Revised June 22, 1999

The Determinants Of Welfare Caseload Decline:
A Brief Rejoinder
by Liz Schott, Robert Greenstein, and Wendell Primus

A recent Heritage Foundation paper by Robert Rector and Sarah Youssef, "The Determinants of Welfare Caseload Decline," analyzes recent decreases in TANF caseloads. The paper recites the familiar statistics that national caseloads have fallen 37 percent during the last three years, with caseload reduction ranging from a decline of 84 percent in Wisconsin to a seven percent increase in Hawaii. The paper reaches two primary conclusions. First, it argues that "the relative vigor of state economies, as measured by unemployment rates, has no statistically significant effect on caseload decline;" it attributes all of the decline in caseloads to differences in state welfare policies and especially to stringent sanctions and immediate work requirements. Second, the paper argues that the "fall in the national AFDC/TANF caseload has not resulted in an increase in child poverty."

The Rector/Youssef paper is weak in three respects. It misses the effect that economic growth and declines in unemployment rates have indeed had in reducing caseloads. Second, it misses the substantial effect, documented in recent research, that non-welfare policy changes — particularly the large expansion in the Earned Income Tax Credit — have had in increasing employment and reducing welfare receipt among single parents. Finally, it misanalyzes the effect of caseload reductions on child poverty rates. Census data suggest that the dramatic caseload declines have, in fact, resulted in child poverty rates being somewhat higher than they otherwise would have been.

 

The Role of the Economy

The paper's conclusion that the economic boom has had no effect on caseload decline is not consistent with more rigorous mainstream analysis. Such analysis finds that declines in unemployment have contributed to this decline, although they account for a relatively small share of the decline.

The problem with Rector and Youssef's analysis in this area appears to stem primarily from the fact that their analysis examines caseload decline only during the 18-month period from January 1997 through June 1998. TANF caseloads began to decline in April 1994, however, 33 months before the beginning of the time period they examine. Almost 50 percent of the caseload decline between April 1994 and June 1998 occurred outside the period the Heritage analysis considers. (In a footnote, Heritage states it performed the same analysis for the 24 months from June 1996 to June 1998 and "achieved virtually the same overall results." This 24-month period still misses, however, the first 27 months of caseload decline. It is unclear why Heritage chose to present the results for the 18-month period rather than the 24-month period, since the longer period would have been a better period to use. While Heritage states that the results are similar for the 24-month period and the shorter 18-month period it uses, it does not provide data to substantiate that claim.)

A more rigorous analysis conducted by economists Geoffrey Wallace and Rebecca M. Blank and issued in February 1999 examines caseload declines and finds "the ongoing decline in unemployment rates can explain about 8 to 19 percent of the AFDC caseload declines since 1994 and about 28 to 44 percent of the Food Stamp caseload declines."⁽¹⁾ Several researchers also have looked at caseload reductions in the period from 1993 through 1996. A study by the Council of Economic Advisers estimated that 31 percent to 45 percent of caseload changes through 1996 were due to economic factors.⁽²⁾ Another study estimated that in the 26 states with the largest caseload reductions, 78 percent of the change during the 1993-to-1996 period was due to economic and seasonal factors.⁽³⁾

There is little question that welfare policies have played a large role in TANF caseload reduction. Broad consensus exists on that point. To argue, however, that neither the economy nor other policy changes have had any role is to miss the evidence to the contrary.

 

The Role of Other Policy Changes

There also is broad agreement among analysts that other factors besides the economy and TANF policies have contributed significantly to welfare caseload declines. The expansion of the Earned Income Tax Credit over the years appears to have played a large role in caseload reduction by increasing the proportion of single mothers participating in the labor force.

An important study by Northwestern University economists Bruce Meyer and Dan Rosenbaum found that the EITC was responsible for more than half of the increase between 1984 and 1996 in the proportion of single mothers who are employed.⁽⁴⁾ These findings are consistent with a study by Nada Eissa of the University of California and Jeffrey Liebman of Harvard University, as well as a study by Stacy Dickert, Scott Hauser, and John Karl Scholz of the University of Wisconsin. Eissa and Liebman found the EITC expansion contained in the Tax Reform Act of 1986 had a significant effect in inducing more single women with children to go to work.⁽⁵⁾ Dickert, Hauser, and Scholz projected that the EITC expansions in the 1993 budget law would induce a sizeable number of non-working single parents to seek employment and, in so doing, generate a reduction in welfare receipt.⁽⁶⁾ The Dickert, Hauser, Scholz study estimated that the 1993 EITC expansions would induce approximately 500,000 families to move from welfare to the workforce.

 

Caseload Decline and Child Poverty Rates

Rector and Youssef analysis also argue that reductions in caseloads have had no adverse impact on child poverty. To support this contention, they cite the fact that when the Earned Income Tax Credit, food stamps and other means-tested benefits are counted as income, the child poverty rate was 13.8 percent in 1997, the lowest such rate since 1980.

That the child poverty rate is lower now than at any time since 1980 is not, however, evidence that the caseload reductions have not affected child poverty rates. The economy was stronger in 1997 than at any time since 1980, and child poverty rates might have been expected to be lower than in any intervening year even in the absence of welfare policy changes. If the child poverty rate would have been below 13.8 percent in the absence of the large caseload reductions — and the caseload reductions raised the rate from what it otherwise would have been to 13.8 percent — then the caseload reductions have raised the child poverty rate. (Note: Most analysts do not use the Rector/Youssef figure of 13.8 percent for the child poverty rate in 1997. That figure is artificially low because it is based on a methodology that counts health insurance provided through Medicaid as income. Most analysts believe benefits such as food stamps and the Earned Income Tax Credit should be counted as income when measuring poverty, but health insurance should not be. Most analysts also agree that if the value of health insurance were to be counted as income, the poverty line would need to be raised.)

To determine whether — and if so, how — caseload reductions have affected child poverty rates entails examining Census data for trends in child poverty rates both before and after government benefits are counted. Such an analysis enables researchers to ascertain whether various categories of government benefit programs are growing stronger or weaker in their effect on child poverty.

Various researchers, including analysts at the Center on Budget and Policy Priorities, have conducted such analyses. The analyses show that the low child poverty rate in 1997 resulted primarily from the combined effects of the strong economy and the large EITC expansions enacted in 1990 and 1993. These analyses also indicate that contrary to Rector and Youssef's contentions, the sharp declines in TANF and food stamp caseloads have raised child poverty rates. The Census data suggest that if these caseloads had not declined as dramatically, the child poverty rate would have been lower in 1997.

Specifically, the Census data show that the diminished coverage of means-tested programs such as cash assistance and food stamps resulted in 460,000 fewer children being lifted out of poverty by means-tested programs in 1997 than in 1994. This figure takes into account the role of improvements in the economy in reducing the number of poor children between 1994 and 1997.

• The Census data show that in 1994, means-tested benefits lifted from poverty 19.1 percent of children who were poor before receipt of these benefits.

• In 1997, these benefits lifted from poverty only 16.0 percent of the children who otherwise would have been poor.

• Had the means-tested programs lifted from poverty in 1997 the same proportion of children who otherwise would be poor as in 1994, some 460,000 fewer children would have been poor in 1997.

Food stamp data underscore this point. Between 1994 and 1997, the number of poor children fell 1.4 million. The number of children receiving food stamps fell 2.5 million, almost twice as much.

 

Judging the Success of Welfare Reform: Is Caseload Decline the Best Measure?

The Census data raise a question: Should welfare caseload reductions be the major criterion on which the success of welfare reform efforts are judged? Or is the success of welfare reform more appropriately judged by whether — and if so, to what extent — the employment of parents has increased and the poverty of children been reduced?

From this perspective, caseload decline achieved by imposing immediate "full-family" sanctions should not necessarily be regarded a welfare reform "success." To some extent, sanctioning a family by cutting the family off welfare entirely (i.e., terminating benefits for children as well as their parents) represents a failure of the state to connect the adult to work or work preparation activities. Recent studies indicate that families that leave welfare due to a sanction are less likely to be employed than families that leave welfare in general.⁽⁷⁾ Studies that track the employment of families that leave welfare due to a sanction have found that from 16 percent to 53 percent of these families have employment after leaving welfare.⁽⁸⁾ In contrast, studies that track the employment of families that leave welfare for any reason have found that from 61 percent to 87 percent of the families have employment after leaving welfare.⁽⁹⁾

A state that persists in working with a parent who is not complying with work activities and ultimately gets the parent to participate has succeeded in furthering a key objective of welfare reform — promoting job preparation and work. State sanction policies and procedures that lead to participation rather than termination are more likely to achieve that objective.

These types of sanction policies in use in some states can help advance this objective.

• Escalating sanctions: Some states which use sanctions that terminate benefits for the entire family impose such "full-family" sanctions only after repeated instances of noncompliance or a continued period of noncompliance. These states impose an initial partial sanction, which provides an opportunity for the state to maintain contact and to continue working with the family to achieve compliance after the initial infraction, while continuing to provide assistance to the children.

• Sanction avoidance: Some states work with families up front to avoid a sanction. Prior to imposing a sanction, these states use a conciliation process to ensure the parent understands what is required, to identify why the parent is not participating in the work activity and to adjust requirements or address barriers to participation as appropriate. If the parent complies with work requirements, the sanction is not imposed.

Some states also review cases prior to closure to avoid erroneously cutting off benefits and to provide another opportunity for parents to comply. For example, Tennessee has established a system under which all cases scheduled for closure due to non-compliance (plus cases scheduled for closure due to other reasons such as time limits) are reviewed by an outside entity prior to case closure. As part of the review, the outside entity (which operates under a contract with the state) checks to make sure the parent has been properly notified, understands the requirements, and is given an opportunity to comply. When Tennessee instituted this system, 30 percent of the cases subject to a Customer Service Review ultimately were not closed. In most of the cases that had been slated for closure due to noncompliance but were not closed as a result of the review, the parents came into compliance.

• Safety net assistance: Connecticut continues working with sanctioned families that have lost cash welfare benefits. It allows these families to participate in a "safety net" program that provides assessments, case management services, and vouchers. This program has achieved notable success in securing compliance and raising employment rates among families that have been sanctioned, many of which have significant barriers to employment. The employment rates for "safety net" program participants have increased from 12 percent of the families being employed at the time of referral to the program to 37 percent of the families being employed while participating in the program.⁽¹⁰⁾

The Rector/Youssef paper would characterize all three of these approaches as weak or less-desirable sanction policies. Yet these approaches further the welfare reform goals of promoting work, strengthening families, and improving the well-being of children. Rather than being weaker approaches to welfare reform, these policies appear to be stronger approaches to achieving compliance and participation in work activities. They also provide more adequate protection to children.

It thus appears to be important for states to continue to work with sanctioned families to achieve compliance, rather than simply cutting off benefits and contacts. These families often face substantial barriers to employment. It may take repeated, lengthy efforts to surmount those barriers.

• A Utah study found that three-fourths of sanctioned families had three or more barriers to employment, most commonly a health or medical problem, lack of transportation and lack of skills. For more than half of the families, a health-related problem — including a mental health problem — was the reason for nonparticipation.⁽¹¹⁾

• A Minnesota study found a significant percentage of sanctioned families had one or more barriers to employment. The study found sanctioned families were four times as likely as the caseload as a whole to report chemical dependency, three times as likely to report a family health problem, and twice as likely to report a mental health problem or domestic violence.⁽¹²⁾

• A recent Delaware study found sanction rates to be higher for those with the least work experience and the least education. The study also found sanctioned individuals were more likely to have trouble understanding TANF rules and the consequences of not participating.⁽¹³⁾

• Data from Connecticut's safety net program, which serves families not eligible for TANF cash assistance because they have used up the maximum 21 months of cash benefits that Connecticut provides to families and cannot receive an extension because of sanction or noncompliance with work, provide a further indication that sanctioned families are among the most vulnerable. As compared to the general population receiving welfare benefits, families in the safety net program had a significantly higher incidence of substance abuse, health or mental health problems, and domestic violence difficulties.⁽¹⁴⁾

---

End Notes:

1. Wallace, Geoffrey and Rebecca Blank. 1999. What Goes Up Must Come Down? Explaining Recent Changes in Public Assistance Caseloads. Paper presented at the conference "Welfare Reform and the Macroeconomy" sponsored by the Joint Center for Poverty Research, February 1999.

2. Council of Economic Advisers. 1997. Technical Report: Explaining the Decline in Welfare Receipt, 1993-1996. A Report by the Council of Economic Advisers, Washington, D.C.

3. Ziliak, James P., David N. Figlio. 1998. Welfare Reform, the Business Cycle, and the Decline in AFDC Caseloads. Unpublished paper, University of Oregon at Eugene.

4. Bruce D. Meyer and Dan T. Rosenbaum. Welfare, The Earned Income Tax Credit, and the Labor Supply of Single Mothers. October 26, 1998. The study estimated that between 1992 and 1996, the EITC was responsible for one-third of the increase in the proportion of single mothers who are employed.

5. Nada Eissa and Jeffrey B. Liebman, "Labor Supply Response to the Earned Income Tax Credit," Quarterly Journal of Economics, May 1996, 112(2), pp. 605-637.

6. Stacy Dickert, Scott Hauser, and John Karl Scholz, "The Earned Income Tax Credit and Transfer Programs: A Study of Labor Market and Program Participation," in James M. Poterba, ed., Tax Policy and the Economy, Vol.9., MIT Press, 1995.

7. Brauner, Sarah and Loprest, Pamela. Where Are They Now? What States' Studies of People Who Left Welfare Tell Us. Urban Institute, May 1999.

8. A recent Utah study of case closures due to sanction found that parents generally did not move into employment after termination of benefits. The adult found work after grant sanctioning in only 16 percent of the cases. Derr, Michelle K. The Impact of Sanctioning on Utah's TANF Families. Paper presented at Association of Public Policy and Management Annual Research Conference. October 1998.

A Tennessee study found that about 40 percent of families whose cases were closed due to sanction were employed. Summary of Surveys of Welfare Recipients Employed or Sanctioned for Noncompliance, University of Memphis. March 1998. The Tennessee study, along with ones from Michigan and Iowa which found 53 percent of families leaving welfare due to sanctions were employed, are discussed in Brauner, Sarah and Loprest, Pamela. Where Are They Now? What States' Studies of People Who Left Welfare Tell Us. Urban Institute, May 1999.

A GAO review of early sanction experiences in three states (Massachusetts, Iowa, and Wisconsin) found that between 23 percent and 32 percent of households whose AFDC cases were closed due to sanction reported wages after termination of benefits. General Accounting Office. Welfare Reform: States' Early Experiences with Benefit Termination. May 1997.

9. General Accounting Office. Welfare Reform: Information on Former Recipient's Status. April 1999. This study thoroughly reviewed state efforts to track families that have left welfare. GAO looked at 18 studies in 17 states and focused on eight studies that cover seven states (two studies were done in Wisconsin). The GAO choose these eight studies because their design included most of the families that had left welfare and these studies contained sufficient data to be representative of the population of families leaving welfare in those states.

10. Welfare Reform in Connecticut. WorkSteps: Connecticut's Safety Net Program, Presentation materials prepared by Donna Campbell, Executive Director, Employment Success Program, February 24, 1999. In addition, about one-third of the families referred to the "safety net" program are able to requalify for TANF cash assistance either because of compliance with work requirements or a determination of exemption from work requirements pursuant to a thorough assessment.

11. Derr, Michelle K. The Impact of Sanctioning on Utah's TANF Families. Paper presented at Association of Public Policy and Management Annual Research Conference. October 1998.

12. Minnesota Department of Human Services. 1996. Internal Memorandum, cited in Pavetti, LaDonna, et al. Welfare-to-Work Options for Families Facing Personal and Family Challenges: Rationale and Program Strategies. Urban Institute. August 1997.

13. Fein, David J. and Wang, S. Lee. Carrying and Using the Stick: Financial Sanctions in Delaware's A Better Chance Program. Abt Associates, Inc. May 1999.

14. Welfare Reform in Connecticut. WorkSteps: Connecticut's Safety Net Program, Presentation materials prepared by Donna Campbell, Executive Director, Employment Success Program, February 24, 1999.