"Leavers" from TANF and AFDC: how do they fare economically?

Ozawa, Martha N. ; Yoon, Hong-Sik

Temporary Assistance for Needy Families (TANF) was enacted as a part of the Personal Responsibility and Work Opportunity Reconciliation Act of 1996 (PRWORA) (P.L. 104-193) and is a drastic departure from the Job Opportunity and Basic Skills Training Program (JOBS), enacted as part of the Family Support Act of 1988 (P.L. 100-485; for information on how TANF is implemented in the states, see Stoltzfus, Burke, & Falk, 2000). This historic departure is reflected in TANF's "work first" orientation, the two-year limit in any spell of receiving cash assistance without engaging in work or work-related activities, and the five-year lifetime limit on receiving cash assistance.

The JOBS policy with regard to work and work-related activities had an enabling rather than mandatory tone. It allowed, but did not require, states to offer job development, job placement and job search services, on-the-job training, work supplementation (wages subsidized by benefits), and community work experience. Moynihan (1990) and others (for example, Lurie & Sanger, 1991) argued that the JOBS program established a mutually understood social contract, under which recipients of Aid to Families with Dependent Children (AFDC) were to assume responsibility for becoming self-sufficient while the government provided necessary education, training, child care, and supportive services.

The JOBS program imposed no time limit on any spell of welfare dependence or a work trigger. If states wanted to experiment with a time limit, they were required to receive waivers from the U.S. Department of Health and Human Services. The JOBS programs prohibited states from terminating family benefits as a penalty for not working.

The TANF two-year limit began in July 1997, and the five-year limit was effective in July 2002. Some states, (for example, New York) implemented TANF in December 1996, and the five-year limit became effective in December 2001.

Across-the-board mandates with regard to work and time limits might have reflected a shift in the U.S. public's attitudes toward recipients of cash assistance. For example, in establishing TANF in 1996, some congressional leaders thought that providing permanent income support for needy children in single-parent families had encouraged family breakup, enabled out-of-wedlock births, and fostered long-term dependence (U.S. House of Representatives, 2000). What really pushed such mandates through Congress, however, was the Clinton administration's co-optation of a conservative agenda for political reasons (Goldberg & Collins, 2001).

Economic theory on rational decision making suggests that such an across-the-board treatment of low-income families headed by single mothers may not have desirable effects. Economists argue that individuals decide to change, or not to change, their life courses to maximize their economic well-being (see McConnell & Brue, 2002). Thus, it would be expected that welfare recipients who choose to exit from welfare do so because they believe they can improve their economic well-being through work. These recipients would take into account the level of wages they think they could receive based on their work experience (economists call such wages "reservation wages"); the level of market wages that may be offered to them; and the level of income they need to support their families, including the cost of child care. Conversely, recipients who decide to stay on welfare conclude that the market wages are not enough to support their families and take care of their children.

The time limits and work-related rules under TANF operate against such a theoretical assumption and prevent individuals from deciding their life courses according to their own sense of rationality. In particular, the two-year limit in any spell of welfare dependence and a five-year lifetime limit in receiving cash assistance do not take into account individual differences in employability and family needs. These rules force TANF recipients with limited marketable skills and many children to take life courses that are not conducive to maximizing their economic well-being. Moreover, these rules do not allow them enough time to find suitable jobs. Thus, they rightly may believe that it is in their best interests to stay on welfare and take care of their children.

The JOBS program, in contrast, offered flexibility. Recipients were encouraged to continue their education and to participate in work-related activities. Because of the absence of time limits and work-related requirements, a smaller proportion of AFDC than TANF recipients worked or participated in work-related activities: 12 percent among AFDC families in 1995 and 38 percent among TANF families in 1999 (Burke, 2000b).

Government data and the related literature indicate that although TANF has succeeded in moving a sizable portion of TANF families off the cash assistance rolls, it has not succeeded in helping them escape poverty. It is of interest, therefore, to investigate which group of women--TANF or AFDC recipients--fared better economically after they exited from TANF and AFDC. Such an investigation can provide insight into the success (or failure) of TANF in improving the economic conditions of TANF leavers. Thus, the major purpose of this study was to compare the degree of change in the income status of these two groups of women from three months before to three months after their exit from TANF and AFDC.

LITERATURE REVIEW

TANF Leavers

Employment Rate. Government data indicate that about 50 percent to 65 percent of recipients who leave TANF have jobs at the time or get them after they leave (U.S. House of Representatives, 2000). On the basis of studies conducted by seven states (Indiana, Maryland, Oklahoma, South Carolina, Tennessee, Washington, and Wisconsin) from 1995 through 1998, the U.S. General Accounting Office (2002) reported that 61 percent to 71 percent of former TANF recipients were employed at the time they were surveyed. Citing studies in 12 states, Anderson and Gryzlak (2002) concluded that during the first year after exit from TANF, between 55 percent and 65 percent of TANF leavers were employed.

Wages and Earnings. Burke (2000a) compiled the results of studies in several states on the hourly wages of TANF recipients. She reported that the quarterly median earnings of ex-TANF recipients in Maryland increased 21 percent, from $2,100 in the first quarter after they left TANF to $2,556 in the ninth quarter, and that the average hourly wage of former TANF recipients in South Carolina was $7 in the sixth quarter after they exited TANF. Burke (2000a) noted similar average hourly wages of single-parent families in other states in 1997 and 1998: $8.42 and $8.09 for ex-TANF recipients in two surveys in Washington; $7.41 for those in Wisconsin; and $6.00 to $6.99 for those in Florida. An Urban Institute study found that 41 percent of TANF leavers were still poor even after the Earned Income Tax Credit (EITC) and the cash value of food stamps, minus payroll taxes, were accounted for (Loprest, 2001).

In a study of the trend in income of female-headed families, Grogger (2001) reported that time limits had no significant effect on earnings or income: that is, they did not result in an increase in earnings or income for these families. Grogger argued that, under great pressure to leave welfare, TANF leavers could not find stable jobs with adequate wages.

On the basis of a comprehensive review of the literature, Hotz and colleagues (2002) concluded that TANF leavers with low skills could not expect to find jobs that provide progressively increasing wages. They questioned the effectiveness of TANF, which emphasizes the placement of recipients in the labor market at all costs, in helping TANF leavers escape poverty, let alone improving their economic conditions enough to join the middle class.

AFDC Leavers

There are few studies on the employment rate of AFDC leavers. Thus, we reviewed the literature on wages and earnings of AFDC leavers. Burtless (1999) projected the earnings potential of TANF recipients who had a two-year limit, on the basis of the work experiences of AFDC recipients from 1979 to 1994 and the general trend in wages received by workers with less than a high school education. He reported that most AFDC recipients worked in poorly paid occupations and typically received no more than $4 per hour (in 1993 dollars). He also pointed out that in 1993, the fulltime average wage paid to 25--to 34-year-old unmarried mothers who were high school dropouts was only $230 a week, or $11,960 per year (in 1993 dollars). With the help of the EITC and with child care expenses taken into account, the living standards of families of three headed by such women would be at the poverty line ($11,522 in 1993 dollars). Burtless also reported that over the past two decades the earnings of workers with inadequate education (especially high school dropouts) had declined rapidly, although the earnings of highly educated workers had increased considerably. On the basis of this trend in wages and the work experiences of AFDC recipients, Burtless warned that many TANF recipients would face economic hardships when they were forced off welfare.

Using 1979-1992 data, Cancian and Meyer (2000) found that the median hourly wages of women who left AFDC, measured in 1996 dollars, increased only 5.8 percent, from $6.36 in the first year after they left AFDC to $6.73 in the fifth year, and that the women's median annual earnings increased gradually from $6,059 to $9,947 (in 1996 dollars) in these five years. Despite this increase, the incomes of 27.2 percent of these families five years after they left AFDC were lower than their incomes at the time they exited AFDC. As a result, 55 percent of all the women who left AFDC were poor in the first year after they left AFDC, and 42 percent were still poor in the fifth year. Cancian and Meyer projected that even with the EITC, the poverty rate of these women would have declined to only 36 percent.

In a study in Wisconsin, Cancian and colleagues (1999) found that although the average earnings of recipients who left AFDC increased, the women's overall income dropped because welfare benefits fell more than their earnings increased.

Our review of the literature indicates that in either the AFDC or TANF policy environment, welfare recipients' lives after exiting from welfare are by no means easy. No empirical study has been conducted to compare how AFDC and TANF leavers fared economically after they exited from AFDC and TANF. This study attempted to fill that void by answering the following questions:

1. What are the characteristics of TANF leavers and AFDC leavers?

2. To what degree does the level of earnings of TANF and AFDC leavers increase from three months before to three months after exit from the cash assistance rolls? To what extent does their total family income increase from three months before to three months after the exit from the cash assistance rolls?

3. Measured by the income-to-needs ratio, to what extent does the economic well-being of TANF and AFDC leavers improve?

METHOD

Data Source

We used the panel data from waves 1 through 9 of the 1993 Survey of Income and Program Participation (SIPP) for AFDC leavers and waves 5 through 12 of the 1996 SIPP for TANF leavers (for details on the SIPP, see U.S. Census Bureau, 2001). The 1993 SIPP interviewed 56,800 individuals in 21,823 households each quarter starting in January 1993 and ending in December 1995; the 1996 SIPP interviewed 95,402 individuals in 40,188 households each quarter starting in November 1995 and ending in February 2000. All respondents were age 15 or older. The SIPP covers the noninstitutionalized population of U.S. residents and provides data on the source and amount of income, labor force information, program participation and eligibility data, and general demographic characteristics and allows researchers to analyze data in terms of individuals, families, and households.

From the 1993 SIPP we identified 156 female heads of families with children who left AFDC from June 1993 to June 1995 and stayed unmarried three months after the exit from AFDC. From the 1996 SIPP, we identified 300 female heads of families with children who left TANF from July 1997 (the first month of the TANF nationwide implementation) to August 1999 and stayed unmarried three months after the exit from TANF.

Data Analysis

The unit of analysis in our study was individuals. In generating descriptive statistics, we used the weight variable developed by the U.S. Census Bureau to adjust for the sampling, poststratification, and nonresponse biases in the SIPP data set.

The women's economic conditions were measured at three months before and three months after exit from welfare. The month when AFDC or TANF payments stopped was considered the transition month and was not counted. Therefore, each observation involved a time frame of seven months (three months before exit + one month [point of exit] + three months after exit). This seven-month period could occur at different times for different AFDC and TANF leavers, within the time frame of the 1993 and 1996 SIPPs. This time frame was chosen to protect income status from contamination by transitional factors. That is, by using this time frame, incomes from particular sources were allowed to "settle down" after the women left welfare. Another reason for choosing this time frame was to identify as many AFCD and TANF leavers as possible. If we had chosen a longer time frame, we would have been unable to mount an adequate number of observations, given that the SIPP provides a relatively short-range panel data--approximately four years.

Variables

Most of the variables to describe the characteristics of the respondents are self-explanatory. However, the variable "welfare dependence" needs explanation. Welfare dependence was measured by dividing monthly AFDC or TANF payments by monthly total family income. We divided welfare dependence into two categories: (1) high welfare dependence, meaning dependence ratios that were .40 or higher, and (2) low welfare dependence, meaning dependence ratios that were lower than .40.

The income-to-needs ratio, which was used as an indicator of income status, was obtained by dividing the family income by the poverty line. The income-to-needs ratio was an appropriate indicator of the income status of AFDC and TANF leavers and their children because it was adjusted for family size and economy of scale, as well as for changes in the cost of living over time. Thus, the income-to-needs ratio measured the income status--thus, the economic well-being--of individuals in the family or of the family as whole at one point in time or over time.

FINDINGS

Characteristics of AFDC and TANF Leavers

TANF leavers differed from AFDC leavers in six respects (Table 1). First, welfare dependence, measured three months before leaving welfare, was considerably lower among TANF leavers than among AFDC leavers. Among TANF leavers, only 40.3 percent had a welfare dependence ratio of .40 or higher, compared with 50.1 percent among AFDC leavers. Second, the mean number of AFDC leavers' children declined considerably (2.2 children three months before to 1.9 children three months after leaving AFDC), but the number of TANF leavers' children was almost identical at these two points in time (2.2 and 2.1 children, respectively). This finding implies that the incidence of children becoming ineligible on account of the age rule, death, or departure from the households was higher among AFDC leavers than among TANF leavers. Third, the proportion of leavers with work experience (full-time or part-time work) three months before exit was considerably larger among TANF leavers than among AFDC leavers: 57.6 percent compared with 49.1 percent. Note, however, that the employment rate (the percentage of those who worked part-time or full-time) of AFDC leavers increased 40.5 percent (from 49.1 percent to 69.0 percent) three months after exit from welfare compared with 18.8 percent (from 57.6 percent to 68.4 percent) among TANF leavers. Fourth, the proportion of recipients who were not white was slightly smaller among TANF leavers than among AFDC leavers: 43.1 percent compared with 45.8 percent. Fifth, the proportion of never-married mothers was slightly larger among TANF leavers than among AFDC leavers: 52.2 percent compared with 48.6 percent, measured three months before they left welfare. Sixth, TANF leavers were about two years older than AFDC leavers.

The backgrounds of AFDC leavers and TANF leavers were similar in one critical respect: The level of education, measured three months before exit from welfare, was almost identical between these two groups. The proportion of TANF leavers with a high school education was 38.0 percent compared with 37.0 percent of AFDC leavers; the proportion of leavers with at least some college was 28.9 percent among TANF and 28.1 percent among AFDC leavers.

Change in Income Status

Income Level, by Source. The rate of increase in the earnings of AFDC recipients after they left AFDC was considerably higher than that of TANF recipients (Table 2). Three months before their exit, AFDC leavers had monthly earnings of $499, which increased 109.8 percent to $1,047 three months after they left. In comparison, three months before leaving TANF, TANF leavers had monthly earnings of $690, which increased 37.7 percent to $950 three months after they left. The higher mean earnings among TANF leavers three months before their exit occurred largely because a larger proportion of TANF leavers than AFDC leavers worked part-time before they left welfare (see Table 1).

For AFDC leavers, child support payments increased 102.0 percent, from $51 to $103, but for TANF leavers, such payments increased only 31.3 percent, from $32 to $42. For AFDC leavers, other welfare payments declined 33.3 percent between these two points in time, but for TANF leavers, the decline was only 17.8 percent. On the other hand, the amount of other income increased 24.2 percent among AFDC recipients, but declined 16.7 percent among TANF recipients. All told, the monthly total family income of AFDC leavers increased 22.0 percent, from $1,108 to $1,352, but that of TANF leavers declined 3.1 percent, from $1,233 to $1,195.

The degree of increase in earnings and the degree of decline in other welfare payments are clear indications that AFDC leavers achieved greater self-sufficiency than TANF leavers, further compounded by a greater increase in child support payments among AFDC leavers.

Change in Income-to-Needs Ratio. The mean income-to-needs ratio of AFDC leavers (.82) was slightly lower than that of TANF leavers (0.90) three months before they left AFDC (NS, t = -.64), but it was significantly higher (1.11) than that of TANF leavers (0.87) (p < .05, t = 2.92) three months after the exit (Table 3). Put another way, AFDC leavers' mean income-to-needs ratio increased 35.4 percent between these two points in time (p < .001, t = 3.43), but TANF leavers' mean income-to-needs ratio declined 3.3 percent (NS, t = -.13).

The greater increase in the income-to-needs ratio among AFDC leavers was due in part to the decline in the number of children on AFDC from three months before to three months after the exit from AFDC (see Table 1): It is generally easy to have a higher income-to-needs ratio when the number of dependents declines. However, the great improvement in AFDC leavers' income-to-needs ratios had a great deal to do with the greater rate of increase in employment and earnings, which are indicators of the economic performance of individual AFDC leavers: From before to after exit from AFDC, the employment rate among AFDC leavers increased 40.5 percent (from 49.1 percent to 69.0 percent) compared with a 18.8 percent increase (from 57.6 percent to 68.4 percent) among TANF leavers (see Table 1). During the same period, AFDC leavers' earnings increased 109.8 percent, but the earnings of TANF leavers increased only 37.7 percent (see Table 2).

To elaborate on the findings shown in Table 3, we created Tables 4 and 5. Table 4 illustrates how the distribution of AFDC and TANF leavers, categorized by their income-to-needs ratios, changed from three months before to three months after exit from the cash assistance rolls. For example, a decline in the proportion of AFDC leavers with income at or below 50 percent of the poverty line from three months before to three months after exit would indicate an improvement in income status of AFDC leavers. The same can be said if the proportion of AFDC leavers with high income-to-needs ratios increased. That is, an increase in the proportion of those with income more than 200 percent of the poverty line between these two points in time would indicate an improvement in income status of AFDC leavers.

Table 5 was created to determine whether the degree of inequality in income (measured by the coefficient of variation) was greater among TANF leavers than among AFDC leavers at two points in time and to show whether the degree of increase in income inequality from three months before to three months after exit was greater among TANF leavers than among AFDC leavers.

Three months before exit, 39.8 percent of AFDC leavers had income at or below 50 percent of the poverty line (that is, an income-to-needs ratio of .50 or less) compared with 37.6 percent of TANF leavers (Table 4). But three months after exit, only 25.2 percent of AFDC leavers had such a low income-to-needs ratio (that is, a 36.7 percent decline from three months before exit) whereas 34.7 percent of TANF leavers still had this low income-to-needs ratio (that is, only a 7.7 percent decline).The distribution of AFDC leavers versus TANF leavers, by the level of income-to-needs ratios, three months before exit was not significant [[chi square] (4, N = 456) = 1.5729]; the distribution became significant at p < .02 three months after exit [[chi square] (4, N = 456) = 12.4106].

In addition, three months before the exit, 73.9 percent of AFDC leavers compared with 67.6 percent of TANF leavers had an income-to-needs ratio of 1.00 or less (that is, at or below the poverty line) (Table 4). Three months after exit, however, 54.9 percent of AFDC leavers had an income-to-needs ratio of 1.00 or less, compared with 66.3 percent of TANF leavers.

Taking the extreme cases of leavers with an income-to-needs ratio of 2.01 or more, the proportion of AFDC leavers with such a high income-to-needs ratio increased 150 percent--from 5.6 percent three months before exit to 14.0 percent three months after exit (Table 4). In contrast, 9.2 percent of TANF leavers had such a high income-to-needs ratio three months before exit and only 8.5 percent of TANF leavers had it three months after exit--or a 7.6 percent decline. The distribution of AFDC leavers three months before compared with three months after the exit was significant at p < .001 [[chi square] (4, N = 456) = 18.9331], but that of TANF leavers at these two points in time was not significant [[chi square] (4, N = 456) = 2.5168].

These findings indicate that AFDC leavers fared better economically after they left AFDC. The proportion of AFDC leavers who stayed poor three months after exit declined more, and the proportion of AFDC leavers who attained an income status of more than twice the poverty line increased more than did the proportion of TANF leavers.

Table 5 shows the coefficients of variation of the income-to-needs ratio for AFDC and TANF leavers three months before and three months after exit from the respective programs. The coefficient of variation is a measure of inequality obtained by dividing the mean income-to-needs ratio among the top quintile of AFDC or TANF leavers by the mean income-to-needs ratio among the bottom quintile of the respective group. The larger the coefficient, the greater the inequality.

Three months before exit from welfare, the coefficient of variation was 8.792 among AFDC leavers, compared with 11.353 among TANF leavers (Table 5).This means that three months before exit, the distribution of income-to-needs ratios was more unequal among TANF leavers than among AFDC leavers. Three months after exit, the coefficient increased to 18.318 among AFDC leavers and to 31.203 among TANF leavers. Again, this finding indicates that three months after exit, the distribution of income-to-needs ratios was more unequal among TANF leavers than among AFDC leavers. Moreover, the rate of increase in inequality in income-to-needs ratios from three months before to three months after exit was greater among TANF leavers than among AFDC leavers.

These findings indicate that the distribution of AFDC leavers' income-to-needs ratios was not only less unequal both three months before and three months after exit than was TANF leavers', but that the rate of increase in the inequality in AFDC leavers' income-to-needs ratios was smaller (a 108 percent increase) than that of TANF leavers (a 175 percent increase). Put differently, the distribution of income-to-needs ratios among AFDC leavers was relatively more homogenous than among TANF leavers at the both points in time, and the income status of AFDC leavers improved more in a lock-step fashion than that of TANF leavers, although the degree of income inequality increased between the two points in time among both groups of leavers.

DISCUSSION AND CONCLUSION

The major finding of this study is that AFDC leavers fared considerably better economically after they left AFDC than did TANF leavers. As a result, three months after they left AFDC, their levels of earnings and total family income, as well as the income-to-needs ratio, were higher than those of TANF leavers. Put differently, starting from only 82 percent of the poverty line three months before exit from AFDC, the mean income status of AFDC leavers increased to 111 percent of the poverty line three months after exit. In comparison, the TANF leavers' mean income status declined 3.3 percent to 87 percent of the poverty line, so that it was below the poverty line three months after exit from TANF. Furthermore, in terms of the percent distribution of leavers by their levels of income status and in terms of income inequality, AFDC leavers fared better economically than did TANF leavers.

We argue that AFDC leavers' considerable economic improvement after leaving AFDC was related to the flexibility allowed under the JOBS regulations, under which the AFDC program operated, for AFDC recipients to decide whether and when to leave AFDC. Given such flexibility, AFDC recipients tended to choose to leave AFDC or stay on AFDC to suit their best interests, based on the wages they could command in the labor market and their family needs.

In contrast, most TANF leavers were compelled to leave TANF, which had the two-year limit for any spell of receiving cash assistance without engaging in work or work-related activities and the five-year lifetime limit in receiving cash assistance.

The findings from this study are even more remarkable, because the five-year lifetime limit was not fully in place at the time of this study. (It became effective in July 2002). Obviously, TANF recipients were informed by the state agency running the program about the two-year and five-year time limits. As these recipients went through a series of meetings with agency officials to discuss their obligations to work or to engage in work-related activities and the five-year lifetime limit in receiving cash payments, they must have heard enough to believe that they were living under different rules from those in the AFDC program. Thus, it is reasonable to assume that TANF recipients were under pressure to leave TANF as soon as possible.

The findings from this and other studies (Grogger, 2001; O'Neill & Hill, 2001) indicate that TANF is succeeding in getting TANF recipients off welfare, but it is not helping TANF leavers improve their economic conditions. TANF's ineffectiveness in improving TANF leavers' economic conditions certainly will be a focal point of the debate on future welfare reform.

Two issues need to be addressed. First, some may argue that TANF leavers could not improve their income status after their exit from TANF as well as AFDC leavers could because they were more disadvantaged to start with. But such a concern is unwarranted. As indicated earlier, TANF leavers were in many ways more advantaged. Three months before exit from TANF, their welfare dependence ratio was lower, the proportion with prior work experience was higher, the proportion of nonwhites was smaller, the proportion of never-married mothers was only slightly larger, and the number of children was identical to that of AFDC leavers. Furthermore, TANF and AFDC leavers had almost identical levels of education, and TANF leavers were about two years older.

Second, the findings may have been affected by different macroeconomic conditions in the two periods studied. Was the unemployment rate higher during the time TANF leavers left the assistance rolls? Government data indicate that the average unemployment rate from 1997 to 1999--when TANF leavers in our study exited--was 4.6 percent, which was considerably lower than the average unemployment rate of 6.2 percent from 1993 to 1995, when AFDC leavers exited (U.S. Census Bureau, 1997; U.S. Census Bureau, 2000). Thus, with the unemployment rate taken into account, the findings of this study would be strengthened, not weakened.

We acknowledge limitations in our study. First, the findings do not explain how education affected the level of wages three months after the exit from AFDC and TANF. However, a finding from another study is instructive. Ozawa and Lee (n.d.) found that the hourly wages of those with a college education was no different from the hourly wages of those with less than a high school education five months after exit from TANF. This finding implies that TANF leavers had to take any job available, regardless of their education level.

Second, the findings could have been enhanced if we could have shown reasons, such as noncompliance with rules, for leaving AFDC or TANF. Because of the limitations in the SIPP data, we could not explore this issue. However, a separate, logistic regression analysis of the data (the same one used in this study) indicated that exit from AFDC was associated, in part, with the youngest child reaching age 18 some time between three months before and three months after exit, but such a phenomenon was nonexistent among TANF recipients. This finding suggests that the exit from TANF was hurried; that is, TANF leavers left before their youngest child reached age 18, whereas some AFDC leavers left the program because of the age-related rules.

Third, because of the limitations in the data, we could not account for the impact of child care expenses. If, for example, AFDC leavers were more likely to pay for child care out-of-pocket than were TANF leavers, then AFDC leavers' economic conditions three months after exit would not have been as favorable as projected in this study.

Fourth, there might have been some overlap between AFDC leavers and TANF leavers under investigation. Nevertheless, all AFDC leavers left the cash assistance rolls under the previous rules and regulations, and TANF leavers left the cash assistance rolls under the new rules and regulations provided by the PRWORA of 1996. As such, the two groups of welfare leavers acted under different policy environments.

Fifth, the reader is cautioned about the findings with regard to work experiences and earnings before leaving AFDC or TANF. As Edin and Lein (1997) suggested, welfare recipients often engaged in work that was not reported to the welfare department. Because we used survey data from the SIPP, there is no way to estimate the extent of unreported work or the amount of earnings from such work.

Finally, this study dealt with the economic conditions of AFDC leavers and TANF leavers during a relatively short period of time--three months before to three months after leaving the cash assistance rolls. If we had data from a longer time frame, the findings might have been different.

With these caveats, the findings of this study call for changes in the welfare reform enacted under the PRWORA of 1996. The uniform requirements and sanctions with regard to time limits, work, and work-related rules are not conducive to deploying TANF recipients effectively. Instead, consideration should be given to TANF recipients' capabilities to earn money through work; their family needs; and other living conditions, such as their and their children's health conditions and the unemployment rates in their areas of residence. For recipients with a larger number of children and who can earn only low wages, it makes sense to help them concentrate on child rearing and becoming good parents, instead of forcing them to join the labor force. As their children grow, these recipients can be helped to acquire more job-related skills, so they will eventually become economically self-sufficient. Forcing them to join the labor force prematurely is not conducive to maximizing the economic well-being of individual families or of the nation as a whole. In short, the future reform of welfare needs to take account of the experiences under both the JOBS and TANF programs and to reintroduce some degree of flexibility in helping the nation's poor families with children so they can live their lives more constructively, which ultimately will lead to greater life opportunities than the TANF program provides.

Because the economy has been sluggish for the past few years, the projected work opportunities for those who leave TANF are not encouraging. Close monitoring of their economic status is needed, and policymakers should act to ensure some economic stability and sufficiency in their lives.

Manuscript received October 14, 2002

Final revision received January 2, 2004

Accepted April 27, 2004

REFERENCES

Anderson, S. G., & Gryzlak, B. M. (2002). Social work advocacy in the post-TANF environment: Lessons from early TANF research studies. Social Work, 47, 301-314.

Burke, V. (2000a). Welfare reform: TANF trends and data (CRS Report for Congress No. 98-369 EPW). Washington, DC: Congressional Research Service.

Burke, V. (2000b). Welfare reform: Work activities and sanctions in state TANF programs (CRS Report for Congress No. RL30767). Washington, DC: Congressional Research Service.

Burtless, G. (1999). The employment experiences and potential earnings of welfare recipients. In J. E. Hansan & R. Morris (Eds.), Welfare reform, 1996-2000: Is there a safety net? (pp. 51-73).Westport, CT: Auburn House.

Cancian, M., Haveman, R., Kaplan, T., & Wolfe, B. (1999). Post-exit earnings and benefit receipt among those who left AFDC in Wisconsin (Special Report No. 75). Madison, WI: Institute for Research on Poverty.

Cancian, M., & Meyer, D. R. (2000).Work after welfare: Women's work effort, occupation, and economic well-being. Social Work Research, 24, 69-96.

Edin, K., & Lein, L. (1997). Making ends meet: How single mothers survive welfare and low-wage work. New York: Russell Sage Foundation.

Family Support Act of 1988, P.L. 100-485, 102 Stat. 2343.

Goldberg, G. S., & Collins, S. D. (2001). Washington's new poor law: Welfare "reform" and the roads not taken, 1935 to the present. New York: Apex Press.

Grogger, J. (2001). The effects of time limits and other policy changes on welfare use, work, and income among female-headed families (Working Paper 815). New York: National Bureau of Economic Research.

Hotz, V.J., Mullin, C. H., & Scholz, J. K. (2002).Welfare reform, employment, and advancement. Focus, 22(1), 51-55.

Loprest, P. (2001). How are families that left welfare doing? A comparison of early and recent welfare leavers (Series B, No. B-36). Washington, DC: Urban Institute.

Lurie, I., & Sanger, B. (1991). The Family Support Act: Defining the social contract in New York. Social Service Review, 65, 43-67.

McConnell, C. R., & Brue, S. L. (2002). Economics: Principles, problems, and policies. New York: McGraw-Hill.

Moynihan, D. P (1990, November 25). The children of the state. Washington Post, p. C1.

O'Neill, J. E., & Hill, A. M. (2001). Gaining ground? Measuring the impact of welfare reform on welfare and work. New York: Manhattan Institute, Center for Civic Innovation.

Ozawa, M. N., & Lee, Y. W. (n.d.). Employment and hourly wages of TANF leavers. Unpublished manuscript, George Warren Brown School of Social Work, Washington University, St. Louis.

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

Stoltzfus, E., Burke, V, & Falk, G. (2000). Welfare reform: State programs of Temporary Assistance for Needy Families (TANF) (CRS Report for Congress No. RL30695).Washington, D C: Congressional Research Service.

U.S. Census Bureau. (1997). Statistical abstract of the United States, 1997. Washington, DC: U.S. Government Printing Office.

U.S. Census Bureau. (2000). Statistical abstract of the United States, 2000. Washington, DC: U.S. Government Printing Office.

U.S. Census Bureau. (2001). Survey of income and program participation: Users' guide (3rd ed.). Washington, DC: U.S. Department of Commerce, Economics and Statistics Administration.

U.S. General Accounting Office. (2002). Welfare reform: States provide TANF-funded work support services to many low-income families who do not receive cash assistance (GAO-02-615T).Washington, DC: Author.

U.S. House of Representatives, Committee on Ways and Means. (2000). Overview of entitlement programs: 2000 green book. Washington, DC: U.S. Government Printing Office.

Martha N. Ozawa, PhD, is Bettie Bofinger Brown Distinguished Professor of Social Policy, George Warren Brown School of Social Work, Washington University, One Brookings Drive, St. Louis, MO 63130-4899; e-mail: ozawa@wustl.edu. Hong-Sik Yoon, PhD, is assistant professor, Department of Social Welfare, College of Social Science, Chonbuk National University, Chonbuk, Korea. An earlier version of this article was presented at the 7th annual conference of the Society for Social Work and Research, January 17, 2003, Washington, DC.

Table 1: Characteristics of the AFDC and TANF Leavers

 AFDC Leavers
 (N = 156)

 3 Months 3 Months
Characteristic before Exit after Exit

Age (Years) 32.0 32.5
Race (%)
 White 54.2 --
 Not white 45.8 --
Marital status (%)
 Divorced 32.3 --
 Separated 15.8 --
 Widowed 3.4 --
 Never married 48.6 --
Education (%)
 Less than high school 34.9 33.9
 High school graduate 37.0 37.2
 At least some college 28.5 29.0
Number of children (%) 2.2 1.9
Work status (%)
 Full-time 33.6 47.9
 Part-time 15.5 21.1
 Nonworking 50.9 31.0
Welfare dependence (%)
 High dependence 50.1 --
 Low dependence 49.9 --

 TANF Leavers
 (N = 300)

 3 Months 3 Months
Characteristic before Exit after Exit

Age (Years) 33.8 34.3
Race (%)
 White 56.9 --
 Not white 43.1 --
Marital status (%)
 Divorced 27.2 --
 Separated 18.4 --
 Widowed 2.3 --
 Never married 52.2 --
Education (%)
 Less than high school 33.1 32.1
 High school graduate 38.0 37.5
 At least some college 28.9 30.5
Number of children (%) 2.2 2.1
Work status (%)
 Full-time 34.6 49.4
 Part-time 23.0 19.0
 Nonworking 42.4 31.6
Welfare dependence (%)
 High dependence 40.3 --
 Low dependence 59.7 --

Notes: -- = not measured at the time indicated. AFDC = Aid to Families
with Dependent Children. TANF = Temporary Assistance for Needy
Families.

Table 2: Mean Monthly Income and Percent Distribution
of Monthly Income, by Source (in 1999 dollars)

 3 Months 3 Months
 before Exit after Exit
 Change
Income Source $ % $ % %

AFDC leavers (n = 156)
 Earnings 499 45.0 1,047 77.4 109.8
 Social security 41 3.7 37 2.7 -9.8
 Child support 51 4.6 103 7.6 102.0
 AFDC 323 29.2 0 0.0 -100.0
 Other welfare payments 132 11.9 88 6.5 -33.3
 Other income 62 5.6 77 5.6 24.2
 Total family income 1,108 100.0 1,352 100.0 22.0
TANF leavers (n = 300)
 Earnings 690 56.0 950 79.5 37.7
 Social security 46 3.7 57 4.8 23.9
 Child support 32 2.6 42 3.5 31.3
 TANF 288 23.4 0 0.0 -100.0
 Other welfare payments 129 10.5 106 8.9 -17.8
 Other income 48 3.9 40 3.3 -16.7
 Total family income 1,233 100.0 1,195 100.0 -3.1

Notes. AFDC = Aid to Families with Dependent Children. TANF = Temporary
Assistance for Needy Families.

Table 3: Mean Income-to-Needs
Ratio before and after Exit
from AFDC and TANF

 3 Months 3 Months %
Group before Exit after Exit Change

AFDC leavers (n = 156) 0.82 1.11 35.4
TANF leavers (n = 300) 0.90 0.83 -3.3

Notes: AFDC = Aid to Families with Dependent Children. TANF = Temporary
Assistance for Needy Families.

Table 4: Percent Distribution of AFDC
and TANF Leavers, by Income-to-Needs
Ratio before and after Exit from
AFDC and TANF

Income-to- 3 Months 3 Months %
Needs Ratio before Exit after Exit Change

AFDC leavers (n = 156)
 0.00 to 0.50 39.8 25.2 -36.7
 0.51 to 1.00 34.1 29.7 -12.9
 1.01 to 1.50 13.0 22.8 75.4
 1.51 to 2.00 7.6 8.2 7.9
 2.01 or more 5.6 14.0 150.0
 All 100.0 100.0
TANF leavers (n = 300)
 0.00 to 0.50 37.6 34.7 -7.7
 0.51 to 1.00 30.0 31.6 5.3
 1.01 to 1.50 16.1 16.2 0.6
 1.51 to 2.00 6.9 9.1 31.9
 2.01 or more 9.2 8.5 -7.6
 All 100.0 100.0

Notes AFDC = Aid to Families with Dependent Children. TANF = Temporary
Assistance for Needy Families.

Table 5: Quintile Distribution and Coefficient of Variation of
Income-to-Needs Ratios among AFDC and TANF leavers before and
after Exit from AFDC and TANF

 AFDC Leavers (n = 156)

 3 Months 3 Months
Quintile before Exit after Exit

1 (lowest) 0.216 0.148
2 0.374 0.573
3 0.676 0.901
4 0.987 1.272
5 (highest) 1.899 2.711

Coefficient of variation 8.792 18.318

 TANF Leavers (n = 300)

 3 Months 3 Months
Quintile before Exit after Exit

1 (lowest) 0.187 0.187
2 0.404 0.404
3 0.639 6.639
4 1.040 1.040
5 (highest) 2.123 2.123

Coefficient of variation 11.353 31.203

Notes: AFDC = Aid to Families with Dependent Children.
TANF = Temporary Assistance for Needy Families.