Moving Up, Moving Out, or Going Nowhere?

The nonpartisan Urban Institute publishes studies, reports, and books on timely topics worthy of public consideration. The views expressed are those of the authors and should not be attributed to the Urban Institute, its trustees, or its funders.

Executive Summary

Conventional wisdom holds that women on welfare will be better off in the long run if they take a job, even if it means initially having less money to spend on their and their children's needs. Underlying this thinking is the belief that women who take low-paying jobs will eventually move up to higher paying jobs either with their current employers or by changing employers. It is certainly true that some women who take "bad" jobs move into "good" jobs (that pay at least $8/hour for at least 35 hours a week). But the extent to which this is true for less educated women, single mothers, and mothers who received welfare is less clear. Indeed, some women may be permanently mired in "bad" jobs, barely earning enough to care for themselves and their children.

To investigate the likelihood that women who turn to the welfare system for support will make the transition from bad to good jobs, we use data from the National Longitudinal Survey of Youth (NLSY) to examine the employment patterns of young women. We construct quarterly employment histories for all the women in our sample, beginning at age 18 and continuing through age 27.

We find that, in general, it is common for young women to make the transition from bad jobs to good jobs and eventually to work steadily in good jobs. However, this is far less the case for women who share the characteristics of women who ever turn to the welfare system for support. For example, although 73 percent of women will work in a good job between the ages of 18 and 27, only 59 percent of mothers and 47 percent of high school dropouts do so. Furthermore, while 41 percent of all women work steadily in good jobs by ages 26 and 27, only 22 percent of mothers and 15 percent of women who have not completed high school do so. While some women who make the transition from a bad job to a good job do so quickly, others do not. On average, women take four years to make the transition from a bad to a good job, and about 25 percent take more than six years.

Assuming that women who ever turn to the welfare system for support have work experiences that mirror those of women with similar characteristics who have never used welfare, we estimate that about one-quarter could be firmly established in a good job by ages 26 and 27. An additional 40 percent would work steadily but would work primarily in bad jobs, and more than one-third would work only sporadically. The employment prospects for the most disadvantaged recipients are far less favorable. For example, we estimate that just 14 percent of recipients who have not completed high school would work primarily in a good job, 34 percent would work steadily in a bad job, and 52 percent would work only intermittently.

I. Introduction

TANF's success will ultimately rest on the ability to move women off the welfare rolls and into jobs. In FY 1997, states are required to engage 25 percent of their single-parent caseload in work activities for a minimum of 20 hours per week. The participation requirement increases steadily through FY 2002, when states must have 50 percent of their single-parent caseload working a minimum of 30 hours per week.¹ The legislation also places a 60-month lifetime limit on the receipt of benefits, substantially increasing the potential adverse consequences for recipients who do not make the transition to steady employment.

In contrast to the Job Opportunities and Basic Skills (JOBS) program, which encouraged states to place program participants in welfare-to-work activities that would increase recipients' skills or education, the TANF work requirements encourage states to place recipients in employment as quickly as possible. This strategy, often referred to as "work first," is based on the premise that the best way to succeed in the labor market is to take any job, even one that may not pay well and may not be full-time (Brown 1997). Underlying this thinking is the belief that women who take low-paying and/or part-time jobs will eventually move up to higher paying jobs and/or full-time jobs. It is certainly true that some women who take "bad" jobs do move into "good" jobs. But the extent to which this is true for less educated women, single mothers, and mothers who have received welfare is less clear. Indeed, some women may be permanently mired in bad jobs, barely earning enough to care for themselves and their children. If former welfare recipients stay in bad jobs for extended periods, a substantial number may need access to resources that make it possible for them to sustain their employment (i.e., child care assistance) or supplement their income from earnings (i.e., child support or the Earned Income Tax Credit) for an extended period of time. However, if bad jobs do lead to good jobs over a relatively short period, such supplementary resources are less critical.

To investigate the likelihood that women who turn to the welfare system for support will make the transition from bad to good jobs, we use data from the National Longitudinal Survey of Youth (NLSY) to examine the employment patterns of young women. We construct quarterly employment histories for all the women in our sample, beginning at age 18 and continuing through age 27. For each quarter we determine whether a woman holds (1) a good job--one job paying at least $8/hour for at least 35 hours a week; (2) a bad job, (3) no job but does not receive welfare; or (4) no job and receives welfare. We then examine the probabilities of moving from one employment state to another over time and develop a microsimulation model to estimate the potential employment outcomes for welfare recipients in a radically changed welfare reform environment.

We find that, in general, it is common for young women to make the transition from bad jobs to good jobs and to eventually work steadily in good jobs. However, this is far less the case for women who share the characteristics of women who ever turn to the welfare system for support. For example, although 73 percent of women will work in a good job between the ages of 18 and 27, only 59 percent of mothers and 47 percent of high school dropouts do so. Furthermore, while 41 percent of all women work steadily in good jobs by ages 26 and 27, only 22 percent of mothers and 15 percent of women who have not completed high school do so. While some women who make the transition from a bad job to a good job do so quickly, others do not. On average, women take four years to make the transition from a bad to a good job, and about 25 percent take more than six years.

Assuming that women who ever turn to the welfare system for support have work experiences that mirror those of women with similar characteristics who have never used welfare, we estimate that about one-quarter could be firmly established in a good job by ages 26 and 27. An additional 40 percent would work steadily but would work primarily in bad jobs, and more than one-third would work only sporadically. The employment prospects for the most disadvantaged recipients are far less favorable. For example, we estimate that just 14 percent of recipients who have not completed high school would work primarily in a good job, 34 percent would work steadily in a bad job, and 52 percent would work only intermittently.

This paper is organized as follows. In the next section, we provide a brief history of work requirements under the old AFDC and the new TANF programs and review research on the employment patterns of less skilled women and welfare recipients. We then describe the data used for our analysis. In subsequent sections we examine how common it is for women to ever work in a good job; present the paths that lead to good jobs; look at the effect of personal characteristics on women's prospects for transitioning from bad jobs to good jobs; and estimate the fraction of former welfare recipients who are likely to make the transition to good jobs. In the final section we summarize our findings and discuss their policy significance.

II. Background and Previous Research

Over time, welfare agencies have also changed the types of work and training requirements placed on recipients. The WIN program was often criticized for being a work registration requirement, focusing mainly on job placement rather than education, training, or employment assistance. When the JOBS program was implemented, states typically emphasized education and training, although activities such as job search and Community Work Experience Programs (CWEP), where recipients must work to receive cash assistance, were also included as program options.

Recently, many states have begun shifting the emphasis of their welfare-to-work programs away from education and training and toward immediate job placement activities. For example, Massachusetts, a state that emphasized long-term education and training under its JOBS program, now requires recipients with a child over the age of six to work for 20 hours a week within 60 days of applying for assistance. Those who cannot find employment on their own are placed in community service jobs in nonprofit organizations (General Accounting Office 1997). Under TANF, states are permitted to have only 30 percent of their assistance caseload participating in education or training. States that have not already shifted their welfare-to-work programs to a rapid employment will have to do so to meet the work participation mandates set forth under the new law.

Research on the work experiences of women who ever turn to the welfare system for support shows that work is quite common among welfare recipients. Recent studies based on the NLSY find that nearly half of all welfare exits occur when women leave welfare for work (Pavetti 1993; Gritz and MaCurdy 1991). Analysis of a more general sample of welfare mothers using the Panel Study of Income Dynamics (PSID) finds an even higher level of employment-related exits -- fully two-thirds of all exits from welfare occurred when a mother found a new job or left welfare after working for more than three months (Harris 1993). Even though employment is common among welfare recipients, steady employment is not. Thirty-nine percent of recipients observed leaving welfare for work in the NLSY and 22 percent observed leaving for work in the PSID returned to welfare within one year (Pavetti 1993; Harris 1996). When women leave welfare for work, they generally find relatively low-paying jobs, usually in the range of $5 to $6 per hour (Riccio, Friedlander, and Freedman 1994; Pavetti, Holcomb, and Duke 1995). However, women who were interviewed by Edin (1995) and her colleagues felt that they would need to earn $8 per hour working full-time to be able to cover their additional work expenses such as child care and transportation.

In contrast to the literature on the transitions from welfare to work, there is limited research on the transitions from bad jobs to good jobs. Burtless (1994) finds that even though women who earn low wages in their early twenties can usually look forward to moderately well-paying jobs when they reach their early or mid-thirties, steady career advancement is much more unusual among women who have done poorly in high school or failed to complete it than among women who have some post-secondary education. Women who had completed one to three years of college by age 24, on average, experienced a $.45 per hour increase each year from age 21 to 29, while women who did not complete high school experienced an average yearly increase of just $.07 per hour (Burtless 1994). Other research shows that part-time work is not frequently used as a transition between nonwork and full-time employment, but is used instead as an alternative to being out of the labor force or working full-time (Blank 1989).

Research on whether low-wage employment is a port of entry into higher paying jobs for young men shows that for the majority of men, low-wage employment does serve as a stepping stone to higher wage employment, though it generally takes high school dropouts longer than high school graduates to make these transitions. During the ten years after leaving school, black high school dropouts spend an average of more than six and a half years in low-paying jobs or no jobs, making them the least likely to make the transition from low-wage to high-wage employment (Gritz and MaCurdy 1992). Comparisons of the experiences of men and women in low-wage employment suggest that low-wage jobs may be less beneficial for women (Pearce 1989). In addition, other research finds that when women switch jobs, they enjoy smaller increases in their wages than men (Loprest 1992). However, the higher increases for men reflect more hours worked rather than higher wage rates.

III. The Data

The detailed information in the NLSYs work history file reveals how much a woman worked and how much she earned at each job she held on a weekly basis. We use this information to categorize her employment status on a quarterly basis. For each of the 40 sample quarters, we designate a woman as having a "good" job, having a "bad" job, or having no job.

We consider a good job to be one in which a woman usually works at least 35 hours a week and earns at least $8 per hour (in real 1993 Consume Price Index dollars). To qualify as a good job quarter, a woman has to work at least 70 hours in the quarter.² If a woman works at all during the quarter but does not meet the good job criteria, we characterize that quarter as a bad job quarter. Because we use a relatively minimal hours requirement to classify a quarter as a good job quarter, only about 1 percent of all bad job quarters are quarters in which a woman earns more than $8 per hour and works part-time. For some parts of the analysis, we also distinguish between quarters in which women do not work at all but receive no money from AFDC (no work, no welfare) and quarters in which women do not work at all and receive AFDC benefits (no work, welfare). Note that under this classification scheme, we make no distinction between women who voluntarily choose to stay home and those who want to work but cannot find jobs. Similarly, we make no distinction between women who work in part-time jobs and/or cannot find full-time work and women who choose to work part-time while they raise a child or attend education or training programs.

IV. The Employment Patterns of Young Women

Table 1 presents the share of women who ever had each type of employment quarter between the ages of 18 and 27. Good jobs are relatively common among this group of women: almost three-quarters (73.5 percent) held a good job at some point during their early working years. But bad jobs are even more common: 98.2 percent of women worked at a bad job. Only 1 percent spend all their time out of work and less than one-tenth of a percent spend all their time out of work and on welfare. Thus, virtually all women spend some time in bad jobs, but the overwhelming majority (about three-quarters) will eventually work in good jobs. Even though good jobs are common, it is important to note that 26.5 percent of all women never find a full-time job paying over $8 per hour through age 27. These women spend their time working part-time and/or in low-paying jobs, some steadily and others intermittently with frequent periods of joblessness.

We now turn to a more detailed analysis of the employment transitions that lead women to good jobs, the extent to which good jobs last for extended periods of time, and the relationship between a woman's characteristics and the likelihood that she ever holds a good job or holds a good job for an extended period of time.

b. The Paths to Good Jobs

To analyze the ability of women to move from not working to working in bad jobs and on to good jobs, as well as their ability to retain good jobs, we examine the probability that women in one employment state in a given quarter shift to a different employment state in the following quarter. For example, we ask "what share of women who hold bad jobs in one quarter make the transition to good jobs by the next quarter?" There are 39 transition periods for each woman in the sample. When we examine transitions over the entire ten-year period, the transition probabilities are averaged over all women and all 39 potential transitions. When we examine transitions over a specified portion of the ten-year period, the transition probabilities are averaged over all women and the transitions that occur during the specified period.³

Table 2 shows that from their late teens to their late twenties, young women progress relatively quickly from joblessness to bad jobs and then gradually move on to steady employment in good jobs. About one out of five women without jobs finds a job from quarter to quarter but most of these jobs are bad jobs. Only 2.5 percent of women with no jobs one quarter jump directly to good jobs the following quarter. Once a woman makes the transition to a bad job, the probability that she still works in a bad job in the following quarter is 85 percent. Only 6 percent of women in bad jobs move into good jobs the following quarter, while 9 percent become jobless. Women who make the transition to a good job tend to keep them from quarter to quarter: 91.4 percent of women in good jobs are still in good jobs the following quarter.

The data in Table 2 also show that employment patterns vary considerably by age, even within our sample of young women. When women are completing their formal schooling and just entering the labor market, they spend the majority of their time working in bad jobs or not working at all (see column 4). As women age, their periods of joblessness decline and the percentage of time spent working in good jobs increases. For example, women are jobless or working in bad jobs in 86.6 percent of the quarters between the ages of 18 and 22, but in just 53.8 percent of the quarters between the ages of 26 and 27. Time spent jobless is not necessarily idle time. Indeed, between the ages of 18 and 22, women are in school during 40 percent of the quarters when they are jobless. Further, during nearly half of their bad job quarters, 18 to 22 year-olds were combining school and work. Once most women have finished their formal schooling, the amount of time spent working in good jobs increases dramatically, from 13.3 percent of all quarters between the ages of 18 and 22 to 40 percent between the ages of 23 and 25, to 46.2 percent between the ages of 26 and 27.

When young women are first entering the labor market, transitions from joblessness to employment--in primarily bad jobs--are common. However, as a woman ages, the likelihood of making the transition from joblessness to employment decreases substantially. A young woman between the ages of 18 and 22 who is jobless has about a one in four chance of finding employment in the subsequent quarter. By the time she reaches her late twenties, if she is still jobless the likelihood she will make the transition into employment falls by one-quarter, to just 15 percent.

For very young women, transitions from bad jobs to joblessness are actually far more common than transitions from bad jobs to good jobs. However, as a woman ages and presumably gains more experience in the labor market, she is more likely to make the transition from a bad job to a good job than to make the transition from a bad job to joblessness. Between the ages of 18 and 22, the likelihood that a woman in a bad job will become jobless in the next quarter is 10.1, percent while the likelihood that she will make the transition to a good job is just 4.7 percent. Between ages 23 and 25, the likelihood of a transition from a bad job to joblessness declines to 7.3 percent and the likelihood of a transition from a bad job to a good job increases to 8.1 percent. The likelihood of making both of these transitions remains the same for women between the ages of 26 and 27.

In general, once a woman finds a good job she tends to keep it, and good jobs become more stable over time: 86 percent of women in good jobs between the ages of 18 and 22 remain in good jobs in the subsequent quarter, compared to 93.6 percent of women in good jobs between the ages of 26 and 27. Between the ages of 18 and 22, a young woman in a good job has a 10.9 percent chance of moving to a bad job and a 3.1 percent chance of becoming jobless in the next quarter. By ages 26 and 27, the chance of moving from a good job to a bad job declines to 5 percent and the chance of becoming jobless declines to 1.4 percent. In general, women in good jobs are also far less likely to become jobless than women in bad jobs. Just 1.9 percent of women in good jobs become jobless in the following quarter, compared to 9 percent of women in bad jobs. The limited transitions out of good jobs show that a substantial fraction of women who make the transition into a good job are able to sustain well-paid employment over an extended period of time.

Although women move into good jobs throughout this time period, initial transitions into good jobs actually occur quite early. Table 3 shows that about one-fifth of the women who ever work in a good job do so by age 19 and three-quarters do so by age 23. Transitions to good jobs that last for at least a year occur more slowly: only 11 percent occur by age 19 and 70 percent by age 23. The transition from a bad to a good job occurs quickly for some, but not for all women. On average, transitions from bad jobs to good jobs occur four years after women are first observed working in a bad job, but one-quarter of women who make the transition from a bad to a good job do not do so until six years after their entry into a bad job.

c. The Importance of Personal Characteristics

A substantial fraction of young women spend some time working in good jobs, and most of these women are firmly established in good jobs by their late twenties. However, the data presented in Table 4 indicate that a woman's personal characteristics play an important role in determining whether or not she ever holds a good job or holds a good job steadily. Three groups of women have especially low probabilities of ever holding a good job and of holding a good job steadily by their late twenties: (1) women with low levels of education, (2) black women, and (3) women with children.

Women who fail to complete high school are very unlikely to make the transition to a good job and work primarily in a good job by the ages of 26 and 27.⁴ Only 47.4 percent of women who have not completed high school ever work in a good job, and only 17 percent work primarily in good jobs by their late twenties. Women who completed high school are far more likely to make the transition to good jobs than women with less than twelve years of schooling. However, just over one-third of women with twelve years of schooling were working primarily in good jobs by their late twenties. It is only among women who have completed some post-secondary education, regardless of whether they received a college degree or not, that one finds a majority (61.4 percent) working steadily in good jobs by the time they reach their late twenties.

Although the majority of women with children work steadily by the ages of 26 and 27, steady work in good jobs is not the norm for this group of women. Only one-quarter of women with one or more children work primarily in a good job by their late twenties. In contrast, more than twice as many women without children (60.4 percent) work primarily in good jobs by ages 26 and 27.

Black women are also substantially less likely than white or other minority women to ever work in a good job or to work steadily in a good job by ages 26 and 27. The gap is especially wide between black and white women. About half of all black women ever work in a good job, and 30.5 percent work steadily in a good job by the ages of 26 and 27. Over three-quarters of white women work in good jobs, and 47.7 percent work primarily in good jobs by their late twenties. More than two-thirds of all Hispanic and other non-black minority women spend some time in good jobs and about 40 percent work steadily in good jobs by ages 26 and 27.

Not surprisingly, relatively few women who ever turn to the welfare system for support ever work in a good job and only a very small fraction are firmly established in a good job by ages 26 and 27. Only 42.9 percent of this group of women ever work in a good job. While almost half work steadily by ages 26 and 27, just 13.2 percent work steadily in good jobs.

Examining work transitions for women with selected characteristics (Table 5), we find that several factors contribute to the lower representation of less educated women, minorities, and mothers among women who ever hold good jobs and among women who work steadily in good jobs by their late twenties. First, these women are less likely to make the initial transition from joblessness to working than other women. Fewer than 15 percent of black or other minority women, women with children, or women with less than a high school degree who are jobless in one quarter make the transition to employment in the following quarter. In contrast, between 20 and 25 percent of women without children, women who have completed high school, and white women make this transition. Second, once black women, women who have not completed high school, and women with children make the transition from joblessness to a bad job, they are at least twice as likely to leave a bad job to return to a state of joblessness as they are to make the transition to a good job. For example, 11.3 percent of black women who work in a bad job in one quarter are jobless in the next quarter, while only 4.8 percent make the transition from a bad to a good job. Among women who are most likely to make the transition to a good job, the likelihood of moving from a bad job to no job or from a bad job to a good job are roughly equal. For example, 6.1 percent of whites in bad jobs make the transition to good jobs, while 8.5 percent become jobless.

In general, once women make the transition to good jobs, they are likely to stay in those jobs, regardless of their characteristics. In addition, there are only modest variations in the speed at which different women make the transition from bad to good jobs. For example, on average, white women make this transition in four years, while black women do so in four and a quarter years. On average, high school dropouts and high school graduates with no post-secondary schooling both make the transition to good jobs in about four years, compared to four and a quarter years for women with some education beyond high school.

d. Prospects for Welfare Recipients

Current initiatives to increase the work effort of welfare recipients differ dramatically from previous attempts. This makes it nearly impossible to rely on the work experiences of current or former welfare recipients to predict how they might fare under a program that requires most recipients to find employment as quickly as possible. Thus, to estimate how current recipients may fare under a newly designed welfare system that emphasizes employment, we developed a microsimulation model based on the work experiences of women who have never received welfare.

To identify the factors that determine whether a woman eventually makes the transition to stable employment in a good job, we use a multivariate framework known as a competing risk model, where the competing risks are the various employment states. This framework allows us to estimate how various characteristics, including demographics, family status, education and mastery of basic skills, local labor market conditions, previous work experience (in good and bad jobs), and length of time in the current employment state affect the probability of making the transition from one employment state to another, while holding other characteristics constant. We estimate three separate competing risk models, one to capture transitions from each of our three employment states. Using the estimates from these models, based on a woman's observed characteristics in quarters when she is jobless, we estimate the probability that she stays jobless, moves to a bad job, or moves to a good job. In quarters when she is in a bad job, we estimate the probability that she stays in a bad job, moves to a good job, or becomes jobless. Finally, if a woman is in a good job we estimate the probability that she stays in a good job or moves to a bad job or to a state of joblessness. We then use the estimates from these models and the observed characteristics of women who ever turn to the welfare system for support to simulate the quarter to quarter transitions for welfare. (The actual coefficients from the competing risk models are difficult to interpret and are, therefore, included in the appendix along with a more detailed discussion of the simulation procedures. The general pattern of the results also described in the appendix parallels the descriptive results presented earlier.)

There are two important assumptions underlying this approach to modeling the potential employment outcomes for welfare recipients. First, we assume that in the face of strict work requirements the employment experiences of welfare recipients will mirror the experiences of women with similar characteristics who never rely on welfare. Thus, if unobserved characteristics associated with a young woman's reliance on welfare reduce the likelihood that she will make the transition from joblessness to a bad job and from a bad job to a good job, our approach will overestimate the employment outcomes for welfare recipients. Second, for our simulations we rely on the observed characteristics of women in our sample who have ever used welfare. Some of these characteristics reflect decisions these young women made about completing school and having children. If these women faced the current welfare laws at the time they were making these decisions, they might have behaved differently. Our simulations, however, do not incorporate potential behavioral changes that may result in response to new program mandates.

The simulation results presented in Table 6 suggest that if women who ever turn to the welfare system for support followed employment paths comparable to those of women who never use welfare, we could expect a substantial increase in employment. But most of these women would be working in bad jobs. We estimate that 61 percent of all welfare recipients would work steadily by the time they reach their late twenties, with 36.9 percent working primarily in bad jobs and about one-quarter working primarily in good jobs. Among the remaining 39 percent of recipients not working steadily, the vast majority spend most of their time out of the labor market.

The data in Table 6 also suggest that we can expect even less positive employment outcomes for welfare recipients with characteristics generally associated with long-term welfare receipt. We predict that only about half of all welfare recipients who have not completed high school will work steadily by their late twenties, and only 15.2 percent will work primarily in good jobs. Among women who gave birth to their first child prior to turning 18, we predict that more than half will work steadily but less than one out of five will work primarily in good jobs. Women with three or more children will also experience high rates of joblessness and low rates of working steadily in a good job: just 43.7 percent will work steadily, with 31 percent working primarily in bad jobs and the remaining 12.7 percent working primarily in good jobs. Similar to the employment patterns of young women in general, black women who ever use welfare are substantially less likely to work primarily in a good job by their late twenties than white non-Hispanic, Hispanic, and other minority women. We estimate that 28 percent of white women, 31.5 percent of Hispanic and other minority women, and just 16.5 percent of black women who ever use welfare potentially could be employed primarily in good jobs by ages 26 and 27.

V. Conclusion

Over time, a substantial fraction of young women do make the transition from bad jobs to good jobs. In general, these transitions tend to occur rather slowly, usually four years after the start of a bad job. But not all women are equally likely to make these transitions. Women with children, women who have not completed high school, and minority women are all far less likely than women in general to ever work in a good job by age 27 and to work steadily in a good job. Since women who ever turn to the welfare system for support are over-represented on all of these dimensions, the prospect of the majority of welfare recipients becoming completely self-supporting in a short period of time does not seem promising. If welfare mothers were to achieve the same employment outcomes as women with similar characteristics, one would expect about a quarter of women who ever use welfare to work primarily in good jobs by their late twenties, leaving the vast majority working primarily in bad jobs or only working sporadically. The employment outcomes for the most disadvantaged recipients are even more discouraging; only 14.4 percent of welfare recipients who have not completed high school can be expected to work steadily in good jobs by the time they reach their late twenties.

This suggests that substantial numbers of women who leave the welfare rolls are likely to need additional support to help them meet work expenses such as child care and transportation. Some may also need additional support to meet these expenses for far longer than they are eligible to receive them under current law, especially if more and more women are required to enter the labor force when their children are very young. In addition, policies that provide additional income to working families such as the Earned Income Tax Credit and child support are likely to be especially important to women who are unable to make the transition from a bad job to a good job or to work steadily in a good job.

It is important to note that, even though a substantial fraction of welfare mothers are unlikely to make the transition from a bad job to a good job, if welfare mothers were to achieve the same employment outcomes as non-welfare mothers with similar characteristics, their employment outcomes would be far better than they are currently. In general, welfare mothers' employment experiences consist of sporadic employment. If this pattern of sporadic employment were replaced with steady employment in a bad job, the likelihood of eventually working steadily in a good job would increase substantially. Nonetheless, steady work is no guarantee of moving from a bad job to a good job. Among women who had not completed high school and who worked steadily from age 18 to 25, only 40.4 percent worked steadily in a good job by ages 26 and 27. In addition, steady work is not the norm among women who share the characteristics of many welfare recipients. Only about half of high school dropouts and 60 percent of women with children worked steadily by their late twenties.

Finally, it is important to note that because this analysis focuses on women's early experiences in the labor market and includes all women who have ever received welfare, it reflects the potential employment outcomes of women who are receiving welfare for the first time rather than women who are currently receiving assistance. Women with the least favorable prospects for making the transition from a bad to a good job are even more over-represented among women currently receiving assistance. This suggest that the employment prospects of current recipients are likely to be even less favorable than those of women who ever receive welfare. Because many current recipients have spent substantial periods of time out of the labor force, they are likely to experience difficulty making the transition both from joblessness to a bad job and from a bad job to a good job.

Appendix

To simulate the work experiences of welfare recipients, we develop a relatively simple probalistic microsimulation model that is based on three multivariate competing risk models: one each for the transitions from joblessness, bad jobs, and good jobs. In each model, the dependent variable is the transition from one employment state to another as indicated by a change in employment status from one quarter to the next. For example, the competing risks for a period of nonemployment are moving into a bad or a good job. For periods of time spent in bad jobs, the competing risks are moving into joblessness or into a good job, and for periods of time spent in good jobs, the competing risks are moving into a bad job or joblessness.

The competing risk models are based only on the observations for women who are never observed on welfare.⁵ For each quarter that a woman is in a particular employment state, she contributes one person-quarter to the data to analyze the duration of and transition from that employment state. The competing risk model for the transition from joblessness is based on 13,571 person-quarters, the model for the transition from bad jobs is based on 28,185 person-quarters, and the model for the transition from good jobs is based on 16,978 quarters.

A competing risk model provides a framework that allows us to estimate how various characteristics, including demographics, family status, education and mastery of basic skills, local labor market conditions, previous work experience (in good and bad jobs), and the length of time in the current employment state affect the probability of making the transition from one employment state to another, while holding other characteristics constant. We include an identical set of variables in each of the three models. The variable means can be found in Table A-1. The model results are presented in Table A-2. Below we provide a description of some of the variables included in the models, then highlight some of these results, most of which mirror the results presented in the body of the paper.

II. Variables Included in the Model

Demographic variables included in the model include a woman's age and her race/ethnicity. To capture the influence of a woman's role models on her employment choices, we include a variable that indicates whether a woman lived in a household with a working parent at the age of 14.

We also include several variables in the model to capture a woman's family status. We include variables that identify both the number of children and the age of the youngest child. We also include variables to include a woman's current marital status.

To measure the impact of the local labor market on the likelihood that a woman stays in her current employment state or moves to an alternative state, we include variables that identify the region in which a woman lives, whether she lives in an urban area, and the local unemployment rate. Because we do not have data on the actual unemployment rate, we classify the unemployment rate in an area as high (greater than 12 percent), medium (between 6 and 12 percent), or low (below 6 percent).

Because the transitions in and out of the labor market and from bad to good jobs often coincide with a woman's transitions in and out of school, we include several variables in the model to capture these transitions. First, we include a variable that indicates whether a woman is in school, distinguishing between whether she is in high school or college. Then, we include a variable to indicate whether a woman is within one or two quarters of completing school, again distinguishing between whether she is in high school or college. We also include a variable that indicates whether a woman is a high school dropout, defined as being out of school and having completed fewer than twelve years of schooling. Finally, we include a variable that identifies whether a woman has completed college.

To capture the importance of the mastery of basic skills, we include a variable that identifies women who fall in the bottom decile or bottom quartile of the distribution of scores on the Armed Forces Qualifying Test (AFQT) for women of the same age group. The AFQT score is based on the Armed Services Vocational Aptitude Battery (ASVAB), a series of vocational aptitude and basic skills tests administered in 1980 to all youth who were surveyed for the NLSY in that year.

III. Highlights of the Model Results

Not surprisingly, a woman's transition between employment states is significantly affected by the amount of education she has and whether or not she is in school. While women are in school, their employment status is relatively unstable. However, once young women are within one to two quarters of completing their education, the likelihood that they will move from joblessness to a bad job or from a bad job to a good job increases substantially. Once a woman has completed her education, more schooling is associated with better employment outcomes. Women who have not completed high school are more likely to lose jobs once they become employed, while women who have completed college have a significantly better chance of making the transition to a good job, either directly from a state of joblessness or from a bad job.

The presence of young children affects women's transitions in and out of the labor market in several important ways. Having a child under the age of five significantly reduces the likelihood that a woman who is jobless will make the transition to employment. Having a child under the age of one significantly reduces the likelihood that a woman will move from a bad to a good job and increases the likelihood that a woman employed in a good or bad job will make the transition to a state of joblessness.

The structure and condition of the local labor market affect the likelihood that a woman will find employment if she is jobless and will make the transition from a bad to a good job once she is employed. Women living in areas of lower unemployment are more likely to find employment if they are jobless and to move from a bad to a good job than women living in areas with higher levels of unemployment. Women living in the Midwest and the South are less likely than women in the Northeast to make the transition from bad to good jobs. Good jobs in the Midwest and the South also tend to be less stable than in the Northeast. Employment outcomes are also somewhat better for women living in urban areas. Women living in or near cities are significantly more likely to find a job if they are without one and to move from a bad job to a good job than are women in non-urban areas.

A woman's transition from her current employment state is also significantly affected by the length of time she has been in her current state and her previous work history. A woman's previous work experience increases the stability of her current employment and increases the likelihood of moving from a bad to a good job. Time spent working in a bad rather than good job reduces a woman's employment stability by increasing the likelihood that she will lose her employment altogether or that she will move from a good to a bad job. In general, transitions from one employment state to another occur quite rapidly. The longer a woman stays in a particular employment state, the less likely she is to leave it.

IV. How the Microsimulation Model Works

To simulate the employment patterns of welfare recipients in the current welfare reform environment, we calculate the probability of moving from one employment state to another in each of the 40 quarters for which we have data. The probabilities are calculated based on the observed characteristics of each woman who has ever used welfare and the results of the competing risk models that are presented in Table A-3. To determine a recipient's employment status for the next quarter, this calculated probability is compared to a number (r) between 0 and 1 that is drawn randomly from a uniform distribution. For transitions from joblessness, the likelihood of leaving joblessness for a bad job (p_bad) or a good job (p_good) are evaluated simultaneously. Thus, a recipient's status is determined according to the following:

Transitions from bad and good jobs follow exactly the same pattern. The simulation is repeated for each quarter and for each recipient.

To test the efficacy of the simulation model, we simulated the employment transitions for women who have never used welfare. Because the competing risk models for this group of women are based on data from this group of women, on average, we would expect the simulation to produce a transition matrix that we actually observe. A comparison of the two transition matrices indicates that the model reproduces the observed employment experiences of this group of women almost exactly. Copies of the simulated transition matrix are available from the authors upon request.

* indicates significance at the 90% level
** indicates significance at the 95% level

The nonpartisan Urban Institute publishes studies, reports, and books on timely topics worthy of public consideration. The views expressed are those of the authors and should not be attributed to the Urban Institute, its trustees, or its funders.

Usage, posting and reprint of materials on the UI web site:

Most publications may be downloaded free of charge from the web site in PDF format. This information may be used and copies made for research, academic, policy or other non-commercial purposes. Proper attribution is required.

Copyright of the written materials contained within the Urban Institute website is owned or controlled by the Urban Institute. Posting UI research papers on other websites is permitted subject to prior approval from the Urban Institute—contact paffairs@urban.org.

If you are unable to access or print the PDF document please contact us or call the Publications Office at (202) 261-5687.