PPIC’s primary focus on California and the Joyce Foundation’s primary interest in the Midwestern region of the United States account for the particular selection of metropolitan areas for this study. 4 recipiency. But given the size and ethnic diversity of the caseload in Los Angeles, implementation of its plan has been relatively slow and not always aggressive (e.g., Quint et al., 1999; Zellman et al., 1999). In contrast, the Wisconsin Works (or W-2) program is widely regarded as being the most aggressive in the country (Wiseman, 1999; Nightingale and Mikelson, 2000). It replaced the entitlement to income benefits with a four-tiered system of employment that rapidly pushes participants toward unsubsidized employment. Generous subsidized child care and transportation assistance are provided, along with a variety of other services. The drop in the welfare caseload in Wisconsin has also been dramatic, with almost 90 percent of the caseload eliminated between 1992 and 1998 (and about 75 percent in Milwaukee County).4 Using our survey data, we analyze the variation in hiring outcomes for welfare recipients across these four areas. Because establishment characteristics vary across these areas, we also examine the extent to which these characteristics account for differences in employment outcomes of welfare recipients across these areas. Finally, we consider differences in local demographics, geography, and policy, as well as their effects on these outcomes. Outline of This Report The outline of the remainder of this report is as follows. Chapter 2 reviews the demographic characteristics of the population in the four metropolitan areas under study as well as some general characteristics of employers and jobs that should be relevant for the analysis of demand for welfare recipients. The employer survey that generated the data for this study will also be discussed in more detail here. Chapter 3 considers data on the demand for welfare recipients. Two kinds of data are considered: prospective demand, or the self-reported willingness of employers to hire ____________ 4The caseload data are available from the Brookings Institution. Caseload reductions in Los Angeles County, Cook County (Chicago), and Cuyahoga County (Cleveland) from their peaks in the early 1990s have been approximately 20 percent, 29 percent, and 32 percent, respectively. In California, a relatively weak economy in the early 1990s and a continued inflow of immigrants caused the rolls to rise until 1995. Consistent with this, analysis of the Los Angeles Jobs-First Greater Avenues of Independence (GAIN) program has demonstrated significant employment increases for participants but fairly small reductions in the welfare rolls (Freedman et al., 1999). 5 recipients currently or in the future, and actual demand, or the hiring of recipients that has occurred to date. Many of the establishment characteristics associated with different levels of demand are considered here as well. Chapter 4 reviews data on what happens after welfare recipients are hired at these establishments. We focus on the most recently hired welfare recipients and consider a wide range of outcomes—demographics of the worker hired, characteristics of the job filled, wages and benefits earned, hours worked, job retention and duration, and performance ratings and problems. Finally, Chapter 5 summarizes the findings of our study and review potential implications for policy in California and elsewhere. 6 2. Workers, Employers, and Jobs in Four Metropolitan Areas Before analyzing the data on employer attitudes toward and experiences with welfare recipients, we begin by reviewing some of the characteristics of workers and employers in the four metropolitan areas: Los Angeles, Chicago, Cleveland, and Milwaukee. First, we review some data on the demographics of the population and employment outcomes of workers in these areas, derived from standard data sources (i.e., the Census and the Bureau of Labor Statistics). We then describe our new survey in greater detail and present summary data from the survey on the characteristics of employers and jobs located in these areas. In particular, we focus on factors such as the current state of overall labor demand (as measured by the job vacancy rate and difficulties reported by employers while seeking new employees), general demands for skills in these establishments, and various hiring practices. These metropolitan area and establishment characteristics will set the stage for the analysis of demand for welfare recipients and experiences with those hired that we present in subsequent chapters. Population and Worker Characteristics We begin in Tables 2.1 and 2.2 by presenting some general characteristics of the four metropolitan areas and their central cities, respectively. Not surprisingly, we find that: • Los Angeles is the most heavily populated of the metropolitan areas, and (along with Chicago) among the greatest in geographic size. • Los Angeles also has the largest minority population of any of these areas, with roughly two-thirds composed of blacks, Hispanics, and Asians and almost half (44 percent) 7 Table 2.1 Metropolitan Area Characteristics Area, square miles, 1990 Los Angeles 4,060 Chicago Cleveland Milwaukee 5,065 2,708 1,460 Population 1997 (March 1)a 1990 (April 1) Change, 1990–1997 9,145,219 8,863,052 .032 7,773,896 7,410,858 .049 2,225,997 2,202,069 .011 1,451,179 1,432,149 .013 Racial characteristics, 1997b White Black Hispanic Asian .340 .623 .758 .795 .083 .201 .171 .125 .439 .127 .047 .057 .127 .043 .010 .017 Population without high school diploma, 1990c .186 .157 .162 .130 White .482 .578 .708 .752 Black .095 .274 .269 .198 Hispanic .600 .205 .024 .054 Asian .074 .024 .005 .014 Households with poor female heads, 1990c White Black Hispanic Asian .034 .041 .042 .045 .327 .194 .325 .296 .323 .705 .640 .638 .472 .137 .044 .072 .057 .010 .004 .008 Metro area unemployment rate, October 1998b .065 .039 .042 .032 Estimated poverty rate, 1993d .238 .137 .150 .137 SOURCES: U.S. Bureau of the Census for relevant years unless noted otherwise. aMarch 1997 Current Population Survey, U.S. Bureau of the Census. bRacial categories include non-Hispanic white, non-Hispanic black, Hispanic, and non-Hispanic Asian. cRacial categories are not exclusive of Hispanic background. Hence, columns do not add to 100 percent. dU.S. Bureau of Labor Statistics, 1999. 8 Table 2.2 Central City Characteristics Los Angeles Chicago Cleveland Milwaukee Population, 1997 (March 1)a 3,553,638 2,721,547 498,246 590,503 Change, 1990–1997 .020 –.022 –.015 –.060 Population, 1997b White Black Hispanic Asian .329 .371 .497 .625 .102 .363 .326 .239 .461 .217 .165 .108 .105 .049 .003 .013 Estimated poverty rate, 1995 .286 .228 .299 .229 Change in jobs, 1993–1996c Central city Suburbs –.048 .020 .004 .045 –.047 .090 .084 .123 SOURCES: U.S. Bureau of the Census for relevant years unless noted otherwise. aMarch 1997 Current Population Survey, U.S. Bureau of the Census. bRacial categories include non-Hispanic white, non-Hispanic black, Hispanic, and non-Hispanic Asian. cDepartment of Housing and Urban Development (HUD) Special Tabulations of County Business Patterns, U.S. Bureau of the Census. composed of Hispanics alone. The central cities in these areas have much larger concentrations of black and Hispanic populations than do the overall metropolitan areas, with minorities accounting for half or more of central city populations everywhere but Milwaukee. Both employment and population growth in the suburbs exceed those of the central cities in each metropolitan area. • Minorities are somewhat more heavily concentrated among high school dropouts than among the overall populations, and much more heavily concentrated among poor female heads of households in all areas. The generally lower educational attainment of minorities than whites, particularly among Hispanics and especially among immigrants, accounts for their 9 greater representation in the population of high school dropouts (Mare, 1995) and the higher fraction of high school dropouts overall in Los Angeles.1 But the very strong majorities of poor female heads of households who are black in the Midwestern areas also reflect the much higher rates of female headship in this population group (e.g., Ellwood and Crane, 1990). • Unemployment and poverty rates are currently higher in Los Angeles than in the Midwest. As of October 1998, Milwaukee had the lowest unemployment rate of these metropolitan areas and Los Angeles had the highest, reflecting the generally tighter labor markets of the Midwestern region relative to that of California. Higher poverty rates in Los Angeles no doubt reflect a longer-term trend, perhaps because of the influx of poor and less-educated immigrants to this area (e.g., Borjas, 1994). The Employer Survey The focus of this study is the survey that was administered to roughly 750 employers in each of these metropolitan areas between July 1998 and June 1999.2 Employers were drawn from lists compiled by Survey Sampling Inc. (SSI), primarily from telephone directories. The sample was random within strata defined by establishment size, and the size strata were chosen so that the size distribution of establishments would resemble that of the overall workforce.3 In each establishment, we ____________ 1Given that overall population characteristics reflect data from 1997 whereas those for high school dropouts and poor female household heads are for 1990, the relative concentration of minorities in the latter groups might actually understate what actually exists today. 2The areas surveyed roughly correspond to the Primary Metropolitan Statistical Areas in each case. They include Cook, DuPage, and McHenry Counties in Chicago; Cuyahoga, Geauga, Lake, and Medina Counties in Cleveland; Milwaukee, Ozaukee, Washington, and Waukesha Counties in Milwaukee; and Los Angeles County in Los Angeles. 3In other words, large establishments are oversampled so that the percentages of establishments in different size groups would reflect the percentages of workers in those groups. This implies a sample of establishments that is already weighted by size to some extent. The size distributions used here were 20 percent in the 1–19 category, 30 percent in the 20–99 category, and 50 percent in the category of 100 and above. Since response rates in small establishments lag behind those of larger establishments, the actual 10 identified and tried to interview the individual responsible for entry-level hiring. Once this person was identified and contacted within these establishments, response rates averaged about 70 percent. The survey instrument, which can be found in Appendix B, was administered over the phone and took roughly 20 minutes to complete. Questions focused on the general characteristics of the employer, including size, industry, and location; composition of the workforce and of current jobs at the establishment, such as skill requirements and job vacancy rates; the demographics of and methods used to hire the worker in the most recently filled noncollege job; the number of welfare recipients hired in the past year or two and characteristics of the most recently hired welfare recipient and the job that he or she filled; experiences with any welfare-to-work agencies that were contacted; and prospective responses to any solicitations that might have come from such agencies. Employer Characteristics Table 2.3 shows some general characteristics of the sample of employers in each metropolitan area, such as their sizes, industries, and locations. The data indicate roughly similar distributions of employers across size and industry categories in these four areas. The results by size and industry indicate that retail trade and services account for about the majority of establishments in each area, and manufacturing accounts for 20 percent or less. The concentrations of employment in manufacturing are somewhat greater in Milwaukee and Cleveland than in Chicago or Los Angeles.4 Somewhat greater variance across metropolitan areas can be found in the locations of establishments. Several measures are presented here, ________________________________________________________ distributions of establishments across these categories are a bit more skewed toward the larger establishments. 4The formula for the standard error on the mean of a dichotomous variable is the square root of p(1 – p)/n, where p is the mean and n is the sample size. Standard errors on the means presented in this and the following chapters are generally in the range of .010–.020, and standard errors on differences in means across categories are generally .020–.025. Any differences discussed below are those that are at least marginally significant by conventional standards. 11 Table 2.3 Characteristics of Establishments and Jobs Firm size 1–19 20–99 100+ Los Angeles Chicago Cleveland Milwaukee .23 .22 .23 .19 .22 .21 .58 .55 .56 .24 .23 .53 Industry Manufacturing Retail trade Services .14 .16 .21 .22 .22 .22 .41 .38 .36 .20 .21 .39 Location Central city Suburbs 1/4 mile or less from public transit 30 min or less from downtown by public transit Both of above Distance from population on public assistance Distance from population not on public assistance Ratio .35 .65 .73 .58 .45 20.25 21.82 .92 .36 .73 .64 .27 .64 .65 .60 .76 .41 .56 19.67 15.75 21.05 19.19 .90 .79 .59 .41 .63 .71 .53 12.10 15.32 .74 Minority ownership .19 .14 .08 .08 Sample size 711 700 718 751 NOTES: Columns under “Firm size” and “Location (central city and suburbs)” sum to approximately one. Those under “Industry” sum approximately to 0.8, as other industries are omitted. Sample sizes are roughly constant throughout the remaining tables in this chapter. such as (a) the percentages of establishments located in the central city versus the suburbs, (b) the percentages located within one-quarter of a mile of a public transit stop, within a 30-minute ride from downtown using public transit, or both,5 and (c) the average distances in miles of ____________ 5The relationship of these locations to public transit is self-reported by the respondents at these establishments. 12 establishments from the populations receiving and not receiving public assistance, as well as the ratios of these average distances.6 The results also show that the majority of employers in Los Angeles and Chicago are in the suburbs, whereas most in Cleveland and Milwaukee are in the central city.7 In all areas, over two-thirds of employers are located near public transit stops, well over half are located within a 30-minute ride of downtown, and between 40 and 60 percent have both attributes. Interestingly, employers in Los Angeles are more likely to be near public transit stops than are those in other areas, reflecting that city’s expansive bus route systems; but the rides are somewhat lengthier, reflecting the greater geographic size and dispersion of the metro area there. As for distances from populations, employers are generally closer to the populations on public assistance than to those that are not; but on average, employers in Los Angeles are relatively farther away from these populations than are employers in the other metro areas.8 Finally, we find larger percentages of minority-owned establishments in Los Angeles than in the other metro areas. Labor Market Tightness and Skill Requirements One of the most salient features of labor markets in the United States today is the degree of tightness that has been achieved in the late 1990s without generating inflation (e.g., Katz and Krueger, 1999). ____________ 6These distances are weighted averages of the distances from the establishments’ Census tracts to every other Census tract in the metro area, weighted by the percentages of each population group that are in those other Census tracts, according to the 1990 Census of Population. For an illustration of these measures and how they affect the hiring practices of establishments, see Holzer and Ihlanfeldt (1996). 7The central city in Los Angeles is defined in such a way as to exclude the San Fernando Valley and to include East Los Angeles. 8The relatively greater proximity of employers to the low-income population parallels their greater proximity to minority groups (Holzer and Ihlanfeldt, 1996). But employment growth is often much more concentrated in outlying suburban areas that are farther away from the minority populations (Hughes and Sternberg, 1992; Raphael, 1997), thus contributing to “spatial mismatch” problems between the locations of employers and minority populations. 13 Tight labor markets have helped generate significant improvements in the employment rates of many groups, particularly minorities (Freeman and Rodgers, 1999). As noted above, economic conditions have also contributed to declining welfare rolls and to the apparent employability of welfare recipients. To gauge the current tightness of labor markets and how it varies within and among these four metro areas, we present several measures of such tightness from the employer’s perspective in Figure 2.1. The first, presented in Panel A, is the job vacancy rate at the establishment, defined as the percentage of all jobs that are vacant, that the employer is currently trying to fill, and that are available for immediate hire. In Panel B, we also show employer responses to questions about how difficult it currently is to find qualified applicants for their positions and how many employers had to fill jobs in the past two years with individuals who were less qualified than usual. Panel A Percentage of jobs vacant Los Angeles Chicago Cleveland Milwaukee 4.0 4.4 4.8 5.2 5.6 6.0 6.4 6.8 7.2 Percentage of employees who had difficulty finding qualified applicants Panel B Somewhat difficult Very difficult Percentage who hired less-qualified applicant in the past two years 10 15 20 25 30 35 40 45 50 55 60 Figure 2.1—Evidence of Labor Market Tightness 14 The results showed that: • Job vacancy rates currently are very high, ranging from just under 5 percent in Los Angeles to 6 percent in Milwaukee; • Roughly 70–80 percent of employers claim it is somewhat or very difficult to find qualified workers; and • One-third to almost one-half have had to hire less-qualified workers than usual in the past two years. All of these results are consistent with the view that labor markets are currently very tight in the United States. Although the more subjective data on hiring difficulties are harder to interpret without some comparisons over time, the job vacancy rate has traditionally been in the range of 1–3 percent and certainly lower than the unemployment rate at most points in time (Abraham, 1983; Holzer, 1989). Thus, it is quite striking to have conventionally measured mean vacancy rates actually exceed local unemployment rates.9 Also, the labor market of Los Angeles appears to be somewhat looser than those in all of the other metro areas, according to the measures examined here. Another potentially important determinant of employer willingness to hire welfare recipients is the overall level of skills required at their establishments. In the survey, we asked respondents about the percentages of all jobs in their establishments that do not explicitly require any particular level of education or experience at the time of hiring; and, of these, the percentages that also do not require the daily use of reading, writing, or arithmetic. We calculated similar measures of skill demand for the most recently filled noncollege jobs at these establishments. These data appear in Table 2.4. The results show that 30–40 percent of all jobs in these establishments do not require education or experience, and roughly 10 percent or more also do not require reading, ____________ 9It is noteworthy that the median job vacancy rate for this sample is just .02, well below the mean that appears in Figure 2.1. The distribution of vacancy rates reported here is heavily skewed to the right, with 5 percent of all establishments reporting vacancy rates of 25 percent or higher and 1 percent reporting them above 50 percent (although these are primarily small establishments). In contrast, about 37 percent of establishments report no vacancies at all. 15 Table 2.4 Employer Skill Demands Jobs that require No education or experience Also no reading, writing, or arithmetic Newly filled jobs that require No education or experience Also no reading, writing, or arithmetic Los Angeles .30 .11 .25 .14 Chicago Cleveland Milwaukee .40 .39 .40 .09 .10 .17 .26 .28 .14 .12 .26 .12 writing, or arithmetic. The generally lower percentages of jobs that require no skills among newly filled ones than among all jobs seem to reflect a secular growth in the demand for these skills among employers, as has been documented elsewhere (e.g., Katz and Murphy, 1992; Murnane and Levy, 1995). The effects of these trends in skill demands on the employment prospects of welfare recipients, who frequently have much lower-than-average cognitive skills and labor market experience, need to be considered in any analysis of employer demand for these recipients. Employer Attitudes and Hiring Behaviors The tendency of employers to hire welfare recipients in any given area should generally be influenced not only by local demographics and employer characteristics but also by their own attitudes toward different groups of workers and their hiring practices and procedures. Attitudes toward groups can encompass potentially discriminatory racial or gender preferences (e.g., Becker, 1971) as well as attitudes toward disadvantaged groups that are frequently disproportionately minorities.10 The relevant hiring practices can include a variety of recruiting and screening ____________ 10The relationship between attitudes toward race and those toward stigmatized groups has been demonstrated in Holzer and Neumark (2000), who show that firms engaging in Affirmative Action express a greater willingness to hire from the various stigmatized groups that we consider below. 16 practices, as well as a tendency to use various kinds of local labor market institutions or intermediaries when hiring (Holzer, 1996). In Figure 2.2 and Table 2.5, we present summary data on various self-reported employer attitudes toward disadvantaged workers, their use of particular hiring screens, and their contacts with local agencies. The attitudinal data in the figure reflect employer responses to questions about whether they would have been willing to hire various kinds of disadvantaged or stigmatized workers into the noncollege job in their establishment that they most recently filled. The categories of workers included welfare recipients, those who had been unemployed for a year or more, those who listed only short-term or part-time work experience, and those with criminal records. In each case, employers were asked whether they would definitely, probably, probably not, or definitely not hire such an individual into the position that was recently filled; either of the first two answers was counted as an affirmative response. Welfare recipient Criminal record Long-term unemployed Los Angeles Chicago Cleveland Milwaukee Only short-term part-time experience 20 30 40 50 60 70 80 90 100 NOTE: “Would hire” indicates that employers would “definitely hire” or “probably hire” someone with the listed attribute into the noncollege job that they most recently filled in their establishment. Figure 2.2—Percentage of Employers Who Would Hire Disadvantaged Workers, by Disadvantage 17 Table 2.5 Employer Attitudes and Hiring Behaviors Would do Drug test Background check Los Angeles Chicago Cleveland Milwaukee .39 .42 .58 .54 .45 .65 .44 .54 When hiring recipient, did Test (including drug test) Background check .45 .49 .36 .40 .25 .52 .39 .34 Had contact with local agency Of these: Employer initiated Hired referral .17 .30 .58 .20 .37 .54 .22 .39 .48 .26 .39 .66 NOTES: “Would do” indicates how employers would prospectively hire welfare recipients referred by a local agency; “When hiring recipient, did” indicates, for the subsample that had hired one or more recipients in the previous two years, actual hiring behavior; and “Had contact with local agency” refers to contact with a local welfare-to-work agency in the past year. The questions about screening used to generate Table 2.5 were limited to whether employers would (or did) use various kinds of tests or background checks on prospective applicants, particularly drug tests and checks for criminal records. The questions about prospective use of such tests were based on a set of questions (described more fully below) on whether employers would hire welfare recipients; another set was asked of those who had recently hired current or former recipients. Finally, a set of questions was asked on whether employers had had any contact with a local welfare-to-work agency, and, if they had, what were the outcomes of that encounter. Here we consider the frequency of such contact, which side initiated it, and whether the employer hired a referral of the agency. The results of Figure 2.2 and Table 2.5 indicate that: • Virtually all employers express a willingness to hire welfare recipients, whereas only about 40 percent express a willingness to hire those with criminal records; 18 • Drug tests or criminal record checks would (or are) being done by roughly 40–60 percent of employers in the hiring of recipients; and • Roughly one-fifth of all firms have had contact with local welfare-to-work agencies, and half or more of these resulted in a referral being hired by the employer. The relative preferences expressed by employers regarding the various disadvantaged groups parallel those found in an earlier study of employers (Holzer, 1996). Although employers are generally quite open to the hiring of welfare recipients or those needing remedial education or training (presumably so long as they met other hiring criteria and skill requirements), they are far more reluctant to consider those with criminal records, even in a very tight labor market.11 The fear of these workers does not bode well for the employment prospects of less-educated young black men, whose high rates of criminal activity and incarceration (e.g., Freeman, 1992) likely generate a reluctance to hire even those young black men without criminal records.12 Indeed, the employment rates of young black men did not grow at all between the cyclical peaks of the late 1980s and the late 1990s, and their labor force participation rates actually declined during that time period—in stark contrast to the dramatic growth in labor force activity that has been observed among young black females and especially single mothers (Aron et al., 2000).13 The greater tendency of firms to engage in background checks than drug tests may be somewhat surprising, given the ease with which the latter can be done. The frequency of contact with agencies also appears ____________ 11The effect of the tight market can still be found in these data; in comparison, in the looser labor market of the early 1990s, only about a third of employers said they would hire such workers (Holzer, 1996). 12See, for instance, Kirschenman and Neckerman (1991) or Kirschenman (1991). Given the imperfect information that many employers have about whether any particular individual has been engaged in criminal activity, they likely discriminate “statistically” against an entire group with high rates of incarceration, such as young black men. 13Although Freeman and Rodgers (1999) focus primarily on improvements in employment rates for less educated young black men over the business cycle during the 1990s, their data are consistent with these observations for the peak-to-peak period. 19 to be rising a bit over time, judging by some comparisons with earlier observations.14 This rise might reflect the very tight labor markets of recent years and a growing tendency of employers to use temporary agencies and other labor market intermediaries (Katz and Krueger, 1999). Finally, although the variance in these attitudes and behaviors across metropolitan areas is not great and shows no particularly consistent pattern, employers in Milwaukee show somewhat greater tolerance of previous criminal activity among their job applicants and greater willingness to use local agencies in the hiring process. In contrast, employers in Los Angeles are relatively less likely to accept candidates with criminal records and have had less contact with local agencies than employers in any of the other metro areas. Conclusion The usefulness of the preceding data will become more apparent in subsequent chapters, but several striking results emerged from this analysis. For one thing, labor markets appear to be extremely tight; mean vacancy rates may exceed unemployment rates in many locations, and most employers note the difficulties of finding qualified job applicants. Although substantial numbers of jobs are available that do not require particular levels of education or experience, most require some minimal cognitive tasks, such as reading, writing, or arithmetic. Further, the number of jobs not requiring skills appears to be dwindling over time. Although most employers are located near public transit, a large number are not easily accessible by short rides from downtown areas, and their proximity to low-income populations varies greatly. Finally, we also note the much greater willingness of employers to hire welfare recipients than other disadvantaged groups, such as young ____________ 14In Michigan during 1997, just under 17 percent of employers had such contact, although a larger fraction of these contacts resulted in welfare recipients being hired (Holzer, 1999a). Although the differences between these results and those presented in Table 2.5 could simply reflect the differences across areas rather than over time, it is noteworthy that the frequency of use in Michigan in 1997 lagged behind all of the rates observed in the other Midwestern metropolitan areas presented here. See Pavetti et al. (2000) for more evidence on the role of intermediaries in the hiring of welfare recipients. 20 men with criminal records. This fact is consistent with the contrast between the dramatic growth of labor force activity among young women (especially single mothers) in the 1990s and the fairly stagnant employment rates for black young men. Also, there has been a growing but still limited tendency of employers to have contacts with welfare-towork agencies in their local areas, as part of a broader trend toward greater use of intermediaries (including temporary agencies) in the labor market. 21 3. The Demand for Welfare Recipients: Prospective and Actual Hiring by Employers The previous chapter considered the characteristics of employers in Los Angeles and the other metropolitan areas, some of which might affect their willingness to hire welfare recipients. In this chapter, we look at data on the demand for welfare recipients by employers in these areas. We distinguish between prospective and actual (or realized) demand for welfare recipients; the former refers to an expressed willingness to hire welfare recipients and the latter measures whether employers have done so to date. Because prospective demand measures a self-reported attitude rather than actual behavior, its reliability might be subject to greater question than actual demand. On the other hand, it might also be a cleaner measure of employers’ demand for the labor of welfare recipients, as actual demand should reflect a mix of demand-side (i.e., employer) and supply-side (i.e., worker) factors that influence hiring. A comparison of outcomes observed for the two kinds of measures should therefore enable us to sort out the relative importance of demand-side and supplyside factors that might be inhibiting the hiring of recipients in some circumstances.1 We begin below by presenting summary measures of prospective and actual hiring of welfare recipients for each of the four metropolitan areas. We then consider how these measures vary by employer characteristics (i.e., employer size, industry, minority ownership, and location within the metropolitan area); measures of labor market tightness and employer skill requirements; and a variety of hiring attitudes and behaviors, ____________ 1See Holzer (1999a) for evidence on prospective and actual demand for welfare recipients from data collected about a year earlier in Michigan. 23 including the use of local intermediary agencies. We conclude the chapter with some discussion of policy implications. Prospective and Actual Demand for Welfare Recipients Figure 3.1 and Table 3.1 present means of several measures of employer demand for welfare recipients by metropolitan area. These include measures of prospective demand, or willingness to hire recipients either currently or over the next year, and of actual demand, indicating whether the firm has actually hired recipients in the previous year or two. The survey questions about prospective hiring begin with, “Suppose you were contacted by an employment agency that was trying to place welfare recipients. We are specifically talking about recipients who do not have a high school diploma or any recent work experience. Do you currently/over the next year have any open positions that you might consider filling with these welfare recipients?” The questions about actual Prospective Panel A. Employers Percentage who have hired recipients in the past year Actual Percentage who would consider hiring recipients currently 0 4 8 12 16 20 24 28 32 36 40 Panel B. Jobs Prospective Percentage that employers would consider filling currently with recipients Actual Percentage filled by recipients in the past year Los Angeles Chicago Cleveland Milwaukee 0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 Figure 3.1—Prospective and Actual Hiring of Welfare Recipients 24 Table 3.1 Prospective and Actual Hiring of Welfare Recipients Prospective hiring Jobs that employers would consider filling with recipients currently Jobs that employers would consider filling with recipients over the next year Los Angeles .013 .055 Chicago Cleveland .014 .020 .037 .037 Milwaukee .016 .048 Actual hiring Jobs filled by recipients in the past year Jobs filled by recipients in the past two years Sample size .015 .029 711 .019 .039 .034 .049 700 718 .030 .045 751 hiring are based on the question, “To your knowledge, has your business in the past year/two years hired any women who had been on welfare?”2 The measure of actual hiring behavior, of course, depends critically on whether employers really know which of their employees are welfare recipients; but most employers in the survey indicated that they were relatively certain of the welfare recipiency status of at least some of these employees when they indicated that they had hired recipients in the recent past.3 ____________ 2The prospective questions impose a stronger skill deficiency on the welfare recipients who would be hired but allow employers to merely consider hiring them. If any employers answer “yes” to either question, they are asked about the numbers that they have hired or would consider hiring. These numbers (including zeros for those who answer “no”) are then calculated as percentages of all current jobs in the establishment, whether filled or vacant. 3For instance, over half of these employers claimed that they were “definitely sure” that their most recently hired recipient had in fact been on welfare, and roughly 80 percent were “definitely” or “fairly” sure. Employers might have an incentive to find out about current or past recipiency to qualify for various federal or state tax credits for the hiring of recipients, although take-up rates on these credits are often quite low (Holzer, 1999a). More important, welfare recipients may need to inform their employers of their status for local Temporary Assistance for Needy Families (TANF) administrators to be able to confirm that they are complying with work requirements in the new program. 25 Figure 3.1 presents two measures of prospective demand: the percentage of employers who claim that they would consider hiring welfare recipients currently and the percentage of all jobs that could be filled by these recipients. Each measure is presented separately by metro area. The measures of actual demand are the percentage of all employers who claim that they have hired welfare recipients during the past year and the percentage of all jobs filled by them. Table 3.1 presents data only on jobs filled but for a broader range of prospective and actual hiring categories: jobs that employers would consider filling either currently or in the next year, and those that they have actually filled in the past year or two. The data indicate that there is substantial demand for welfare recipients in these metropolitan areas, both prospectively and in terms of actual hiring. We find that: • About a third of employers claim that they would consider hiring recipients currently and would potentially fill 1–2 percent of all current jobs this way. Over half also claim that they would consider doing so over the next year and would fill roughly 4–5 percent of all jobs this way.4 • 30–40 percent of employers also claim that they believe they have hired welfare recipients over the past two years, and 18–24 percent believe that they have done so in the previous year. As percentages of current jobs, these new hires would constitute roughly 3–5 percent in the previous two years and roughly 2–4 percent in the previous year. These latter measures reflect total hiring activity (or the flow of workers over the particular period), whereas the prospective measures capture a level of employment (or the stock of workers) at any point in time. In other words, the quantity of new hires over the past year would ________________________________________________________ Finally, firms are likely to have good information about the recipient status of workers hired through welfare-to-work agencies. 4Questions about hiring over the next year presumably involve expectations not only of the welfare recipients and their characteristics as potential employees but also of the employer’s expectations regarding his/her hiring needs over the next year, based on anticipated turnover and net employment growth. 26 have to be adjusted for turnover, which would generate a stock of employment for welfare recipients of over 1–2 percent of all jobs over the previous year.5 These data suggest that the overall level of employer demand for the labor of welfare recipients is quite substantial, as has been suggested by Burtless (1999) and others. To more fully assess the magnitude of demand, we note that, as of 2000, the welfare rolls had fallen by over two million adults since 1994. Studies of welfare “leavers” (e.g., Loprest, 1999; U.S. GAO, 1999) suggest that 70–80 percent have experienced some employment within a year of leaving the rolls, and roughly 50–60 percent are employed at any point in time. This suggests that, out of an overall labor force of about 140 million workers, welfare recipients have recently added over 1 percent to the labor force, in addition to those recipients who had been employed earlier or who were employed but still on the rolls. Those who remain on the rolls and will still enter the labor force over the next few years could constitute an extra 1 percent as well. In comparison, Table 3.1 implies that 1–2 percent of all jobs have already been filled by recipients at some point in the past two years; and roughly 1.5 percent of all current jobs are potentially available for new hires right away, with significantly more available over the coming year. Thus, overall demand for recipients appears sufficiently large to handle their inflow into the labor market, at least in the aggregate. This assumes, of course, that the prospective demand translates into actual demand if the recipients actually appear as applicants to the firm. Prospective demand might overstate actual demand if employers find welfare recipients relatively deficient in skills and other attributes when they appear and apply for work, especially relative to other applicants; but the great difficulty that employers currently have in finding acceptable applicants (as noted in Chapter 2) renders this concern less serious, at least in the context of the very tight labor markets of the late 1990s. Alternatively, employer responses to questions about hypothetical willingness to hire welfare recipients might reflect some desire to appear ____________ 5This calculation assumes a turnover rate of 40 percent, which transforms the 2–4 percent gross hire rate into a 1.2–2.4 percent rate net of turnover. The turnover rate is based on calculations that appear in Chapter 4. 27 “politically correct.” But it is not clear why this desire would surface in a confidential survey and why it would apply more to welfare recipients than to other kinds of disadvantaged applicants.6 Some variation in the levels of prospective and actual demand is also apparent across metropolitan areas. In particular, we find that employer demand for recipients is relatively low in Los Angeles and higher in Cleveland and Milwaukee. These differences across areas will be further explored in the analyses presented below. Variation in Demand by Characteristics of Employers There are no doubt other dimensions along which the demand for welfare recipients varies. Table 3.2 presents measures of demand by various employer characteristics, such as establishment size, industry, and whether the establishment is minority-owned. We limit ourselves to the variables that measure the percentage of all jobs filled rather than the percentage of employers who would fill them. The results of Table 3.2 show that both actual and prospective demand for welfare recipients is higher in the retail trade sector and in minority-owned businesses than elsewhere, whereas demand is particularly low in manufacturing establishments. The greater inclination of minority-owned businesses to hire welfare recipients parallels their greater tendency to hire minorities, all other factors being equal (Bates, 1998; Raphael et al., 2000). Differences by industry and establishment size might well reflect variables such as skill requirements, which are considered below; evidence of lower skill needs in small establishments appears in Brown et al. (1990), and the recent rise in skill requirements in the manufacturing sector has been linked to technological innovations there (Berman et al., 1994).7 ____________ 6As we noted in Chapter 2, employers were clearly willing in this survey to reveal their reluctance to hire applicants with criminal records. 7Larger firms are more likely to have hired at least one recipient, given their larger volume of hiring activity; but the jobs so filled constitute a smaller percentage of all of their jobs than they do in smaller establishments. 28 Table 3.2 Prospective and Actual Hiring of Welfare Recipients, by Industry Size, Type, and Minority Ownership Prospective hiring Jobs that employers would consider filling with recipients currently Jobs that employers would consider filling with recipients over the next year 1–19 Size 20–99 100+ Industry Manufac- Retail turing Trade Minority Ownership Service Yes No .030 .015 .008 .009 .022 .016 .080 .042 .026 .032 .070 .039 .024 .015 .073 .041 29 Actual hiring Jobs filled by recipients in the past year Jobs filled by recipients in the past two years Sample size .047 .022 .016 .008 .042 .034 .069 .032 .029 .016 .056 .057 648 1,014 1,143 504 601 1,054 .042 .022 .070 .033 341 2,414 Some further evidence of variation in employer demand for recipients by location appears in Figure 3.2 and Table 3.3. We analyze measures of employer demand for welfare recipients by whether employers are in the central city or suburbs (in Figure 3.2); whether they are “near” public transit, defined as being within a quarter mile of a transit stop and also within a 30-minute ride of the downtown area; and whether they are located relatively “near” the public assistance population, defined as establishments in which the ratio of their distance to this population relative to the nonassisted population is below the mean (Table 3.3). The results indicate that willingness to hire welfare recipients does not vary substantially by the employer’s location within the metropolitan area, whereas actual demand varies considerably. More specifically, firms located in suburban areas or in areas that are less accessible to welfare recipients are just as willing to hire them as their counterparts who are more accessible, but they actually hire fewer of them. These data indicate that spatial and transportation factors affect the ability of recipients to access potentially available jobs, as has often been suggested in the literature on “spatial mismatch” (Kain, 1992; Ihlanfeldt Prospective Percentage of jobs that employers would consider filling currently with recipients Central City Suburb Actual Percentage of jobs filled by recipients in the past year 0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 Figure 3.2—Prospective and Actual Hiring of Welfare Recipients in Central Cities and Suburbs 30 Table 3.3 Prospective and Actual Hiring, by Location Within Metropolitan Areas Prospective hiring Jobs that employers would consider filling with recipients currently Jobs that employers would consider filling with recipients over the next year Near Public Transit .015 .047 Not Near Public Transit .019 .048 Near Population on Public Assistance .015 .041 Not Near Population on Public Assistance .016 .046 Actual hiring Jobs filled by recipients in the past year Jobs filled by recipients in the past two years .032 .049 .011 .037 .023 .056 .019 .031 Sample size 1,183 1,344 1,018 1,239 NOTE: “Near public transit” refers to employers located within a quarter mile of a transit stop and a 30-minute ride from downtown; “near population on public assistance” refers to employers whose ratio of distance of public assistance population to nonassistance population is below the mean of the entire sample (0.83). and Sjoquist, 1998). These results mirror earlier ones involving the black population’s (Holzer and Ihlanfeldt, 1996) or the welfare population’s (Blumenberg and Ong, 1998) access to employers, as would be expected for groups that frequently lack automobile transportation or information about the broader metropolitan labor market.8 ____________ 8This interpretation assumes, of course, that the lower rate of hiring of welfare recipients in the suburbs reflects a relatively lower rate at which they apply for jobs there, as opposed to employers’ lower rates of hiring these applicants. What little evidence we have in the survey on the tendency of employers to accept or reject welfare recipient applicants is based on a question of how many recipient referrals they have received from agencies and how many they have hired. The data suggest a somewhat greater tendency of central-city employers to hire one or more of their referrals (.58 versus .46), which is partly accounted for by their relatively greater size. The differences in tendency to hire from the applicant pool, therefore, do not appear large enough to fully account for the observed geographic differences in actual hiring, as has also been noted in the literature on minority employment differences (e.g., Holzer and Ihlanfeldt, 1996). 31 These data also imply that there is some unmet prospective demand for these workers in geographic areas that are not accessible to the welfare population. In periods when most welfare recipients face few constraints on their ability to get job offers nearby when they seek them, the existence of this unmet demand may not matter greatly. But in other times or places where nearby job availability may be insufficient to ensure quick employment for these workers (e.g., when labor markets are not as tight, or where larger numbers of recipients enter the market quickly), improving the access of welfare recipients to suburban employers could be an important part of the relevant policy mix for policymakers and intermediaries. Figure 3.3 and Table 3.4 reveal two more dimensions of employer characteristics that might affect their demand for welfare recipients: labor market tightness and skill demands. The data compare employer demand for welfare recipients when their job vacancy rate is above and below the mean for these metropolitan areas (Figure 3.3), and when the percentage of unskilled jobs in their establishments (defined as jobs that Prospective Percentage of jobs that employers would consider filling currently with recipients Job vacancy rate below 5 percent Job vacancy rate at or above 5 percent Actual Percentage of jobs filled by recipients in the past year 0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 Figure 3.3—Prospective and Actual Hiring of Welfare Recipients, by Vacancy Rate of Firm 32 Table 3.4 Prospective and Actual Hiring of Welfare Recipients, by Skill Demands Prospective hiring Jobs that employers would consider filling with recipients currently Jobs that employers would consider filling with recipients over the next year Proportion of Unskilled Workers in Firm Below Mean Above Mean .014 .023 .032 .066 Actual hiring Jobs filled by recipients in the past year Jobs filled by recipients in the past two years .017 .031 .043 .061 Sample size 1,621 1,143 NOTE: “Unskilled workers” refers to the percentage of jobs in the establishment that require no specific levels of education or experience. require no education or experience) is above or below the mean (Table 3.4). The results in Figure 3.3 indicate that labor market tightness, as measured by the job vacancy rate, is a strong determinant of both prospective and actual demand for recipients. These differences are quite large—in particular, prospective demand for welfare recipients currently is five times as high in establishments with higher vacancy rates than in those with lower ones, and even over longer time periods, demand for recipients (both prospective and actual) is roughly twice as high in the highervacancy establishments. This pattern is consistent with the well-known tendency for the employment of less-skilled and less-experienced employees to be more sensitive to business cycles than that of moreskilled or more-experienced groups (e.g., Freeman and Rodgers, 1999). Not surprisingly, Table 3.4 shows that establishments with large percentages of unskilled workers have higher demand (both prospective and actual) for recipients than do those with lower percentages of these workers. Thus, both overall hiring needs and skill requirements at establishments are important influences on an employer’s willingness to hire welfare recipients. 33 These findings suggest that changes in overall labor market tightness or in the relative demand for unskilled workers over time could have important effects on job availability for welfare recipients. During an economic downturn, for example, the hiring rate for welfare recipients should decline substantially as well.9 Whether this affects their employment rates or income levels may depend on a variety of other factors (such as the severity of the downturn or how much work experience they have accumulated at that time), but the potential effects of a downturn should be a source of concern as we consider policy issues for this population. Effects of Employer Attitudes and Hiring Behavior In Chapter 2 we considered data on a variety of employer attitudes and hiring practices that might potentially affect their willingness to hire welfare recipients. These included (1) their willingness to hire workers with some type of stigma, such as a criminal record or limited work history (at least for the last noncollege job that they had to fill), (2) whether they would conduct drug tests or do background checks, and (3) whether they had contact with a local welfare-to-work agency, and if they did, whether they initiated such contact. Tables 3.5 and 3.6 analyze our four measures of demand for welfare recipients by each of these measures of attitudes or behaviors. We find that: • Both prospective and actual demand are higher among employers who are willing to hire job candidates with criminal records or weak employment histories; • Demand is higher among those who do not conduct drug tests or background checks; and • Actual demand is considerably higher among those who have had contact with local agencies, although prospective demand is not. ____________ 9Mean vacancy rates in recessions have generally been roughly 1–2 percent (Abraham, 1983; Holzer, 1989). Thus, the vast majority of employers would be distributed at the low end of the lower-vacancy category during a downturn. 34 Table 3.5 Prospective and Actual Hiring of Welfare Recipients, by Employer Attitudes Prospective hiring Jobs that employers would consider filling with recipients currently Jobs that employers would consider filling with recipients over the next year Criminal Record Yes No Would Hire Long-Term Unemployed Yes No Short-Term Experience Yes No . .020 .014 .017 .010 .018 .011 .051 .044 .048 .038 .051 .034 Actual hiring Jobs filled by recipients in the past year Jobs filled by recipients in the past two years .030 .026 .029 .022 .028 .021 .050 .039 .044 .034 .044 .031 Sample size 926 1,375 2,133 468 1,836 815 NOTE: “Would hire” indicates that employers would “definitely hire” or “probably hire” someone with the listed attribute into the noncollege job they most recently filled in their establishment. The fact that contact with agencies influences actual but not prospective demand indicates that these agencies might help transform potential into realized demand by bridging gaps in access of welfare recipients to available jobs for them. Interestingly, actual demand for recipients is higher when the agency has contacted the firm, although the basic finding holds up in both cases—when either the agency or the employer initiates contact, employer demand for recipients is still higher than when no contact occurs at all. Two other statistical or interpretive questions can be raised about these results. First, it is possible that the very subjective answers on questions about attitudes might be measured with considerable error, especially since the most recently filled jobs to which they refer vary a great deal. However, if the measurement error is random, it is likely to bias the results toward a finding of no difference across attitudinal 35 Table 3.6 Prospective and Actual Hiring of Welfare Recipients, by Employer Hiring Behavior Prospective hiring Jobs that employers would consider filling with recipients currently Jobs that employers would consider filling with recipients over the next year Drug Test Yes No .012 .019 .040 .048 Would Do: Criminal Check Use Agency Yes No Yes No Firm-Initiated Contact Yes No .013 .020 .016 .016 .021 .014 .041 .050 .041 .047 .042 .042 36 Actual hiring Jobs filled by recipients in the past year Jobs filled by recipients in the past two years .021 .036 Sample size 1,128 NOTE: See Table 2.5 for definition of “Would Do.” .030 .042 1,531 .026 .037 1,489 .027 .041 1,121 .037 .019 .060 .027 580 1,920 .036 .026 201 .038 .063 349 categories, which implies that the true differences are likely to be even greater than those observed here. A second question involves the extent to which we are measuring attributes of employers that really have a causal effect on the demand for welfare recipients, as opposed to merely reflecting that demand and its underlying determinants. For example, employers who conduct drug tests or do background checks may be less inclined to hire welfare recipients. Alternatively, drug tests or background checks may objectively screen out some welfare recipients from employment. In truth, we cannot really distinguish between the two interpretations of these results; but even if the effects are strictly correlative and not causal, it is useful to know the extent to which the hiring of welfare recipients is correlated with these attitudes. Given the large number of employer attributes that appear to be correlated with the demand for welfare recipients, it seems reasonable to investigate these issues through multiple regression analysis.10 In Appendix Table A.1, we present results from estimated regression equations in which the dependent variables are the percentage of jobs filled by welfare recipients, either prospectively or actually. The independent variables in the first specification are only dichotomous variables for the metropolitan statistical area of the employer; the second specification also includes dummies for one-digit industry, establishment size, minority ownership, location within the metro area (measured by distance to the public assistance population and to local transit stops), measures of labor market tightness (i.e., the job vacancy rates as well as whether employers have difficulty finding qualified workers), demand for unskilled workers, and agency contact.11 Comparisons between the first and second specification in each case will indicate the extent to which the regressors added in the second equation can account for differences across metro areas in hiring behavior observed in the first specification. ____________ 10Readers who are less familiar with or interested in these more technical research findings can skip to the Conclusion of this chapter. 11We have omitted the attitudinal and behavioral variables in Table 3.5 from these specifications, given the many questions that exist about their accuracy and exogeneity. 37 All equations are estimated using the tobit specification, given the large concentrations of hiring outcomes at zero.12 The results indicate that the relatively low prospective demand of employers for welfare recipients in Los Angeles is little explained by the regression equations, although their actual demand is partly explained. In particular, the lower job vacancy rate in Los Angeles, the greater distance of employers from the low-income population, and their lessfrequent contact with local welfare-to-work agencies all help to account for the relatively low rate of hiring welfare recipients there. Thus, a looser labor market and the geographic features of Los Angeles account for some of the lower tendency of employers there to hire welfare recipients. But the relatively less-frequent contact of employers with local agencies seems to be part of a broader phenomenon of slower and less-aggressive implementation of welfare reform there that has been noted elsewhere (e.g., Zellman et al., 1999). The enormous size and ethnic diversity of the welfare caseload there seems to partly account for the slower implementation, and certain provisions of the CalWorks program also allow much greater discretion to local officials in how strictly they enforce work rules.13 These features of CalWorks stand in sharp contrast to the aggressive Wisconsin Works (or W-2) program, in which entitlement to cash benefits without work has been eliminated and participation in the four-tiered employment program has been quite strictly enforced (Nightingale and Mikelson, 2000). Furthermore, almost all of the effects of industry characteristics on such demand continue to be observed here, suggesting that many of their effects are independent of one another. In particular, the effects of industry (relative to retail trade), the job vacancy rate, and the demand for unskilled workers continue to be significant determinants of both ____________ 12The tobit specification is most appropriate when there is a latent variable that is potentially negative but is “censored” by the zero limit on its values. In this case, such a latent variable might be net employer willingness to hire new recipients, which could be negative in a situation where employers are discharging or laying off more than they are hiring. 13For instance, the five-year time limit on welfare recipiency in California began in January 1998, later than virtually every other state; when sanctions are imposed, they limit the payments of benefits for adults but not for children in their households; and exemptions from work rules can be given more liberally than in most other states. 38 prospective and actual demand. Other factors, such as relative distance from populations receiving public assistance, contact with agencies, and minority ownership, affect actual demand more strongly than prospective demand. These factors suggest that the access of workers receiving public assistance to employers who might demand their services varies across establishments as well, and that improving workers’ access to a broader range of employers might be a useful goal for policy interventions in some cases. It also seems noteworthy that the effect of establishment size on demand, which appeared to be positive in earlier tables, now appears to be negative. In other words, once we control for industry, job vacancy rates, skill demands, and the like, small establishments are less likely to hire welfare recipients than are other employers. This seems consistent with other evidence (e.g., Holzer, 1998) that small firms hire fewer blacks than do others and are perhaps more discriminatory in their hiring behavior. Conclusion The results indicate that, in the aggregate, the demand for welfare recipients seems relatively high, and sufficiently so to absorb the number of recipients entering the labor market recently and in the near future. But there is also substantial variation across establishments in their demand for welfare recipients. Firms in retail trade, those with high job vacancy rates, and those with generally higher demand for unskilled workers have hired more welfare recipients to date (as percentages of their workforces) and are prospectively more willing to do so currently or in the near future. Those near public transit or populations receiving public assistance, are minority-owned, or have had contact with local agencies have also been more able to translate their prospective demand for recipients into actual hiring, indicating that these factors affect the access of welfare recipients to jobs that potentially exist for them in many places. Lower job vacancy rates and less frequent contact with local agencies account for some, but not all, of the lower hiring of welfare recipients in Los Angeles than in Milwaukee or Cleveland. The slower 39 implementation of welfare reform in Los Angeles than elsewhere likely helps to account for this lower level of activity among employers and agencies there as well. The findings above also imply that location, transportation, information, and contacts regarding potential job openings are key determinants of the access of welfare recipients to these jobs. These factors often limit the access of minorities (especially blacks) to jobs in the labor market more broadly, and they might limit employment opportunities for minorities who have also been welfare recipients. Indeed, data presented in the next chapter suggest that minorities have more difficulty gaining employment than white welfare recipients, and perhaps these factors play some role in their relative disadvantage. Although we will discuss implications for policy more fully in Chapter 5, at least a few are worth noting here. For one thing, the strong effect of the job vacancy rate on demand for recipients suggests that, once an economic downturn occurs, the hiring of welfare recipients might be significantly lower than it is today. For welfare recipients (and other lowincome workers) who do not qualify for Unemployment Insurance or TANF (perhaps because of time limits), some other type of “safety net” must be found—whether it is a special Unemployment Insurance program, suspension of time limits on TANF, or public employment programs. Likewise, in areas or time periods in which overall demand for labor is more slack, the differential access of recipients to potentially available jobs (because of location or transportation, information, etc.) may be more critical determinants of their ability to find jobs quickly or easily. In these cases, improving their access to a wider range of establishments could be an important policy goal as well. Because local welfare-to-work agencies and other intermediaries seem to be relatively successful at placing recipients into jobs, strengthening their role in the job-search process (perhaps through greater outreach of such agencies to the employer community or greater public funding of their efforts) might be an important component of a broader effort to improve access to jobs among current or former welfare recipients. The data in this chapter confirm the findings on self-reported employer attitudes toward welfare recipients that we observed in the 40 previous chapter—namely, that employers are very open to hiring welfare recipients. But we also learned above that employers are much more wary about hiring other groups of applicants who are stigmatized in their eyes, especially those with criminal records. As more and more lowincome men in poor neighborhoods have become incarcerated in recent years and are becoming (or will some day become) “ex-cons,” we need to make greater efforts to integrate them into labor markets and to overcome employer concerns over their skills, work experience, and general reliability. More broadly, the lack of recent labor market progress among less-educated young black men needs to be addressed as well. Beyond merely placing welfare recipients into jobs, we must also be concerned with the quality of these jobs—in terms of wages paid, benefits provided, and hours of work available—as well as the characteristics of recipients who obtain them and their ability to retain their employment and perhaps advance in the workplace. We turn to a fuller consideration of these issues in the next chapter. 41 4. What Happens When Welfare Recipients Are Hired? The previous chapter indicated that the demand for welfare recipients, both prospective and actual, is quite extensive, even though the access of recipients to many of these jobs is limited. This finding also raises a number of important questions about which welfare recipients are being hired, into what kinds of jobs, and with what kinds of observed outcomes. More specifically, are minorities within the welfare population being hired at the same rate as their white counterparts, and are those with little formal education or work experience being hired at the same rate as their more educated and experienced counterparts? Do the jobs that welfare recipients fill require a great deal of cognitive, social, or verbal skill? What wages and benefits do they earn, and with what hours of work? From the employers’ point of view, has their performance been satisfactory? Is job turnover a major problem? And how do these outcomes vary according to the characteristics of the workers hired or when they were hired? This chapter explores many of these issues. For those employers who report that they have hired at least one welfare recipient during the previous two years, we focus on the last welfare recipient hired and on the job that person has filled. We begin by examining the demographics of these workers, in terms of race, educational attainment, and recent work experience. We then consider the characteristics of jobs they have filled, along with wages, benefits, and hours worked. Finally, we consider a variety of measures of worker outcomes, such as turnover rates, particular problems such as absenteeism, and overall ratings of worker performance by the employer. These outcomes will determine the extent to which welfare recipients can obtain the kinds of stable employment and work 43 experience that can lead to earnings growth and perhaps self-sufficiency over time. Recently Hired Recipients: Who Are They? Figure 4.1 and Table 4.1 present data on the characteristics of the most recently hired welfare recipients in these establishments. The vast majority of these recipients were hired in 1998 and early 1999, although a small fraction were hired before then.1 We limit the sample to individuals whom the employers were definitely or fairly sure had been on welfare, either currently or in the past.2 Figure 4.1 presents the race of the most recently hired welfare recipients, along with Census data on the race of poor female heads of households (from Table 2.1). Table 4.1 presents data on whether welfare recipients were high school graduates (not including those with General Equivalence Degrees (GEDs)), and whether or not they had any recent work experience when they were hired. The data are presented separately for Los Angeles and each of the other three metropolitan areas from the Midwest. The results of Figure 4.1 and Table 4.1 indicate the following: • The majority of recently hired welfare recipients are black or Hispanic in each metro area. In Los Angeles, almost threefourths of the recipients were black or Hispanic, whereas over half of the recipients hired in Chicago and Milwaukee were black. Only in Cleveland do whites constitute nearly half of those hired.3 ____________ 1The percentages hired in 1997, 1998, and early 1999 are .15, .59, and .22, respectively. The rest (about .03) of the sample was hired before 1997. 2In roughly 42 percent of the cases, employers believed that these employees were still on welfare. In most of the remaining cases, employers believed that the employee had been on welfare in the recent past. 3Because of the smaller sample sizes of employers with recently hired welfare recipients (about 800 in all and somewhat fewer for variables with missing values), the standard errors on means of dichotomous variables in this chapter are generally .02–.04, and standard errors on differences are roughly .03–.05. 44 Los Angeles Percentage of most recent recipients hired Chicago Cleveland Milwaukee Los Angeles Percentage who are poor female heads Chicago Cleveland Milwaukee White Black Hispanic 0 10 20 30 40 50 60 70 80 SOURCE: 1990 U.S. Census. NOTES: “Most recent recipient hired” refers only to cases where employers are “definitely sure” or “fairly sure” of their employees’ recipiency status. The racial background characteristics of those in the category are mutually exclusive and sum to approximately one. The racial background characteristics of poor female heads are not exclusive of Hispanic background, and thus sum to a bit over 100. Figure 4.1—Recently Hired Welfare Recipients and Poor Female-Headed Households, by Race Table 4.1 Education and Work Experience of Recently Hired Welfare Recipients High school graduate Any recent work experience Sample size Los Angeles .69 .65 148 Chicago .60 .60 161 Cleveland .65 .70 214 Milwaukee .57 .68 225 NOTE: “Hired welfare recipients” refers only to cases where employers are “definitely sure” or “fairly sure” of their employees’ recipiency status. • Roughly two-thirds of the recipients hired have high school diplomas, and roughly two-thirds have some recent work experience. Somewhat surprisingly, the recipients hired in Los 45 Angeles have the highest rates of high school completion, despite the relative concentration of Hispanics and especially immigrants among them and the relatively low educational attainment among Hispanics nationwide (Mare, 1995). Educational attainment and work experience among recipients are relatively high in Cleveland and somewhat lower in Chicago and Milwaukee. • Although the hiring rates of minorities and those with little education or experience are significant, they appear to fall somewhat below the representation of these groups in the poor female household head or welfare populations of these areas.4 Comparing the hiring data by race in Figure 4.1 to those on poor female-headed households, the percentage of hired recipients who are black is lower than the percentage of poor female heads of households who are black in each metropolitan area in 1990 except for Los Angeles, whereas the percentage of hired recipients who are Hispanic in Los Angeles lags behind the comparable percentage of the poor female-headed household population there as well.5 Furthermore, the percentages of those hired who have no high school diplomas are lower than the comparable nationwide figure (43–44 percent) among welfare recipients at a point in time (Polivka, 1998). Although caseload data by race or education were generally not available to us at the county level in most cases, what data we did find for the ____________ 4Given that recent welfare hires in our data are from 1997 to 1999 and that poor female-headed households reflect data from 1990, if minorities have become much more concentrated in the latter category over time, these comparisons are likely to understate the underrepresentation of minority recipient hires relative to their representation of the poor female-headed household population. 5The employer and household data by race are not exactly comparable, since Hispanic ethnicity is not mutually exclusive with racial identifications of white or black in the Census household data (see the note to Table 2.1). However, since Hispanics in the Midwestern and Western region of the United States are vastly more likely to self-identify as white rather than black, the discrepancy between the two data sources could be roughly corrected by deducting the entire overlap (8 percentage points in Los Angeles and much less in the other metro areas) from the white category. This would not change any of the results described above. 46 largest counties generally support the notion of lower employment relative to welfare population among minority recipients.6 Thus, it appears that minorities and high school dropouts are being hired somewhat less frequently than whites or high school graduates within the welfare population. These results are consistent with evidence that the employment rates of those on welfare have recently been somewhat lower among blacks than whites and among high school dropouts (Danziger et al., 1999), and that welfare rolls have declined more slowly for blacks than others and in heavily urban and minority counties than in other areas (Katz and Carnevale, 1998). The data also indicate some variation across areas in the extent to which the least-skilled welfare recipients are being hired. The relatively high rate of high school completion among the recipients hired in Los Angeles, despite the high concentration of Hispanics and immigrants in that population, suggests that employers there have been somewhat less willing to hire the least-educated than those in Milwaukee and other areas. Among these recent hires in Los Angeles, the low level of immigrants, who often have English language as well as skill deficiencies, strongly supports this contention.7 Whether the lower rate of hiring high school dropouts in Los Angeles than elsewhere is due to differences in economic factors or to the policy environment is discussed more fully below. ____________ 6For instance, data for Los Angeles County indicate that Hispanics accounted for 53 percent of the welfare population there in 1997, and data for Cook (Chicago) and Milwaukee Counties indicate that blacks accounted for 74 and 73 percent, respectively, of welfare recipients there. All of these percentages are higher than the percentages of recent hires accounted for by these groups in our survey, although the population data would need to be adjusted for the racial compositions of the much smaller outlying counties for the comparison to be exact. Data on the racial composition of caseloads in the 20 largest counties are available from the Brookings Institution at www.brookings.edu/ es/urban/welfarecaseloads/2000report.htm. 7Immigrants account for just 14 percent of the welfare recipients hired in Los Angeles in our survey data, even though they accounted for roughly 40 percent of participants in the GAIN program in Los Angeles as of 1997 (Flaming et al., 1999). Some of this discrepancy, although probably not all of it, might be accounted for by lower participation rates in our survey of Spanish-speaking and informal employers in Los Angeles; unfortunately, we have no direct data on this issue. 47 What Kinds of Jobs Do They Fill? Table 4.2 presents evidence on the kinds of jobs filled by recently hired welfare recipients, by broad occupational category and by whether each of a set of tasks is performed daily on these jobs. The tasks include reading or writing reports, memos, or lengthy instructions; doing arithmetic (including making change); using a personal computer; speaking directly with customers either in person or on the telephone; filling out forms; or keeping a close watch on gauges, dials, or instruments. The tasks thus encompass a wide range of basic cognitive (or “hard”) skills as well as some social and verbal (or “soft”) ones (e.g., Moss and Tilly, 1996). The results indicate that welfare recipients most frequently fill clerical and service jobs; they fill fewer sales jobs and many fewer blue- Table 4.2 Occupations Filled and Tasks Performed Daily by Recently Hired Welfare Recipients Occupations Sales Clerical Craft Operator Laborer Service Los Angeles .20 .46 — .08 .01 .24 Chicago .15 .35 .04 .08 .06 .31 Cleveland .22 .20 .03 .09 .05 .40 Milwaukee .16 .32 .06 .13 .02 .31 Tasks Reading/writing Arithmetic Computer Serving customers Gauges/instruments Completing forms .62 .50 .58 .50 .52 .36 .83 .70 .27 .19 .67 .57 .56 .51 .40 .68 .28 .56 .52 .58 .41 .62 .28 .57 Sample size 143 115 150 166 NOTES: Columns under “Occupations” sum to approximately one. “Tasks” are those that must be performed on a daily basis. 48 collar jobs.8 The semi-skilled and unskilled blue-collar categories (e.g., operatives and laborers) generally account for about 15 percent of these jobs, although less than 10 percent in Los Angeles. In contrast, clerical jobs account for more of these hires in Los Angeles than elsewhere. Together, sales and clerical jobs account for 40 to 50 percent of all jobs filled recently by welfare recipients in the Midwestern areas and almost two-thirds in Los Angeles. The relatively high representation of clerical and sales positions among these jobs is consistent with the frequent use of various cognitive and social tasks on these jobs that we find in the bottom half of the table. For instance, dealing with customers daily is required in about 70 percent of the jobs in each metro area. Reading or writing fairly lengthy material or completing forms is required in over half of all jobs in each metro area, as is the use of arithmetic. Computers are used daily in about 40 percent of these jobs on average (and in over half in Los Angeles), although gauges and instruments are read less frequently. Comparing these job characteristics to broader samples of jobs filled recently in the survey that require relatively little education helps put them in context. Compared to the entire sample of recently filled jobs that do not require college degrees, fewer jobs filled by welfare recipients are managerial or blue-collar and somewhat more fall into the service category, although the daily tasks required by all such jobs are quite comparable.9 Compared to all noncollege jobs filled by women, those filled by welfare recipients include fewer managerial and clerical jobs, more service jobs, and jobs with somewhat lower task requirements.10 But when compared to jobs filled by women that do not require a high ____________ 8More specific occupations that are frequently filled include receptionist, cashier, food service, cleaning, and child care jobs. 9Among all recently filled jobs not requiring college degrees, the percentages in managerial and blue-collar occupations were .03 and .28, respectively, whereas service jobs accounted for .23. The percentages requiring customer contact, reading/writing, arithmetic, and computer use were .64, .54, .55, and .42, respectively. 10Of all noncollege jobs filled by women, .04 are managerial, .44 are clerical, and .21 are in services. The percentages requiring the tasks listed in the previous note are .78, .60, .59, and .53, respectively. 49 school degree, the task requirements on the jobs filled by recipients are actually a bit higher.11 Overall, the data indicate that the jobs filled by welfare recipients generally require at least some minimal amount of cognitive, social, or verbal skills. Recipients who have difficulty performing these tasks will have greater difficulty finding or keeping employment, as has been demonstrated elsewhere (e.g., Danziger et al., 1999). Given the high percentage of individuals with very poor cognitive skills among longterm welfare recipients (e.g., Burtless, 1995; Pavetti, 1997), and particularly the concentration of immigrants with English language difficulties and other skill problems in California (Johnson and Tafoya, 1999), the fraction of these women who should have difficulty becoming or remaining employed in jobs like these may not be trivial at any point in time.12 Furthermore, these low skills are likely correlated with other characteristics, such as disabilities and physical or emotional problems, that limit the kind of “job readiness” that virtually all employers seek when they hire workers (e.g., Regenstein and Meyer, 1998). These facts likely help to explain the high rates of nonemployment among former welfare recipients at any moment in time that we noted in Chapter 1. Wages, Benefits, and Hours Worked Although some of these jobs may be inaccessible to the least-skilled welfare recipients, the better occupations and frequent use of tasks might be associated with higher earnings and benefits or more regular hours of work. Figures 4.2, 4.3, and Table 4.3 present some evidence on these measures by metropolitan area. In particular, the figures present means and medians of starting hourly wages and hours worked, and the table presents frequency distributions on these measures as well as the ____________ 11The percentages in the tasks listed in the footnote above are now .59, .30, .40, and .23, respectively. The percentage of these jobs that are clerical falls to just .12 of the sample. 12 For instance, Danziger et al. (1999) report that about 20 percent of the welfare recipients whom they surveyed had used fewer than four out of nine job skills in previous jobs that they measured, and that the fraction of this group working at least 20 hours a week was much lower than among those with four or more skills (.34 versus .63). 50 Hourly wage, $ 8.5 8.0 7.5 7.0 6.5 6.0 5.5 5.0 Los Angeles Chicago Mean Median Cleveland Milwaukee Figure 4.2—Starting Hourly Wages Paid to Recently Hired Welfare Recipients 45 Mean Median 40 35 30 25 20 Los Angeles Chicago Cleveland Milwaukee Figure 4.3—Hours Worked per Week by Recently Hired Welfare Recipients Hours worked per week 51 Table 4.3 Starting Wages, Hours Worked, and Health Insurance Coverage for Recently Hired Welfare Recipients Starting hourly wages Mean Distribution $5.50 $6.00 $6.50 $7.00 $7.50 $8.00 Los Angeles Chicago Cleveland Milwaukee $7.83 $7.53 $6.94 $7.39 .01 .17 .32 .28 .43 .37 .54 .49 .57 .60 .63 .68 .16 .36 .49 .62 .75 .82 .11 .23 .38 .46 .58 .72 Hours per week Mean Distribution 15 20 25 30 35 40 Employer contributes to health insurance 34.4 34.8 .03 .08 .14 .16 .20 .19 .32 .24 .39 .32 .96 .95 .59 .72 34.0 .03 .14 .20 .35 .44 .95 .65 35.2 .03 .11 .21 .28 .34 .93 .72 Sample size 149 122 155 175 percentages of jobs in which employers contribute to the provision of health insurance for the workers or their families. The results indicate that: • Median wages are generally in the vicinity of $7.00 per hour, and mean wages are somewhat higher; • The median number of hours worked per week is 40; and • Roughly two-thirds of these jobs provide health insurance. The average starting wages earned by recipients and percentage working full-time are somewhat higher than has been reported elsewhere (e.g., Polivka, 1998; Holzer, 1999b; Loprest, 1999), although these data are consistent with other evidence on Milwaukee (Nightingale and 52 Mikelson, 2000).13 Wages and employer contributions to health insurance are lower on these jobs than on all noncollege jobs filled by women recently but somewhat higher in jobs that do not require high school diplomas.14 The frequency with which 35 to 40 hours are worked per week or where health insurance is provided suggests that most of these jobs are “regular” and provide stable hours.15 The use of the Earned Income Tax Credit and a variety of publicly provided subsidies for health care or child care to supplement these earnings would enhance their adequacy and attractiveness even more (Acs et al., 1999; Danziger et al., 2000), as would consistency of such employment over a significant period of time. Indeed, employment at the starting wages and hours per week that we have observed for the average newly hired welfare recipient would generate annual earnings of $12,000–$14,000 per year, or $16,000– $18,000 after including the Earned Income Tax Credit, if such employment could be sustained for the entire year. The issue of job retention and turnover is therefore critical, and we return to it below.16 On the other hand, the data also indicate significant fractions of jobs that pay lower wages and provide fewer hours of work or no health insurance. In fact, roughly one of every three jobs recently filled by a welfare recipient pays $6.00 per hour or less; roughly one-third provide 30 or fewer hours of work per week; and roughly one-third provide no health insurance. In fact, over half (54 percent) of the jobs filled by welfare recipients fall into one or more of these categories; and jobs that ____________ 13Other studies indicate median starting wages of about $6.00–6.50 for welfare recipients, whereas Polivka finds 45 to 47 percent of current or former welfare recipients working fewer than 35 hours per week. 14Mean and median wages on all noncollege jobs filled by women are $8.73 and $8.00, respectively, whereas those on jobs not requiring high school diplomas and filled by women are $6.87 and $6.00, respectively. The percentages with employer contributions to health insurance are .73 and .59, respectively, in the two categories. 15Employers may “provide” health insurance but still require a substantial contribution from employees. In some such cases, the employees might choose to forgo the insurance because of its high cost to them. 16The sufficiency of any given level of earnings will also vary across metropolitan areas with the cost of living, implying potentially greater material hardship for working recipients in Los Angeles than elsewhere. 53 are weaker along any one dimension are likely to be weaker along the others as well.17 Thus, overall earnings and benefits in a large fraction of these cases may not be high enough on their own to enable families to achieve selfsufficiency, as has been stressed by Edin and Lein (1997). Worker, Employer, and Job Characteristics To what extent do the demographics of the workers and their pay vary by establishment characteristics and occupational categories? Table 4.4 provides data on worker demographics, wages, and benefits by Table 4.4 Characteristics and Wages of Recently Hired Welfare Recipients, by Employer Size and Industry Race White Black Hispanic Asian 1–19 Size 20–99 100+ Industry Manufac- Retail turing Trade Service .35 .35 .24 .31 .39 .23 .41 .47 .53 .42 .43 .55 .20 .14 .17 .21 .14 .16 .01 .01 .01 .02 — .01 High school graduate .61 .61 .64 .54 .59 .64 Any recent work experience .69 .67 .64 .62 .70 .68 Starting hourly wages Mean Median $7.37 $7.10 $7.66 $7.79 $7.00 $6.75 $7.25 $7.50 $5.87 $7.67 $5.75 $7.47 Employer contributes to health insurance .55 .74 .63 .85 Sample size 80 205 295 102 NOTE: Columns under “Race” sum approximately to one. .50 .66 148 260 ____________ 17For instance, among jobs that provide 30 or fewer hours per week, median wages are just $6.00 per hour (as opposed to $7.50 for those with more hours) and 66 percent of jobs provide no health insurance. 54 employer size or industry; Table 4.5 presents these by the location of the establishment; and Table 4.6 provides these data by occupation. The data in Tables 4.4–4.6 indicate that: • Black welfare recipients are more likely to be hired in large establishments, in the service sector, and in establishments that are more easily accessible to minority populations (i.e., those located in central cities or near transit stops/low-income populations), and • Wages and the provision of health insurance are highest in large establishments and lowest in the retail trade sector. Both findings are consistent with other research evidence on where blacks are most likely to be hired (e.g., Stoll et al., 2000; Holzer, 1996, Table 4.5 Characteristics and Wages of Recently Hired Welfare Recipients, by Employer Location Race White Black Hispanic Asian Central City Suburb Near Public Transit Not Near Not Near Near Population Population Public on Public on Public Transit Assistance Assistance .22 .40 .22 .39 .61 .40 .58 .45 .16 .19 .18 .16 .01 .00 .01 .01 .21 .62 .15 .01 .36 .43 .19 .01 High school graduate .59 .68 .61 .64 .57 .67 Any recent work experience .66 .66 .68 .64 .67 .66 Starting hourly wages Mean $7.32 Median $7.00 $7.51 $7.00 $7.33 $7.00 $7.41 $7.00 $7.31 $7.00 Employer contributes to health insurance .67 .68 .67 .68 .67 Sample size 311 264 388 300 317 NOTE: Columns under “Race” sum approximately to one. $7.47 $7.00 .67 390 55 Table 4.6 Characteristics of Recently Hired Welfare Recipients, by Type of Jobs Filled Race White Black Hispanic Asian Clerical Sales Craft Operator Laborer Service .28 .30 .52 .30 .45 .30 .51 .53 .42 .44 .40 .55 .19 .17 .05 .26 .15 .15 .02 — — — — .01 High school graduate .77 .63 .65 .52 .40 .56 Any recent work experience .68 .67 .63 .56 .39 .67 Starting hourly wages Mean Median $8.35 $6.74 $7.97 $7.87 $8.00 $6.00 $8.00 $7.25 $6.58 $6.65 $6.25 $6.50 Employer contributes to health insurance .71 .54 .75 .87 Sample size 178 101 20 55 NOTE: Columns under “Race” sum approximately to one. .86 .60 21 175 1998) and on wage and benefit levels by establishment size and industry (e.g., Brown et al., 1990; Krueger and Summers, 1987). They suggest that there is significant variation across establishments in their accessibility not only to the population receiving public assistance but also to minorities versus whites within that population. Spatial factors, size, and industry all seem to play key roles in determining the availability of jobs to minority welfare recipients and the quality of these jobs. The data in Table 4.5 on welfare recipient demographics and job characteristics also suggest the following: • Blacks are hired most frequently into service jobs and Hispanics are relatively more likely to obtain semi-skilled blue-collar (i.e., operative) jobs; and • Wages are highest in clerical and craft jobs, although health insurance coverage is highest in the blue-collar jobs. 56 Thus, minority welfare recipients have mixed access to the betterpaying jobs, with some limitations because of the skill requirements of these jobs. On the other hand, the provision of health insurance in the blue-collar jobs to which Hispanic welfare recipients seem to have greater access than black recipients (even within metropolitan areas) suggests once again that access to good jobs varies across demographic groups in ways that are not fully explained by their relative skill attainments. Overall, cognitive skill requirements seem to pose greater barriers to Hispanic employment, whereas blacks are particularly underrepresented in the better blue-collar categories. Finally, a lack of a high school diploma and work experience prevents at least some welfare recipients from attaining higher-skilled jobs. Turnover and Job Performance Once welfare recipients are hired into jobs, how well do they succeed? Many authors have expressed concerns about issues of job turnover and performance among these recipients (e.g., Burtless, 1995; McMurrer et al., 1997). Indeed, high turnover has been observed among workers with low educational attainment, weak cognitive skills, and low wages—especially among women (Holzer and Lalonde, 2000). High turnover among welfare recipients could potentially lead to frequent and perhaps lengthy periods of unemployment, which would reduce their annual earnings substantially and weaken their ability to enjoy the wage growth that comes with work experience over time.18 In the next few figures and tables, we examine evidence on turnover and job performance among hired welfare recipients in Los Angeles and elsewhere. Figures 4.4 and 4.5 present data on the percentages of recently hired welfare recipients who are still with their employers and on their employment durations (as measured in months). Table 4.7 ____________ 18Taber and Gladden (2000) provide clear evidence that stagnant wage growth among less-skilled workers often results from their limited work experience over time; for those who manage to gain work experience, wage growth is comparable to the rates enjoyed by more-skilled workers. This limited work experience also reflects lengthy durations of nonemployment as well as the high frequencies of nonemployment that result from worker turnover (e.g., Clark and Summers, 1982; Ballen and Freeman, 1986). 57 Percentage 90 85 80 75 70 65 60 55 50 Los Angeles Chicago Cleveland Milwaukee Figure 4.4—Percentage of Recently Hired Welfare Recipients Still with the Firm 12 Mean 10 Median 8 6 4 2 0 Los Angeles Chicago Cleveland Milwaukee Months of employment Figure 4.5—Months of Employment of Recently Hired Welfare Recipients 58 Table 4.7 Employment Duration and Turnover of Recently Hired Welfare Recipients Los Angeles Still with the firm .80 Employment duration (months) All recently hired recipients Median 8 Distribution 3 .17 4 .24 5 .33 6 .38 7 .49 8 .52 Chicago .77 6 .18 .30 .37 .50 .57 .62 Cleveland .76 7 .22 .31 .39 .44 .50 .62 Milwaukee .66 8 .26 .28 .34 .41 .48 .52 Those still with the firm (incomplete spells) Median Distribution 4 5 6 7 8 86 .24 .30 .33 .37 .38 .50 .49 .57 .52 .62 7 .31 .39 .44 .50 .62 8 .28 .34 .41 .48 .52 Those no longer with the firm (completed spells) Median 76 Distribution 3 .32 .25 4 .42 .38 5 .42 .44 6 .42 .50 5 .37 .37 .53 .53 4 .45 .58 .70 .78 Of those who left the firm, reason Quit .70 .54 Discharge .20 .36 Laid off/job disappeared .10 .08 .65 .28 .06 .61 .33 .06 Sample size 167 177 224 241 NOTE: Columns under “Of those who left the firm, reason” sum to approximately one. 59 presents additional detail on these measures, and on the reason for separation among those who have left their jobs. Figure 4.6 also presents data on their performance relative to other workers in the same positions, as perceived by the employers, who were asked how the welfare recipients compare to the “typical one that you hire into this position.” Employers were also asked whether they had experienced each of a set of problems with the welfare recipients they had hired—such as absenteeism (and its particular causes), attitudes toward work, basic and job-related skills, substance abuse, and relations with coworkers (Figure 4.7). The causes of absenteeism problems are presented as well (Table 4.8). All of these results appear separately by metropolitan area. The results show the following: • On average, roughly three-fourths of the most recently hired welfare recipients are still with the firm. The retention rates are actually highest in Los Angeles (at 80 percent) and lowest in Milwaukee (at 66 percent). • Mean durations of employment to date average about eight months and median durations about seven months. Similar differences across metropolitan areas can be observed in these measures as in those described above. • High turnover and short employment durations can be observed among some subset of the recipients hired. For instance, a fourth to a third of recipients have had employment durations of four months or less.19 ____________ 19Incomplete spells of employment are truncated by the interview, which implies a downward bias in their average length. But they are, on average, longer than completed spells, which implies an upward bias in average length if one looks only at those spells. An unbiased estimate of overall spell length can be obtained by analyzing a set of completed spells that all begin within some time period (e.g., Bane and Ellwood, 1986). However, all spells begun in any time period in this study still contain some incomplete spells. The hazard functions estimated and described below incorporate the information from both kinds of spells while analyzing the determinants of employment duration. The data on incomplete spells (i.e., those in progress at the time of the survey) differ somewhat (and suggest longer periods of employment) than those on completed spells. But both of these measures need to be considered when analyzing turnover rates and employment durations. 60 Percentage 100 80 60 40 20 0 Los Angeles Chicago Cleveland Milwaukee Better Similar Worse Figure 4.6—Performance of Recently Hired Welfare Recipients Relative to Others in the Same Position • Welfare recipients are most frequently considered to be as good as or even better than other employees in similar positions. The percentage who are considered better than the typical employee in this position ranges from 40 percent in Los Angeles to 25 percent in Milwaukee. Roughly half are considered comparable in each metro area, and just 10 to 20 percent are considered worse. • Problems with absenteeism are experienced relatively frequently and are often related to child-care and transportation problems. The percentage of employers reporting problems with absenteeism ranges from about a fourth in Los Angeles to over half in Milwaukee. Child care seems to be the single greatest cause of absenteeism, being reported as a problem by over half of employers in each metro area. But transportation problems are frequently mentioned as well, especially in Los Angeles (by 58 percent of respondents), as are health problems (by about 30 percent on average). 61 Absenteeism Attitude toward work Basic skills Job skills Coworker relations Los Angeles Chicago Cleveland Milwaukee 0 5 10 15 20 25 30 35 40 45 50 55 60 Figure 4.7—Percentage of Employers Observing Problems with Recently Hired Welfare Recipients Table 4.8 Absenteeism Problems of Recently Hired Welfare Recipients Absenteeism Los Angeles .24 Chicago Cleveland Milwaukee .35 .39 .52 Due to Physical health Mental health Child care Transportation Domestic violence .26 .28 .07 .03 .63 .62 .58 .41 .05 .05 .35 .05 .58 .39 .07 .37 .06 .67 .39 .12 Sample size 162 176 223 239 NOTE: Columns do not sum to one, since multiple absenteeism problems could be listed by the employer. • Problems with attitudes toward work and basic skills appear somewhat less frequently. The percentages of employers reporting attitude problems range from 12 percent (in Los Angeles) to 23 percent (in Milwaukee), whereas problems with 62 basic skills or coworkers appear in about 10 to 20 percent of cases. Problems related to substance abuse or job-related skills appear even less frequently in these data. Overall, these data present a fairly positive picture of the performance of welfare recipients who are hired. For instance, average retention rates of almost 75 percent over this time period imply job turnover rates that are not dramatically higher than those of the overall workforce and certainly are lower than those of younger and less-skilled workers on average.20 Job performance among these workers is rated fairly positively relative to other workers, as has been noted elsewhere (e.g., Welfare-to-Work Partnership, 1999). Such data need to be interpreted very carefully. Comparisons with the employees most frequently hired into low-skill and low-wage jobs might well lead us to overstate the success of welfare recipients, as those employees often have their own turnover and performance difficulties (especially if they are young workers). What matters for welfare recipients is not how they compare with other unskilled workers but whether their retention rates and performance are good enough to lead to sufficient earnings levels and growth over time. At least for some of the workers considered here, retention and performance do not yet appear high enough to allay these concerns. Significant percentages of these recipients are having trouble keeping jobs for more than three to four months, and problems with absenteeism are quite frequent. On the other hand, not all job turnover is negative. Voluntary turnover or turnover from one job to another (as opposed to the transition from a job to nonemployment) is often associated with wage growth, particularly among the young (Taber and Gladden, 2000; Holzer and Lalonde, 2000). Because the majority of turnover here appears to be voluntary, the extent to which its effects are positive or negative will depend on whether and how quickly these individuals move ____________ 20Given that roughly eight months elapsed, on average, between the time of hire and the interview date, an average turnover rate of .26 in this period implies a rate of about .39. Anderson and Meyer (1994) report that 40 percent of jobs experience turnover in a year, and with multiple incidents of turnover for some jobs, a turnover rate of over 40 percent is implied. But much of the turnover appears to be heavily concentrated among young and inexperienced workers (Holzer and Lalonde, 2000). 63 to other jobs and also on the extent to which even voluntary turnover is associated with child care or related concerns. The differences in results between Los Angeles and the Midwestern metro areas, particularly Milwaukee, are also quite striking. Along virtually every dimension, outcomes for welfare recipients are most positive in Los Angeles and least positive in Milwaukee. The relatively high wages of workers in Los Angeles is consistent with the somewhat higher skills demanded there as well as the higher cost of living, as we noted above. But they are also consistent with the relatively low rate of hires for welfare recipients there compared to the other areas and the high rate observed in Milwaukee. In other words, employers in Los Angeles have not had to dig as deeply into the pool of welfare recipients there (in terms of worker quality) as in Milwaukee and the other areas because of a looser labor market and less-aggressive implementation of welfare reform. The implied tradeoff between the overall quantity and quality of welfare recipients hired also suggests that employers in Los Angeles may face more difficulties from these workers in the coming months and years, as reforms are implemented more forcefully, and driving a larger fraction of welfare recipients (especially less-skilled immigrants) into the workforce. Of course, questions might well be raised about the reliability of the results presented here, especially since the sample sizes of the duration measures in the tables and figures indicate many missing values on these variables; also, the performance measures are subjective and thus likely to be measured with considerable error. One way to ascertain the usefulness of these data is to see whether the different measures of performance are consistent with one another, which would imply that they contain somewhat more real information relative to random error. Table 4.9 presents evidence on performance measures by retention status with the firm and by reason for departure among those who have left. We also present evidence on median durations of employment by retention status and relative performance category. The results of Table 4.9 show that perceived performance of welfare recipients on the job is strongly related to retention status. Specifically, those who are still with the firm are rated much more positively than those who have left. In fact, over 90 percent of those still with the 64 Table 4.9 Performance and Employment Duration of Recently Hired Recipients, by Retention Status Performance relative to average Much better A little better Similar A little worse Much worse Median employment duration All recently hired recipients Those with Better performance Similar performance Worse performance Still with Not with Firm Firm All Quits Discharges .19 .07 .07 .21 .09 .11 .52 .44 .48 .06 .23 .18 .01 .17 .15 .05 .08 .29 .34 .23 7 54 8 99 7 44 8 43 5 9 4 5 Sample size 593 203 122 60 NOTE: Columns under “Performance relative to average” sum to approximately one. employer are rated similar to or better than typical employees in these jobs, whereas almost 40 percent of those who have left are rated worse. Furthermore, those who have been discharged are rated considerably worse than those who have quit, although even the latter are viewed considerably more negatively than those who stayed.21 Also, among those with completed spells, those with relatively better performance have considerably longer median durations of employment than those with similar or worse performance. Indeed, median durations of employment among those who have left the firm after performance that was similar to or worse than average is just four months. The consistency of performance measures with one another thus gives us some reason to believe that these data are capturing real dimensions of performance, at least to some extent. ____________ 21Of course, the relatively negative views of “quitters” may be reactions to their quits rather than unbiased assessments of their performance while working. 65 There may be other reasons as well to be concerned about measurement error or biases in one direction or another. For instance, the relatively high average retention rates (as well as the percentages of welfare recipients working full-time or receiving health insurance) might reflect the fact that this survey does not capture information on jobs for welfare recipients in the irregular or informal sector of the economy, thus leading to upward biases in all of these measures (or downward biases in turnover). Higher response rates to the survey among larger employers and older employers could have the same effect. Also, the large numbers of missing values on durations of employment could systematically lead to upward bias in retention or performance measures, if the less successful recipients on the job are less well-remembered by the employer (or if they chose to reveal their status as welfare recipients to employers less frequently). Of course, this latter bias could just as easily go in the opposite direction, if shorter-term employees are more recent and therefore more easily remembered. Furthermore, focusing on the last recipient hired might also generate downward biases in retention (or upward biases in turnover), since the most recent hire is more likely to be in a high-turnover job category than the average employee at the establishment. Thus, potential biases certainly exist here, but their net effects on the direction of our estimates as well as their magnitudes are unclear. Variation in Performance by Demographics and Time Period To what extent do observed measures of job performance differ according to the demographic characteristics of the welfare recipients hired? Tables 4.10 and 4.11 present data on job retention and performance measures, respectively, by race, education, and experience of the last welfare recipient hired. The results show at least some evidence of variation in performance by demographic group. For instance, durations of employment among whites are longer than those among minorities, even though blacks and Hispanics are more likely to still be working with their employers. Whites are somewhat more likely to be considered better than the typical 66 Table 4.10 Turnover of Recently Hired Welfare Recipients, by Characteristics of Recipients Still with the firm White .65 Race Black Hispanic .72 .80 High School Graduate Yes No .73 .75 Any Recent Work Experience Yes No .75 .73 Duration of employment (months) Mean Median 9.56 8 8.16 7 8.59 8.41 8.42 8.15 8.26 7 77 7 7 Sample size 173 296 97 442 264 496 252 Table 4.11 Performance of Recently Hired Welfare Recipients, by Characteristics of Recipients Rating of employee relative to others in this position Much better A little better Similar A little worse Much worse White Race Black Hispanic High School Graduate Yes No Any Recent Work Experience Yes No .16 .15 .21 .18 .44 .48 .10 .13 .08 .05 .18 .18 .16 .19 .12 .16 .16 .19 .18 .16 .58 .49 .51 .47 .55 .05 .09 .10 .10 .12 .02 .07 .03 .05 .04 Problems with Absenteeism Attitude Basic skills Job skills Substance abuse Coworker relationships .45 .21 .09 .07 .04 .18 .44 .20 .12 .10 .02 .16 .25 .44 .36 .40 .41 .10 .20 .16 .18 .18 .20 .16 .09 .10 .14 .05 .09 .09 .09 .10 .03 .03 .02 .03 .01 .07 .18 .13 .15 .15 Sample size 173 296 97 442 264 496 252 NOTES: Columns under “Rating of employee” sum to approximately one. Columns under “Problems with” do not sum to one since multiple problems could be listed by the employer. 67 employee than are minorities, although they are also more likely to be rated as much worse. Hispanics are the least likely group to show significant problems related to poor relative performance, absenteeism, and relationships with coworkers, although they suffer from greater problems with basic skills than do other groups. High school graduates are a bit more likely than dropouts to be rated positively and are less likely to have problems with absenteeism, attitudes, and basic skills. Recent work experience has little effect on observed outcomes, except that those without experience have more problems with basic skills. Although we see differences in performance across demographic groups, these differences are often modest. This pattern could reflect measurement error in the demographic variables (which will bias results toward a finding of no differences across groups) or selection effects at the hiring stage that tend to equalize the attributes of those who are actually hired across demographic groups.22 The higher retention rates but shorter employment durations of minorities than whites also suggest that, on average, the former groups were hired more recently than the latter. This might have occurred if, for example, the relatively more skilled or white recipients left the rolls or sought employment most quickly, whereas minorities or those with greater difficulties did so somewhat less quickly. Alternatively, employers might have chosen to hire white recipients first and then minority or lessskilled recipients only as labor markets grew increasingly tight over time and their hiring needs were increasingly unmet (as we saw in the chapters above). We analyze this last issue somewhat more systematically in Figures 4.8 through 4.10 and Table 4.12, which present race, wages, hours of work, and retention/performance measures by the time period in which the individual worker was hired. The figures compare those hired in 1997 with those hired in 1998–1999. In Table 4.12, we cut the data into three time periods: those hired between January and September of 1997, those hired between October 1997 and June 1998, and those hired ____________ 22In other words, the hiring screen eliminates more potential minority or high school dropout candidates with performance problems, so that those who are actually hired are more similar to their white counterparts than those in the general population. 68 Percentage 60 Black 50 Hispanic 40 30 20 10 0 1997 1998–1999 Figure 4.8—Race of Recently Hired Welfare Recipients, by Year Hired 8.0 Mean Median 7.5 7.0 6.5 6.0 5.5 5.0 1997 1998–1999 Figure 4.9—Starting Hourly Wages Paid to Recently Hired Welfare Recipients, by Year Hired Starting hourly wage, $ 69 100 80 Percentage 60 40 20 0 1997 1998–1999 Better Similar Worse Figure 4.10—Performance of Recently Hired Welfare Recipients Relative to Others in the Same Position, by Year Hired since July of 1998. The benefit of the latter cut is that it enables us to isolate the effects of other factors, such as the most recent increase in the federal minimum wage in October 1997, and still have two other periods of relatively similar length to compare to one another. Using either cut of the time period, we find that those hired relatively later in this time period are more likely to be minorities and to have lower median starting wages, shorter durations of employment, and lower performance ratings than those hired relatively earlier. Despite the increase in the minimum wage after October 1997, starting wages still deteriorated for those hired more recently, implying a potentially larger decline in the absence of that mandated wage increase. Also, the more recently hired workers are more likely to still remain with the employer, and their employment durations up to the survey date are shorter. This suggests that more welfare recipients have been hired relatively recently, closer to the interview dates of these firms, 70 Table 4.12 Worker Characteristics and Job Performance of Recently Hired Recipients, by Month and Year Hired Race Black Hispanic Month and Year Hired 1/97–9/97 10/97–6/98 7/98–6/99 .41 .57 .49 .11 .17 .18 Starting hourly wages Mean Median 7.56 7.47 7.36 7.50 7.00 7.00 Hours Median Below 30 40 40 40 .27 .27 .31 Still with the firm .58 .70 .78 Median employment durations All recently hired recipients Those still with the firm Those no longer with the firm 19 22 14 11 13 5 5 5 3 Performance relative to average All recently hired recipients Better .45 .36 .32 Similar .33 .43 .52 Worse .19 .20 .13 Those still with the firm Better .54 .46 .37 Similar .38 .43 .54 Worse .05 .11 .06 Those no longer with the firm Better .32 .14 .13 Similar .26 .45 .45 Worse .39 .40 .38 Sample size 62 132 399 NOTE: Columns under “Performance relative to average” sum to approximately one. which also implies an increasing rate at which recipients are hired over time (perhaps because more of them enter the labor force over time, or labor market tightness grows among employers leading them to hire 71 recipients at a faster rate).23 Since performance tends to be positively correlated with retention, a higher retention rate over time might suggest higher performance as well. But, within the samples of those who remained employed as well as those who did not, the decline in performance in more recent time periods is quite striking.24 The evidence thus suggests that the quality of welfare recipients hired has been falling over time, consistent with the notion that the easiest to hire were employed fairly quickly, whereas those with more problems made up a growing part of this pool of workers over time. For this growing group, outcomes will not be as positive as those observed for the overall sample of recently hired welfare recipients. Before concluding the chapter, we consider some evidence from multiple regression analysis on these topics.25 In Appendix Table A.2, we present results of regression equations in which we explore the determinants of three primary outcome measures presented above: wages, relative performance, and duration of employment.26 Three specifications are presented for each of the three outcome measures. In the first, we begin only with independent variables for the metropolitan areas (relative to the omitted category of Milwaukee); the second specification then adds worker demographics, establishment size ____________ 23In these data, the date of interview is strongly correlated with the date of hiring, so that the most recently hired workers are frequently in the most recently interviewed establishments. But the average time elapsed between date of interview and date of most recent hire is also declining with more recent interviews, implying a greater frequency of new hires on welfare over time. 24An alternative interpretation of these results is that the most recently hired welfare recipients entered the labor market at the same time as those hired earlier but experienced more job turnover and therefore have been hired more recently by their current employers. We have no way of distinguishing between these interpretations with the current data. 25Readers who are less familiar with or interested in these more technical research findings can skip to the Conclusion of this chapter. 26The wage equation uses the log of real hourly wages as the dependent variable, where wages have been deflated by the gross domestic product deflator for Personal Consumption Expenditures. For relative performance, the dependent variable is simply a dichotomous variable equaling one if the individual’s performance is deemed better than the typical employee in that position and zero otherwise. Finally, the duration equation is an estimated hazard function, using the Weibull functional form, which accounts for the truncation of duration values among incomplete spells. 72 and location, occupation, and task performance as independent variables; the third also includes industry dummies. Thus, we can ascertain the extent to which these observable characteristics of employers and their workers can help to account for average differences between Los Angeles and the other metropolitan areas in these outcomes. These regression equations indicate, as we noted above, that employers in Los Angeles pay higher real wages and have workers with higher average performance and longer durations of employment than do those in other metropolitan areas, particularly Milwaukee. Although much of the wage differential favoring Los Angeles can be accounted for by the characteristics of employers and workers, the Los Angeles advantage for the other dependent variables cannot be easily accounted for and might be attributed to the policy differences on implementation of welfare reform that we noted above. The regressions also confirm the effects of many groups of worker and establishment/job characteristics on observed outcomes. Thus, being a high school dropout and working in smaller establishments, in the laborer/service occupations, or in jobs with no computer use are associated with lower wages among welfare recipients.27 Reported relative performance is lower among minorities, higher for those with previous experience, higher in smaller establishments, and lower in establishments located relatively farther from populations on public assistance. Also, employment durations are lower for minorities, lower in smaller establishments, lower for those establishments located farther from populations on public assistance, and lower in the sales, blue-collar, and service occupations than in clerical or white-collar jobs. The results thus confirm many of our findings noted above and indicate that a wide range of personal, employer, and job characteristics influence the success of welfare recipients on the job. Overall, significant variation in these employment outcomes has been observed among welfare recipients, and this variation can be partly, but not fully, explained by these characteristics. ____________ 27The effect of a high school diploma is roughly .10 before including occupational and task dummies. The effect of a high school diploma on turnover is also somewhat reduced by including these dummies. 73 Conclusion The data in this chapter indicate some positive outcomes for welfare recipients who have been recently hired into the workplace. For example, significant numbers of minorities are being hired; many recipients wind up with fairly good jobs, and most work full-time; average wages (at least in these areas) are somewhat higher than previously reported, with a median of $7.00 per hour or more; and turnover as well as relative performance are often as good as or better than that achieved by other groups of low-wage workers. But certain findings also suggest some areas of concern. In particular (1) minorities and high school dropouts are not hired as frequently as might be implied by their representation in the welfare population, suggesting somewhat greater problems for these groups at the hiring stage, (2) skill requirements are not trivial in many jobs filled, indicating potential problems with hiring or retention for the very unskilled, (3) low wages and hours of work, high turnover, and poor performance plague significant percentages of those hired, and (4) absenteeism, especially related to child care and transportation, is a pervasive problem. Furthermore, the quality of recipients hired and their performance on the job have been deteriorating, suggesting that the relatively more employable recipients were hired more quickly and those with somewhat greater problems are now entering the workforce or being hired in greater numbers. Virtually all outcomes among hired recipients in Los Angeles are more positive than those observed in other metropolitan areas, especially Milwaukee. Although this is certainly good news at some level, it likely reflects the lower hiring rates of welfare recipients in Los Angeles that we noted in Chapter 3—especially those welfare recipients whose are less “work-ready”—and the slower implementation of welfare reform there. In contrast, the aggressive strategy of placing welfare recipients into jobs, which is central to the Wisconsin Works program, has resulted in somewhat greater employer difficulties with more marginal workers there. Eventually, as implementation of welfare reform proceeds, Los Angeles employers will need to grapple with particular skill deficiencies 74 plaguing welfare recipients (especially those who are immigrants), and other employment barriers recipients may face. In the next chapter, we discuss at greater length some of the implications of these findings for public policy, but a few of these implications are briefly noted here. In particular: • Transportation and child care problems remain barriers to successful job retention, if not to job placement, in large numbers of cases involving welfare recipients in the workplace. These issues, and performance/retention issues more broadly, deserve greater attention from state and local workforce boards and agencies in all of these locations. • The low wages and benefits earned by significant fractions of welfare recipients imply a continuing need to enhance their returns from work, perhaps through the expansion of the Earned Income Tax Credit and subsidies or credits for the provision of health insurance to workers and their families, at the federal or state levels. • Finally, the lower employment rates observed among recipients who are minorities or high school dropouts, as well as the nontrivial cognitive and social skills required on most of the jobs filled, suggest a need for greater skill development among the most disadvantaged recipients or perhaps more use of public service jobs to enhance their transitions to private sector employment. Both options are considered more extensively in the next chapter. 75 5. Conclusion The recent reforms of the welfare system in the United States, beginning with state-level experiments approved under federal vouchers in the early 1990s and culminating in the 1996 passage of the Personal Responsibility and Work Opportunity Reconciliation Act, remain controversial. Since then, welfare rolls have declined dramatically and employment rates among recipients (and single females more broadly) have risen substantially, partly because of welfare reform and partly because of other factors (such as economic conditions and other policyinduced changes in net earnings and benefits for the working poor). But many questions remain unanswered. For instance, are employers willing to hire even the most unskilled welfare recipients? Do minorities on welfare have fairly easy access to available jobs, including those located in suburban areas, and are employers as willing to hire them as white recipients? What role does the current strength of the economy play in those decisions, and what will happen during an economic downturn? What is the ability of recipients to retain jobs once they are hired, and what determines performance and retention rates? Finally, are the levels of pay, benefits, and hours of work sufficient for families to achieve independence? To address many of these questions, we have analyzed data from a new survey of employers, administered to 750 establishments in each of four metropolitan areas: Los Angeles, Chicago, Cleveland, and Milwaukee. The survey was administered by phone between July 1998 and June 1999 and focused on employers’ attitudes toward the hiring of welfare recipients and any experiences they have had to date with this population. In particular, we asked about the prospective willingness of employers to hire very unskilled recipients, the numbers actually hired in the past year or two, the characteristics and performance of the last such recipient hired, and any contact the employer might have had with a local welfare-to-work agency. 77 Summary of Findings Our most important findings can be summarized as follows: • The current demand for welfare recipients among employers is quite strong, at least in the aggregate. Employers express much greater willingness to hire welfare recipients than members of other stigmatized groups, such as those with criminal records. Hiring rates for welfare recipients in the aggregate seem sufficient to absorb the numbers of recipients seeking employment in the current period and the near future. • The rate at which recipients are hired is quite sensitive to the job vacancy rate across establishments, at a time when the aggregate job vacancy rate is quite high and labor markets are extremely tight. This suggests that a good deal of employer demand for recipients will diminish as well when these vacancy rates decline during an economic downturn. • Although many establishments are prospectively open to hiring welfare recipients, actual hiring is much stronger at establishments that are easily accessible to recipients, such as those near public transit stops or near their neighborhoods. Access to jobs for welfare recipients also seems related to minority ownership and to employer contact with local welfareto-work agencies, regardless of who initiates the contact. • Although many of the welfare recipients being hired are minorities, they appear to be hired somewhat less frequently (relative to their representation in the welfare population) than are whites, perhaps because of their lower access to jobs. High school dropouts are also hired less frequently than high school graduates, consistent with the fact that the cognitive and social tasks demanded on the jobs filled by welfare recipients are generally not trivial. • The median job filled by a recently hired welfare recipient in these metropolitan areas pays about $7.00 per hour and provides 40 hours of work a week. Employers contribute to the provision of health insurance in approximately two-thirds of these jobs. 78 But a significant portion of these jobs pay low wages, provide few hours, or provide no health insurance. • On average, turnover rates for recently hired recipients are near the national average, and most recipients are rated as being as good as or better than other workers in the jobs that they fill. But high turnover and weak job performance are problems in a significant fraction (i.e., a fourth to a third) of cases. Absenteeism is particularly pervasive, often linked to child care and transportation issues. • The average quality of welfare recipients hired also appears to be deteriorating somewhat over time, which implies that the part of the welfare caseload with more significant employment-related problems is entering the workforce now in greater numbers. • Finally, there are some noteworthy differences between hiring patterns among employers in Los Angeles and those in the Midwest, particularly Milwaukee. Those in Los Angeles hire fewer welfare recipients, but the average quality of those hired seems higher there. Thus, the better performance of recipients hired in Los Angeles seems to come at the expense of relatively more disadvantaged recipients who do not get hired there. Differences in employment outcomes across these metropolitan areas partly reflect differences in job vacancy rates or geographic accessibility of establishments, but other differences are not fully accounted for and may well reflect policy differences as well. All in all, these findings indicate many positive employment outcomes for welfare recipients in the workplace—such as strong demand for their work, average wages well above the minimum, significant health care provision, and job performance and retention that is above the average of many other less-skilled groups of workers. However, some groups (e.g., blacks and high school dropouts) still seem to face difficulties becoming employed, which likely contributes to the fact that over 40 percent of former welfare recipients are not employed at any point in time. Furthermore, at least some of those employed (i.e., the lowest one-fourth or one-third along most dimensions) experience a variety of serious difficulties on the job. Results over the coming months 79 or years might also deteriorate, as relatively less-employable recipients enter the workforce or when an economic downturn finally occurs. Before proceeding to policy implications that flow from these results, it is worth restating a few caveats about the data. Specifically, some of our results might be more positive than the real outcomes experienced by welfare recipients in the labor market, as our data likely miss the informal sector of the market and underrepresent smaller or newer establishments. But other potential biases (such as an overrepresentation of high-turnover jobs among the most recently hired) actually go in the other direction, and the net effects of these biases remain unclear. Policy Implications The discussion below describes the implications of these findings for policy at the federal, state, or local level regarding welfare reform in general, and especially for the goal of encouraging the success of welfare recipients in the labor market. Preparing for a Downturn There is no doubt that the strong state of the U.S. economy and the relative tightness of the labor market have helped welfare recipients find jobs. Because employers are very open to hiring them and working with local agencies to promote their success, this is a very good time to invest in the training of recipients and in supportive services (such as child care and transportation). Any work experience gained in this environment should help lead to wage growth and employment stability over time. On the other hand, at least some of this demand will no doubt disappear during the next cyclical downturn. Some recipients who cannot find employment at that time will be eligible for Unemployment Insurance, and many others will return to the welfare rolls (the growth of which will likely be financed out of TANF surpluses that many states are now accumulating). But some will not be eligible for either program— because of time limits in the case of TANF and insufficient prior earnings or other reasons in the case of Unemployment Insurance.1 The ____________ 1For a discussion of the Unemployment Insurance eligibility of welfare recipients see Gustafson and Levine (1998) and Vroman (1998). Most states require minimum levels 80 eligibility for either program among low-income males with little recent labor market experience will also be quite low, thus denying the children of these (often noncustodial) fathers an additional source of support. Just how many will fall into this category is unclear and will depend on factors such as the severity and duration of the downturn. Nevertheless, other arrangements will need to be made to preserve some “safety net” for such individuals. These might include: • A special Unemployment Insurance program for those who do not qualify for the usual program, as well as reforms in the Unemployment Insurance system that might make it easier for low-wage workers with limited experience to qualify; • More community service jobs for welfare recipients during this time period; and • Temporary suspensions of time limits or work requirements under the current TANF system, perhaps linked to state-level unemployment rates.2 This approach and its implementation raise several questions. For instance, exactly who might be eligible for the community service jobs or special Unemployment Insurance funds? Who would administer the program? How would state and federal funding be balanced? At a minimum, there needs to be more discussion of and planning for this period, at both the federal and state levels. Discussions regarding the reauthorization of TANF during the next year or two would be a perfect occasion for addressing these issues. ________________________________________________________ of earnings in a recent four-quarter “base period” to establish eligibility, with the current and most recent quarter excluded from that period. Many others are disqualified if they quit their jobs or were discharged for cause. Finally, those who work in the informal sector will generally not be covered by Unemployment Insurance programs. 2Reforms in Unemployment Insurance that might improve eligibility among lowwage workers include the use of an “alternative base period,” which includes the most recent quarter of earnings in calculations of prior earnings (which is potentially important for those who have entered the labor market recently); and providing some eligibility for part-time workers. A precedent for the development of a special Unemployment Insurance program can be found in the federal Special Unemployment Assistance program, funded by general revenues, that was put in place during the 1974–1975 recession. All of these possibilities are discussed in Vroman (1998). 81 Improving Access to Jobs When the labor market develops more slack, the relatively limited access of many welfare recipients (especially inner-city minorities) to suburban jobs may be more of a problem than it is currently. Improving access through innovative approaches to transportation and job placement might become a more pressing need. In fact, several federal initiatives (such as the “Bridges to Work” and “Access to Jobs” programs run by the U.S. Department of Housing and Urban Development) have already been implemented, as well as several at the state level. However, the scale of such efforts remains limited, and the problems associated with providing public transit access to outlying suburban employers remain substantial. We thus need continuing efforts to improve transportation to jobs by state and local agencies. More generally, the data presented here suggest that local welfare-towork agencies and other intermediaries can help recipients overcome some of the difficulties in gaining access to employers and achieving basic employability. But since relatively few employers in any given area have contact with these agencies, improving funding for their efforts and raising their outreach to local employer communities might be worthwhile goals. Improving Job Performance and Retention The need to improve retention and especially to reduce absenteeism among at least some current or former welfare recipients in the workplace remains strong. Improved access to reliable transportation and child care will certainly help deal with these problems. But local workforce boards and agencies need to focus more broadly on retention issues, rather than simply job placements for recipients. Approaches to job retention often include the provision of support services and counseling to both employers and employees in the workplace. Unfortunately, we have little clear evidence to date on successful models for improving job retention. The results in the Post Employment Services Demonstration (sponsored by the U.S. Department of Health and Human Services) have not been encouraging to date, though other strategies offer somewhat greater promise (Rangarajan, 1998). Continued experimentation with and evaluation of 82 different approaches in this area should remain a high priority for state and local agencies. Employment for the Hardest-to-Serve Given that skill requirements on jobs filled by welfare recipients are generally not trivial, and that many long-term recipients have poor skills and other personal characteristics that limit their employability, we may need to either improve their skills or develop various public sector employment options (at least in the short run) for the most disadvantaged portion of the caseload. The need to improve the skills of welfare recipients is sometimes seen as contrary to the “work first” notion that is embodied in most state-level welfare reform efforts. Indeed, programs that emphasize “work first” have generally proven to be more cost-effective than stand-alone training programs for this population. On the other hand, programs that combine some elements of both have often proven to be the most successful at raising the subsequent earnings of disadvantaged recipients (Michalopoulos et al., 2000). Public sector employment options might be useful for dealing with the most disadvantaged welfare recipients, as well as for time periods and local areas in which job availability is limited (such as during an economic downturn). These options might include various kinds of community service jobs or “sheltered workshops,” either as transition to the private sector or as ends in themselves. Although these approaches have largely fallen out of favor in recent years, Ellwood and Welty (2000) review the evidence on past programs and argue convincingly that such programs can be structured to provide useful public services, limited substitution for private sector employment, and subsequent gains for the individuals employed. Indeed, evidence on the Supported Work program for welfare recipients indicates strong positive postprogram effects on employment retention and earnings for very disadvantaged welfare recipients (Ham and Lalonde, 1996). And, although most states and local areas are not currently using these approaches, both New York City and Wisconsin 83 have developed relatively large-scale programs of public employment as part of their welfare reform efforts.3 Serving Low-Income Fathers Many other groups face barriers in the labor market similar to or even greater than those faced by welfare recipients. Indeed, employer reluctance to hire from some of these other stigmatized groups, such as those with criminal records, is substantially stronger than for welfare recipients, and they are the focus of much less assistance in the labor market today. Thus, efforts to expand various kinds of training and assistance to low-income fathers and other disadvantaged groups are critical (Committee for Economic Development, 2000). In fact, low-income young men and especially noncustodial fathers are now eligible to participate in local welfare-to-work programs as well as the new Youth Opportunity grants to high-poverty neighborhoods across the country (both funded by the U.S. Department of Labor), and strong outreach is needed to ensure their participation in these efforts. Given the very diverse and decentralized nature of these programs, strong evaluation efforts to identify effective interventions are critical as well. In addition, efforts to deal specifically with the employment issues faced by those with criminal records, who are rapidly becoming a much larger part of the population, need to become a higher priority. Finally, we should consider making the Earned Income Tax Credit and other wage/income supplements available to noncustodial fathers and other low-income men or women without children. ____________ 3Ellwood and Welty (2000) emphasize that New York’s Work Experience Program (“WEP”) has been a “workfare” model for welfare recipients, as opposed to the wageearning model used in Wisconsin and elsewhere. Nevertheless, both have the potential of providing needed work experience for recipients as they transition to the private sector labor market. 84 Appendix A Regression Results These tables present results from multiple regression equations that were estimated using the employer survey data. Table A.1 is discussed in Chapter 3 and Table A.2 in Chapter 4. Table A.1 Regression Results: Determinants of Prospective and Actual Hiring Metro area Chicago Cleveland Los Angeles Industry Agriculture Construction Manufacturing Transportation/communications Wholesale trade Finance Services Size 1–19 20–49 50–99 Prospective Demand: Current (1) (2) –.014 (.015) –.003 (.015) –.043** (.015) –.028* (.020) .001 (.017) –.047** (.020) — .027 (.070) — –.057* (.041) — –.111** (.022) — –.049* (.030) — –.077** (.033) — –.108** (.030) — –.062** (.017) — –.060** (.019) — –.047** (.018) — –.014 (.019) Actual Demand: Last Year (1) (2) –.043 (.024) .020 (.023) –.080 (.025) –.023 (.030) .033 (.026) –.050* (.031) — –.217* (.163) — –.133** (.070) — –.106** (.032) — –.178** (.053) — –.174** (.056) — –.042 (.043) — –.007 (.026) — –.076** (.028) — –.032 (.027) — .012 (.028) 85 Table A.1 (continued) Relative distance to public assistance population Distance to transit < .25 miles .25 – 1.0 miles Vacancy rate Difficulty finding qualified workers Percentage of employees unskilled Contact with agency Minority ownership Log L Prospective Demand: Current (1) (2) — .117** (.039) — –.003 (.020) — –.011 (.025) — .380** (.060) — .017* (.013) — –.031 (.376) — .027** (.016) — .009 (.020) 708.60 Actual Demand: Last Year (1) (2) — –.187** (.060) — — — — — — — 408.40 .037 (.032) .039 (.034) .215** (.093) .016 (.019) .030** (.004) .078** (.023) .063** (.029) 994.19 NOTES: Statistical significance: *p < .10, **p < .05, one-tailed test. Equations are estimated by tobit. Dependent variables in both cases are percentages of jobs, rather than employers. Reference groups for categorical variables are Milwaukee (metro area), retail trade (industry), 100 and above (size), and greater than one mile (relative distance to public assistance population). Standard errors are in parentheses. 86 87 Metro area Chicago Cleveland Los Angeles Industry Race Black Hispanic High school graduate General experience Size 1–19 20–49 50–99 Table A.2 Regression Results: Determinants of Wages and Relative Performance ln (Real Wages) (1) (2) (3) Better Performance Than Average (1) (2) (3) Duration (*Weibull Hazard Function) (1) (2) (3) .016 (.036) –.091** (.033) .051* (.033) No .021 (.046) –.065** (.038) .046 (.047) No .008 (.043) –.076** (.036) .016 (.044) Yes .122** (.048) .087** (.045) .173** (.049) No .118* (.080) .022 (.067) .205** (.081) No .133 (.079) .033 (.066) .211 (.080) Yes .457** (.259) .585** (.246) .569** (.245) No 1.034** (.391) .798** (.307) .848** (.348) No 1.012** (.381) .808** (.299) .755** (.336) Yes — .019 .005 — –.130** –.156** — –.523** –.564** (.035) (.033) (.062) (.062) (.301) (.304) — –.010 –.021 — –.124* –.145** — –.165 –.173 (.048) (.045) (.081) (.080) (.381) (.369) — .052** .037* — –.036 –.026 — .223 .187 (.031) (.029) (.055) (.055) (.261) (.259) — .001 .005 — .117** .136** — .010 –.051 (.031) (.029) (.054) (.054) (.244) (.244) — –.081** –.063* — .174** .151** — –.594** –.201 (.042) (.040) (.075) (.076) (.295) (.307) — –.118** –.089** — .084 .097* — –.363 –.105 (.039) (.038) (.070) (.071) (.309) (.301) — –.128** –.093** — .072 .049 — –.050 .280 (.042) (.040) (.073) (.073) (.359) (.394) 88 Relative distance Occupation Sales Clerical Craft Operative Laborer Service Tasks Read/write Arithmetic Computer Customer Table A.2 (continued) ln (Real Wages) (1) (2) (3) — –.029 –.005 (.087) (.082) Better Performance Than Average (1) (2) (3) — –.266** –.277** (.149) (.149) — –.129* .046 — –.065 –.091 (.091) (.091) (.139) (.149) — .055 .063 — –.074 –.028 (.087) (.082) (.133) (.132) — –.040 –.028 — –.109 .002 (.115) (.112) (.180) (.184) — –.077 –.057 — –.172 –.047 (.100) (.100) (.156) (.167) — –.125 .011 — .085 .091 (.120) (.116) (.189) (.196) — –.168** –.092 — –.082 –.087 (.087) (.082) (.133) (.133) — –.000 .010 — .039 .071 (.034) (.032) (.059) (.059) — –.004 .038 — –.044 –.063 (.032) (.032) (.056) (.058) — .063** .042 — .022 .018 (.038) (.036) (.067) (.067) — –.071** –.023 — .018 .006 (.041) (.040) (.071) (.073) Duration (*Weibull Hazard Function) (1) (2) (3) — –1.94** –2.02** (.736) (.724) — –1.194 –1.297* (.857) (.884) — –.652 –.640 (.838) (.823) — –1.57** –1.85** (.936) (.926) — –.661 –1.155 (.945) (.958) — –1.615* –1.75** (.993) (1.000) — –.110 –.063 (.857) (.840) — –.125 .005 (.265) (.259) — .278 .237 (.236) (.237) — .415* .499** (.273) (.277) — –.489 –.442** (.348) (.339) Table A.2 (continued) Forms Gauges R2 Log L ln (Real Wages) (1) (2) (3) — .024 –.015 (.037) (.035) — –.001 .010 (.034) (.032) .035 .205 .312 —— — Better Performance Than Average (1) (2) (3) — .031 .012 (.063) (.064) — .134** .152** (.059) (.059) .017 .102 .149 — —— Duration (*Weibull Hazard Function) (1) (2) (3) — .071 .026 (.289) (.285) — –.080 –.055 (.274) (.271) —— — 292.5 155.9 152.2 NOTES: Statistical significance: *p < .10, **p < .05, one-tailed test. The equation for “Better Performance Than Average” is estimated as a linear probability model. Coefficients in the hazard function can be interpreted as marginal effects of characteristics on the log of the expected duration of employment (in months). Reference groups for categorical variables are Milwaukee (metro area), white (race), 100 and above (size), and technical/managerial. 89 Appendix B Survey of Employers The actual text of the survey that was administered to employers by phone is reproduced below. U1 Before we begin, let me tell you that all the information that you give me will be kept strictly confidential. Should we come to any question that you do not want to answer, just let me know and we will go on to the next question. PART A: GENERAL BUSINESS QUESTIONS A2 (I’d like to start by asking you some general questions about your company.) Is this a for-profit or a nonprofit company? 1 FOR PROFIT 5 NONPROFIT 8 DO NOT KNOW 9 REFUSED A3 Does this company operate at more than one site? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED 91 A4a Is this a minority-owned company? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED A5a (I’d like to start by asking you a couple general questions about your organization.) About how far is your location from the nearest public transit stop? 0–9 TENTHS OF MILE 25 QUARTER OF A MILE 50 HALF OF A MILE 75 THREE QUARTERS OF A MILE 10 1 MILE OR MORE 97 NO PUBLIC TRANSPORTATION[goto A6a] 98 DO NOT KNOW[goto A6a] 99 REFUSED[goto A6a] A5b About how long do you think it would take for a worker to get to your business/organization from the center of the downtown business area if they used public transportation? 0–59 MINUTES 60 ONE HOUR OR MORE 98 DO NOT KNOW 99 REFUSED A6a Thinking about your customer or client base, what percent of your customers or clients are African American (or black)? 0–100 PERCENT AFRICAN AMERICAN 998 DO NOT KNOW 92 999 REFUSED A6b (What percent of your customers or clients are) Asian? 0–100 PERCENT ASIAN 998 DO NOT KNOW 999 REFUSED A6c (What percent of your customers or clients are) Hispanic? 0–100 PERCENT HISPANIC 998 DO NOT KNOW 999 REFUSED PART B: SITE EMPLOYMENT B1 All of the questions from now until the end of the interview refer only to the business (name) located at (address) (in city). How many employees currently work at (business name) (at this location?) 1–9996 EMPLOYEES AT SITE 9998 DO NOT KNOW 9999 REFUSED B2 How many people worked there approximately one year ago (one year ago would be (month) 1997)? 0–9996 EMPLOYEES YEAR AGO 9998 DO NOT KNOW 9999 REFUSED B4 What percentage of your current employees are covered by a collective bargaining agreement (belong to a union)? 93 0–100 PERCENT 998 DO NOT KNOW 999 REFUSED B6 How many of your (current) employees are in jobs that do not require any particular skills, education, previous training, or experience when they are hired? 0 EMPLOYEES[goto B7] 1–9996 EMPLOYEES 9998[goto B7] DO NOT KNOW[goto B7] 9999[goto B7] REFUSED[goto B7] B6b Of these employees, how many perform no significant reading, writing, or arithmetic on the job? 0 NO READING/MATH SKILLS[goto B7] 1–9996 NO READING/MATH SKILLS 9998 DO NOT KNOW/NOT APPLICABLE 9999 REFUSED B6c Of these employees (employees that perform no significant reading, writing, or arithmetic on the job), how many have been hired in the past year? 0 EMPLOYEES 1–9996 NUMBER HIRED PAST YEAR 9998 DO NOT KNOW/NOT APPLICABLE 9999 REFUSED B6d Of these [fill nskl] employees (those that perform no significant reading, writing, or arithmetic), how many are women? 0 NO WOMEN 1–9996 WOMEN 94 9998 DO NOT KNOW 9999 REFUSED B7 Now, thinking about all the different types of positions that you have at (business name), approximately how many vacancies are you currently trying to fill? 0–9996 JOB VACANCIES 9998 DO NOT KNOW 9999 REFUSED all3 During the past year, have you used the Internet when trying to fill job vacancies? 1 YES 5 NO[goto C1] 8 DO NOT KNOW[goto C1] 9 REFUSED[goto C1] all4 Have you listed job openings with America’s Job Bank on the Internet? [r]IWER: FYI: America’s Job Bank is an Internet Resource in which employers post job openings and people looking for jobs can search for job openings that meet their qualifications all over the nation 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED all5 Have you considered job applicants from America’s Talent Bank? 95 [r]IWER: FYI: On America’s Talent Bank, people post their resumes for viewing by prospective employers who are looking for people to fill positions 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED ml13 Have you used “JobNet”? [r]IWER: FYI: JobNet is Wisconsin’s job search system. Both employers and job seekers can use the system to either find workers or work. It contains all the jobs posted with Wisconsin Job Service 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED ml14 Have you participated in a Job Fair sponsored by a state job center or a public agency? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED[goto C1] PART C: LAST WORKER HIRED C1 This next section focuses entirely on the last worker you hired in the past year (that is since (month), 1997) into a position that does not require a college degree. 96 What type of work does this person do, that is, what is their position with your business/organization? 0 SPECIFY POSITION[#specify] 1–9996 SOC CODE 95 HAVE NOT HIRED SINCE [fill cMON] 1997 9998 DO NOT KNOW 9999 REFUSED C2a For this position, how necessary is a high school diploma? Would you say it is absolutely necessary, strongly preferred, mildly preferred, or not at all? 1 ABSOLUTELY NECESSARY 2 STRONGLY PREFERRED 3 MILDLY PREFERRED 4 NOT AT ALL 8 DO NOT KNOW 9 REFUSED C2b (How necessary is) previous experience in this line of work? (Would you say absolutely necessary, strongly preferred, mildly preferred, or not at all)? 1 ABSOLUTELY NECESSARY 2 STRONGLY PREFERRED 3 MILDLY PREFERRED 4 NOT AT ALL 8 DO NOT KNOW 9 REFUSED 97 C2c (How necessary is) some previous training or skill certification? (Would you say absolutely necessary, strongly preferred, mildly preferred, or not at all)? 1 ABSOLUTELY NECESSARY 2 STRONGLY PREFERRED 3 MILDLY PREFERRED 4 NOT AT ALL 8 DO NOT KNOW 9 REFUSED C3a Next, I would like to ask you some questions about the tasks this employee performs on a daily basis. Does this position involve speaking directly with customers in person or over the phone on a daily basis? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED C3b (Does this position involve) reading or writing reports, memos, or lengthy instructions on a daily basis? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED C3c (Does this position involve) doing arithmetic including making change (on a daily basis)? 1 YES 5 NO 98 8 DO NOT KNOW 9 REFUSED C3d (Does this position involve) using a personal computer (on a daily basis)? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED C5a Next, I would like to ask you some general questions about the last person you hired. Is this person male or female? 1 MALE 5 FEMALE 8 DO NOT KNOW 9 REFUSED C5b What is this employee’s racial or ethnic background? (Are they black, Asian, Native American, white, or Hispanic?) 1 AFRICAN-AMERICAN OR BLACK 2 ASIAN OR PACIFIC ISLANDER 3 NATIVE AMERICAN 4 WHITE OR CAUCASIAN 5 HISPANIC 0 OTHER: SPECIFY[#specify] 7 OTHER 8 DO NOT KNOW 9 REFUSED TO ANSWER 99 C5ba Is this employee an immigrant? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED C5d What is the highest level of education they have completed? 1 8TH GRADE OR LESS 2 SOME HIGH SCHOOL 3 GED 4 HIGH SCHOOL GRADUATE 5 SOME COLLEGE 6 ASSOCIATE DEGREE 7 COLLEGE GRADUATE 8 GRADUATE OR PROFESSIONAL DEGREE 9 TECHNICAL OR SKILLS TRAINING 0 OTHER: MISCELLANEOUS[#specify] 98 DO NOT KNOW 99 REFUSED all6 Did you use the Internet while recruiting this person? 1 YES 5 NO[goto C8ca] 8 DO NOT KNOW[goto C8ca] 9 REFUSED[goto C8ca] all7 Did the Internet generate this particular employee? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED[goto C8ca] 100 C6b When this person was first hired, were they hired through a temporary agency? 1 YES[goto C8ca] 5 NO 8 DO NOT KNOW 9 REFUSED[goto C8ca] C7a How was this employee recruited? (Was it through state employment agency, a private employment agency, referral from another current employee, referral from an acquaintance, walk-in from the street, newspaper ad, or a referral from a school, union, or community agency?) 1 STATE EMPLOYMENT AGENCY 2 PRIVATE EMPLOYMENT AGENCY 3 REFERRAL FROM ANOTHER EMPLOYEE 4 REFERRAL FROM AN ACQUAINTANCE/FRIEND 5 WALK-IN FROM THE STREET/HELP WANTED SIGN 6 NEWSPAPER AD 7 SCHOOL REFERRAL 8 UNION REFERRAL 9 COMMUNITY AGENCY 10 JOB LINE 11 PROMOTED FROM WITHIN/ALREADY AN EMPLOYEE 12 TRAINING PROGRAM 0 OTHER: SPECIFY[#specify] 98 DO NOT KNOW 99 REFUSED[goto all6] C8ca Did you have the applicants take any kind of tests? 1 YES 5 NO[goto C10a] 101 8 DO NOT KNOW[goto C10a] 9 REFUSED[goto C10a] C8cb Which kinds of test(s) did you administer? 1 BASIC SKILLS APTITUDE 2 BOOK-KEEPING TEST 3 CIVIL SERVICE EXAM 4 COMPUTER TEST 5 DATA ENTRY 6 DRIVING TEST 7 DRUG TEST 8 ENGLISH TEST 9 LOGIC TEST 10 MATH TEST 11 MECHANICAL SKILLS/APTITUDE 12 READING TEST 13 PSYCHOLOGICAL, PERSONALITY TESTS 14 PHYSICAL EXAM 15 PHYSICAL AGILITY, PHYSICAL APTITUDE 16 SPECIFIC JOB SKILL TEST (receptionist, bar tender, waitress, map reading, field test, company tests) 17 WRITING TEST/SAMPLES 18 TYPING TEST 19 VERBAL TEST 98 DO NOT KNOW 99 REFUSED C10a Of the applications you received for this position, roughly what percentage were from African American males? 0-100 PERCENT AFRICAN AMERICAN MALES 998 DO NOT KNOW 999 REFUSED 102 C10b (Roughly, what percentage were from) African American females? 0–100 PERCENT AFRICAN AMERICAN FEMALES 998 DO NOT KNOW 999 REFUSED C10c (Roughly, what percentage were from) Hispanics? [r]IWER IF NECESSARY USE: Your best estimate is fine. 0–100 PERCENT HISPANICS 998 DO NOT KNOW 999 REFUSED C10d (Roughly, what percentage were from) Asians? [r]IWER IF NECESSARY USE: Your best estimate is fine. 0–100 PERCENT ASIANS 998 DO NOT KNOW 999 REFUSED W1 What was their starting wage or salary for this position? [r]IWER: ENTER THE DOLLARS ONLY AT THIS SCREEN [r]IT IS POSSIBLE TO ENTER UP TO 100,000 AT THIS SCREEN 0–999997 DOLLARS 999998 DO NOT KNOW 999999 REFUSED 103 W1a (What is the starting wage?)[allow 3] [r]IWER: ENTER ONLY THE CENTS AT THIS SCREEN 00–99 CENTS 998 DO NOT KNOW 999 REFUSED W1b (Is this figure hourly, weekly, monthly, or yearly)? [r]IWER: YOU DO NOT HAVE TO ASK THIS QUESTION, JUST RECORD WHAT THE RESPONDENT HAS SAID TO YOU ABOUT THE WAGE STRUCTURE 1 HOURLY 2 WEEKLY 3 MONTHLY 4 YEARLY 5 SOMETHING ELSE[#specify] 6 BI-WEEKLY 7 COMMISSION 0 DAILY 8 DO NOT KNOW 9 REFUSED C12 In addition to their wage, does this employee get tips, commission, or profit sharing? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED C13a Do you provide health insurance to the employee in this position? 104 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED C14a How many hours per week does this person usually work? 0–90 HOURS PER WEEK 98 DO NOT KNOW 99 REFUSED C16a Did this employee receive any formal training after being hired? (By formal training, we mean with an instructor or in a classroom)? 1 YES 5 NO[goto C20] 8 DO NOT KNOW[goto C20] 9 REFUSED[goto C20] C16b Approximately how many hours of formal training did he or she receive? 1–996 HOURS OF TRAINING 998 DO NOT KNOW 999 REFUSED C20 Would you say it is easy, somewhat difficult, or very difficult to find qualified applicants for this type of position at the present time? 1 EASY 3 SOMEWHAT DIFFICULT 5 VERY DIFFICULT 105 8 DO NOT KNOW 9 REFUSED C21a In the past two years, have you hired someone into a non-college degree position who did not meet all of the qualifications you usually require because you really needed the employee? 1 YES 5 NO[goto C22a] 8 DO NOT KNOW[goto C22a] 9 REFUSED[goto C22a] Ca21 How many have you hired in the past two years? 1–9996 HIRES 9998 DO NOT KNOW 9999 REFUSED C22a Next, I am going to list a few different types of applicants. Please tell me if you would accept each type of applicant for a position in your business/organization. A person who is or has been on welfare? (Would you definitely, probably not, or definitely not accept this applicant)? 1 DEFINITELY WOULD 2 PROBABLY WOULD 3 PROBABLY NOT 4 DEFINITELY NOT 8 DO NOT KNOW 9 REFUSED C22b A person who is or has been in a government employment program or had a GED instead of a high school diploma? 106 (Would you definitely, probably not, or definitely not accept this applicant)? 1 DEFINITELY WOULD 2 PROBABLY WOULD 3 PROBABLY NOT 4 DEFINITELY NOT 8 DO NOT KNOW 9 REFUSED C22c A person who has a criminal record? (Would you definitely, probably not, or definitely not accept this applicant)? 1 DEFINITELY WOULD 2 PROBABLY WOULD 3 PROBABLY NOT 4 DEFINITELY NOT 8 DO NOT KNOW 9 REFUSED C22d A person who only lists short term or part time jobs for work experience? (Would you definitely, probably not, or definitely not accept this applicant)? 1 DEFINITELY WOULD 2 PROBABLY WOULD 3 PROBABLY NOT 4 DEFINITELY NOT 8 DO NOT KNOW 9 REFUSED 107 C22e A person who has been unemployed for a year or more? (Would you definitely, probably not, or definitely not accept this applicant)? 1 DEFINITELY WOULD 2 PROBABLY WOULD 3 PROBABLY NOT 4 DEFINITELY NOT 8 DO NOT KNOW 9 REFUSED PART D: WELFARE RECIPIENTS D1 The next set of questions focus on hiring individuals who have received welfare. Does your business/organization have any type of a formal welfare to work program at this workplace? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED Da1 To your knowledge, has your business/organization in the past two years (that is, since (month) 1996) hired any women who had been on welfare? 1 YES 5 NO[goto E1] 8 DO NOT KNOW[goto E1] 9 REFUSED[goto E1] Da2 How many (women who have been on welfare) has your business/organization hired in the past two years? 108 1–500 HIRES PAST 2 YEARS 998 DO NOT KNOW 999 REFUSED Da3 How many (women who have been on welfare) has your business/organization hired in the past year? 0–500 HIRES PAST 1 YEAR 998 DO NOT KNOW 999 REFUSED Da4 I’d like to ask you some questions about the last woman you have hired that had been a welfare recipient. First, how sure are you that she had been a welfare recipient? Would you say you are definitely sure, fairly sure, or not really sure? 1 DEFINITELY SURE 2 FAIRLY SURE 3 NOT REALLY SURE[goto E1] 8 DO NOT KNOW[goto E1] 9 REFUSED[goto E1] Da5 Was she still on welfare, had she just come off of welfare, or had she been off welfare for some time when she was hired? 1 STILL ON WELFARE 2 JUST COME OFF WELFARE 3 BEEN OFF WELFARE FOR SOME TIME 8 DO NOT KNOW 9 REFUSED 109 Da6 Did she have a regular high school diploma? 1 YES[goto Da8] 5 NO 8 DO NOT KNOW 9 REFUSED[goto Da8] Da7 Did she have a GED? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED Da8 Was she recruited through some type of agency or by some other means? 1 AGENCY 5 OTHER MEANS OF RECRUITMENT[goto same] 8 DO NOT KNOW[goto same] 9 REFUSED[goto same] Da9 Did you contact the agency from which you hired this person, or did the agency contact you? 1 CONTACTED THE AGENCY 5 AGENCY CONTACT BUSINESS 8 DO NOT KNOW 9 REFUSED Is this the person that we have already discussed above? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED 110 Da10 What is this employee’s racial or ethnic background? (Are they black, Asian, Native American, white, or Hispanic)? 1 AFRICAN-AMERICAN OR BLACK 2 ASIAN OR PACIFIC ISLANDER 3 NATIVE AMERICAN 4 WHITE OR CAUCASIAN 5 HISPANIC 7 OTHER 8 DO NOT KNOW 9 REFUSED TO ANSWER Da11 Is she an immigrant? 1 YES 5 NO[goto D1c] 8 DO NOT KNOW[goto D1c] 9 REFUSED[goto D1c] Da12 Does this employee speak English fluently? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED D1c In what month and year was this person hired? 1–12 MONTH 98 DO NOT KNOW 99 REFUSED 111 Dc1 In what month and year was this person hired? 90–99 YEAR HIRED[goto sm3] 998 DO NOT KNOW 999 REFUSED D2 Into what position was she hired? 1–9996 SOC CODE 9998 DO NOT KNOW 9999 REFUSED W3 What was the starting wage or salary for her position? 0–999997 DOLLARS 999998 DO NOT KNOW 999999 REFUSED W3a (What is the starting wage?)[allow 3] 00–99 CENTS 998 DO NOT KNOW 998 REFUSED W3b (Is this figure hourly, weekly, monthly, or yearly)? 1 HOURLY 2 WEEKLY 3 MONTHLY 4 YEARLY 5 SOMETHING ELSE [#specify] 6 BI-WEEKLY 7 COMMISSION 8 DO NOT KNOW 9 REFUSED 112 D2a In addition to their wage, does this employee get tips, commission, or profit sharing? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED D4 Do you provide health insurance to her and her family? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED D4a How many hours per week does this person usually work? 0–80 HOURS PER WEEK 98 DO NOT KNOW 99 REFUSED D4b Thinking about this employee’s work schedule, is it at all flexible or are the hours strictly fixed? 1 FLEXIBLE 5 FIXED 8 DO NOT KNOW 9 REFUSED D6a Did she have any recent work experience before you hired her? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED 113 D6b (Did she have) previous experience in this particular type of work? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED D6c Had she received any previous training that was relevant to this position or this line of work? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED D7a Before hiring her, did you do a criminal background check? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED Db7 Did you have the applicants take any kind of tests? 1 YES 5 NO[goto D7b] 8 DO NOT KNOW[goto D7b] 9 REFUSED[goto D7b] Dc7 Which kinds of test(s) did you administer? 1 BASIC SKILLS APTITUDE 2 BOOK-KEEPING TEST 3 CIVIL SERVICE EXAM 4 COMPUTER TEST 114 5 DATA ENTRY 6 DRIVING TEST 7 DRUG TEST 8 ENGLISH TEST 9 LOGIC TEST 10 MATH TEST 11 MECHANICAL SKILLS/APTITUDE TEST/SAMPLES 12 READING TEST 13 PSYCHOLOGICAL, PERSONALITY TESTS 14 PHYSICAL EXAM 15 PHYSICAL AGILITY, PHYSICAL APTITUDE 16 SPECIFIC JOB SKILL TEST (receptionist, bar tender, waitress, map reading, field test, company tests) 17 WRITING 18 TYPING TEST 19 VERBAL TEST 98 DO NOT KNOW 99 REFUSED D7b How much did each of the following factors weigh into your decision to hire her for this position? Her overall appearance, dress, and demeanor? Would you say a lot, some, or none at all? 1 A LOT 2 SOME 3 NONE AT ALL 8 DO NOT KNOW 9 REFUSED D7c Her recommendations and referrals? (Would you say a lot, some, or none at all)? 1 A LOT 2 SOME 115 3 NONE AT ALL 8 DO NOT KNOW 9 REFUSED D7g How many other applicants for this position did you pass over to hire her? 0–500 APPLICANTS 998 DO NOT KNOW/NOT APPLICABLE 999 REFUSED D8a Next, I would like to ask you some questions about the tasks this employee performs on a daily basis. Does this position involve speaking directly with customers in person or over the phone on a daily basis? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED D8b (Does this position involve) reading or writing reports, memos, or lengthy instructions on a daily basis? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED D8c (Does this position involve) doing arithmetic including making change (on a daily basis)? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED 116 D8d (Does this position involve) using a personal computer? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED[goto D8e] de8 Next, I have a couple of questions about other tasks this employee performs on a daily basis. D8e (Does this position involve) filling out forms on a daily basis? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED D8f (Does this position involve) keeping a close watch over gauges, dials, or instruments of any kind (on a daily basis)? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED D9a For this position, did she receive any formal training after being hired? 1 YES 5 NO[goto D9c] 8 DO NOT KNOW[goto D9c] 9 REFUSED[goto D9c] 117 D9b Was any of the training she received after being hired remedial of basic skills? Basic skills include the employee’s ability to read or write, do arithmetic, or speak English. 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED D9c After hiring her, did she pursue any additional education on her own? 1 YES 5 NO[goto D9f] 8 DO NOT KNOW[goto D9f] 9 REFUSED[goto D9f] D9d Did you support her with tuition reimbursements? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED D9e Did you support her with reduced or flexible work hours? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED D9f Do you provide child care assistance to this employee? 1 YES 118 5 NO 8 DO NOT KNOW 9 REFUSED D9g (Do you provide) transportation assistance? 1 YES 5 NO[##md20] 8 DO NOT KNOW 9 REFUSED D10a For the next set of questions, please tell me if you had problems with this employee in each of the following areas. Absenteeism or tardiness? 1 YES 5 NO[goto D10g] 8 DO NOT KNOW[goto D10g] 9 REFUSED[goto D10g] D10b Were her problems with absenteeism or tardiness due to any of the following? Physical health? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED D10c (Were her problems with absenteeism or tardiness due to) mental health or depression? 1 YES 5 NO 119 8 DO NOT KNOW 9 REFUSED D10d (Were her problems with absenteeism or tardiness due to) child care? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED D10e (Were her problems with absenteeism or tardiness due to) transportation? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED D10f (Were her problems with absenteeism or tardiness due to) domestic violence? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED D10g Did you have any problems with this employee’s attitude toward work? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED 120 D10h (Did you have any problems with this employee’s) basic verbal, math, or reading skills? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED D10i (Did you have any problems with this employee’s) other job-related skills? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED D10j (Did you have any problems with this employee’s) substance abuse (alcohol or drug abuse)? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED D10k (Did you have any problems with this employee’s) relationships with co-workers, supervisors, or customers? 1 YES 5 NO 9 REFUSED D13 If this person performs well, what are the chances that she could be promoted? Would you say excellent, good, fair or poor? 121 1 EXCELLENT 2 GOOD 3 FAIR 4 POOR 8 DO NOT KNOW 9 REFUSED D14 Overall, how would you rate the employee relative to the typical one that you hire into this position? Would you say she was much better, a little better, about the same, a little worse, or much worse? 1 MUCH BETTER 2 A LITTLE BETTER 3 ABOUT THE SAME 4 A LITTLE WORSE 5 MUCH WORSE 8 DO NOT KNOW 9 REFUSED D5 Is she still with your firm? 1 YES[goto all8] 5 NO 8 DO NOT KNOW[goto all8] 9 REFUSED [goto all8] D5a In what month and year did she leave your business/ organization? 1–12 MONTH 98 DO NOT KNOW 99 REFUSED D5b (In what year did she leave your business/organization)? 122 19 94–99 YEAR 98 DO NOT KNOW 99 REFUSED D5c Did she voluntarily quit, get discharged, get laid off, or something else? 1 VOLUNTARILY QUIT 2 DISCHARGED 3 LAID OFF 4 POSITION ENDED 0 OTHER: SPECIFY[#specify] 8 DO NOT KNOW 9 REFUSED[goto all8] E1 Have you had any contact during the past year with any agency that is trying to place welfare recipients into jobs? 1 YES 5 NO[goto F1] 8 DO NOT KNOW[goto F1] 9 REFUSED[goto F1] Ea1 Did you contact the agency, or did the agency contact you? 1 CONTACTED THE AGENCY 5 AGENCY CONTACT BUSINESS 8 DO NOT KNOW 9 REFUSED[goto fl2] 1e Next, I have some questions about contacts you may have had with agencies who find work for welfare recipients. E1a Were they a(n)(name of agency) agency or contractor? 1 YES 123 5 NO 8 DO NOT KNOW 9 REFUSED m11 Have you claimed the subsidy for “Trial Jobs”? 1 YES 5 NO[goto skp5] 8 DO NOT KNOW[goto skp5] 9 REFUSED[goto skp5] ml12 For how many employees have you claimed the subsidy at any time? 1–500 EMPLOYEES 998 DO NOT KNOW 999 REFUSED 1000 [goto skp5] ml1 I would like to read you a list of agencies in the Milwaukee area. Please tell me if you have had any contact with the agency during the past year. YW Works? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED ml2 (Have you had any contact with) Opportunities Industrialization Center? 1 YES 5 NO 124 8 DO NOT KNOW 9 REFUSED ml3 (Have you had any contact with) United Migrant Opportunities Services? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED ml4 (Have you had any contact with) Workforce Solutions? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED ml5 (Have you had any contact with) Maximus? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED ml6 (Have you had any contact with) The Milwaukee Private Industry Council? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED ml7 (Have you had any contact with) Waukesha Workforce Development Center? 125 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED ml8 (Have you had any contact with) Washington County Department of Social Services? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED ml9 (Have you had any contact with) Ozaukee County Workforce Development Center? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED[goto E3] E3 Did you agree to consider an application or interview at least one of these referrals? 1 YES 5 NO[goto E3b3] 8 DO NOT KNOW[goto E3b3] 9 REFUSED[goto E3b3] E3a How many applications did you consider or interviews did you conduct? 0–996 APPLICANTS/INTERVIEWS 998 DO NOT KNOW/NOT APPLICABLE 999 REFUSED 126 E3b Did you hire any of their referrals? 1 YES 5 NO[goto E3b3] 8 DO NOT KNOW[goto E3b3] 9 REFUSED[goto E3b3] E3b1 How many (referrals) did you hire? 0–996 REFERRALS 998 DO NOT KNOW 999 REFUSED E3b2 How many (referrals) still work for you? 0–996 REFERRALS 998 DO NOT KNOW 999 REFUSED Ea3 Did the agency provide any of the following assistance to the applicants you hired? General work experience not related to the type of work at your business/organization? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED Eb3 Experience in the particular type of work? (Did the agency provide any assistance to the applicants you hired)? 1 YES 127 5 NO 8 DO NOT KNOW 9 REFUSED Ec3 Training relevant to the type of work your business/organization does? (Did the agency provide any assistance to the applicants you hired)? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED Ed3 Did the agency provide any services to your business/organization after you hired referrals to assist in retaining referrals or dealing with problems that arise with referrals? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED E3b3 Would you consider any future referrals from this agency? 1 YES 3 MAYBE 5 NO 8 DO NOT KNOW 9 REFUSED F1 Suppose you were contacted by an employment agency that was trying to place welfare recipients. We are specifically talking 128 about welfare recipients who do not have a high school diploma or any recent work experience. Do you currently have any open positions that you might consider filling with these welfare recipients? 1 YES 5 NO[goto F1a] 8 DO NOT KNOW[goto F1a] 9 REFUSED[goto F1a] 1F How many of them would you consider employing right away? 1–996 IMMEDIATE HIRES 998 DO NOT KNOW 999 REFUSED F1a Do you think you will have open positions during the next year that you might consider filling with these welfare recipients? 1 YES 5 NO[goto Fa5] 8 DO NOT KNOW[goto Fa5] 9 REFUSED[goto Fa5] 2F How many of them would you possibly employ at any time during the next year? 1–996 HIRES DURING THE NEXT YEAR 998 DO NOT KNOW 999 REFUSED Fa5 Did you know that tax credits for hiring welfare recipients are now available from the federal government? 1 YES 129 5 NO 8 DO NOT KNOW 9 REFUSED Fa51 Did you know that you can receive these credits per employee for up to 2 years? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED ml10 Did you know that you can receive a subsidy from the state of Wisconsin for putting people into “Trial Jobs”? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED[goto Fa6] Fa6 Do you think this will make you any more likely than before to hire poorly educated young women who are single mothers? 1 YES 3 MAYBE 5 NO 8 DO NOT KNOW 9 REFUSED Fa10 If you discover that an employee is eligible, will you or your company actually claim the credit next year? 1 YES 3 MAYBE 5 NO 8 DO NOT KNOW 130 9 REFUSED[goto all1] all1 Have you claimed the Federal Tax Credits for any employee during the past year? 1 YES 5 NO[goto skp5] 8 DO NOT KNOW[goto skp5] 9 REFUSED[goto skp5] all2 For how many employees have you claimed the credits at any time? 1–500 EMPLOYEES 998 DO NOT KNOW 999 REFUSED all8 When hiring this person, did you remind them that they may be eligible for the Earned Income Credit (on their Federal Income Taxes)? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED ml15 (When hiring this person) did you remind them about child care subsidies from the state? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED[goto skp4] 131 skp5 [if F1 ge 5][if F1a ge 5][goto F11a][endif][endif] would not hire right away and would not hire in the next year skip to question F11a. F5 Is there some specific job into which you would most likely hire a worker who was a welfare recipient who did not have a high school diploma or recent work experience? 1000–9996 SOC CODE 1 YES, SPECIFY POSITION(S)[#specify] 5 NO, NO SPECIFIC POSITION[goto F11a] 8 DO NOT KNOW, UNSURE OF POSITION[goto F11a] 9 REFUSED[goto F11a] W5 What is the starting wage or salary for this position? 0–999997 DOLLARS 999999 DO NOT KNOW 999998 REFUSED W5a (What is the starting wage?) 00–99 CENTS 998 DO NOT KNOW 999 REFUSED W5b (Is this figure hourly, weekly, monthly, or yearly)? 1 HOURLY 2 WEEKLY 3 MONTHLY 4 YEARLY 5 SOMETHING ELSE[#specify] 6 BI-WEEKLY 7 COMMISSION 0 DAILY 132 8 DO NOT KNOW 9 REFUSED F6 In addition to their wage, would this employee get tips, commission, or profit sharing? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED F7 Would this position provide health benefits to the employee and the employee’s family? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED F8 How many hours per week would this person usually work? 0–60 HOURS PER WEEK 98 DO NOT KNOW 99 REFUSED F10 If this person performs well, what are the chances that this person could be promoted? Would you say excellent, good, fair or poor? 1 EXCELLENT 2 GOOD 3 FAIR 4 POOR 8 DO NOT KNOW 9 REFUSED 133 F11a Would you be more likely to hire welfare recipients, especially those without a high school diploma or recent work experience, if the agency could provide the following. A 50% tax credit against their wages for at least one year? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED F11b (Would you be more likely to hire welfare recipients, especially those without a high school diploma or recent work experience, if the agency could provide) A 100% tax credit against their wages for at least one year? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED F11c (Would you be more likely to hire welfare recipients, especially those without a high school diploma or recent work experience, if the agency could provide) Applicants with any kind of recent work experience? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED 134 F11d (Would you be more likely to hire welfare recipients, especially those without a high school diploma or recent work experience, if the agency could provide) Work experience in your particular line of work? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED F11e (Would you be more likely to hire welfare recipients, especially those without a high school diploma or recent work experience, if the agency could provide) Training in your particular line of work? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED F11f (Would you be more likely to hire welfare recipients, especially those without a high school diploma or recent work experience, if the agency could provide) Assurances that applicants have basic math or reading skills? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED F11g (Would you be more likely to hire welfare recipients, especially those without a high school diploma or recent work experience, if the agency could provide) 135 Assurances that applicants have good attitudes toward work? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED F11h (Would you be more likely to hire welfare recipients, especially those without a high school diploma or recent work experience, if the agency could provide) (Assurances that the applicant) has no criminal record? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED F11i (Would you be more likely to hire welfare recipients, especially those without a high school diploma or recent work experience, if the agency could provide) (Assurances that the applicant) does not have any substance abuse problems? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED F11j (Would you be more likely to hire welfare recipients, especially those without a high school diploma or recent work experience, if the agency could provide) (Assurances that the applicant) has good dress and appearance? 136 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED F11k (Would you be more likely to hire welfare recipients, especially those without a high school diploma or recent work experience, if the agency could provide) (Assurances that the applicant) has good social skills? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED F12a Before hiring a referral, would you do a drug test? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED F12b (Before hiring a referral), would you check into their background for a possible criminal record? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED F13a Which of the following types of assistance would you be willing to provide, if any, in order to improve a welfare recipient’s ability to hold a job with your business/organization? 137 Transportation? 1 YES 3 MAYBE 5 NO 8 DO NOT KNOW 9 REFUSED F13b Child Care? 1 YES 3 MAYBE 5 NO 8 DO NOT KNOW 9 REFUSED F13c Remedial basic skills training? 1 YES 3 MAYBE 5 NO 8 DO NOT KNOW 9 REFUSED F13d Specific job skills training? 1 YES 3 MAYBE 5 NO 8 DO NOT KNOW 9 REFUSED F13e Tuition reimbursements for school? 1 YES 3 MAYBE 138 5 NO 8 DO NOT KNOW 9 REFUSED F13f Reduced or flexible hours for schooling? 1 YES 3 MAYBE 5 NO 8 DO NOT KNOW 9 REFUSED F14a Would you be more likely to provide some training if you could get a tax credit against at least some of the costs? 1 YES 3 MAYBE 5 NO 8 DO NOT KNOW 9 REFUSED F14b Would you be more likely to provide some training if you could get some technical assistance with the training? 1 YES 3 MAYBE 5 NO 8 DO NOT KNOW 9 REFUSED H1 Would you be more likely to hire someone from an agency trying to place workers such as the ones we just spoke about if they provided services to help you retain the employees or to deal with problems that arose? 1 YES 139 5 NO 8 DO NOT KNOW 9 REFUSED H2 Would you be willing to work with such an agency in developing relevant work experience or training for a client that you might hire afterward? 1 YES 5 NO 8 DO NOT KNOW 9 REFUSED G1 Finally, I have two questions about another group of workers. Suppose you were contacted by an employment agency that was trying to place young males with criminal records. Do you currently have any open positions that you might consider filling with this group of workers? [r]IWER: IF THE RESPONDENT SAYS, DEPENDS OR DO NOT KNOW, USE THE FOLLOWING PROBE: “Just assume that they are young males who have committed crimes that have lead to their incarceration” 1 YES 5 NO[goto MOD7] 8 DO NOT KNOW[goto MOD7] 9 REFUSED[goto MOD7] 1G How many of them would you consider employing right away? 1–996 IMMEDIATE HIRES 998 DO NOT KNOW 999 REFUSED 140 Bibliography Abraham, Katherine, “Structural/Frictional v. Demand-Deficient Unemployment,” American Economic Review, Vol. 73, 1983, pp. 708– 724. Acs, Gregory, Norma Coe, Keith Watson, and Robert Lerman, Does Work Pay? An Analysis of the Work Incentives Under TANF, The Urban Institute, Washington, D.C., 1999. Anderson, Patricia, and Bruce Meyer, “The Extent and Consequences of Job Turnover,” Brookings Papers on Economic Activity— Microeconomics, 1994, pp. 177–248. Aron, Laudan, Robert Lerman, and Stephanie Riegg, Youth in 2010: Trends and Policy Implications, The Urban Institute, Washington, D.C., 2000. Ballen, John, and Richard Freeman, “Transitions from Employment to Nonemployment,” in R. Freeman and H. Holzer (eds.), The Black Youth Employment Crisis, University of Chicago Press, Chicago, 1986. Bane, Mary Jo, and David Ellwood, “Slipping Into and Out of Poverty—The Dynamics of Spells,” Journal of Human Resources, Vol. 21, 1986, pp. 2–23. Bane, Mary Jo, and David Ellwood, Welfare Realities, Harvard University Press, Cambridge, 1994. Bates, Timothy, “Job Creation Through Improved Access to Markets for Minority-Owned Businesses,” in W. Leigh and M. Simms (eds.), The Black Worker in the 21st Century: Job Creation Prospects and 141 Strategies, The Joint Center for Political and Economic Studies, Washington, D.C., 1998. Becker, Gary, The Economics of Discrimination, University of Chicago Press, Chicago, 1971. Berman, Eli, John Bound, and Zvi Griliches, “Changes in the Demand for Skilled Labor Within U.S. Manufacturing: Evidence from the Annual Survey of Manufacturers, Quarterly Journal of Economics, Vol. 109, 1994, pp. 367–397. Blumenberg, Evelyn, and Paul Ong, “Accessibility and Welfare Usage Rates: Evidence from Los Angeles,” Journal of Policy Analysis and Management, Vol. 17, No. 4, 1998. Borjas, George, “The Economics of Immigration,” Journal of Economic Literature, Vol. 32, 1994, pp. 1667–1717. Brown, Charles, James Hamilton, and James Medoff, Employers Large and Small, Harvard University Press, Cambridge, 1990. Burtless, Gary, “The Employment Prospects of Welfare Recipients,” in D. Nightingale (ed.), The Work Alternative, The Urban Institute Press, Washington, D.C., 1995. Burtless, Gary, “Can the Labor Market Absorb Three Million Welfare Recipients?” The Low Wage Labor Market: Challenges and Opportunities for Economic Self-Sufficiency, Office of the Assistant Secretary for Policy and Evaluation, U.S. Department of Health and Human Services, Washington, D.C., 1999. Cancian, Maria, Robert Haveman, Thomas Kaplan, Daniel Meyer, and Barbara Wolfe, “Work, Earnings and Well-Being after Welfare,” in S. Danziger (ed.), Economic Conditions and Welfare Reform, W. E. Upjohn Institute for Employment Research, Kalamazoo, MI, 1999. 142 Clark, Kim, and Lawrence Summers, “The Dynamics of Youth Unemployment,” in R. Freeman and D. Wise (eds.), The Youth Labor Market Problem: Its Nature, Causes and Consequences, University of Chicago Press, Chicago, 1982. Committee for Economic Development, Welfare Reform and Beyond: Making Work Work, Washington, D.C., 2000. Council of Economic Advisers, Technical Report: Explaining the Decline in Welfare Receipt, 1993–96, Executive Office of the President, Washington, D.C., 1997. Council of Economic Advisers, Technical Report: Economic Expansion, Welfare Reform, and the Decline of Welfare Caseloads: An Update, Executive Office of the President, Washington, D.C., 1999. County of Los Angeles, Evaluating CalWorks in Los Angeles County, Urban Research Division, Chief Administrative Office, 1999. Danziger, Sandra, Mary Corcoran, Sheldon Danziger, Colleen Heflin, Ariel Kalil, Judith Levine, Daniel Rosen, Kristin Seefeldt, Kristine Siefert, and Richard Tolman, Barriers to the Employment of Welfare Recipients, Poverty Research and Training Center, University of Michigan, 1999. Danziger, Sheldon, Colleen Heflin, and Mary Corcoran, “Does It Pay to Move from Welfare to Work?” Mimeo, University of Michigan, 2000. Edelman, Peter, “The Worst Thing Bill Clinton Has Done,” The Atlantic Monthly, March 1997, pp. 43–58. Edin Kathryn, and Laura Lein, Making Ends Meet: How Single Mothers Survive Welfare and Low-Wage Work, Russell Sage Foundation, New York, 1997. 143 Ellwood, David, “The Impact of the Earned Income Tax Credit and Other Social Policy Changes on Work and Marriage in the United States,” Working Paper, Harvard University, 1999. Ellwood, David, and Jonathan Crane, “Family Change among Black Americans: What Do We Know?” Journal of Economic Perspectives, Vol. 4, 1990, pp. 65–84. Ellwood, David, and Elizabeth Welty, “Public Sector Employment and Mandatory Work: A Policy Whose Time Has Come and Gone and Come Again?” in D. Card and R. Blank (eds.), Finding Jobs: Work and Welfare Reform, Russell Sage Foundation, New York, 2000. Figlio, David, and James Ziliak, “Welfare Reform, the Business Cycle, and the Decline in AFDC Caseloads,” in S. Danziger (ed.), Economic Conditions and Welfare Reform, 1999. Flaming, Daniel, Mark Drayse, and Peter Force, On the Edge: A Progress Report on Welfare to Work in Los Angeles, Economic Roundtable, Los Angeles, 1999. Freedman, Stephen, Marisa Mitchell, and David Navarro, The Los Angeles Jobs-First GAIN Evaluation, New York Manpower Demonstration Research Corporation, 1999. Freeman, Richard, “Crime and the Employment of Disadvantaged Youth,” in G. Peterson and W. Vroman (eds.), Urban Labor Markets and Job Opportunities, The Urban Institute Press, Washington, D.C., 1992. Freeman, Richard, and William Rodgers, “Area Economic Conditions and the Labor Market Outcomes of Young Men in the 1990’s Expansion,” National Bureau of Economic Research Working Paper, 1999. 144 Gustafson, Cindy, and Philip Levine, “Less-Skilled Workers, Welfare Reform, and the Unemployment Insurance System,” NBER Working Paper, 1998. Ham, John, and Robert Lalonde, “The Effect of Sample Selection and Initial Conditions in Duration Models: Evidence from Experimental Data on Training,” Econometrica, Vol. 64, No. 1, 1996. Holzer, Harry, Unemployment, Vacancies and Local Labor Markets, W. E. Upjohn Institute for Employment Research, Kalamazoo, MI, 1989. Holzer, Harry, What Employers Want: Job Prospects for Less-Educated Workers, Russell Sage Foundation, New York, 1996. Holzer, Harry, “Why Do Small Employers Hire Fewer Blacks Than Larger Ones?” Journal of Human Resources, Vol. 33, 1998, pp. 896– 914. Holzer, Harry, “Will Employers Hire Welfare Recipients? New Survey Evidence from Michigan,” Journal of Policy Analysis and Management, 1999a. Holzer, Harry, “Employer Demand for Welfare Recipients and the Business Cycle,” in S. Danziger (ed.), Economic Conditions and Welfare Reform, W. E. Upjohn Institute for Employment Research, Kalamazoo, Michigan, 1999b. Holzer, Harry, and Keith Ihlanfeldt, “Spatial Factors and the Employment of Blacks at the Firm Level,” New England Economic Review, May/June 1996, pp. 65–86. Holzer, Harry, and Robert Lalonde, “Job Change and Job Stability Among Less-Skilled Young Workers,” in R. Blank and D. Card (eds.), Finding Jobs: Work and Welfare Reform, Russell Sage Foundation, New York, 2000. 145 Holzer, Harry, and David Neumark, “What Does Affirmative Action Do?” Industrial and Labor Relations Review, Vol. 53, 2000, pp. 240– 271. Hughes, Mark, and Julie Sternberg, The New Metropolitan Reality: Where the Rubber Meets the Road in Antipoverty Policy, The Urban Institute, Washington, D.C., 1992. Ihlanfeldt, Keith, and David Sjoquist, “The Spatial Mismatch Hypothesis: A Review of Recent Studies and Their Implications for Welfare Reform,” Housing Policy Debate, Vol. 9, 1998, pp. 849–892. Johnson, Hans, and Sonya Tafoya, The Basic Skills of Welfare Recipients: Implications for Welfare Reform, Public Policy Institute of California, San Francisco, 1999. Kain, John, “The Spatial Mismatch Hypothesis Three Decades Later,” Housing Policy Debate, Vol. 3, 1992, pp. 371–460. Katz, Bruce, and Kate Carnevale, The State of Welfare Caseloads in American Cities, Center on Urban and Metropolitan Policy, The Brookings Institution, Washington, D.C., 1998. Katz, Lawrence, and Alan Krueger, “The High-Pressure U.S. Labor Market of the 1990’s,” Brookings Papers on Economic Activity, No. 1, 1999. Katz, Lawrence, and Kevin Murphy, “Changes in Relative Wages, 1963– 87: The Role of Supply and Demand Factors,” Quarterly Journal of Economics, Vol. 107, 1992, pp. 35–78. Kirschenman, Joleen, “Gender Within Race in the Labor Market,” Unpublished Paper, University of Chicago, 1991. 146 Kirschenman, Joleen, and Kathryn Neckerman, “We’d Love to Hire Them But . . . ,” in C. Jencks and P. Peterson (eds.), The Urban Underclass, The Brookings Institution, Washington, D.C., 1991. Krueger, Alan B., and Lawrence Summers, “Reflections on the InterIndustry Wage Structure,” in Kevin Lang and Jonathan Leonard (eds.), Unemployment and the Structure of Labor Markets, Basil Blackwell Press, London, 1987, pp. 17–47. Loprest, Pamela, “Families Who Left Welfare: Who Are They and How Are They Doing?” Assessing the New Federalism Discussion Paper, The Urban Institute, Washington, D.C., 1999. Mare, Robert, “Changes in Educational Attainment and School Enrollment,” in R. Farley (ed.), State of the Union, Vol. 1, Russell Sage Foundation, New York, 1995. McMurrer, Daniel, Isabel Sawhill, and Robert Lerman, “Welfare Reform and Opportunity in the Low-Wage Labor Market,” Opportunity in America, No. 5, The Urban Institute, Washington, D.C., 1997. Meyer, Bruce, and Dan Rosenbaum, “Welfare, the Earned Income Tax Credit, and the Labor Suppy of Single Mothers,” NBER Working Paper, 1999. Meyer, Bruce, and Dan Rosenbaum, “Making Single Mothers Work: Recent Tax and Welfare Policy and Its Effects,” NBER Working Paper, 2000. Michalopoulos, Charles, and Christine Schwartz, with Diana AdamsCiardullo, What Works Best for Whom: Impacts of Twenty Welfare-toWork Programs by Subgroup, prepared for the U.S. Department of Health and Human Services, 2000. Moss, Phil, and Chris Tilly, “Soft Skills and Race,” Working Paper, Russell Sage Foundation, New York, 1996. 147 Murnane, Richard, and Frank Levy, Teaching the New Basic Skills, Free Press, New York, 1995. Nightingale, Demetra, and Kelly Mikelson, An Overview of Research Related to Wisconsin Works (W-2), The Urban Institute, Washington, D.C., 2000. Pavetti, LaDonna, How Much More Can They Work? Setting Realistic Expectations for Welfare Mothers, The Urban Institute, Washington, D.C., 1997. Pavetti, LaDonna, “What Will the States Do When Jobs Are Not Plentiful?” in S. Danziger (ed.), Economic Conditions and Welfare Reform, W. E. Upjohn Institute for Employment Research, Kalamazoo, Michigan, 1999. Pavetti, LaDonna, Michelle Derr, Jacquelyn Anderson, Carole Trippe, and Sidnee Paschal, The Role of Intermediaries in Linking TANF Recipients with Jobs, prepared for the U.S. Department of Health and Human Services, 2000. Polivka, Anne, “Notes on the Possible Effects of Welfare Reform on Labor Market Activities: What Can Be Gleaned from the March CPS,” Unpublished manuscript, Bureau of Labor Statistics, Washington, D.C., 1998. Primus, Wendell, Lynette Rawlings, Kathy Larin, and Kathryn Porter, The Initial Impacts of Welfare Reform on the Incomes of Single-Mother Families, Center on Budget and Policy Priorities, Washington, D.C., 1999. Quint, Janet, Kathryn Edin, Maria Buck, Barbara Fink, Yolanda Padilla, Olis Simmons-Hewitt, and Mary Valmont, Big Cities and Welfare Reform: Early Implementation and Ethnographic Findings from the Project on Devolution and Urban Change, Manpower Demonstration Research Corporation, New York, 1999. 148 Rangarajan, Anu, Keeping Welfare Recipients Employed: A Guide for States Designing Job Retention Services, Mathematica Policy Research Inc., Princeton, NJ, 1998. Raphael, Stephen, “The Spatial Mismatch Hypothesis and Black Youth Joblessness: Evidence from the San Francisco Bay Area,” Journal of Urban Economics, Vol. 43, 1997, pp. 79–111. Raphael, Stephen, Michael Stoll, and Harry Holzer, “Do Suburban Employers Discriminate More Against African-Americans?” Journal of Urban Economics, forthcoming (2000). Regenstein, Marsha, and Jack Meyer, “Job Prospects for Welfare Recipients: Employers Speak Out,” Assessing the New Federalism Occasional Papers, The Urban Institute, Washington, D.C., 1998. Stoll, Michael, Harry Holzer, and Keith Ihlanfeldt, “Within Cities and Suburbs: Neighborhood Composition and Employment Opportunities for Whites and Minorities,” Journal of Policy Analysis and Management, forthcoming (2000). Taber, Christopher, and Tricia Gladden, “Wage Progressions Among Less-Skilled Workers,” in R. Blank and D. Card (eds.), Finding Jobs: Work and Welfare Reform, Russell Sage Foundation, New York, 2000. Turner, Mark, “The Effects of Minimum Wages on Welfare Recipiency,” Final Report, The Urban Institute, Washington, D.C., 1999. U.S. Department of Labor, Futurework: Trends and Challenges for Work in the 21st Century, Office of the Secretary, Washington, D.C., 1999. U.S. General Accounting Office (GAO),Welfare Reform: Information on Former Recipients’ Status, Washington, D.C., 1999. 149 Vroman, Wayne, “Effects of Welfare Reform on Unemployment Insurance,” The New Federalism: Issues and Options for States, Series A, No. 22, The Urban Institute, Washington, D.C., 1998. Wallace, Geoffrey, and Rebecca Blank, “What Goes Up Must Come Down?” in S. Danziger (ed.), Economic Conditions and Welfare Reform, 1999. Welfare-to-Work Partnership, “Member Survey: Promotion and Partnership,” No. 1, Washington, D.C., 1999. Wiseman, Michael, “In the Midst of Reform,” Assessing the New Federalism Discussion Paper, The Urban Institute, Washington, D.C., 1999. Zellman, Gail L., Jacob A. Klerman, Elaine Reardon, Donna Farley, Nicole Humphrey, Tammi Chun, and Paul Steinberg, Welfare Reform in California: State and County Implementation of CalWorks in the First Year, RAND, Santa Monica, California, 1999. 150 About the Authors HARRY J. HOLZER Harry J. Holzer is a professor of Public Policy at Georgetown University and a visiting fellow at the Urban Institute in Washington, D.C. He is a former chief economist of the U.S. Department of Labor. He is also a senior affiliate of the Joint Center for Poverty Research (University of Chicago and Northwestern University), research affiliate of the Institute for Research on Poverty (University of Wisconsin), and a national fellow of the Program on Inequality and Social Policy (Harvard University). He received his Ph.D. in economics from Harvard University in 1983. Dr. Holzer has written extensively on the labor market problems of minorities and the urban poor. His books include The Black Youth Employment Crisis (coedited with Richard B. Freeman, University of Chicago Press, 1986); What Employers Want: Job Prospects for LessEducated Workers (Russell Sage Foundation, 1996); and Detroit Divided (with Reynolds Farley and Sheldon Danziger, Russell Sage Foundation, 2000). MICHAEL A. STOLL Michael A. Stoll is an assistant professor of policy studies in the School of Public Policy and Social Research at UCLA. He is a former visiting scholar at the Russell Sage Foundation, and a research associate at the Center for the Study of Urban Poverty at UCLA. He received his Ph.D. in urban planning from MIT in 1996. Dr. Stoll has written on the interplay of labor markets, race, and geography. In particular, his work explores the explanations of the unemployment problems experienced by low-skilled workers in urban areas, especially blacks. His recent book is titled Race, Space and Youth Labor Markets (Garland Publishing, 1999). 151" ["post_date_gmt"]=> string(19) "2017-05-20 09:34:35" ["comment_status"]=> string(4) "open" ["ping_status"]=> string(6) "closed" ["post_password"]=> string(0) "" ["post_name"]=> string(8) "r_101hhr" ["to_ping"]=> string(0) "" ["pinged"]=> string(0) "" ["post_modified"]=> string(19) "2017-05-20 02:34:35" ["post_modified_gmt"]=> string(19) "2017-05-20 09:34:35" ["post_content_filtered"]=> string(0) "" ["guid"]=> string(50) "http://148.62.4.17/wp-content/uploads/R_101HHR.pdf" ["menu_order"]=> int(0) ["post_mime_type"]=> string(15) "application/pdf" ["comment_count"]=> string(1) "0" ["filter"]=> string(3) "raw" ["status"]=> string(7) "inherit" ["attachment_authors"]=> bool(false) }