Fall 2018 Journal: Welfare-to-Work: Work Requirements under TANF

By Hinnaneh A. Qazi

Edited by Randall Tran and Chenchen Zhang

Mandatory work requirements require that social safety net recipients meet certain workforce-related criteria in order to continue receiving benefits. This article evaluates various alternatives available to states under federal Temporary Assistance to Needy Families (TANF) mandatory work requirements — including work activity, sanction, and time limit approaches. It then proposes recommendations for how states can more efficiently allocate government resources and incentivize labor market participation without exacerbating cycles of poverty among disadvantaged groups historically supported by welfare. Since 1996, there has been a national overemphasis on embedding direct employment-related activities in mandatory work programs. A more ideal strategy is a mixed program approach, which leverages both employment- and education/training-focused options. Evidence suggests this alternative can maximize payoffs to both governments and recipients. Furthermore, research points to the fact that sanctions are disproportionately harming some of the most vulnerable groups within the welfare system, and that partial sanctions should be considered over full sanctions. With regard to time limits, state variation should be exploited to determine what duration of time limits is optimal. In the end, this article will also illustrate that comprehensive federal reform is imperative. The analysis narrows in on state-level approaches to take into consideration present-day issues of political feasibility at the federal level, but limitations facing states and other systemic issues within TANF are urgently vital to address.

BACKGROUND

An overhaul led by Republicans and President Bill Clinton, the Personal Responsibility and Work Opportunity Reconciliation Act (PRWORA) of 1996 aimed to strengthen America’s work ethic, reduce dependency on the social safety net, and incentivize recipients into labor market participation. The bill, which established the block grant program Temporary Assistance for Needy Families (TANF), was a response to major growth in the Aid to Families with Dependent Children (AFDC) entitlement program caseload. Between 1981 and 1994, the AFDC basic caseload (not including the smaller Unemployed Parent program) grew from 3.6 million to 5.1 million cases, a 42 percent increase in a little over a decade.1At the same time, federal funds to states for benefits were not keeping up with inflation. From 1970 to 1994, typical benefits in the median state fell 47 percent after adjusting for inflation.2In an effort to rein in caseloads and address budgeting shortfalls, many states began to implement designs to increase workforce participation prior to the 1996 law.3

TANF replaced ADFC and was modeled after many of these state initiatives. Instead of setting mandatory work requirements (also known as “workfare,” mandatory work requirements are policies which specify that social safety net recipients must meet certain participation criteria to continue receiving benefits),4on individual recipients, however, TANF set a work participation standard on states. The standard requires that states engage a percentage of TANF recipients in designated work-related activities for a minimum threshold of hours per week and/or a minimum number of weeks per year. Under the provision, states have a different standard to meet for the overall population of recipients versus two-parent families. The performance standards vary by state based on previous caseload reductions the state has experienced.5For example, as Utah had significantly reduced its caseload in fiscal year 2015, the state’s overall participation standard was reduced from 16.3 percent in 2015 to 11.5 percent in 2016 (the overall participation standard average across all states in 2016 was 40 percentage points higher at 51.9 percent).6In essence, the structure rewards states that effectively reduce TANF dependency over time.

Ultimately, by placing higher opportunity costs on non-market time (the price of not working) and tightening available budget constraints under welfare, the law sought to incentivize welfare recipients to subsidize away from non-market time toward market time. With some exceptions, PRWORA also implemented time limits on federal funds and required states to place sanctions on recipients who did not comply with work requirements. The law placed a two-consecutive-year time limit on benefits: recipients are expected to be employed and active in the labor force two years after first receiving assistance. The bill also instituted a lifetime benefit limit of at most five years (or less at a state option).7

Despite being subject to the above requirements, states maintain considerable flexibility in structuring approaches to meet their work participation standard, and have autonomy to establish more stringent time limits and sanctions than set by federal law.

OUTCOMES OF THE 1996 REFORM

Since enactment, there has been widespread consensus that the law dramatically reduced state caseloads and significantly increased employment and earnings among single mothers, the largest population served by welfare. Research has revealed several indicators of the reform’s success:

Between 1994 and 2005, the federal welfare caseload declined by 60 percent.8

The employment rate among single mothers jumped markedly from 60 percent in 1994 to 72 percent in 1999, an unprecedented change in the labor force by historical standards. The conclusion that welfare reform played a significant role in this increase is supported by multiple studies that have controlled for the strong economy of the 1990s and the Earned Income Tax Credit (EITC).9

Earnings as a share of income (as compared to welfare payments, food stamps, housing support, and school lunch) for low-income, female-headed families leaped from 30 percent in 1993 to 60 percent in 2000. During the same period, welfare payments went from constituting 55 percent of income to only 23 percent.10

Caseload totals and employment and earnings estimates do not tell the whole story — despite the increase in employment, 40 percent of welfare leavers were found to not work regularly after leaving the rolls and nearly 30 percent were found to have not worked at all since exiting welfare,11 alluding to a larger problem with employment opportunity, retention, and stability.

The evidence also illustrates that the total income of women leaving welfare has remain relatively unchanged.12Mothers in the bottom 10 percent of single earners actually lost income in the years following PRWORA13and in 2015, TANF was reaching fewer jobless single mothers than it had ever before.14

THE PROBLEM

This article provides insight into a longstanding problem facing policymakers and administrators: how to provide the necessary financial support low-income families seriously need, while also reducing dependence on assistance programs and helping lift recipients out of poverty in the long-term. My analysis evaluates approaches available to states under PRWORA, and prioritizes analysis of work activity, sanction, and time limit options, state approaches for which there is the greatest body of evidence. It does so by investigating which alternatives to-date are the most efficient, effective, and equitable means to achieve the above ends.

CRITERIA

Efficiency:Does the approach maximize net benefits from society’s perspective? Does it also reduce transaction costs and administrative burden?

Equity:Is there equity in terms of the process and outcome? Does the approach avoid exacerbating negative outcomes for America’s most disadvantaged families? Is any group disproportionately harmed?

Effectiveness:Is the alternative successful in bringing desired results? On the recipient side, this includes self-sufficiency, employment, earnings, and income. On the government side, this includes decreasing welfare receipts and expenditures.

STATE APPROACHES

Under TANF’s block grant financing structure, states must meet their annual workforce participation standard in order to qualify for future federal funds but are allowed some decision-making freedom including the ability to instate exemptions (e.g. exemptions for mothers with infant children).15

Recipient participation in 12 activities counts toward the performance measure: unsubsidized employment, subsidized private employment, subsidized public sector employment, on-job training, job search and readiness, work experience, community service, vocational educational training, job skills training, education directly related to employment, completion of secondary school, and providing child care to a community service participant.16States can select some or all activities to include in their welfare-to-work programs (many states make all 12 available), and in some states, counties have been granted discretion. Activities are often grouped into two categories: Labor Force Attachment (LFA), which emphasizes job search and finding a job as quickly as a recipient can, and Human Capital Development (HCD), which emphasizes education or training prior to job search.17However, because of PRWORA’s overarching focus on employment and the high costs of classroom instruction and specialized training, most states have opted for programs that prioritize LFA initiatives.18

Mixed strategies, which are a combination of LFA and HCD tactics, can vary state-by-state depending on the range and duration of activities provided, but are generally defined as programs in which “individuals are assigned to participate initially in either an education or training activity or in a job search activity, depending on an assessment of their needs.”19Participants can begin with an education or training-focused activity, if necessary, and then move onto employment strategies, or vice versa. They may spend the duration of their benefit on one approach or alternate between activities until they have obtained employment.

LFA, HCD, AND MIXED STRATEGIES

The most rigorous evidence of the reform’s impacts on participant earnings, employment, income, dependency, and government receipts and expenditures comes from the National Evaluation of Welfare-to-Work Strategies (NEWWS) in the 1990s. Directed by The Manpower Demonstration Research Corporation (MDRC), NEWWS was a five-year randomized control trial study that surveyed alternative approaches to help welfare recipients find jobs, advance in employment, and leave public assistance.20It included a sample of 55,000 adults and 11,000 children in seven program sites across the U.S (Atlanta, Grand Rapids, Riverside, Columbus, Detroit, Oklahoma City, and Portland, Oregon). In the experiment, control groups were not subject to any work requirements (similar to AFDC), while treatment groups received one of the following interventions: a mandatory LFA, HCD, or mixed strategy program.

In terms of participant economic well-being, the study found that nearly every program increased employment and earnings relative to what was achieved by control groups. Impacts for employment-focused programs (LFA) were particularly noteworthy. LFA programs produced effects almost immediately, whereas education-focused programs (HCD) took on average about a year after random assignment to reveal any increases. Five years after random assignment, average effects for education-focused programs were still smaller than employment-focused programs. This was true when comparing all unique LFA sites to HCD sites in the study, and for sites that had both LFA and HCD programs. Thus, contrary to what intuition might suggest about the long-term return on investment of education in this context, HCD programs did not close the gap in later years.21This was, and continues to be, a surprising and far-reaching implication from the evaluation.

Figure 1: MDRC experiment of impacts on total earnings in years 1 to 5 for sample members with and without a high school diploma or GED at random assignment. Note: Detroit, Oklahoma City, and Portland were mixed strategy programs.
Source: Hamilton, Gayle et. al. (2001)

Among LFA programs, earnings gains from year one to five ranged from $1,500 to $2,500 and employment gains ranged from 0.7 to 1.1 additional quarters of employment over the five-year span. For high-enforcement education-focused programs, employment gains ranged from 0.3 to 0.8 quarters and earnings gains ranged from $800 to about $2,000 over five years.22The differential between LFA and HCD programs was especially pronounced among high school nongraduates (Figure 1).23LFA was illustrated as an alternative that might better support individuals without any degree, a group that may face significant barriers to employment.

MDRC also examined changes in reliance on public assistance, which included reduction in both welfare receipts and benefit amounts. After five years, program group members across all approaches spent less time on welfare and received smaller welfare payments than control group members. In addition, treatment members in all program types received less food stamp assistance than control group members, indicating that recipients did not just shift to another subsidy under work requirements. Welfare savings tended to be larger and more persistent than any earnings gains across both LFA and HCD types, yet LFA approaches generally led to lower welfare use and expenditures than HCD approaches.24

The MDRC experiment also studied impacts on recipient income (including earnings, welfare payments, Earned Income Tax Credit, etc.) over time. Income is perhaps the most important indicator of economic well-being because programs that raise income also decrease the probability of future welfare (or other public assistance program) reliance, unemployment, and economic hardship. In fact, existing literature illuminates that programs that increase employment without increasing income do not translate into benefits for children.25In the end, the experiment found that even though earnings as a percentage of income increased, neither LFA or HCD yielded income gains above control group levels.26

However, analysis from 20 program studies from 1985 to 1999, including MDRC’s NEWWS experiment, illustrated that mixed strategy programs actually resulted in larger employment and earnings gains than pure LFA or HCD approaches. Perhaps most significantly, mixed strategy programs were the only approach proven to improve the government’s budgetary position and increase participant income.27Mixed programs that focused on employment also had welfare savings that were on par with job-search-first strategies. Lastly, for two of three mixed strategy programs in NEWWS, there were similar gains across both high school graduate and high school nongraduate subgroups (Figure 1).28In sum, these findings indicate this alternative’s potential on multiple fronts.

Benefit-Cost-Analysis

Nearly a decade after its NEWWS experiment, MDRC synthesized findings from multiple large-scale studies (including NEWWS) to comprehensively estimate the benefits and costs of various welfare-to-work program approaches. The estimates include earnings, government expenditures, tax revenues, EITC, food stamps, and more. The synthesis included programs from 11 states and two Canadian provinces and over 100,000 research sample members.29

From the government perspective, employment-focused approaches that required job search as the initial activity yielded the highest payoff for governments. The median program payoff per recipient was $2,266 (Figure 2). Education-focused programs cost significantly more than employment-focused programs that emphasized job search.30At the same time, education-focused program returns were slightly lower than those of mandatory work experience programs.31

From the perspective of the welfare recipient, the mixed strategy approach generated the greatest payoff. The benefits outweighed the costs by $1,422 per program group member. This was a significantly larger benefit than mandatory employment or education program benefits.

Note: Work experience and job search are both considered  LFA programs.
Source: Greenberg, David, Deitch, Victoria, and Hamilton, Gayle. (2009).
Source: Author’s own representation

Considering the benefit-costs from a social perspective, employment-focused programs (both mandatory work experience and job-first) and the mandatory mixed strategy presented benefits that outweigh the costs. Mandatory work experience programs had the highest social return on investment (ROI)33with a benefit-cost of $1,261 and mixed programs had the second highest social ROI with a return of $744. With a negative benefit-cost of -$2,510, it would appear that education-focused programs are not an economical way to incentivize recipients into work. If efficiency is a priority, policymakers should prioritize the other three alternatives.

Recommendations

All alternatives demonstrated some success compared to the previous status quo of AFDC (control groups were not subjected to any work requirements). More specifically, the three strategies increased employment rates and earnings and reduced dependence on welfare, with no evidence of recipients completely shifting to other forms of government subsidization.

However, certain work activities have proved to fare better than others. Figure 3 weighs the alternatives against three criteria (efficiency, effectiveness, and equity) to determine which solution best addresses the underlying problem. Evaluating the alternatives reveals that pure employment strategies perform well on an efficiency basis and mixed strategy approaches score highly based on their effectiveness and equitableness. HCD approaches, on the other hand, score low or medium-low on all three criteria counts. Considering ample literature on the high return on investment of education in a broader context, this was an unexpected finding.

This analysis and conclusions from multiple studies suggest that there is no evidence to support a rigid education or training strategy. However, they also do not justify a rigid employment-focused strategy while mixed strategy approaches show so much promise. Nonetheless, allowable work activities across states indicate that many states have prioritized stricter LFA approaches at the expense of more holistic LFA-HCD alternatives.34

Given the more significant reduction in receipts and greater effectiveness of LFA programs in comparison to education/training programs, and the fact that pure job-search programs produce the highest returns to government budgets, it is no surprise that states have allocated institutional resources in this manner.

The District of Columbia and 24 out of 50 of states limit the extent to which some education or training programs can count toward the majority of the requirement, regardless of recipient needs. They refer to these as “non-core activities.”35Other states require that recipients engage in employment-related activities for a set amount of time before they are permitted to enroll in education or training.

On the extreme end of this scheme, in states such as Texas, job search is the only allowable activity that may count toward required hours. In contrast, basic education, high school/GED, ESL, vocational training, and life skills training can all count toward the main requirement in the State of Washington.36

If mixed strategy approaches are successful because they allow flexibility for recipients to receive services that are more tailored to their needs, or because of some interaction effect between LFA and HCD approaches, then many states are failing to realize potential gains. States may be missing opportunities to maximize payoff for both recipients and government by not supplementing employment-focused programs with education or training plans.

Mixed strategy approaches vary on two dimensions:

Programs that vary by initial activity required based on an individual’s needs assessment (e.g. programs that require participants who do not have high school degrees to complete several months of basic or remedial education prior to partaking in any other activity).

Programs can also vary based on the combination or duration of employment or education/training services activities allowed (e.g. programs that require participants to engage in both LFA and HCD activities while on TANF).

Further analysis should be done to determine what is driving better outcomes: Is the interaction effect between both types of services or the needs assessment aspect most impactful? Are there other dimensions underlying mixed programs that make them particularly effective? Additionally, little research has been done to parcel out the potentially differential impacts of education versus training. Being able to narrow in on what within mixed strategies is driving more efficient, effective, and equitable outcomes will aid policymakers and administrators in implementation.

Moreover, in an increasingly globalized and de-industrialized America, perhaps the work activity approaches discussed in this article, which were largely devised in the years prior to and immediately following PRWORA’s passage, are no longer appropriate. A paltry amount of literature exists regarding new approaches and it is unclear whether current alternatives addressed in this article would best equip recipients to find stable and quality employment in a 21st century economy.

OBAMA ADMINISTRATION’s “WELFARE WAIVER”

In 2012, the Obama administration issued a memo permitting states to apply for waivers for federal work participation standards if they are able to propose “alternative and innovative strategies, policies, and procedures that are designed to improve employment outcomes for needy families.”37The initiative allowed states to test new welfare-to-work program approaches by engaging recipients in work activities not currently designated under TANF.38While the policy is no longer active, it would be beneficial for the federal government to explore this strategy again in the future.

SANCTIONS

Work requirements should not be analyzed in isolation. Instead, analysts should examine their effects in tandem with two related policies implemented during the 1996 welfare reform: financial sanctions for noncompliance with work requirements and time limits.39

TANF increased the frequency and severity with which states can place sanctions on individuals who fail to comply with welfare requirements. All states are now required to levy at least partial sanctions on recipients who fail to comply with mandatory work requirements, and most states actually levy sanctions above the minimum federal requirement.40While sanctions have not been evaluated to the extent that work activity options have been, data do illustrate that they are a state choice that can have deleterious impacts on a recipient’s livelihood.

Impacts

For sanctions to be effective and efficient, recipients must act consistently in making decisions that provide the greatest utility and have perfect and complete information. In the aggregate, it would appear that these assumptions have held and that sanctions have worked. Randomized control trials demonstrate that the most successful work programs have strong enforcement mechanisms such as sanctions,41and that low-enforcement programs which have little to no financial penalties rarely raise employment rates above control group levels.42A 2000 evaluation of state penalty approaches even revealed that cash and penalty incentives were associated with higher income. In fact, the analysis concluded that “it is the more lenient states with softer penalties where income seems to have grown the least.”43

Over time, however, disaggregated data has revealed that sanctions have had adverse consequences on mothers within the lowest end of the income distribution. Research has also found that mothers who were sanctioned off TANF rolls faced significant barriers to finding and maintaining employment. When compared to other mothers on welfare, they were found to be less likely to have a high school degree or job experience and more likely to have substance addictions, mental health problems, or three or more children.44,45Women who leave welfare because of sanctions were also found to be much less likely to have jobs after leaving the rolls.46

Furthermore, nearly 14 percent (260,000 out of 1.8 million cases) of families under TANF in 2010 were sanctioned because of noncompliance with work requirements and therefore lost some or all of their benefits. This is compared to only 1.3 percent (25,000 out of 1.8 million cases) of families who lost benefits because of time limits in the same year.47Researchers have found this to be an interesting imbalance given that at the time of PRWORA’s passage there was significantly greater media coverage on the major effects time limits would have. In any event, these results show the need for additional research on sanctions, and beg the question of what can be done to prevent any harmful unintended consequences.

Full Sanctions v. Partial Sanctions

Though broader reform at the federal level may be necessary, this article will continue to focus on state-level decisions. As it currently stands, states have some flexibility to determine the severity of sanctions they levy.  States have the option to impose either full sanctions or partial sanctions.

Over time states have increasingly adopted full-family sanctions, ending benefits entirely for families who do not comply with work requirements. Today, most states levy full sanctions.48In fact, as of 2015, 46 states ended benefits for families that failed to comply. In most of these cases, families must reapply for benefits, and in some cases, repeated noncompliance can lead to a lifetime ban on benefits. California, New York, Vermont, Missouri, and D.C. are the only jurisdictions to place partial rather than full sanctions on participants.49

This trend makes it challenging to assess the true impacts of full versus partial sanctions. This is especially the case because 39 percent of the nation’s caseloads are from California and New York alone.50As a result there have been few comparable evaluations, but the arguments for one approach over the other are as follows:

Merits of Full Sanctions:51

Because of the severity of the penalty, full-family sanctions may be more likely to generate changes in recipient behavior.

When partial sanctions are not as effective of an incentive, recipients may end up exhausting months of eligibility prior to the time limit without obtaining needed employment, education, or training services.

Most states have implemented extensive reconciliation or review processes to minimize risk of error, and full sanctions better help to recoup these administrative costs.

Merits of Partial Sanctions:52

Since some of the most disadvantaged groups in welfare disproportionately experience sanctions, partial sanctions can decrease the likelihood that families with the most serious employment barriers will simply exit from welfare without receiving needed services and support.

Reduced sanctions may be just as effective in generating changes in recipients’ behavior.

There has not been a cost-benefit analysis to determine to what extent full sanctions actually recoup transaction costs and other administrative costs above and beyond partial sanctions.

Since PRWORA went into effect, there has been one major study on sanctions that provides greater insight into the arguments above. Conducted by the University of Maryland Department of Family Studies in 2001, the evaluation ultimately found that there was no association between stricter sanction policies and work-related welfare exits.53This indicates that both approaches are effective at encouraging recipients to meet requirements, but that these approaches differ in terms of their efficiency and equitableness as a means to achieving ends.

Recommendations

Figure 4 summarizes the tradeoffs between these two approaches. In conclusion, while full sanctions and partial sanctions are both effective, full sanctions have limitations from an equity perspective and partial sanctions have limitations from an efficiency perspective.

Given conclusive evidence pointing to the fact that certain groups have been disproportionately impacted by sanctions, this article recommends states consider less harsh penalties.

TIME LIMITS

Time limits are important to analyze in conjunction with work activities and sanctions. Proponents argue that time limits promote self-sufficiency by providing a signal that benefits are meant to be temporary in nature.54To review, PRWORA set the following time limit restrictions on federal funds (not state funds):

With few exceptions, recipients must work after two years on assistance. PRWORA also placed a lifetime limit of five years (or less at state option) on benefits.

As there has been minimal quantitative analysis on time limit alternatives available to states, this analysis will touch the surface and propose future research questions. While there has been little variation in consecutive time limits, there has been significant state variation in lifetime limits on benefits. Many have selected time limits more restrictive than the national requirement: 12 states have lifetime limits between 13 and 24 months, 3 states have limits between 25 and 36 months, 5 states have limits between 37 and 48 months, and 30 states have limits between 49 and 60 months.55Arizona is the only state to have a limit of 12 months, the strictest limit in the nation. It is too soon to tell, however, the impacts of this time limit as Arizona slashed its lifetime limit from 24 to 12 months in 2015.56

Optimal time limits remain unknown. Nevertheless, research does show that even in highly successful programs, substantial shares of the caseload remain unemployed after time limits pass. A multi-state analysis 10 years after PRWORA indicated that 30 percent of mothers who leave welfare have not worked at all since leaving welfare.57Moreover, in MDRC’s control trials, it was estimated that less than one-third of program group members worked in all four quarters of the second year of follow-up.58

These findings raise questions with respect to whether existing time limits are sufficient for recipients to achieve stable employment and for families to obtain the necessary support needed to robustly transition out of basic assistance initiatives.

Recommendations

In summary, this article makes the following recommendations regarding work activity options, sanctions, and time limits:

There is strong evidence that work requirements and sanctions as an accompanying incentive mechanism are better than the prior alternative (AFDC) in reaching important outcomes. Thus, the welfare system should continue to utilize work programs as one policy within the broader system.

If it is a government priority to alleviate the likelihood of future economic hardship, mixed strategy approaches were shown to be the only pathway which could lead to gains in recipient income. More states should test mixed strategies as an alternative to better maximize the payoff to both government and recipients.

States should seek to understand why education has failed in previous contexts, and what specific education or training activities as part of mixed strategies are most impactful.

Much of the evidence and literature on work activities is outdated. States should explore more innovative options to better equip future workers to find employment in an increasingly deindustrialized and globalized world.

Partial sanctions are a more equitable means of achieving compliance than full sanctions. Since these approaches are both effective and relatively efficient, states should opt to pursue partial sanction approaches when politically feasible.

There is little existing literature regarding the various approaches states have taken under federal time limits, but there is substantial state variation and there may be natural experiments worth analyzing.

Lastly, work requirements do not operate in a vacuum. The most successful work program in MDRC’s experiment was that of Portland, Oregon. It was successful not only as a result of its mixed strategy approach (as indicated in Figure 1), but also because its activities were very much in tune with local labor market needs. The City at the time also had a healthy economy and significant labor demand, indicating potentially serious underlying problems with programs that aim to incentivize work in the midst of poor local economic conditions.

CONCLUSION

Mandatory work requirements and accompanying policies set under The PRWORA of 1996 have met many goals set under welfare reform. On the whole, the policies have increased employment rates and earnings among targeted populations, and improved self-sufficiency among many recipients. The reforms have also helped to alleviate major budgetary concerns facing both state and federal governments and have enhanced efficiency within the welfare system. For instance, state and federal expenditures on TANF/AFDC decreased from $41 billion in 1993 to $32 billion in 2015, adjusting for inflation.59,60

Based on states’ current decisions around approaches to take under TANF, however, there is considerable room for improvement. Feasibility will undoubtedly depend on the state’s political will and some states may be more likely than others to pick up recommendations in this analysis. In any event, greater attention needs to be paid to some of the equity consequences of PRWORA, such as effects on income and disparities across subgroups.

States are constrained in many ways, and there are bounds to what states will likely do under current incentive structures. A major issue continues to be funding. The TANF block grant has been frozen since 1996 and has not adjusted for inflation.61The system as it stands also leads to many problems of accountability, with many states controversially spending only half of their TANF dollars on core welfare reform areas.62Additionally, there are insufficient incentive structures in place at the state-level to innovate and improve recipient experience. Perhaps a different and better balance needs to be struck between state accountability and flexibility.

Figure 5: TANF’s role as a safety net continues to decline
Source: Center on Budget and Policy Priorities

Over time, TANF has also increasingly served fewer families who qualify for benefits (Figure 5), raising criticisms regarding its effectiveness and resolve as a program that actually helps families who need basic assistance.63As a result of these concerns, many advocates, administrators, and politicians alike say that TANF is long overdue for comprehensive federal reform.

This article leaves many future questions for discussion. Not only should the ideas in this article be further reviewed, but there are also many additional policies as part of TANF and other social safety net programs — including earned income disregards, eligibility, and exemptions —  that should be thoroughly evaluated to understand how America’s social safety net system as a whole can better serve the nation’s low-income individuals and families.

Appendices
Appendix A: MDRC SSynthesis of Welfare-to Work Program Benefits and Costs,Overall Performance of Each Program Type64

Appendix B: MDRC Synthesis of Welfare-to-Work Program Benefits and Costs, ROI per Program Group Member, by Type and Perspective65 
Appendix C: TANF’s Role in Serving Poor Families by State and Over Time66
TANF-to-Poverty Ratio (number of families receiving TANF benefits for every 100 families with children in poverty in 2015)

 

ENDNOTES
1. Stapleton, David; Livermore, Gina; and Tucker, Adam (1997). “Determinants of AFDC Caseload Growth.” The Lewin Group, Inc: for the Office of the Assistant Secretary for Planning and Evaluation, Department of Health and Human Services.
2. Greenberg, Mark, and Jim Baumohl. (1996) “Income Maintenance: Little Help Now, Less on the Way,” in Homelessness in America, Oryx Press. National Coalition for the Homeless, 2201 P St. NW, Washington, DC 20037; 202/462-4822. 
3. Zedlewsk, Sheila R. and Giannarelli, Linda. (1997) “Diversity among State Welfare Programs
4. “From Welfare to Workfare.” The Economist. July, 27, 2006. https://www.economist.com/node/7226120 
5 .Falk, Gene; Maggie McCarty; and Randy Alison Aussenberg (2014). “Work Requirements, Time Limits, and Work Incentives in TANF, SNAP, and Housing Assistance.” Congressional Research Service. 
6. U.S. Department of Health and Human Services, Office of Family Assistance Resource Library. Temporary Assistance for Needy Families: Changes in Combined Work Participation Rate From FY 2015 to FY 2016. https://www.acf.hhs.gov/ofa/resource/work-participation-rates-fiscal-year-2016. 
7. Institute for Policy Research (2008). “A Look Back at Welfare Reform” 30 (1). Northwestern University. 
8. “The Outcomes of 1996 Welfare Reform”. Brookings Institution. July 19, 2006. Retrieved December 31, 2016. https://www.brookings.edu/testimonies/the-outcomes-of-1996-welfare-reform/.
9. Moffit, Robert A. (2002) “From Welfare to Work: What the Evidence Shows.” Brookings Institution. https://www.brookings.edu/research/from-welfare-to-work-what-the-evidence-shows/. 
10. “The Outcomes of 1996 Welfare Reform”. Brookings Institution. July 19, 2006. Retrieved December 31, 2016. https://www.brookings.edu/testimonies/the-outcomes-of-1996-welfare-reform/.
11. Ibid.
12. Moffit, Robert A. (2002) “From Welfare to Work: What the Evidence Shows.” Brookings Institution. https://www.brookings.edu/research/from-welfare-to-work-what-the-evidence-shows/. 
13. “The Outcomes of 1996 Welfare Reform”. Brookings Institution. July 19, 2006. Retrieved December 31, 2016. https://www.brookings.edu/testimonies/the-outcomes-of-1996-welfare-reform/.
14. “Chart Book: Temporary Assistance for Needy Families,” Center on Budget and Policy Priorities. August 16, 2017. https://www.cbpp.org/research/family-income-support/chart-book-temporary-assistance-for-needy-families. 
15. Falk, Gene; Maggie McCarty; and Randy Alison Aussenberg (2014). “Work Requirements, Time Limits, and Work Incentives in TANF, SNAP, and Housing Assistance.” Congressional Research Service.
16. Ibid. 
17. Hamilton, Gayle and Gueron, Judith M. (2012) “The Role of Education and Training in Welfare Reform”. Brookings Institution. 
18. Ibid.
19. Greenberg, David, Deitch, Victoria, and Hamilton, Gayle. (2009) “Welfare to-Work Program Benefits and Costs A Synthesis Of Research.” MDRC.
20. Hamilton, Gayle et. al. (2001) “How Effective Are Different Welfare-to-Work Approaches?  Five-Year Adult and Child Impacts for Eleven Programs.” MDRC. https://aspe.hhs.gov/report/how-effective-are-different-welfare-work-approaches-five-year-adult-and-child-impacts-eleven-programs.
21. Ibid. 
22. Low enforcement education programs, that did not institute any sanctions on participants, did not yield any significant earnings or employment gains.
23. Ibid.
24. Ibid.
25. Duncan, G. J., Gennetian, L., & Morris, P. (2007). Effects of welfare and antipoverty programs on participants’ children. Focus, 25(2), 3–12. 
26. Hamilton, Gayle et. al. (2001) “How Effective Are Different Welfare-to-Work Approaches?  Five-Year Adult and Child Impacts for Eleven Programs.” MDRC. https://aspe.hhs.gov/report/how-effective-are-different-welfare-work-approaches-five-year-adult-and-child-impacts-eleven-programs.
27. Greenberg, David, Deitch, Victoria, and Hamilton, Gayle. (2009) “Welfare to-Work Program Benefits and Costs A Synthesis Of Research.” MDRC. 
28. Hamilton, Gayle et. al. (2001) “How Effective Are Different Welfare-to-Work Approaches?  Five-Year Adult and Child Impacts for Eleven Programs.” MDRC. https://aspe.hhs.gov/report/how-effective-are-different-welfare-work-approaches-five-year-adult-and-child-impacts-eleven-programs.
29. Greenberg, David, Deitch, Victoria, and Hamilton, Gayle. (2009) “Welfare to-Work Program Benefits and Costs A Synthesis Of Research.” MDRC. 
30. Hamilton, Gayle. (2002) “Moving People from Welfare to Work. Lessons from the National Evaluation of Welfare-to-Work Strategies.” MDRC. https://aspe.hhs.gov/report/moving-people-welfare-work-lessons-national-evaluation-welfare-work-strategies. 
31. Greenberg, David, Deitch, Victoria, and Hamilton, Gayle. (2009) “Welfare to-Work Program Benefits and Costs A Synthesis Of Research.” MDRC. 
32. To prevent estimates being skewed by outliers on both low- and high-ends of the spectrum, this analysis references the median net benefit-cost by program group member.
33. Social return on investment (ROI) is a metric used to evaluate the economic efficiency/profitability (or bang per buck) of a policy alternative. In other words, it is the gains on an investment. 
34. Hamilton, Gayle and Gueron, Judith M. (2012) “The Role of Education and Training in Welfare Reform”. Brookings Institution.
35. The following states place a limit on the extent to which education or training can count toward the core work requirement: Alabama, Arkansas, California, Washington D.C., Florida, Georgia, Indiana, Kansas, Kentucky, Louisiana, Maine, Maryland, Montana, Nebraska, Nevada, New Hampshire, New Jersey, New Mexico, Ohio, Rhose Island, South Dakota, Tennessee, Virginia, West Virginia, Wyoming.
36. Urban Institute, Welfare Rules Database. 2015 Activities Requirements. http://wrd.urban.org/wrd/Query/query.cfm. 
37. Falk, Gene; Maggie McCarty; and Randy Alison Aussenberg (2014). “Work Requirements, Time Limits, and Work Incentives in TANF, SNAP, and Housing Assistance.” Congressional Research Service.
38. Ibid.
39. There are several exemptions to sanctions. Parents who receive SSI or Social Security disability may be exempted and disregarded in measuring state performance. A caretaker of a disabled family member may be exempt. Single parents caring for an infant (under age 1) may be exempted and disregarded in measuring state performance. In addition, parents and caretakers with a child under the age of 6 cannot be sanctioned if child care is unavailable.
40. Ibid.
41. Hamilton, Gayle and Gueron, Judith M. “The Role of Education and Training in Welfare Reform”. Brookings Institution. April 1, 2002. https://www.brookings.edu/research/the-role-of-education-and-training-in-welfare-reform/.
42. Hamilton, Gayle et. al. (2001) “How Effective Are Different Welfare-to-Work Approaches?  Five-Year Adult and Child Impacts for Eleven Programs.” MDRC. https://aspe.hhs.gov/report/how-effective-are-different-welfare-work-approaches-five-year-adult-and-child-impacts-eleven-programs.
43. Schoeni, Robert F. and Rebecca M. Blank. (2000) “What Has Welfare Reform Accomplished? Impacts on Welfare Participation, Employment, Income, Poverty, and Family Structure.” National Bureau of Economic Research Working Paper 7627. Cambridge, MA: NBER.
44. Andrew J. Cherlin and others. (2002) “Operating within the Rules: Welfare Recipients’ Experiences with Sanctions and Case Closings,” Social Services Review 76: 387-405. 
45. “The Outcomes of 1996 Welfare Reform”. Brookings Institution. July 19, 2006. Retrieved December 31, 2016. https://www.brookings.edu/testimonies/the-outcomes-of-1996-welfare-reform/. 
46. Moffit, Robert A. (2002) “From Welfare to Work: What the Evidence Shows.” Brookings Institution. https://www.brookings.edu/research/from-welfare-to-work-what-the-evidence-shows/.
47. Falk, Gene; Maggie McCarty; and Randy Alison Aussenberg (2014). “Work Requirements, Time Limits, and Work Incentives in TANF, SNAP, and Housing Assistance.” Congressional Research Service.
48. Ibid.
49. Urban Institute, Welfare Rules Database. 2015 Financial Penalty Requirements. http://wrd.urban.org/wrd/Query/query.cfm. 
50. Falk, Gene; Maggie McCarty; and Randy Alison Aussenberg (2014). “Work Requirements, Time Limits, and Work Incentives in TANF, SNAP, and Housing Assistance.” Congressional Research Service.
51. Bloom, Dan and Winstead, Don. (2002) “Sanctions and Welfare Reform.” Brookings Institution, https://www.brookings.edu/research/sanctions-and-welfare-reform/. 
52. Ibid.
53. Hofferth, Sandra L., et. al. (2001) “Exiting Welfare in the 1990s: Did Public Policy Influence Recipients’ Behavior?” College Park: Department of Family Studies, University of Maryland.
54. Freedman, Stephen, et. al. (2000) “Evaluating Alternative Welfare-to-Work Approaches:  Two-Year Impacts for Eleven Programs.” https://www.mdrc.org/sites/default/files/full_93.pdf. 
55. Smith, Nathan. (2016). “State Welfare Reforms: TANF Time Limits by the Numbers.” The Council of State Governments. http://knowledgecenter.csg.org/kc/content/state-welfare-reforms-tanf-time-limits-numbers. 
56. Ibid.
57. “The Outcomes of 1996 Welfare Reform”. Brookings Institution. July 19, 2006. Retrieved December 31, 2016. https://www.brookings.edu/testimonies/the-outcomes-of-1996-welfare-reform/.
58. Freedman, Stephen, et. al. (2000) “Evaluating Alternative Welfare-to-Work Approaches:  Two-Year Impacts for Eleven Programs.” https://www.mdrc.org/sites/default/files/full_93.pdf.
59. Department of Health and Human Services, Office of Family Assistance: Office of the Administration for Children and Families, State TANF Spending in FY 2015. August 15, 2016. https://www.acf.hhs.gov/ofa/resource/state-tanf-spending-2015-factsheet. 
60. Page, Stephen B., and Larner, Mary B. “Introduction to the AFDC Program.” The Future of Children: WELFARE TO WORK Vol. 7: 1. https://www.princeton.edu/futureofchildren/publications/docs/07_01_01.pdf 
61. Floyd, Ife; Pavetti, LaDonna; and Schott, Liz. (2017) “TANF Reaching Few Poor Families.” Center on Budget and Policy Priorities. https://www.cbpp.org/sites/default/files/atoms/files/6-16-15tanf.pdf. 
62. Schott, Liz, and Floyd, Ife. (2017) “How States Use Funds Under the TANF Block Grant.” Center on Budget and Policy Priorities. https://www.cbpp.org/research/family-income-support/how-states-use-funds-under-the-tanf-block-grant. 
63. Floyd, Ife; Pavetti, LaDonna; and Schott, Liz. (2017) “TANF Reaching Few Poor Families.” Center on Budget and Policy Priorities. https://www.cbpp.org/sites/default/files/atoms/files/6-16-15tanf.pdf.
64. Greenberg, David, Deitch, Victoria, and Hamilton, Gayle. (2009) “Welfare to-Work Program Benefits and Costs A Synthesis Of Research.” MDRC.
65. Ibid.
66. Floyd, Ife; Pavetti, LaDonna; and Schott, Liz. (2017) “TANF Reaching Few Poor Families.” Center on Budget and Policy Priorities. https://www.cbpp.org/sites/default/files/atoms/files/6-16-15tanf.pdf.