Printer Friendly

Welfare recipients: how do they become independent?

This research used data concerning recipients' employment, receipt of Aid to Families with Dependent Children (AFDC), receipt of Temporary Assistance for Needy Families (TANF), and poverty status to develop a typology of adaptations by welfare recipients. Using U.S. Department of Labor survey data, a sample of AFDC/TANF recipients was analyzed through event history analysis. The results show that welfare reforms launched in 1996 moved dependent recipients out of welfare but had no effect on working recipients' chances of leaving welfare. New two-year limits on unbroken program participation (and a five-year lifetime limit) pushed many unprepared recipients into poverty, working or not. Economic conditions became worse for working poor people than for those on welfare. The study also found that some former welfare recipients did go to work and eventually leave welfare and poverty. Occupational skills, work experience, child support, marriage, and experience in dependency or supplementation were among the factors promoting such a change.

Key words: autonomy; dependency; self-reliance; supplementation; welfare recipients


Since the Personal Responsibility and Work Opportunity Reconciliation Act (P.L. 104-193) was passed in 1996, "reformed" welfare programs have been experimenting with new policies. Passage of this cornerstone legislation replaced decades-old Aid to Families with Dependent Children (AFDC) with Temporary Assistance for Needy Families (TANF). The TANF program emphasizes the "welfare-to-work" principle. To help welfare recipients leave welfare and join the workforce, state TANF programs generally provide employment support services (for example, child care, limited medical coverage) and enforce aid restrictions (for example, time-limited benefits, work quotas, and sanctions). Welfare recipients who are able-bodied adults are required to be employed within two years of enrolling in a TANF program. Those who do not find employment lose all benefits. Furthermore, TANF benefits are limited to five years over a recipient's lifetime. The new approach is controversial. Not only has the sufficiency of the child care and medical benefits been called into question, the new time restrictions have been said to force many unprepared recipients to leave welfare. After leaving welfare, many people are likely to continue living in poverty, despite being employed (McCrate & Smith, 1998; Riemer, 1997). Therefore, it is important to understand the effect of welfare reform on recipients' employment and welfare use, as well as on their financial status after leaving welfare.


Financial independence means succeeding economically to the point of supplying one's own needs for shelter, food, clothing, and other necessities. Financial independence is achieved typically through employment; it is impeded by welfare use. Financial independence constitutes a continuum. Some people are employed for higher wages than others, for a variety of reasons. Many who are employed live well above the poverty line, but many live below that line, despite having regular employment. Some of these "working poor" individuals need public assistance (or other aid) to pay ordinary living costs. Whereas some are employed and still must participate in welfare, others rely entirely on public aid. This article presents a "typology of adaptations" that helps explain the differential behavior of welfare recipients.

This study used concepts such as welfare use, employment, and poverty status to define four types of adaptation exhibited by welfare recipients--dependency, supplementation, self-reliance, and autonomy (see Figure 1). Dependency on welfare is defined as reliance on AFDC or TANF to meet all financial needs. Utterly dependent welfare recipients are unemployed, generate no earnings, and live in poverty. Some welfare recipients live in poverty yet engage in supplementation (symbolized by the area overlapped by the "employment" and "welfare use" blocks in Figure 1). They are workers who earn a regular wage, yet still require AFDC/TANF to care for their families. Self-reliance, describes the many welfare recipients who leave AFDC/TANF but do not leave poverty, despite being employed. They are generally called the "working poor," and they live in deprivation. Autonomy describes recipients who leave AFDC/ TANF and leave poverty by becoming employed. Families demonstrating any of the four modes of adaptation may use public assistance other than AFDC/TANF--such as food stamps and housing benefits. The study reported in this article focused on the change from AFDC/TANF dependency or supplementation to other adaptation modes and on the factors that contributed to that change.



Dependence on welfare is a controversial topic. Many factors may be related to the mode or modes of adaptation (see Figure 1). Occupational skills, personal earnings, and educational level are usually associated with an individual's employment. Possessing education and occupational skills has been shown to enable recipients to become employed and eventually achieve autonomy (Bora, Caudill, Spera, & Kunz, 1998; Harris, 1993; Kroch & Sjoblom, 1994; Meyer & Cancian, 1998; Nichols, 1979; Yoshikawa, 1999). Limited education has been found to be associated with longer periods of supplementation or dependency (Bane & Ellwood, 1983; Blank & Ruggles, 1996; Coe, 1981; Goodwin, 1983; Petersen, 1995), and with a reduced likelihood of exit from dependency or supplementation (Gault, Hartmann, & Yi, 1998; Hardina, 1999; Taylor, 1999). However, some studies (Cheng, 1995; Hutchens, 1981; Plotnick, 1983) showed no significant effect of educational level on AFDC recipients' chances of leaving welfare (that is, exiting dependency or supplementation). One plausible reason for these contradictory findings is that none of the studies examined interventions to increase educational level. The role of personal earnings is also influential in the adoption of an adaptive mode following exit from dependency. Almost 39 percent of welfare recipients who exited dependency or supplementation did so because of significantly improved wages (Bane & Ellwood, 1994).

Under typical welfare policy, receipt of AFDC/ TANF provides access to other assistance and services. Many argue that the need for a range of assistance (for example, public housing, food stamps, and Medicaid) encourages families to remain in dependency or supplementation rather than seek autonomy or self-reliance. Losing these auxiliary benefits is a threat to the survival of families and children. Many families who have left AFDC regularly run out of food and forgo medical care (Kohlert, 1989; Moscovice, Craig, & Pitt, 1987; U.S. General Accounting Office, 1985; Wodarski, Parham, Lindsey, & Blackburn, 1986).

Child support enforcement and, especially, the institution of the 1996 welfare reforms may promote certain modes of adaptation over others. Child support enforcement was one of the important elements of welfare reform in the 1990s. Although only 20 percent of welfare mothers receive child support, the new cap on cash assistance will likely encourage them to actively seek fathers' support (Wolk & Schmahl, 1999). After 1996 and its reforms, some state programs moved up to 30 percent of TANF recipients into self-reliance or autonomy in a dramatically short period (Carter, Dunbar, Micks, Brown, & Davis-Johnson, 1997).

Personal experience in welfare programs as well as employment probably influences the mode of adaptation. Because as many as 50 percent of individuals who leave welfare will return to it (Bane & Ellwood, 1994; Meyer & Cancian, 1998), the "welfare history" of recipients affects their mode of adaptation. Spending long periods in welfare programs increases the likelihood recipients will remain on welfare or return to it if they do leave (Bane & Ellwood, 1994; Harris, 1996). Moreover, chronic unemployment and lack of work experience significantly reduce the chances of AFDC recipients leaving dependency and supplementation (Hutchens, 1981; Petersen, 1995; Plotnick, 1983; Taylor, 1999; U.S. Bureau of the Census, 1989; White, 1996).

Individual characteristics--single motherhood, need for child care, and ethnicity--have been shown to be related to the dependency and supplementation modes of adaptation. Although 19 percent to 30 percent of AFDC recipients are married women (Moffitt, Reville, & Winkler, 1998), studies have shown consistently that single mothers are likely to stay longer in supplementation or dependency than married people (Bane & Ellwood, 1983; Cheng, 1995; Coe, 1981; Dickinson, 1986; Osmond & Grigg, 1978). Employed welfare mothers with preschool-age children are likely to use or pay for child care (Gault et al., 1998; Meyers & Heintze, 1999; Meyers & van Leuwen, 1992). Finally, most studies have indicated that white recipients are more likely to leave dependency or supplementation than ethnic minority recipients (Bane & Ellwood, 1983; Boskin & Nold, 1975; Coe, 1981; Osmond & Grigg, 1978; Piskulich, 1993; U. S. Bureau of the Census, 1989). However, a recent study (Taylor, 1999) using data from an experimental regional program found that ethnicity had no significant effect on the dependency or supplementation modes.



This study is a secondary data analysis of participants in the National Longitudinal Survey of Youth 1979 (NLSY79). The NLSY79 collected information about the labor force experience of 12,686 nationally representative people, from 1979 to 1998 (U. S. Department of Labor, 1999). All participants in the NLSY79 were interviewed yearly from 1979 to 1994, and then biennially. There were 18 interviews per participant. In each interview, the survey researchers collected from the participants' information for the preceding year. The information in NLSY79 provides an empirical base for examining exits from AFDC/ TANF and relevant social changes. The 18 interviews resulted in a longitudinal record for each individual.

Because data in the NLSY79 were measured on a yearly basis, the discrete-time method was used to process data for the study (Allison, 1984; Yamaguchi, 1991). First, all individual records were separated into two samples--one for dependency spells and the other for supplementation spells. Each dependency spell began with the dependency mode and ended with an exit to another adaptation mode. Similarly, a supplementation spell began with the supplementation mode and ended in a different mode.

However, some spells, known as fight-censored spells, did not end with exits at the end of survey. Right-censored spells were included in this study because the date at the end of the survey was assumed to be independent of the occurrence or nonoccurrence of the exit (Yamaguchi, 1991). Furthermore, one aim of this study was to explore the modes of adaptation individuals exhibited only after reaching age 18. Thus, only the dependency and supplementation spells of respondents after they turned 18 were analyzed. Recipients who were older than 18 in 1979 generated the study's "left-censored" spells. Left-censored spells were excluded from analysis, because information for the period prior to the survey was unknown.

Each dependency or supplementation spell was divided into person-years, for which outcome and explanatory variables were measured. Some explanatory variables changed over time and were recorded in each person-year. Other explanatory variables were constant over time. Ethnicity or race, for instance, does not change. Samples in the present study included individuals who were African American, white, and Hispanic. Some recipients reported more than one spell (whatever the particular adaptation mode) within the NLSY79 survey period. Recurrent spells were included in this study for their potential to enrich the understanding of welfare dynamics. A respondent's recurrent spells did not always end in a single type of exit. For example, the explanatory variable for one spell could differ greatly from that of a subsequent spell. Thus, the history of prior supplementation or dependency describing each spell served as an explanatory variable and provided a connection or control between recurrent spells.

The sample for dependency comprised 978 individuals, involved in 1,579 dependency spells, who met the age criterion. The ultimate sample for data analysis of the dependency spells was composed of 4,705 person-years. In contrast, 926 people 18 and older involved in 1,345 supplementation spells provided 2,939 person-years for analysis of supplementation.


Outcome Variable. The outcome variable is the type of adaptation demonstrated by AFDC/TANF recipients. In analysis of exits from dependency, the outcome variable could be continued dependency on AFDC/TANF or one of three types of exit from dependency: (1) An "exit to autonomy" was noted if an individual reported receiving some AFDC/TANF payment and had no earnings in year t-1 but had earnings and reported no AFDC/ TANF payment and was not poor, in year t. (The estimation of a family's poverty status was based on the federal poverty threshold, in light of the family's size and its rural or urban location.) (2) An "exit to self-reliance" was noted if an individual received some AFDC/TANF payment and had no earnings in year t-1 but reported no payment and had some earnings, although he or she remained poor, in year t. (3) An "exit through supplementation" was noted if an individual reported some AFDC/TANF payment and no earnings in year t-1 but reported some earnings as well as AFDC/TANF payment in year t. Individuals who registered "no exit from dependency" were those who reported having AFDC/TANF payment and no earnings in both year t-1 and year t. This "nonoccurrence of the exit" served as a reference category in the analysis.

In the analysis of exits from supplementation, the reference category was a "nonoccurrence" of exit from a supplementation spell. A nonoccurrence was noted if recipients received AFDC/ TANF payment and had earnings in both year t-1 and year t. An exit to autonomy was noted when working recipients had earnings but received no AFDC/TANF payment and were not poor in year t. Exit to self-reliance was noted if working recipients continued to have earnings, received no AFDC/TANF payment in year t, and lived below the poverty line in year t. An exit to dependency was noted when recipients received AFDC/TANF benefits and had earnings in year t-1, but in year t continued to receive benefits and no longer had any earnings.

Explanatory Variables. The study's hypotheses included an assumption that the following four groups of explanatory variables would influence the likelihood of AFDC/TANF recipients changing from dependency or supplementation to other modes of adaptation.

1. Employment--Occupational skills, modified from a scale developed by Rexroat and Shehan (1986), denotes the type of occupational skills exhibited by an individual on the job. Four dummy variables--professionals/managers, craftsmen, clerical and sales workers, and operatives and farmers--were created. Service workers and unemployed people were combined in the reference category, because most people without specialized experience or training can perform skills required in service. "Unemployed" people were never employed during the sample period. In this study, recipients who became dependent on AFDC/TANF, although unemployed, were nevertheless assigned the occupational skills suggested by their employment in the year prior to the dependency spell. The variable educational level was the number of years of schooling completed by a respondent.

2. Welfare Policies--The variable food stamps equaled the value of the food stamps received by a family in a year. Public housing or rent subsidy indicated receipt of housing or rent subsidy by the family. Child support was the annual amount of payment received by the custodial parent from the noncustodial parent. The welfare reform variable indicated whether the person-year in question occurred after the launching of TANF in 1996.

3. Personal Experience--Years of work experience was the total number of years of employment prior to each interview. The following variables provided a connection or history between recurrent spells. Number of times in dependency was the number of dependency spells experienced prior to the current dependency spell. Number of times in supplementation was the number of supplementation spells experienced prior to the current supplementation spell. Total number of years in dependency in the past was the accumulated number of years an individual spent in dependency prior to the current year. Total number of years in supplementation in the past was the accumulated number of years an individual spent in supplementation prior to the current year.

4. Individual Characteristics--Race or ethnicity represented which of the three major racial or ethnic groups--African American, Hispanic, and white--a recipient of AFDC/TANF belongs to. "White" was the designated reference category--that is, statistical comparisons were made between African Americans and whites, as well as between Hispanics and whites. Single mother indicated whether a recipient was a single mother. Need for child care indicated that the youngest child in the family was below age 6.


There were the inevitable inconsistencies over the years in the NLSY79 survey's questions and variables. For instance, in 1993 no information was sought about the age of the family's youngest child. To address this problem in the present study, that variable was estimated from relevant information collected in 1992 and 1994.

The biennial interviews in 1996 and 1998 created a problem for the research. Information collected in 1995 and 1997 was used in this study because certain variables were not recorded in 1994 and 1996. Hence, there are two gaps (the years 1994 and 1996) in each longitudinal record in this study. On the other hand, assessments of the effect of TANF rely heavily on information from 1997, the year after the reforms.

The NLSY79 data used in this study differs in important ways from data sets such as the Michigan Panel Study on Income Dynamics (PSID) and the Survey of Income and Program Participation (SIPP). The SIPP and recent PSID obtained data from respondents monthly, increasing the accuracy of collected information. Nevertheless, the 18-year information from the respondents in NLSY79 is an invaluable source for understanding welfare recipients' adaptation dynamics.


Event history analysis was applied to analyze the probability of the occurrence of an event (Allison, 1984). The two events of interest were exit from dependency and exit from supplementation to another mode of adaptation. Separate analyses were made for exits from dependency and exits from supplementation. Because the outcome variable comprised four categories, multinomial logistic regression (instead of ordinary least squares or linear probability models) was selected for analysis of the samples (Aldrich & Nelson, 1984; Hosmer & Lemeshow, 1989; Menard, 1995). Under a logit transformation the regression equations used to test the hypothesis became ln(P(Y=h|X)/P(Y=0|X)) = [a.sub.h] + [SIGMA] [b.sub.hi] [x.sub.hi] (t) + [SIGMA] [b.sub.hj] [x.sub.hj]. On the left-hand side of the equation, the log-odds was a logarithm of the ratio of the probability of occurrence of an event, P(Y=h|X), to the probability of nonoccurrence of the event, P(Y=0|X). The probability of occurrence ranges from 0 to 1. The equation presents a linear relationship between the log-odds and the explanatory variables. The time-varying explanatory variables, ([x.sub.hi](t), could change over the course of the survey. In fact, all variables except one could change over time. "Race or ethnicity" was the only constant explanatory variable ([x.sub.hj]) that remained unchanged in all cases during consecutive person-years. The likelihood-ratio test statistic (G) between the hypothesized model and the null model was used to determine the statistical significance of the model. The null model consisted of only an intercept ([a.sub.h]). The test statistic used to compare the two models was G = -2(log [L.sub.0]-log [L.sub.1]); [L.sub.1] was the likelihood function for the hypothesized model, and [L.sub.0] was the likelihood function for the null model (Menard, 1995).

The focus of analysis was on statistical significance of each explanatory variable in predicting the log-odds for each event. The sign and value of the coefficient (b) of a predictor indicated the direction and strength of its relationship with the log-odds of the event. Describing the same information from a different angle, the value of the odds ratio [exp(b) (greater or less than 1)] indicated whether the odds of an exit increased or decreased with an increase in the explanatory variable.


The correlation coefficients and tolerance statistics showed no multicollinearity problem for independent variables pertaining to dependency or supplementation spells.

Dependency Spells

Of 1,579 dependency spells, 112 ended with autonomy, 142 ended with self-reliance, and 579 ended with supplementation. Examination of the dependency spells found that 29.3 percent of recipients were white; 50.7 percent were African American, and the remainder were Hispanic. Moreover, 77.2 percent of the recipients were single mothers whose average age was 25.7 years. On average, recipients had two children and 11 years of formal schooling.

Respondents in the autonomy mode had personal earnings that were triple the earnings of those in supplementation spells, and total family income was 2.5 times that for those in dependency spells (see Table 1). However, for individuals in the autonomy mode, personal earnings made up only 35.4 percent of the total family income, and child support constituted only a small portion of total income. Obviously, income earned by a spouse or from some other source typically produced the financial achievement leading to autonomy. Although personal earnings represented the major proportion (58.3 percent) of total family income for families in the self-reliance mode, the total income of such families was only 67 percent of the income of families who remained in dependency. Entering self-reliance meant a great reduction of the total family income, and potential economic hardship. Because the average number of years in the current dependency spell was two-plus for those remaining in dependency and for those exiting it, enforcement of the two-year limit on TANF participation is likely to push many dependent recipients out of the program unprepared. They are more likely to adapt through self-reliance than through autonomy.

The overall test statistic (G = 2084.17, p < .01) for analysis of dependency spells confirmed that the hypothesized model is significantly different from the null model (see Table 2). This study thus showed welfare reforms launched in 1996 were related significantly to moving dependent recipients into autonomy (b = 1.869, p < .01) or self-reliance (b = 1.385, p < .01). The obtained odds ratio (exp(b)) also showed that the chance of entering autonomy after the introduction of the welfare reforms was 6.48 times the chance of entering autonomy before the reforms. The reforms did not simply help dependent recipients get jobs; they actually moved recipients out of welfare. However, 56 percent of the group that exited from dependency adapted through self-reliance, meaning economic hardship.

Dependent recipients who had experienced multiple dependency spells in the past were more likely to leave dependency. In other words, earlier experience making an exit from a dependency spell helped move recipients out of the current spell. This finding points to the history of recurrent dependency. However, the results show that the number of years in dependency in the past has no effect on the likelihood of leaving dependency. One plausible explanation is that the reforms of 1996 forced dependent recipients to leave TANF whatever their history of dependency.

More years of work experience improved the chance that a dependent recipient would take up a different adaptation mode requiring employment. In other words, years already spent in employment was crucial to obtaining a new job. In general, occupational skills helped dependent recipients to leave dependency. Among these skills, professional or managerial skills had an especially great effect on the chance of substituting autonomy for dependency. Possession of these kinds of skills did more to ensure an entrance into autonomy than the welfare reforms did. Receipt of child support payments was another variable significantly related to the likelihood of entering autonomy. Furthermore, the odds of marriage moving recipients to autonomy proved greater than the odds of marriage moving them to self-reliance or supplementation.

Unfortunately, education was found to be related to the likelihood of entering supplementation only. This result is not a surprise: The average number of years of education for this sample was fewer than 12, regardless of the mode of adaptation. In addition, the current policy's apparent emphasis on immediate employment leads to very few dependent recipients turning to higher education (high school or college) as a means to financial independence. Maintaining a job, accumulating work experience, and being promoted to the managerial ranks are other possible paths to autonomy for dependent recipients.

The study demonstrated that receiving fewer food stamps was significantly related to a better chance of exiting dependency for some other mode. This does not necessarily imply that using fewer food stamps promoted the exits, however. Recipients probably became eligible for fewer food stamps as their personal earnings rose. Similarly, ineligibility for housing benefits, not use of these benefits, was related to exits to autonomy. Food stamps and housing benefits were essential for recipients in dependency spells.

Interestingly, a need for child care was found to be significantly related to exits to supplementation rather than to continued dependency. This finding implies that having young children for whom child care can be found is an inducement for dependent recipients to seek employment--that is, to enter supplementation. Available child care was not related to the move from dependency to self-reliance or autonomy, however. The study results also show that Hispanics and African Americans were likely to remain in the dependency adaptation; white Americans were likely to leave dependency and enter the supplementation or autonomy mode.

Supplementation Spells

There were 1,345 supplementation spells, comprising 2,939 person-years. Almost 43 percent of recipients adapting by supplementation were African American; 17.9 percent were Hispanic. Nearly 61 percent were single mothers. The recipients adopting the supplementation mode of adaptation on average were 26.6 years old, had two children, and had completed almost 12 years of schooling. Of the supplementation spells, 427 eventually ended in dependency, 223 in self-reliance, and 336 in autonomy.

In 40 percent of cases in which a supplementation recipient left welfare, he or she entered self-reliance. Individuals who left supplementation to enter self-reliance made the bulk of their total family income through personal earnings--over 68 percent (see Table 3). This represents a strong reliance on employment. Although the personal earnings and child support payments received by such recipients went up, their income was on average only 52 percent of the income of families adapting through supplementation. The families involved experienced significant losses of food stamps, and the economic hardships they faced were enormous. On the other hand, the total family incomes of former recipients who left supplementation for autonomy typically doubled from the supplementation amount. However, improved personal incomes while in autonomy constituted only 42.6 percent of the total family incomes. Again, a spouse's additional income may explain these families' financial independence.

As many as one-quarter of the recipients in the supplementation mode spent more than five years in a welfare program before leaving it. Under the new five-year lifetime limit, these usually unprepared recipients became ineligible for TANF and likely entered self-reliance. Among the recipients who continued to adapt through supplementation, about 20 percent violated the five-year rule and became ineligible for TANF. They probably entered the self-reliance mode, and not by choice. The overall test statistic (G = 2377.99, p < .01) for analysis of supplementation spells confirmed that the hypothesized and null models are significantly different (see Table 4). This study showed that welfare reforms did not affect recipients' chances of leaving supplementation. This result is logical, since the two-year limit supposedly targeted dependent recipients, not working recipients. Recipients in the supplementation mode could continue to receive TANF as long as they did not violate the five-year lifetime limit. Examining these recipients' recurrent spells and welfare histories became very important in understanding changes in their adaptation. Many who left supplementation had experienced multiple spells in supplementation in the past. Their past experiences of leaving supplementation helped them exit from the current spell. Furthermore, the fewer years spent in supplementation in the past, the greater the likelihood of entering some other mode. The findings suggest that those who had multiple short spells of supplementation in the past were likely to exit from the current spell.

Recipients in supplementation had jobs; they were able to leave welfare thanks to work experience and occupational skills. Those with relatively more work experience were likely to enter self-reliance or autonomy. Interestingly, each kind of occupational skill was equally strong in predicting an entrance to autonomy. However, child support payments and a spouse's income also were involved in the financial independence obtained. Comparison of odds ratios confirmed that marriage provided a better chance to enter autonomy than occupational skills. Recipients who received larger child support payments had better chances of entering autonomy.

Because marriage is linked to exits to self-reliance, a spouse's earnings seem insufficient to lift a recipient family out of poverty after leaving welfare. In addition, work experience helps recipients leave welfare, but not poverty. Working recipients were likely to be operators or farmers, to have little education, and to be African American. As to the insulting myth of ethnic minority groups depending inordinately on welfare, the present study found that African Americans were more likely than white Americans to enter self-reliance, standing up to the economic hardship entailed by that adaptation mode.

As argued earlier, receiving fewer food stamps or nonreceipt of housing benefits did not necessarily promote shifts to autonomy or self-reliance. Again, these findings imply that higher personal earnings led to ineligibility for the maximum food stamp and housing benefits. The need for food stamps was not likely to be diminished for families entering self-reliance, yet many of these families reduced their use of food stamps. Whereas one study (Daponte, Sanders, & Taylor, 1999) found that ignorance of eligibility status and complicated criteria can play a role in decreased participation in food stamp programs, another study (Thompson & Gais, 2000) suggested that severing food stamp eligibility from TANF receipt may be the cause of declining participation.

As expected, those who were single mothers, who received less child support, and who had limited occupational skills and work experience were most likely to adapt through dependency. Food stamps and housing benefits had negligible effects on shifts to the dependency mode, challenging the myth that it is for the auxiliary benefits that welfare recipients sign onto the programs. The implication of the finding is that some uncontrollable factor--perhaps the loss of job--is largely behind an individual's entry into dependency.


In this study of AFDC/TANF recipients, a typology of adaptation was developed featuring four modes: dependency, supplementation, self-reliance, and autonomy. The typology provided a further angle for understanding welfare dynamics. It outlined the differences among adaptation modes, using factors such as income sources and experience in dependency and supplementation. The study presented factors related to the change from dependency to other modes, from supplementation to other modes, and from supplementation to complete welfare dependency. The research included an exploration of recurrent spells of adaptive modes, providing some understanding of cycles that occur in welfare dynamics. Future research might examine welfare dynamics among former recipients who adopted autonomy as an adaptation mode and were able to stay in it, or those who adopted self-reliance before ultimately entering autonomy.

The results of the present study offer some implications for welfare reform policy. Welfare reforms enacted in 1996 not only encourage dependent recipients to be employed, but also lift a number of families out of welfare and even out of poverty. On the other hand, economic growth contributed, although debatably, to as much as 33 percent of increased employment and decreased enrollment among welfare recipients (Kaushal & Kaestner, 2001). In addition, child care programs need to be sustained and broadened so that all dependent recipients are free to seek jobs.

Because having a job is no guarantee of leaving welfare, subsequent welfare reforms might do better to focus their attention on working recipients. Employed welfare recipients should be the subjects of strong efforts to help them, as well as their spouses, find better-paid jobs that eventually allow financial independence. For instance, programs could include (at a minimum) follow-up with working recipients, offering up-to-date information about employment opportunities. Moreover, further enforcement of child support decrees as begun under current policy can help even greater numbers of recipients move toward financial independence.

On the other hand, welfare reforms seem unlikely to stop supplementers from becoming completely dependent, or to boost supplementers out of welfare. In fact, the 1996 reforms pushed many welfare recipients out of TANF and into the ranks of the working poor, where sharply reduced incomes mean difficult lives. To ease economic hardships among this group and better enable them to improve their circumstances, continued provision of food stamps, child care benefits, Medicaid coverage, and mental health care after leaving welfare might be prudent. Rules such as the two-year and five-year limits might be amended to be slightly more generous, to avoid pushing desperate families into hardship. Extending the time limits might give many recipients needed time to stabilize employment, gain work experience, and earn promotions.

Many may tend to weight education heavily as a means of leaving AFDC/TANF programs. This study's findings cast doubt on that assumption. Aside from the fact that recipients in this study generally were not high school graduates, the new two-year limit gives adult recipients little opportunity for educational advancement. Two years' study might accomplish a GED or high school diploma, possibly an associate's degree. If the limit were four years, however, recipients could more realistically pursue education as a path to independence. Finally, programs that help recipients obtain higher education should be available to them even (perhaps especially) after they leave welfare.
TABLE 1--Descriptive Statistics for Welfare Dependency and Exits

 Remain in Exit to
 Dependency Supplementation
 (n = 3,872) (n = 579)

Annual Income Sources $ % $ %

Personal earnings 0 2,840.26 23.0
AFDC/TANF benefits 3,617.79 36.5 3,405.32 27.5
Food stamps 1,679.31 16.9 1,551.86 12.5
Child support 128.27 1.3 165.34 1.3
 Total family income 9,923.68 12,369.35
Number of consecutive years
 in a single dependency spell
 Mean 2.9 3.6
 More than two years (%) 40.1 57.7

 Exit to Exit to
 Self-Reliance Autonomy
 (n = 142) (n = 112)

Annual Income Sources $ % $ %

Personal earnings 3,905.23 58.3 8,665.68 35.4
AFDC/TANF benefits 0 0
Food stamps 506.43 7.6 430.75 1.8
Child support 315.54 4.7 270.52 1.1
 Total family income 6,692.99 24,452.08
Number of consecutive years
 in a single dependency spell
 Mean 3.5 2.9
 More than two years (%) 51.4 56.2

TABLE 2--Determinants of the Log-Odds of an Exit from Welfare
Dependency, by Adopting Supplementation, Self-Reliance, or
Autonomy (N = 4,705)

 (y = 1)

Variable b exp(b)

 Educational level 0.106 ** 1.11 **
 Operators/farmers 1.524 ** 4.59 **
 Clerks/salespersons 2.254 ** 9.52 **
 Craftsmen 1.413 ** 4.11 **
 Professionals/managers 2.731 ** 15.35 **
Welfare policies
 Food stamps -2.51x[10.sup.-4] ** 1.00 **
 Public housing or rent subsidy 0.086 1.09
 Welfare reforms 0.040 1.04
 Child support 9.71x[10.sup.-5] 1.00
Personal experience
 Years of work experience 0.792 ** 2.21 **
 Number of times in dependency 1.797 ** 6.03 **
Total number of years in
 dependency in the past -0.04 0.96
Individual characteristics
 Single mother -0.284 * 0.75 *
 Need for child care 0.721 ** 2.06 **
 Hispanic -0.661 ** 0.52 **
 African American -0.516 ** 0.60 **
Constant -4.722 **
G 2084.17 **

 (y = 2)

Variable b exp(b)

 Educational level -0.121 0.89
 Operators/farmers 1.857 ** 6.41 **
 Clerks/salespersons 2.285 ** 9.82 **
 Craftsmen 3.152 ** 23.39 **
 Professionals/managers 3.125 ** 22.76 **
Welfare policies
 Food stamps -0.001 ** .99 **
 Public housing or rent subsidy 0.231 1.26
 Welfare reforms 1.385 ** 3.99 *
 Child support 3.84x[10.sup.-4] ** 1.00 **
Personal experience
 Years of work experience 0.945 ** 2.57 **
 Number of times in dependency 1.966 ** 7.14 **
Total number of years in
 dependency in the past -0.096 0.91
Individual characteristics
 Single mother -0.538 * 0.58 *
 Need for child care 0.374 1.45
 Hispanic -0.318 0.73
 African American 0.072 1.07
Constant -3.119 **
 (y = 3)

Variable b exp(b)

 Educational level -0.110 1.12
 Operators/farmers 1.15 ** 3.17 **
 Clerks/salespersons 2.45 ** 11.57 **
 Craftsmen 2.533 ** 12.59 **
 Professionals/managers 3.386 ** 29.54 **
Welfare policies
 Food stamps -0.001 ** 0.99 **
 Public housing or rent subsidy -0.744 * 0.48 *
 Welfare reforms 1.869 ** 6.48 **
 Child support 3.60x[10.sup.-4] * 1.00 *
Personal experience
 Years of work experience 1.022 ** 2.78 **
 Number of times in dependency 2.034 ** 7.64 **
Total number of years in
 dependency in the past -0.1 0.91
Individual characteristics
 Single mother -1.463 ** 0.23 **
 Need for child care 0.438 1.55
 Hispanic -1.097 ** 0.33 **
 African American -0.467 0.63
Constant -5.051 **

* p < .05.

** p < .01.

TABLE 3--Descriptive Statistics for Welfare Supplementation and Exits

 Remain in Exit to
 Supplementation Dependency
 (n = 1,953) (n = 427)

Annual Income Sources $ % $ %

Personal earnings 4,428.86 31.0 0
AFDC/TANF benefits 2,865.44 20.1 4,103.09 48.3
Food stamps 1,368.31 9.6 1,833.29 21.6
Child support 193.59 1.4 154.70 1.8
 Total family income 14,290.28 8,492.15
Total number of years on welfare
Mean 3.0 4.4
More than 5 years (%) 19.9 31.4

 Exit to Exit to
 Self-Reliance Autonomy
 (n = 223) (n = 336)

Annual Income Sources $ % $ %

Personal earnings 5,050.68 68.1 12,245.49 42.6
AFDC/TANF benefits 0 0
Food stamps 573.61 7.7 338.01 1.2
Child support 344.33 4.6 938.56 3.3
 Total family income 7,417.52 28,724.34
Total number of years on welfare
Mean 4.0 3.7
More than 5 years (%) 25.6 20.2

TABLE 4--Determinants of the Log-Odds of an Exit from Welfare
Supplementation by Adopting Dependency, Self-Reliance, or
Autonomy (N = 2,939)

 (y = 1)

Variable b exp(b)

 Educational level -0.042 0.96
 Operators/farmers -0.888 ** 0.41 **
 Clerks/salespersons -2.021 ** 0.13 **
 Craftsmen -0.733 0.48
 Professionals/managers -2.027 * 0.13 *
Welfare policies
 Food stamps -5.38x[10.sup.-5] 1.00
 Public housing or rent subsidy -0.025 0.98
 Welfare reforms -0.387 0.68
 Child support -1.18x[10.sup.-4] 1.00
Personal experience
 Years of work experience -1.685 ** 0.19 **
 Number of times in
 supplementation 2.456 ** 11.65 **
 Total number of years in
 supplementation in the past -0.279 ** 0.76 **
Individual characteristics
 Single mother 0.358 * 1.43 *
 Need for child care 0.302 1.35
 Hispanic 0.028 1.03
 African American -0.235 0.79
Constant -1.779 **
G 2377.99 **

 (y = 2)

Variable b exp(b)

 Educational level -0.168 ** 0.85 **
 Operators/farmers 0.708 ** 2.03 **
 Clerks/salespersons 0.376 1.46
 Craftsmen 0.033 1.03
 Professionals/managers 0.357 1.43
Welfare policies
 Food stamps -9.91x[10.sup.-4] ** 1.0 **
 Public housing or rent subsidy -0.185 0.83
 Welfare reforms 0.110 1.12
 Child support 2.05x[10.sup.-4] ** 1.00 *
Personal experience
 Years of work experience 0.101 * 1.11 *
 Number of times in
 supplementation 2.583 ** 13.23 **
 Total number of years in
 supplementation in the past -0.206 ** 0.81 **
Individual characteristics
 Single mother -0.428 * 0.65 *
 Need for child care 0.006 0.99
 Hispanic 0.203 1.23
 African American 0.491 ** 1.63 **
Constant -1.459 *

 (y = 3)

Variable b exp(b)

 Educational level 0.056 1.06
 Operators/farmers 0.803 ** 2.23 **
 Clerks/salespersons 0.888 ** 2.43 **
 Craftsmen 0.714 * 2.04 *
 Professionals/managers 0.799 ** 2.22 **
Welfare policies
 Food stamps -0.001 ** 1.0 **
 Public housing or rent subsidy -0.505 * 0.60 *
 Welfare reforms 0.456 1.58
 Child support 3.94x[10.sup.-4] ** 1.00 **
Personal experience
 Years of work experience 0.241 ** 1.27 **
 Number of times in
 supplementation 2.929 ** 18.71 **
 Total number of years in
 supplementation in the past -0.305 ** 0.74 **
Individual characteristics
 Single mother -0.150 ** 0.22 **
 Need for child care -0.186 0.83
 Hispanic 0.030 1.03
 African American 0.160 1.17
Constant -3.564 **

* p < .05.

** p < .01.


Aldrich, J. H., & Nelson, F. D. (1984). Linear probability, logit, and probit models. Newbury Park, CA: Sage Publications.

Allison, P. D. (1984). Event history analysis: Regression for longitudinal event data. Newbury Park, CA: Sage Publications.

Bane, M. J., & Ellwood, D. T (1983). The dynamics of dependence: The routes to self-sufficiency. Cambridge, MA: Urban Systems Research and Engineering.

Bane, M. J., & Ellwood, D. T (1994). Welfare realities: From rhetoric to reform. Cambridge, MA: Harvard University Press.

Blank, R. M., & Ruggles, P. (1996). When do women use Aid to Families with Dependent Children and food stamps? The dynamics of eligibility versus participation. Journal of Human Resources, 31, 57-89.

Bora, C. E., Caudill, P. J., Spera, C., & Kunz, J. K. (1998). A look at life after welfare. Public Welfare, 56(1), 32-37.

Boskin, M. J., & Nold, F. C. (1975). A Markov model of turnover in Aid to Families with Dependent Children. Journal of Human Resources, 10, 467-481.

Carter, C., Dunbar, D., Micks, S., Brown, M., & Davis-Johnson, M. (1997). Virginia: Promoting independence and employment. Public Welfare, 55(3), 21-32.

Cheng, T. C. (1995). The chances of recipients leaving AFDC: A longitudinal study. Social Work Research, 19, 67-76.

Coe, R. D. (1981). A preliminary empirical examination of the dynamics of welfare use. In M. S. Hill & J. N. Morgan (Eds.), Five thousand American families: Patterns of economic progress (Vol. 1, pp. 121-168). Ann Arbor: University of Michigan.

Daponte, B. O., Sanders, S., & Taylor, L. (1999). Why do low-income households not use food stamps? Journal of Human Resources, 34, 613-628.

Dickinson, N. S. (1986). Which welfare work strategies work? Social Work, 31, 266-272.

Gault, B., Hartmann, H., & Yi, H. Y. (1998). Prospects for low-income mothers' economic survival under welfare reform. Publius: Journal of Federalism, 28, 175-193.

Goodwin, L. (1983). Causes and cures of welfare: New evidence on the social psychology of the poor. Lexington, MA: Lexington Books.

Hardina, D. (1999). Employment and the use of welfare among male and female heads of AFDC households. Affilia, 4, 217-234.

Harris, K. M. (1993). Work and welfare among single mothers in poverty. American Journal of Sociology, 99, 317-352.

Harris, K. M. (1996). Life after welfare: Women, work, and repeat dependency. American Sociological Review, 61, 407-426.

Hosmer, D. W., & Lemeshow, S. (1989). Applied logistic regression. New York: John Wiley & Sons.

Hutchens, R. M. (1981). Entry and exit transitions in a government transfer program: The case of Aid to Families with Dependent Children. Journal of Human Resources, 16, 217-237.

Kaushal, N., & Kaestner, R. (2001). From welfare to work: Has welfare reform worked? Journal of Policy Analysis and Management, 20, 699-719.

Kohlert, N. (1989). Welfare reform: A historic consensus. Social Work, 34, 303-306.

Kroch, E., & Sjoblom, K. (1994). Schooling as human capital or a signal: Some evidence. Journal of Human Resources, 29, 156-179.

McCrate, E., & Smith, J. (1998). When work doesn't work: The failure of current welfare reform. Gender and Society, 12, 61-80.

Menard, S. (1995). Applied logistic regression analysis. Thousand Oaks, CA: Sage Publications.

Meyer, D. R., & Cancian, M. (1998). Economic well-being following an exit from Aid to Families with Dependent Children. Journal of Marriage and the Family, 60, 479-492.

Meyers, M. K., & Heintze, T. (1999). The performance of the child-care subsidy system. Social Service Review, 73, 37-64.

Meyers, M. K., & van Leuwen, K. (1992). Child-care preferences and choices: Are AFDC recipients unique? Social Work Research, 28(1), 28-34.

Moffitt, R. A., Reville, R., & Winkler, A. E. (1998). Beyond single mothers: Cohabitation and marriage in AFDC program. Demography, 35, 257-278.

Moscovice, I., Craig, W., & Pitt, L. (1987). Meeting the basic needs of the working poor. Social Service Review, 61, 420-431.

Nichols, A. C. (1979). Why welfare mothers work: Implications for employment and training services. Social Services Review, 53, 378-391.

Osmond, M. W., & Grigg, C. M. (1978). Correlates of poverty: The interaction of individual and family characteristics. Social Forces, 56, 1099-1120.

Personal Responsibility and Work Opportunity Reconciliation Act of 1996, P.L. 104-193, 110 Stat. 2105.

Petersen, C. D. (1995). Female-headed families on AFDC: Who leaves welfare quickly and who doesn't. Journal of Economic Issues, 29, 619-628.

Piskulich, C. M. (1993). Toward a comprehensive model of welfare exits: The case of AFDC. American Journal of Political Science, 37, 165-185.

Plotnick, R. (1983). Turnover in the AFDC population: An event history analysis. Journal of Human Resources, 18, 63-81.

Rexroat, C., & Shehan, C. (1986). Differential effects of industrial and workers resources on women's wages. Social Science Research, 15, 1-27.

Riemer, F. J. (1997). Quick attachments to the workforce: An ethnographic analysis of a transition from welfare to low-wage jobs. Social Work Research, 21, 225-232.

Taylor, M. J. (1999). Race and regional unemployment as predictors of exit from AFDC. Journal of Social Services Research, 25, 1-18.

Thompson, F. J., & Gais, T. L. (2000). Federalism and the safety net: Delinkage and participation rates. Publius: Journal of Federalism, 30(1/2), 119-142.

U.S. Bureau of the Census. (1989). Characteristics of persons receiving benefits from major assistance programs (P-70, No. 14). Washington, DC: U.S. Government Printing Office.

U.S. Department of Labor. (1999). NLSY79 user's guide 1999: A guide to the 1979-1998 National Longitudinal Survey of Youth Data. Columbus: Ohio State University.

U.S. General Accounting Office. (1985). An evaluation of the 1981 AFDC changes: Final report (GAO/PEMD-85-4). Washington, DC: Author.

White, S. B. (1996). An employment approach to urban poverty alleviation: Employment patterns of AFDC recipients--the Milwaukee experience, 1989-93. Urban Studies, 33, 1923-1933.

Wodarski, J. S., Parham, T.M.J., Lindsey, E. W., & Blackburn, B. W. (1986). Reagan's AFDC policy changes: The Georgia experience. Social Work, 31, 273-279.

Wolk, J. L., & Schmahl, S. (1999). Child support enforcement: The ignored component of welfare reform. Families in Society, 80, 526-530.

Yamaguchi, K. (1991). Event history analysis. Newbury Park, CA: Sage Publications.

Yoshikawa, H. (1999). Welfare dynamics, support services, mothers' earnings, and child cognitive development: Implications for contemporary welfare reform. Child Development, 70, 779-801.

Original manuscript received September 29, 2000 Final revision received December 4, 2001 Accepted February 11, 2002

Tyrone Cheng, PhD, LISW, is assistant professor, Department of Government and Public Service, University of Alabama at Birmingham, U238 1530 Third Avenue S, Birmingham, AL 35294-3350; e-mail: tyronecheng@
COPYRIGHT 2002 National Association of Social Workers
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2002 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Cheng, Tyrone
Publication:Social Work Research
Geographic Code:1USA
Date:Sep 1, 2002
Previous Article:Subsidized guardianship: testing an idea whose time has finally come.
Next Article:Postsecondary education and the well-being of women in retirement.

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters