January 30, 2004

Implications for TANF Reauthorization and State TANF Policies
By Shawn Fremstad

PDF of analysis

View Related Reports

If you cannot access the files through the links, right-click on the underlined text, click "Save Link As," download to your directory, and open the document in Adobe Acrobat Reader.

This paper summarizes some of the key research findings from recent welfare reform research — primarily research released in the last year — and discusses the policy implications of these findings. [1]  Key findings from this new set of research include:

Current TANF proposals that are pending in Congress would do little to address most of these challenges.  When the Senate takes up TANF reauthorization, it should take these recent research findings into account and ensure that states have the resources and flexibility to address these issues.

U.S. Poverty Rate Remains High Relative to Other Industrialized
Nations and Has Increased Since 2000

As the economy expanded and unemployment declined in the 1990s, poverty also declined significantly.  Despite these declines, the child poverty rate in the United States remained higher than the rate in Canada, most western European countries, and many other developed counties (Jesuit and Smeeding, 2002).  Poverty remains disproportionately high in the United States even through it has the highest average income in the industrialized world (except for Luxembourg) (Smeeding, Rainwater, and Burtless, 2001).

In the last few years, poverty in the United States has increased.  Trends in child poverty and severe poverty — income below 50 percent of the poverty line — are particularly relevant to welfare reform policy.  Between 2000 and 2002, the number of children in poverty increased by nearly 600,000.  More than 12 million children were poor in 2002 (Center on Budget and Policy Priorities, 2003).  Between 2000 and 2002, the number of individuals in severe poverty increased by nearly 1.5 million.  Severe poverty is now back to the level it was at in 1996-97.

At the same time, the number of people who experience food insecurity during the year has increased (U.S. Department of Agriculture, 2003).  Between 1999 and 2002, the number of households that experienced food insecurity increased from 10.5 million to 12.1 million (Center on Poverty and Hunger, 2003).  (USDA defines “food insecurity” as meaning that “a household had limited or uncertain availability of food, or limited or uncertain ability to acquire acceptable foods in socially acceptable ways . . . .”).


Families that Left Welfare Recently are Less Likely To Have Jobs Than Prior TANF “Leavers” 

Most studies of families who left welfare in the 1990s found that a majority of welfare leavers found jobs.  More recent research shows a decline in the share of welfare leavers with jobs and an increase in “disconnected” leavers who do not have income from work or welfare.

  • Recent studies show that families that left welfare recently (2000 or later) are less likely to be working than families that left welfare in the 1990s.  A recent Urban Institute report shows that the proportion of families that leave welfare and are not employed rose from 50 percent in 1999 to 58 percent in 2002 (Loprest, 2003).  Similarly, a recent study of welfare leavers in Cuyahoga County, Ohio (Cleveland) found that employment rates for families leaving welfare were relatively constant between 1998 and 2000 but have fallen steadily since 2001 (Colton, Bania, Martin, and Lalich, 2003).  In New York City, the job placement rate of public assistance recipients has declined markedly between 2000 and 2003, even though the city continues to engage a very high percentage of welfare recipients in welfare-to-work activities (New York City Human Resources Administration, 2004).
  • The share of families that leave welfare and are not working and do not have another stable source of support has increased.  A recent study by the Urban Institute shows that the share of leavers who were “disconnected” — that is, leavers who were not working, did not have a working spouse, and were not receiving TANF or SSI — rose from 9.8 percent in 1999 to 13.8 percent in 2002 (Loprest, 2003).
  • “Disconnected” welfare leavers faced significant barriers to employment and difficulty making ends meet.  Disconnected leavers are significantly more likely to have health problems, less likely to have completed high school, and more likely to have been jobless in the past three years (Loprest, 2003).  Nearly two-thirds of disconnected leavers reported running out of money to buy food.
  • A New Jersey study of TANF leavers who do not have jobs found that about 40 percent of these jobless leavers have not worked recently, do not receive SSI or unemployment benefits, and do not live with an employed spouse or parent.   Almost one in five of these “least stable” leavers experienced hunger in the past year, and two-thirds say that they are “barely making it from day to day.” (Wood and Rangarajan, 2003).
  • The overall unemployment rate of low-income single mothers increased from 9.8 percent in 2000 to 12.3 percent in 2002.  After decreasing at a faster rate than the unemployment rate for the overall population in the last half of the 1990s, low-income single mothers’ unemployment rate increased at a faster rate between 2000 and 2002 than the national unemployment rate (Chapman and Bernstein, 2003).


Poverty Rates Among Families that Leave TANF are Very High and Remain High Over Time

Studies conducted in the 1990s found high poverty rates among welfare leavers — most studies have found that between 50 to 75 percent of welfare leavers remain poor two to three years after leaving welfare (Blank, 2002).[2]  Leavers’ poverty rates appear to decline over time, but at a very modest pace.  For example, one of the few studies using national data to track leavers for more than two to three years found that 42 percent of welfare leavers remain poor five years after leaving welfare compared to a 55 percent poverty rate in the first year after leaving welfare (Cancian, Haveman, Kaplan, Meyer, and Wolfe, 1999).  Most welfare leavers with incomes above the poverty line still have very low incomes — state-level leaver studies have found that about 90 percent of leavers have income below 185 percent of the poverty level (Acs and Loprest, 1999).

Recent research continues to show high poverty rates and very modest growth in income for families that leave welfare.

  • A recent study of Michigan women who received TANF in 1997 found that by the fall of 2001, only one-quarter were working in “good jobs.”  The researchers defined “good jobs” as full-time jobs that pay at least $7 per hour and offer health insurance, or full-time jobs that do not offer health insurance but pay at least $8.50 per hour.  The researchers also found that the recent economic downturn substantially reduced mothers' probability of getting a "good job" (Johnson and Cochrane, 2003).  These findings are similar to estimates in earlier research that only about one-quarter of young women who ever use welfare would work primarily in a “good job” by their late 20s (Pavetti and Acs, 2001).
  • A longitudinal study of welfare recipients in Philadelphia found that the share of women who worked full time in jobs that paid at least $7.50 per hour and offered health insurance increased from 1998 to 2001, but that only about two out of five working women held such a job in 2001 (Michalopoulos, 2003).
  • A California study found that half of welfare leavers saw no growth in income in the year after they left welfare.   The HHS-supported study surveyed welfare leavers in the late 1990s at two points in time:  five to 10 months after leaving welfare and 11 to 16 months after they left welfare.  Between these two periods, income gains for all leavers averaged only $60 to $70 per month (MaCurdy, Marrufo, and O’Brien-Strain, 2003).  Although their income growth is modest, it should be noted that poverty rates for welfare leavers in California are significantly lower on average than the poverty rates found in other national and state studies.  This lower poverty rate is likely due to a deliberate policy choice California made in its TANF program.  California is one of a handful of states that allow TANF recipients who find jobs to continue receiving TANF assistance as an income supplement until they reach the poverty level.  Rigorous evaluations have shown that such a policy results in higher income and lower poverty rates than programs that simply mandate employment (Berlin, 2002).
  • A HHS-funded study of welfare reform in Wisconsin — a state often cited as having a particularly innovative welfare reform program — found that the net income of welfare leavers in the year after they exited welfare is lower than their income prior to leaving (Cancian, et al., 2003).  The researchers found that recipients who left Wisconsin’s TANF program (W-2) experienced significant increases in earned income, but on average their benefits declined by more than their earned income increased.


Families That Lose Assistance Because of Sanctions or Time Limits are More Likely to Experience Hardship than Other Welfare Leavers

Recent state-level studies provide new evidence on the extent to which families that lose TANF cash assistance due to sanctions or time limits face serious material hardships, including problems securing housing and affording food.  Early research conducted in a few states, tended to show that time-limit leavers struggled financially but that they didn’t necessarily experience more material hardship than other TANF leavers (Bloom, Farrell, and Fink, 2002).  More recent time limit research, however, finds that time limit-leavers have lower employment rates, higher poverty rates, and higher levels of material hardship than other TANF leavers.

  • A study conducted by Minnesota’s Department of Human Services found that families that lost welfare due to the 60-month time limit were more likely to be poor than families leaving welfare for other reasons.  Families terminated because of the time limit were less likely to have jobs, and more likely to experience hardships such as food insecurity, problems with housing and utilities, and unmet health care needs (Crichton, 2003).
  • A study of families leaving welfare in Cuyahoga County (Cleveland) found that families that left due to time limits had much higher poverty rates, lower employment rates, and more housing-related hardships than families that left for other reasons (Colton, et al., 2003).
  • Among TANF recipients in Virginia who lost eligibility due to the state’s time limit, more than 81 percent had household income below the poverty line when interviewed a year and a half after hitting the time limit.   This high level of poverty persisted even though nearly 90 percent worked at some point after hitting the limit.  When these families were employed, they worked an average of 37 hours per week (Wemmerus, Kuhns, and Loeffler, 2003).

The impact of a state’s time limit policy likely depends in part on recipients’ understanding of the policy.  Recent studies suggest that many recipients have a limited understanding of time limit rules.

  • Nationally, recipients with two or more work barriers and Latino recipients whose primary language is Spanish are much less likely to understand their state’s time limit policy than other recipients.   Almost 40 percent of TANF recipients overall report that they were not told they had a time limit or did not know when they would reach the limit.  About 73 percent of Spanish-speaking TANF recipients and 50 percent of recipients with two or more work barriers did not understand the time limit (Zedlewski and Holland, 2003).
  • Research conducted in Philadelphia found that most recipients knew there was a time limit, but that only half of recipients knew the length of the limit.  Many welfare recipients confused Pennsylvania’s 60-month time with the state’s requirement that recipients engage in work activities after receiving benefits for 24 months.  Others didn’t know extensions and exemptions were available (Michalopoulos, et al., 2003).

Prior research has generally shown that large shares of families that have been sanctioned face significant barriers to employment — such as health problems, children with health problems, low basic skill levels, and substance abuse problems.  (For a review of the literature, see Pavetti (2003)).  More recent research provides additional evidence to support these findings.

  • A national survey of mothers interviewed when their child reached age one found that mothers who left welfare after being sanctioned were more than three times as likely to have experienced material hardship — homelessness or eviction, hunger, or moving in with others — as mothers of infants who stayed on welfare.  These results held even after the researchers controlled for a range of demographic and other variables, including the level of material hardship the mother experienced around the time the child was born.  When the researchers looked at the relationship between hunger and sanctioning, they found that sanctioned mothers were more than six times as likely as mothers staying on welfare or leaving welfare without being sanctioned to have experienced hunger.  The overall level of hardship among sanctioned mothers with one-year children was quite high — about 45 percent of sanctioned leavers had experienced at least one material hardship (Reichman, Teitler, and Curtis, 2003).  Similarly, Cook, et al. (2002) found that households with children age 3 or younger that had been sanctioned were more likely to experience food-related hardships than other comparable households and that the children in these families were more likely to have been hospitalized than other similar children.
  • A major study conducted in three large cities — Boston, Chicago, and San Antonio — found that children in current and former welfare families that had been sanctioned generally had higher rates of serious behavioral and emotional problems than children in other TANF families.  One of the more notable findings was that 56 percent of preschoolers with mothers who had been sanctioned and left welfare scored in the “range of concern” for serious behavioral and emotional problems (Chase-Lansdale, Coley, Lohman, and Pittman, 2002).

Taken together, these new studies and previous research provide unassailable evidence that sanction and time limit policies are leading to real hardship for a substantial number of poor children.


Recent Research Continues to Show that Families with Health Problems Often have Great Difficulty Moving from Welfare to Work

A considerable body of research has examined the prevalence of health problems among TANF participants and leavers, and the correlation between health problems and employment and sanction rates (See Butler, 2002; Goldberg, 2002).  Recent research in this area continues to find a strong relationship between health and employment (Zedlewski, 2003).

A recent University of Michigan study found that the prevalence of health-related barriers to employment is higher than the rates in previous studies when families’ circumstances are examined over a significant period of time, rather than at one point in time.  This research builds on earlier research by examining the prevalence of health problems — and their impacts on employment outcomes — among TANF recipients in part of Michigan over a multi-year period, rather than simply measuring the proportion of recipients at any point in time that have health problems.  The researchers found that physical and mental health problems and child health problems each are related to lower employment durations over a nearly five-year period, even after controlling for a range of factors that effect employability including job skills, prior work experience, and access to transportation (Corcoran, Danziger, and Tolman, 2003).

Mothers caring for children with health problems are less likely to be employed and work fewer hours when they are employed.  Research conducted using pre-welfare reform data has fairly consistently found that poor child health has a negative impact on mother’s employment.  (For a review of this research, see Powers, 2003).    Recent research conducted using post-reform data also finds that poor child health limits mother’s employment (Corman, Reichman, and Noonan, 2003; Bednarek and Hudson, 2003).[3]

Many state officials and welfare-to-work providers have reported that the share of TANF recipients with substantial work limitations or barriers have increased over time as TANF caseloads declined.  Research conducted on this question in the late 1990s generally did not find an increase, although typically the number of work limitations examined was limited.  Some recent research, however, suggests that the share of TANF recipients with health problems may be increasing.  An increasing share of TANF recipients reported work-limiting health conditions in two different national surveys between 1993 and 2001 (Bavier, 2003).  (One other national survey, however, does not show an increase in very poor mental or physical health between 1999 and 2002, although the level of health problems reported in this survey is higher than in the other two (Zedlewski, 2003)).  The same researcher found that returns to welfare among leavers without work-limiting conditions declined after 1996, but remained stable for leavers with work-limiting conditions.


A Large Share of Very Poor Families Do Not Receive TANF Assistance

TANF cash assistance caseloads declined and single-parent employment increased in the last half of the 1990s.  Poverty also declined, but at a much slower rate than the declines in welfare caseloads.  While fewer families received TANF benefits and more families had earnings, many new workers remained poor because they were paid low wages and, in many cases, did not receive benefits such as cash assistance or food stamps.  More recently, poverty has increased, but the TANF caseload has not increased nationally as one would expect in a time of rising need.

  • HHS data show that about half of TANF-eligible families do not receive assistance and that non-receipt among eligible families is increasing.  In 2000, only about half of families poor enough to qualify for TANF received monthly cash assistance, down from nearly 8 in 10 eligible families in 1996 (HHS, 2003).  The continued decline in caseloads since 2002 despite rising poverty suggests that the share of eligible families receiving TANF cash assistance has continued to fall (Fremstad, 2003).
  • Other national surveys also find that about half of TANF-eligible families are not receiving assistance.  A survey of new parents in 20 U.S. cities found that 45 percent of TANF-eligible mothers did not receive TANF between child’s birth and the child’s first birthday (Reichman, Teitler, Garfinkel, and Garcia, 2003).
  • Research using data from the initial years of welfare reform (1996 and 1998) suggests that at least 500,000 to 1 million single-parent families that were eligible for TANF, but not participating, could have received important assistance and supports if they had participated in the program.  These families would have had higher incomes because they would have received cash aid and would have benefited from important employment services and work supports provided by state TANF programs.  About half of these non-participants were extremely poor and many had poor health or other work barriers (Zedlewski, 2002).
  • Single-mothers who were potentially eligible for welfare were much less likely to enter the TANF system after enactment of the 1996 welfare law than in the first half of the 1990s.  The decline in entry rate is not explained by changes in the characteristics of low-income mothers or improvement in the economy, and falling benefit levels (measured in real terms) had only a modest impact on entry rates (Acs, Phillips, and Nelson, 2003).

Food Stamp and Medicaid Programs Respond to Meet Increased Need

Welfare caseloads nationally are essentially flat, even as poverty has increased.  Participation in food stamps and health care programs, however, has increased as one would expect in a weakened economy.

  • Recent data from the Center for Law and Social Policy shows that national TANF cash assistance caseloads were essentially flat between June 2002 and June 2003, the last month for which data are available.  Since the beginning of the recession, national caseloads have fallen 4 percent, though some 28 states have seen caseload increases (Rahmanou, Richer, and Greenberg, 2003).  Among these states, caseloads have risen by an average of 16 percent since March 2001.

  • Food stamp caseloads, by contrast, have increased markedly.  Between July 2000 (when food stamp rolls fell to their lowest level) and October 2003, the number of households receiving food stamps increased by 38 percent.

  • A recent report by the Urban Institute found that the number of children receiving health insurance through Medicaid or SCHIP increased by 4.8 million between 1999 and 2002.  This reflects the impact both of the economy and outreach efforts designed to enroll children eligible in these programs (Kenney, Haley, and Tebay, 2003).

It is likely that informal and formal “diversion” policies have contributed to the decline in TANF participation among poor families.  There has been very little research on diversion policies.  Two recent studies, however, provide important new information on the impact of diversion.

  • Using national data, London (2003) finds that TANF applicants who are “diverted” from participating are more likely to have low education levels than TANF recipients, other TANF applicants who didn’t enter TANF, and TANF leavers.  Interestingly, more educated applicants also are overrepresented in the diverted population.  Diverted applicants also are less likely to have jobs than other TANF leavers and more likely to speak Spanish as their primary language.  These findings suggest that many of the potential TANF participants who are deterred by diversion programs would be better served by participating in TANF, even though diversion may be targeted appropriately in some cases.
  • A study of TANF “non-entrants” in three major cities found that non-entrants were significantly more likely than to be disabled or have other health problems than TANF entrants (Moffitt, et al., 2003).  Overall, the researchers found that families with greater need were more likely to obtain assistance, but that the targeting of diversion often was problematic.  For example, black families were more often discouraged from applying than other families, and families who were diverted from welfare did not have higher income levels over time than families who were not diverted.
  • A study of TANF applicants in Milwaukee County, Wisconsin found that the earnings of applicants who ended up participating in TANF were about the same as the earnings of applicants who did not participate, but that the total income of participants was significantly higher (but still below the poverty line) over a 12-month period.   The median income of non-participants was only $3,380 compared to $8,583 for participants.  Participants fared better because they received cash assistance and were more likely to receive food stamps (Dworsky, Courtney, and Piliavin, 2003).


Child Care Assistance Programs Increase Full-Time Work among Low-Income Families

Most mainstream welfare analysts and researchers agree that the large increase in single parents’ labor force participation and the steep declines in welfare caseloads in the 1990s were due to a combination of factors, including a strong economy, strengthened supports for working families, and an increased emphasis on work within state TANF programs.  Unfortunately, the relative importance of these factors — as well as the interactive effects among them — is difficult to disentangle.  Based on a review of research in this area, one conservative analyst, Doug Besharov of the American Enterprise Institute and the University of Maryland, recently estimated that the economy accounted for 35 to 45 percent of the decline in the welfare caseloads, increased aid to the working poor accounted for 20 to 30 percent, increases in the minimum wage for 0 to 5 percent, welfare reform for 25 to 35 percent, and erosion of the value of cash benefits for 5 to 10 percent (Besharov and Germanis, 2003).

Surprisingly, the role of the expansion of child care subsidies to low-income parents in the 1990s in increasing employment rates and hours worked has just begun to receive significant research attention.[4]  (There are several studies from 1990s that examine the impact of child care costs on employment, but they use pre-welfare reform data and generally do not estimate the impact of providing child care assistance on employment.[5]  More recent child care research has focused on the impact of care on the well-being of children (Fuller, Kagan, Caspary, and Gauthier, 2002)).  A growing body of new research, however, looks more closely at this area and consistently finds strong linkages between child care and employment increases.

  • Providing subsidized child care increases the likelihood that current and former welfare recipients leave welfare and work more than part-time.  A study of changes in Rhode Island’s child care program found that policies that expanded access to child care subsidies significantly increased the probability that parents would leave welfare for work and work more than 20 hours per week (Witte and Queralt, 2003).  Similarly, studies of the impact of increases in child care in Massachusetts and Miami-Dade County, Florida found that increased funding per poor child for child care is associated with increases in single-parent employment (Lemke, Witt, and Dryden Witte, 2001).
  • A study of TANF recipients in Michigan found that receipt of child care subsidies was associated with higher employment rates.  After controlling for demographic and other factors shown to affect work, recipients who had subsidized child care worked more months and had increased earnings — having a subsidy was associated with a 50 percent increase in months worked and a more than 100 percent increase in earnings (Danziger, Ananat, and Browning, 2004).
  • Families that leave welfare and receive child care assistance are less likely to return to the rolls than families that do not receive child care assistance.  A national study found that 28 percent of welfare leavers who didn’t receive child care assistance returned to welfare within three months after leaving, compared to only 19.5 percent of welfare leavers who receive child care assistance (Loprest, 2003).


Immigrant Families Face Increased Hardships Because of Eligibility Restrictions

The 1996 welfare law made many legal immigrants ineligible for TANF, Medicaid, Food Stamps, and SSI.  Research conducted over the past several years has found that food insecurity rose significantly among immigrant-headed households most likely to be subject to the restrictions while declining among most other households (this research pre-dates the recent rise in food insecurity among all households), and that the proportion of immigrants who lack health insurance has increased (Fremstad 2002).  More recent research on the impact of the restrictions on the well-being of immigrant families is consistent with these earlier findings.  In addition, while most of the research to date has focused on immigrants’ participation in food stamps and health care programs, a growing body of new research looks at immigrants’ participation in the TANF program.

Low-income immigrant children are much less likely than citizen children to have health insurance despite recent expansions in children’s health insurance coverage.  This disparity grew much wider in the wake of restrictions Congress placed on many legal immigrants’ eligibility for Medicaid in 1996 (Ku, 2003). 

 Low-income immigrant families are less likely to receive TANF assistance than other low-income families.  However, legal immigrants who do receive TANF have significant barriers to work, and if they are working, are more likely than other recipients to have jobs that provide little opportunity for speaking English, gaining skills, and achieving self-sufficiency (Tumlin and Zimmermann, 2003; Fremstad, 2003).


Federal Policy Implications

The President’s TANF proposal and both versions of TANF legislation pending in the House and Senate on his proposal would impose numerous new restrictions on states — and particularly on state welfare-to-work programs — while providing no new TANF block grant funding.  Both bills provide a very small increase in child care funding that is insufficient even to allow states to maintain their current levels of child care funding.  The pending proposals would do little to address the challenges uncovered by recent research, such as the increase in joblessness among welfare leavers, the growing share of very poor children who do not receive help from TANF, the limited income mobility for many low-income families, and the increase in material hardships faced by immigrants.  In fact, as most states have made clear, these proposals would make it more difficult for states to provide the assistance and support families need to move to stable employment and advance in their careers.

Such proposals might make sense if states had done little since 1996 to move families from welfare to work.  But just the opposite is true.  States aggressively implemented a “work first” approach to welfare reform that included strong mandatory work requirements.  They used the savings generated by significant declines in the cash assistance caseload to increase access to work supports — particularly child care — for both welfare recipients and low-income families not on welfare.  Employment among welfare recipients and single mothers generally reached record high levels and the length of time families spend on welfare declined substantially.

The current proposals seem to assume that the problems states faced in the early 1990s — including increasing caseloads and low employment rates — are the same problems they face today.  Of course, as more recent experience and research shows, today’s challenges are vastly different from those of the early 1990s.  Welfare caseloads remain low but families who are leaving welfare increasingly do not have stable employment.  Even when they have jobs, their earnings remain extremely low — many remain below the poverty level — in both the short- and long-term.

Another one of today’s concerns involves the significant share of TANF recipients that are disabled or have other serious barriers to employment.  Too often these barriers are not addressed.

Some of the reforms implemented over the past decade — particularly full-family sanction policies, time limits, and certain types of work requirements and diversion policies that are not designed to meet the needs of recipients with health and other barriers — have been counter-productive for these families, a point apparently conceded by some of the strongest proponents of the 1996 reforms.  As three Republicans who were leaders in the 1996 welfare reform debate recently wrote:  “some poor families are worse off as a result of the reforms . . . usually because of depression, addictions, or other serious personal problems . . . .”[6]  Yet the current TANF bills under consideration will make it harder in many cases for states to craft programs that can help address these shortfalls in current policies.

To help states better address today’s challenges, the following changes should be made to the current versions of TANF reauthorization legislation:

  • The restrictions the bills place on state flexibility to provide rehabilitative and other services to TANF recipients with disabilities and other barriers to employment should be loosened.  Under both the House and Senate Finance Committee, the amount of time that a state could place a TANF recipient with health or other barriers in rehabilitative services and count that placement toward the state’s participation rate would be limited (to three months in the House bill and six months in the Senate Finance bill).  The limits seem to be motivated by a concern that states would somehow “abuse” their use of such placements.   There is little reason to fear such abuse.  States currently place only a small percentage of TANF recipients in such activities — this is the case even in states that have had the flexibility to count recipients placed in rehabilitative and other activities toward their work rates without any limitations.  Moreover, states have strong financial incentives to limit the length of such placements.
  • Congress should provide funding for “transitional jobs” programs.  Transitional jobs programs provide paid work experience to welfare recipients with the greatest barriers to employment.  Such programs may offer a uniquely effective approach to helping recipients with severe work limitations leave welfare and move into stable employment (Hill, Kirby, and Pavetti, 2002).  Targeted funding would ensure that such programs are implemented and evaluated in several states.
  • The limits that current law and both bills — but particularly the House bill — impose on job training and education for TANF recipients should be lifted.   These restrictions will not help states address the issues of stagnant earnings for former recipients and the increasing numbers of former recipients without jobs.  The opposition to job training and education seems to be driven in part by research from the 1980s and early 1990s showing that welfare-to-work approaches that stressed basic education in classroom settings without strong linkages to employment had modest impacts on employment and earnings.[7]  In designing current education and training approaches, states, educational institutions and providers have taken this research into account and developed new approaches increasing welfare recipients’ skills and earnings that are combine education and job skills training, and have strong linkages to employers and employment services.  Evidence to date suggests that such “mixed services” approaches are more effective than welfare-to-work programs that focus solely on job search and related activities (Martinson and Strawn, 2003).
  • A provision in the House bill that would mandate that all states adopt “full-family” sanctions should not be included in the final legislation.  As the research discussed in this paper shows, sanctioned families often have significant barriers to employment and face significant hardships as a result of being terminated from TANF programs.  In states without full-family sanctions, a family’s grant is reduced but their case remains open.  As a result, it is less likely that they end up disconnected from the TANF system and services that could help them obtain stable employment.  Instead of increasing the harshness of sanction policies and forcing all states to adopt a single national full-family sanction policy, decisions about the severity of sanctions should be left to states.  Moreover, states should be required to develop programs that help families comply with work requirements both before and after being sanctioned.  (See Goldberg and Schott, 2000).
  • At a minimum, Congress should provide states with the child care funding necessary to pay for the cost of any increased welfare to work mandates imposed by TANF reauthorization legislation and maintain the current number of child care slots.  The pending legislation falls some billions of dollars short of meeting even this minimum requirement.  As a result, the number of subsidized child care slots available nationwide would drop considerably over the next five years (Parrott and Mezey, 2003).  As the research discussed in this paper shows, this likely would result in reduced employment and increased welfare participation by single mothers.
  • The current-law provisions that prohibit states from providing Medicaid or federally funded TANF benefits to most legal immigrants during their first five years in the United States should be lifted.  Research has now demonstrated the harm such restrictions have caused to immigrant families, including many citizen children living in these families.  Legal immigrants have nearly all the same civil obligations as U.S. citizens.  And, like U.S. citizens, they are not immune from job loss or economic hardship.  There is little justification for welfare reform rules that treat them differently from U.S. citizens, and particularly for rules that do not even give states the option to use federal funds to provide them with the same employment-related and children’s services as U.S. citizens.


Implications for State TANF Policies and Future Research

The research findings detailed here provide important lessons for state-level TANF policymakers.  Many of the concerns raised by recent research — including the low levels of participation among eligible families, the struggles of families with barriers to employment, and the high rate of poverty among welfare leavers — can be addressed through changes in state TANF policies.

  • States should try to determine whether some of their TANF policies have had the effect of deterring participation by needy families.  Such policies may include diversion programs, job search and other requirements that needy parents must participate in before they are approved for TANF, full-family sanction policies, time limit policies, and welfare-to-work programs that do not take the specific needs and capabilities of individual families into account.
  •  If they are not already doing so, states should tailor work requirements to the needs of those families with barriers to employment, and review families’ circumstances before imposing sanctions to ensure that families unable to comply due to a health or other barrier are not being sanctioned.
  • Similarly, states should reexamine their time limit policies and assess whether their exemption or extension policies should be revamped to ensure that very needy families unable to find and retain employment that can support their families are not left without the ability to make ends meet. 
  • States should consider revamping TANF rules that make working families ineligible for TANF before they reach the poverty line.  Nearly all states have eased the financial penalties on work that were part of the AFDC program, but only a few states ensure that most families who leave welfare with a job also leave poverty.

Many of the research findings discussed in this report make clear that the results of welfare reform research conducted during the strong economic period of the 1990s may differ from results obtained from research during a weak economy.  Thus, there is a role for continuing research even on questions that some think may have been answered by earlier research.  For example, although several leaver studies provide information on income and hardship levels of families that have left welfare in the 1990s, the results of such studies might be very different if they had been conducted during an economic downturn.  Finally, considerable gaps in knowledge on the impacts of certain welfare reform policies remain.  In particular, more research is needed on TANF “non-entrants,” the impact of diversion policies, the long-term economic well-being of TANF leavers, and the long-term impact of time limits.


Greg Acs, Katherin Ross Phillips, and Sandi Nelson, “The Road Not Taken? Changes in Welfare Entry during the 1990s,” December 2003.

Gregory Acs and Pamela Loberts with Tracy Roberts, “Final Synthesis Report of Findings from ASPE “Leavers” Grants,” Urban Institute, November 2001.

Jay Bainbridge, Marcia K. Meyers, and Jane Waldfoger, “Child Care Policy Reform and the Employment of Single Mothers,” Social Science Quarterly, vol. 84, December 2003.

Richard Bavier, “Conditions Limiting Work Among TANF Entries, Stayers, and Leavers,” presentation at APPAM Conference, August 26, 2003

Heather Bednarek and Julie Hudson, “Child Disability and Mothers’ Labor Supply,” presented at American Economic Association Meeting, 2003.

Gordon Berlin, “What Works in Welfare Reform:  Evidence and Lessons to Guide TANF Reauthorization,” MDRC, 2002.

Douglas J. Besharov and Peter Germanis, “Welfare Reform and Caseload Decline,” in Family and Child Well-being After Welfare Reform, 2003.

Rebecca M. Blank, “Evaluating Welfare Reform in the United States,” Journal of Economic Literature, vol. XL, December 2002.

Dan Bloom, Mary FarrellBarbara Fink with Diana Adams-Ciardullo, “Welfare Time Limits:  State Policies, Implementation, and Effects on Families,” Manpower Demonstration Research Corporation, July 2002.

David Butler, Testimony of David Butler, Vice President, Manpower Demonstration Research Corporation on Temporary Assistance for Needy Families and the Hard-to-Employ before the United States Senate Committee on Finance, April 25, 2002.

Maria Cancian, Robert H. Haveman, Daniel R. Meyer, and Barbara Wolfe, “Before and After TANF:  The Economic Well-being of Women Leaving Welfare,” Social Science Review, vol. 76:4, December 2002.

Maria Cancian, Robert Haveman, Daniel R. Meyer, and Barbara Wolfe, “The Employment, Earnings, and Income of Single Mothers in Wisconsin Who Left Cash Assistance:  Comparisons among Three Cohorts,” Institute for Research on Poverty, January 2003.

Maria Cancian, Robert Haveman, Thomas Kaplan, Daniel R. Meyer, and Barbara Wolfe, “Work, Earnings, and Well-being After Welfare:  What do we Know?” Economic Conditions and Welfare Reform (Sheldon Danziger, ed.), 1999.

Center on Budget and Policy Priorities, “Poverty Increases and Median Income Declines for Second Consecutive Year,” September 2003.

Center on Hunger and Poverty Heller School for Social Policy and Management, “Food Security Institute Bulletin, November 2003,” Brandeis University.

Jeff Chapman and Jared Bernstein, “Falling through the Safety Net: Low-income Single Mothers in the Jobless Recovery,” April 11, 2003.

Lindsey Chase-Landsdale, Rebecca Levine-Coley, Brenda J. Lohman, and Laura D. Pittman, “Welfare Reform:  What About the Children?” 2002.

Claudia Colton, Neil Bania, Toby Martin, and Nina Lalich, “How Are They Managing? A Retrospective of Cuyahoga County Families Leaving Welfare,” Center on Urban Poverty and Social Change, Case Western Reserve University, June 2003.

John Cook, Deborah Frank, Carol Berkowitz, Maureen Black, Patrick Casey, Diana Cutts, Alan Meyers, Nieves Zaldivar, Anne Skalicky, Suzette Levenson, and Timothy Heeren, “Welfare Reform and the Health of Young Children: A Sentinel Survey in Six United States Cities,” Archives of Pediatric and Adolescent Medicine, vol 156:7, July 2002.

Mary Corcoran, Sandra Danziger, and Richard Tolman, “Employment Duration of African-American and While Welfare Recipients and the Role of Persistent Health and Mental Health Problems,” The Michigan Program on Poverty and Social Welfare Policy, July 2003. 

Hope Corman, Nancy E. Reichman, and Kelly Noonan, “Mothers’ Labor Supply in Fragile Families:  The Role of Child Health,” Center for Research on Child Wellbeing, Working Paper #03-20-FF, December 2003.

Leslie I. Crichton, “The Welfare Time Limit in Minnesota: A survey of families who lost MFIP eligibility as a result of the five-year time limit,” Program Assessment and Integrity Division, Minnesota Department of Human Services, July 2003.

Sandra K. Danziger, Elizabeth Oltmans Ananat, and Kimberly G. Browning, “Childcare Subsidies and the Transition from Welfare to Work,” forthcoming in Family Relations, vol. 52, no. 2, March 2004.

Amy Dryden Witte and Magaly Queralt, “Impacts of Eligibility Expansions and Provider Reimbursement Rate Increases on Child Care Subsidy Take-Up Rates, Welfare Use, and Work,” May 2003.

Amy Dworsky, Mark E. Courtney, and Irving Piliavin, “What Happens to Families Under W-2 in Milwaukee County, Wisconsin?  Report from Wave Two of the Milwaukee TANF Applicant Survey,” September 2003.

Alison Earle and S. Jody Heymann, “What Causes Job Loss Among Former Welfare Recipients:  The Role of Family Health Problems,” Journal of American Medical Women’s Association, vol. 57:1, Winter 2002.

Shawn Fremstad, “Immigrants and Welfare Reauthorization,” Center on Budget and Policy Priorities, January 2002.

Shawn Fremstad, “Immigrants, Limited-English Proficient Persons, and TANF:  What Do We Know?” Center on Budget and Policy Priorities, March 2003.

Shawn Fremstad, “Falling TANF Caseloads Amidst Rising Poverty Should Be a Cause of Concern,” Center on Budget and Policy Priorities, September 2003.

Bruce Fuller, Sharon L. Kagan, Gretchen L. Caspary, and Christiane A. Gauthier, “Welfare Reform and Child Care Options for Low-Income Families,” The Future of Children, Winter\Spring 2002.

Heidi Goldberg, “Improving TANF Program Outcomes for Families with Barriers to Employment,” Center on Budget and Policy Priorities, January 2002.

Heidi Goldberg and Liz Schott, “A Compliance Oriented Approach to Sanctions in State and County TANF Programs,” Center on Budget and Policy Priorities, October 2000.

Jeffrey Grogger, Lynn A. Karoly, and Jacob Alex Klerman, “Consequences of Welfare Reform: A Research Synthesis,” RAND, July 2002.

Heather Hill, Gretchen Kirby, and LaDonna Pavetti, Transitional Jobs Programs: Stepping Stones to Unsubsidized Employment, Mathematica Policy Research, Inc., April 2002.

David Jesuit and Timothy M. Smeeding, “Poverty Levels in the Developed World,” July 2002.

Rucker C. Johnson and Mary E. Cochrane, “The Road to Economic Self-Sufficiency:  Job Quality and Job Transition Patterns After Welfare Reform,” April 2003.

Ariel Kalil, Kristin Seedfelt, and Hui-chen Wang, “Sanctions and Material Hardship under TANF,” Social Science Review, vol. 76:4, December 2002.

Genevieve M. Kenney, Jennifer M. Haley, Alexandra Tebay, “Children's Insurance Coverage and Service Use Improve,” Urban Institute, July 2003.

Leighton Ku, “Report Documents Growing Disparities in Health Care Coverage Between Immigrant and Citizen Children as Congress Debates Immigrant Care Legislation,” Center on Budget and Polity Priorities, October 2003.

Robert J. Lemke, Robert Witt, and Ann Dryden Witte, “Child Care and the Welfare to Work Transition,” March 2001.

Rebecca London, “Which TANF Applicants are Diverted and What are their Outcomes?” February 2003.

Pamela Loprest, “Fewer Welfare Leavers Employed in Weak Economy,” Urban Institute, August 2003.

Pamela Loprest, “Disconnected Welfare Leavers Face Serious Risks,” Urban Institute, August 2003.

Pamela Loprest, “Use of Government Benefits Increases among Families Leaving Welfare,” Urban Institute, September 2003.

Thomas MaCurdy, Grecia Marrufo, and Margaret O’Brien-Strain, “What Happens to Families When they Leave Welfare?” Public Policy Institute of California, 2003.

Karen Martinson and Julie Strawn, “Build to Last:  Why Skills Matter for Long-run Success in Welfare Reform,” Center on Law and Social Policy, April 2003.

Marcia K Meyers, Theresa Heintze, and Douglas A. Wolf, “Child Care Subsidies and the Employment of Welfare Recipients,” August 1999.

Charles Michalopoulos, Kathryn Edin, Barbara Fink, Mirella Landriscina, Denise Polit, Judy Polyne, Lashawn Richburg-Hayes, David Seith, and Nandita Verma, Welfare Reform in Philadelphia: Implementation, Effects, and Experiences of Poor Families and Neighborhoods, MDRC, October 2003.

Cynthia Miller, “Leavers, Stayers, and Cyclers: Analysis of the Welfare Caseload,” Manpower Demonstration Research Corporation, November 2002.

Robert Moffitt, Katie Winder, Linda M. Burton, Alan F. Benjamin, Tera R. Hurt, Stacy L. Woodruff, and Amy Kolak, “A Study of TANF Non-Entrants,” November 2003.

New York City Human Resources Administration, Office of Program Reporting, Analysis, and Accountability, “Job Placements (FA & SNA): Trend; % Change from Previous Month; % Change from Previous Year,” January 2004.

LaDonna Pavetti, Review of Sanction Policies and Research Studies, Mathematica Policy Research, Inc., March 2003.

LaDonna Pavetti and Greg Acs, “Moving Up, Moving Out, or Going Nowhere? A Study of the Employment Patterns of Young Women and the Implications for Welfare Mothers,”
Journal of Policy Analysis and Management, vol. 20:4, Fall 2001.

Elizabeth T. Powers, “Children’s Health and Material Work Activity: Static and Dynamic Estimates under Alternative Disability Definitions,” Journal of Human Resources, vol. 38:3, 2003.

Hedieh Rahmanou, Elise Richer, and Mark Greenberg, “Welfare Caseload Remains Relatively Flat in Second Quarter of 2003,” Center for Law and Social Policy, October 2003.

Nancy E. Reichman, Julien O. Teitler, Irwin Garfinkel, and Sandra Garcia, “Variations in Maternal and Child Wellbeing Among Financially Eligible Mothers by TANF Participation Status,” Center for Research on Child Wellbeing, Working Paper #03-13-FF, April 2003.

Nancy E. Reichman, Julien O. Teitler, and Marah A. Curtis, “Hardships Among Sanctioned Leavers, Non-Sanctioned Leavers, and TANF Stayers,” Center for Research on Child Wellbeing, Working Paper #03-17-FF, December 2003.

Elise Richer, Steve Savner, and Mark Greenberg, “Frequently Asked Questions About Working Welfare Leavers,” Center on Law and Social Policy, 2002.

Lauren A. Smith, Juliet L. Hatcher, and Richard Wertheimer, “The Association of Childhood Asthma with Parental Employment and Welfare Receipt,” Journal of the American Medical Women’s Association, vol. 57:1, Winter 2002.

Karin Tumlin and Wendy Zimmermann, “Immigrants and TANF:  A Look at Immigrant Welfare Recipients in Three Cities,” Urban Institute, October 2003.

U.S. Department of Agriculture, Household Food Security in the United States, 2002 (2003).

U.S. Department of Health and Human Services, Indicators of Welfare Dependence, Annual Report to Congress 2003.

U.S. General Accounting Office, “Child Care:  Child Care Subsidies Increase Likelihood that Low-Income Mothers will Work,” 1994.

Alan Weil, “Ten Things Everyone Should Know about Welfare Reform,” Urban Institute, May 2002.

Nancy Wemmerus, Carole Kuhns, and Renee Loeffler, Experiences of Virginia Time Limit Families After Case Closure: 18-Month Follow-up with Cases Closed in Early 1998, 1999, and 2000, Mathematica Policy Research, Inc., November 2003.

Robert G. Wood and Anu Rangarajan, “What’s Happening to TANF Leavers Who Are Not Employed?,” Mathematica Policy Research, Inc., October 2003.

Sheila R. Zedlewski, “Left Behind or Staying Away?  Eligible Parents Who Remain Off TANF,” September 2002.

Sheila R. Zedlewski, “Work and Barriers to Work among Welfare Recipients in 2002,” Urban Institute, August 2003.

Sheila R. Zedlewski and Jennifer Holland, “How Much Do Welfare Recipients Know about Time Limits?,” Urban Institute, November 2003.

End Notes:

[1]  This analysis does not provide a comprehensive review of all recent welfare reform research, but instead focuses on research that breaks new ground or is particularly relevant to current federal and state debates over the future of welfare reform.  For recent reports that provide broad surveys of research from the 1990s, see Blank (2002) and Grogger, Karoly, and Klerman (2002).  A short paper by Weil (2002) provides an excellent summary of research, primarily conducted by the Urban Institute, on a broad array of welfare reform issues.

[2]  One limitation of most studies of welfare leavers’ income is that they generally do not take child care expenses and most other work-related expenses, except for payroll taxes, into account.

[3]  Other recent studies with similar findings (although they do not use post-welfare reform data exclusively) include Earle and Heymann (2002) who find that health limitations are associated with a significantly increased risk of job loss among welfare leavers even after controlling for other factors, and Smith, Hatcher, and Wertheimer (2002) who find that single parents of young children with asthma are less likely to have full-time work than other single parents of young children.

[4]  For example, an otherwise excellent 300-page report by Grogger, Karoly, and Klerman (2002) that synthesizes the current state of knowledge from the large body of research literature on the impacts of welfare reform includes no discussion of the impacts of child care assistance on employment and poverty. 

[5]  For a review of this literature, see Meyers, Heintze, and Wolf (1999).  A few studies from the early 1990s have attempted to estimate the impact of providing child care subsidies on employment.  For example, a 1994 report by the U.S. General Accounting Office estimated that subsidizing all of the costs of child care for poor women would increase their employment rates by 50 percent (GAO, 1994).  A more recent report finds that the expansion of child care funding from 1991 to 1996 had a substantial positive impact on the employment of single mothers with young children (Bainbridge, Meyers, and Waldfogel, 2003).

[6] “Compelling case to reauthorize welfare reform,” Bill Archer, Rep. E. Clay Shaw Jr., and Ron Haskins, Houston Chronicle, January 19, 2003.

[7]  Another factor is concern that states will dilute the work-focused nature of TANF by placing too many participants in stand-alone education and training activities.  As is the case with rehabilitative services, the concern is not borne out by anything states are actually doing.  Only a small percentage of TANF recipients are placed in vocational education activities even though these activities currently count toward work rates for up to 12 months.  In addition,  nearly every state could place a significant number of recipients in such activities for longer periods of time without jeopardizing their ability to meet federal work rate requirements.