Work Requirements and Work Supports for Recipients of Means-Tested Benefits

At a Glance

In this report, the Congressional Budget Office analyzes the effects of work requirements and work supports on employment and income of participants in Temporary Assistance for Needy Families (TANF), the Supplemental Nutrition Assistance Program (SNAP), and Medicaid. The agency also assesses how changing work requirements and work supports in those programs would affect the federal budget. In many cases, the size of those effects is highly uncertain.

Effects of Work Requirements on Employment, Income, and the Federal Budget

  • Making the receipt of benefits contingent on working or preparing to work has substantially increased the employment rate of the targeted recipients in TANF during the year after they enter the program and by a smaller amount in later years. Work requirements in SNAP have increased employment less; in Medicaid, they appear to have had little effect on employment.
  • Although some people have higher income because they work more to meet the programs’ requirements, other people do not meet the work requirements and are left with little income from in-kind benefits, cash payments, earnings, or other sources. Overall, the increase in total earnings from TANF’s work requirements is about equal to the reduction in benefits. In contrast, work requirements in SNAP and Medicaid have reduced benefits more than they have increased people’s earnings.
  • In general, tightening work requirements would reduce federal spending by decreasing the amount of benefits provided; the extent of the budgetary savings would depend on the details of the policy. If lawmakers used the savings from tightening work requirements to increase work supports that helped recipients meet those requirements, the federal budget would change little (or perhaps not at all).

Effects of Work Supports on Employment, Income, and the Federal Budget

  • Subsidized child care, job-search assistance, and subsidized employment have increased the employment of recipients, whereas job training has had mixed results.
  • In addition to boosting recipients’ earnings, federal funding for work supports has freed up income that recipients would have spent on those supports to instead be spent on other goods and services.
  • If policymakers chose to expand work supports, they would need to provide additional funding. Child care subsidies can cost several thousand dollars per recipient, whereas less intensive services (such as assisting people who are searching for a job by providing access to literature and online tools) generally cost less.

Notes and Definitions

To complete and publish this report promptly, the Congressional Budget Office used its July 2021 baseline for the analysis. CBO expects that the findings would be similar had the analysis used the agency’s most recent baseline (published in May 2022).

Unless this report indicates otherwise, all years referred to are federal fiscal years, which run from October 1 to September 30 and are designated by the calendar year in which they end.

Numbers in the text, tables, and figures may not add up to totals because of rounding.

Projected future costs and the other dollar figures in the options are in nominal dollars. All other dollar figures are expressed in 2019 dollars, using the price index for personal consumption expenditures from the Bureau of Economic Analysis to remove the effects of inflation.

Some of the figures in this report use shaded vertical bars to indicate periods of recession. (A recession extends from the peak of a business cycle to its trough.)

References to states include the District of Columbia.

A TANF recipient is a person who receives recurring cash payments through the Temporary Assistance for Needy Families program. Data on the recipients of other forms of assistance provided by TANF, such as job training, are limited.

Intensive job-search assistance is generally provided in person, such as through workshops or one-on-one career counseling.

An able-bodied adult is a person over the age of 17 who does not receive disability benefits (either through Supplemental Security Income or Disability Insurance).

A parent is a person who lives with one or more dependents under the age of 18.

A child is a dependent who is under the age of 18.

Cash income generally consists of earnings, business income, income from savings, child support, and cash payments from means-tested programs, Social Security, and unemployment insurance. That measure of economic well-being is similar to what the Census Bureau uses to determine the official poverty rate.

Income includes cash income, in-kind benefits from the Supplemental Nutrition Assistance Program, and refundable tax credits. That measure of economic well-being is similar to what the Census Bureau uses to determine the Supplemental Poverty Measure.

Poverty thresholds are used by the Census Bureau to determine the official poverty rate. The thresholds are based on family income and differ for families of different sizes.

Deep poverty is when a family’s cash income is less than half of the applicable poverty threshold.

Welfare reform is the term for a series of actions that policymakers took in the mid-1990s to encourage employment among benefit recipients and shorten the duration of benefit receipt. It consisted of executive actions that introduced work requirements in Aid to Families With Dependent Children (the program that preceded TANF) and implementation of provisions of the Personal Responsibility and Work Opportunity Reconciliation Act.

Summary

The federal government has many programs that are designed specifically to help people who have relatively low income obtain food, health care, housing, and other goods and services that they might not otherwise be able to afford. Those means-tested programs provide cash payments or other assistance to qualified recipients. If recipients’ earnings rise, then their benefits typically decline. To counter that incentive for participants to work less, the programs often incorporate work requirements (which make the receipt of benefits contingent on working or preparing to work) and work supports (which make working more feasible and profitable for participants). Work support programs usually also have work requirements. (For example, only people who work are eligible for the earned income tax credit.)

This report focuses on work requirements and work supports in three federal programs: Temporary Assistance for Needy Families (TANF), the Supplemental Nutrition Assistance Program (SNAP), and Medicaid. Of those programs, TANF is the smallest, providing monthly cash payments to about 1 million families each year. By law, most able-bodied parents who receive those benefits must participate in work-related activities. SNAP and Medicaid serve much broader and larger populations than TANF; they also have included work requirements at times, but typically applied them only to able-bodied adults without dependents. To support recipients who are working or searching for work, policymakers have made subsidized child care and workforce development services available to some participants in those programs, particularly TANF.

In this report, the Congressional Budget Office analyzes how work requirements and work supports affect the employment and income of former, current, and potential participants. In addition, the agency estimates how changes to those requirements and supports would affect the federal budget over the next nine years. Even though research suggests that means-tested programs can continue to affect former participants for many years, this report focuses on the programs’ more immediate effects.

How Do Work Requirements Affect People’s Employment and Income?

TANF’s work requirements have generally increased employment while having little effect (on net) on average income. Some recipients have earned more by getting a job, but others have lost benefits without finding work, which probably increased the number of people in deep poverty. Work requirements in SNAP and Medicaid have also reduced the benefits that people receive but have increased their employment or earnings less (for SNAP recipients) or maybe not at all (for Medicaid recipients). TANF recipients facing work requirements have been provided with strong work supports, unlike SNAP and Medicaid recipients.

TANF

Most of the research on work requirements focuses on single mothers who received recurring cash payments through TANF or its predecessor, Aid to Families With Dependent Children. The imposition of work requirements in the 1990s boosted the employment of those single mothers but had little effect on their average income, mainly because the increase in earnings for those who worked was about equal to the reduction in cash payments for those who lost benefits. The number of people receiving cash payments has continued to decline over the past two decades; by 2019, the number of recipients was about 2 million, or one-seventh of what it had been in 1993.

Although TANF’s work requirements have probably had little effect on average income among single mothers, those requirements have probably changed how income is distributed among that group. The mothers who gained employment often saw their income boosted by higher earnings and receipt of additional tax credits, but many mothers who lost benefits because they did not meet the work requirements were left in deep poverty. TANF continues to be the primary source of recurring cash assistance for able-bodied single mothers without education beyond high school, but the percentage of those mothers who are receiving assistance has fallen along with the total number of recipients.

Before work requirements were imposed, nearly all less-educated single mothers who did not work for an extended period received cash payments. By calendar year 2019, though, about one-seventh of less-educated single mothers had no earnings and no cash assistance, CBO estimates. Many of those families report having almost no income beyond benefits from SNAP. By removing families from TANF before they found work—and by deterring families from entering the program—work requirements have probably played a role in increasing the number of families in deep poverty.

SNAP

SNAP’s work requirement has probably boosted employment for some adult recipients without dependents but has reduced income, on average, across all recipients. Earnings increased among recipients who worked more, but far more adults stopped receiving SNAP benefits because of the work requirement. Most of the adults who had their SNAP benefits terminated for failing to comply with the work requirement have very low income because few of them have earnings or receive cash payments.

Medicaid

Evidence of the effect of work requirements on Medicaid recipients is limited to Arkansas, the only state where a work requirement was imposed on recipients for more than a few months. There, many of the targeted adults lost their health insurance as a result of the work requirement. Employment did not appear to increase, although the evidence is scant. Research indicates that many participants were unaware of the work requirement or found it too onerous to demonstrate compliance.

How Do Work Supports Affect People’s Employment and Income?

Most work supports increase employment and income. Subsidized child care, for instance, benefits single parents (its main recipients) by boosting their employment and increasing their resources substantially. Parents who find employment benefit from higher earnings, and those who would have purchased child care even without a subsidy have more income to spend on other goods and services. (The resources of some working parents are unaffected because they would have had a relative or friend watch their children in the absence of the subsidy.)

Job-search assistance and subsidized employment provided by workforce development programs have increased employment and income, but job training provided by those programs has had mixed results. According to a study conducted by the Department of Labor in the early 2010s on two of the largest workforce development programs, intensive job-search assistance (such as career counseling) increased participants’ average earnings by about $2,200 in the year following the receipt of assistance. In contrast, job training could reduce employment and income by causing participants to delay their job search or by reducing the number of hours they work. In the Department of Labor’s study, participants in job-training programs worked less while in training and did not work more or earn more afterward, on average.

Lawmakers modified the large workforce development programs in 2014 to better align job training with the demands of local employers. A comparable evaluation of those programs’ effectiveness since then might show increased employment and income from job-training programs if the alignment between those programs and labor market demand has improved lately. Recent research demonstrates that job training provided by smaller programs can increase employment and income when it focuses on the demands of local employers.

How Might Policymakers Change Work Requirements and Work Supports?

To increase employment, raise income, or reduce federal spending, the Congress could pursue various options that would change work requirements and work supports. Those options fall into four broad categories. Options in a category would typically accomplish similar objectives, although the effects of any particular option on employment, income, and the federal budget would depend on its details (see Table S-1).

Table S-1.

Typical Effects of Changing Work Requirements and Work Supports, by Broad Category of Options

Source: Congressional Budget Office.

Work supports consist of subsidized child care and workforce development services, such as job-search assistance, job training, and subsidized employment.

The effects of a particular option would depend on its details.

  • Options that expanded work requirements would typically increase employment. But they would have little effect on recipients’ average income (net of medical expenses, in the case of Medicaid) or would reduce it because the overall loss in benefits would equal or exceed the total gain in earnings from increases in employment. Some participants’ earnings would increase more than their benefits decreased; other participants would be left without any earnings or benefits. If the work requirements were tied to receipt of SNAP or Medicaid, the loss in benefits would reduce federal spending because the government pays for a fixed portion of the cost of those benefits. In this analysis, CBO assessed expanding work requirements in SNAP and Medicaid. The SNAP option that CBO discusses in this report might not reduce federal spending because it includes additional spending on work supports. The other option would substantially increase the amount that people who lost Medicaid coverage pay out of pocket for medical services.
  • Options that reduced work requirements would typically increase federal spending on benefits and reduce employment. They would decrease the number of people with very low income and would either raise average income or change it little. Reducing work requirements during periods of high unemployment could boost recipients’ income substantially while having little effect on employment. In this analysis, CBO assesses reducing work requirements in SNAP and TANF. The TANF option would not increase federal spending because that program’s funding is set at a fixed total.
  • Options that expanded work supports would typically lead to larger gains in income than options that expanded work requirements, but they would also push up federal spending. In addition, they would increase employment and reduce recipients’ expenses for work supports, leaving them with more income to spend on other goods and services. In this analysis, CBO assesses one option that would provide more workforce development services to TANF recipients when jobs are scarce and a second option that would generally increase funding for subsidized child care.
  • Options that reduced work supports would typically lead to less employment, lower income, and reduced federal spending. In this analysis, CBO assesses an option that—rather than reducing federal spending—would shift funding from work supports to cash payments for nonworking families. Such a policy would reduce employment but could raise recipients’ income and make deep poverty less prevalent.

Chapter 1An Overview of Work Requirements, Work Supports, and Means-Tested Programs

The federal government has many programs that provide benefits to help people who have relatively low income. Such means-tested programs—some of which are administered jointly with the states—help the people who receive those benefits obtain goods and services that they might not otherwise be able to afford. Three of those programs are Temporary Assistance for Needy Families (TANF), which funds monthly cash payments along with many other services; the Supplemental Nutrition Assistance Program (SNAP), which assists people in purchasing food; and Medicaid, which provides health insurance.

Although the people served by those programs differ, they generally have cash income near or below the poverty line (see Table 1-1). Most families in TANF, for instance, are headed by single mothers who have no education beyond high school; despite the cash payments they receive from the program, they have very low cash income—in many cases, it is so low that they are considered to be in deep poverty. Income for SNAP and Medicaid recipients is still low but not as low, and a greater share of adults in those programs are married or childless. (In addition, many of those adults are elderly or disabled.)

Table 1-1.

Characteristics of Selected Means-Tested Programs, 2019

Source: Congressional Budget Office, using data from the Urban Institute and the Department of Health and Human Services. See www.cbo.gov/publication/57702#data.

In 2019, the federal poverty threshold for a family of three (a single parent and two children) was $1,716 per month.

SNAP = Supplemental Nutrition Assistance Program; TANF = Temporary Assistance for Needy Families.

a. These amounts are gross income limits for initial eligibility, which typically disregard some earnings and other sources of income. Applicants must also meet a net income standard, which further restricts the eligibility of people who do not have much income that qualifies for exclusion.

b. CBO used benefits in Washington to represent TANF benefits because that state has an income threshold, average benefit, and relationship between benefits and earnings that are close to the average of all states.

c. Medicaid benefits are measured in terms of the payments that federal and state governments make for the medical care of recipients. In most states, Medicaid benefits do not change as earnings increase until families reach the eligibility threshold. At that point, families lose eligibility for Medicaid but can get heavily subsidized insurance through the marketplaces established under the Affordable Care Act.

In 2019, TANF provided cash payments of about $500 a month, on average, to participating families consisting of an able-bodied parent and two children—an amount equal to roughly 30 percent of the poverty threshold. SNAP provided a monthly food allowance of $372, on average.1 Average government spending on health care for such a family in Medicaid was higher—at roughly $800 per month—but was concentrated among recipients with severe medical conditions.

By design, benefit amounts in TANF and SNAP decline rapidly as participants’ earnings increase. That structure can create an incentive for participants to work less, although additional tax credits can offset a substantial portion of that decline.2 The decrease in benefits tends to be largest for families in TANF, who receive about 50 cents less in cash payments for each additional dollar they earn.

Work Requirements and Work Supports in Recent Years

To encourage participants in TANF, SNAP, and Medicaid to work, policymakers have added requirements to those programs. To receive benefits, participants must demonstrate that they are working or preparing to work. Most adults in TANF are required to work or participate in related activities, such as searching or training for a job. A far smaller portion of adults in SNAP are subject to such a requirement: Many SNAP recipients are elderly or disabled and thus not expected to work, and the requirement is not imposed on households that include dependent children. Medicaid does not have work requirements, although Arkansas imposed a work requirement on some childless adults who receive Medicaid benefits in 2018. (Several other states began implementing work requirements for Medicaid recipients, but Arkansas was the only state that terminated Medicaid benefits because of insufficient employment.)3

To make work more feasible and profitable, the federal government provides funds (often directly to the states) for work supports, such as subsidized child care and workforce development services. Those services are primarily funded by the Child Care Development Fund (CCDF), Title 1 of the Workforce Innovation and Opportunity Act (WIOA), and TANF. (All those funding sources focus assistance on people with low income, although not exclusively on TANF, SNAP, and Medicaid recipients.)

In 2019, the CCDF funded $8 billion of subsidized child care; by providing state-run care for dependent children, those programs enabled the parents to work, train, or search for a job.4 That same year, Title 1 of WIOA funded $5 billion of workforce development services, including job-search assistance, job training, and subsidized employment. TANF provided $3 billion for subsidized child care and $3 billion for workforce development services in 2019. Other programs that provide work supports to families in TANF, SNAP, and Medicaid and that also affect people’s incentives to work (such as Head Start, the child tax credit, and the earned income tax credit, or EITC) are beyond the scope of this report.5 (CBO used 2019 as a proxy for what work requirements and supports will look like after the public health emergency declaration related to the coronavirus pandemic is lifted.)

Changes to Work Requirements and Work Supports During the 1990s

The most significant changes to employment of participants in means-tested programs occurred in the 1990s, when work requirements and work supports were substantially expanded for many participants. The changes were implemented through a series of executive actions followed by enactment of the Personal Responsibility and Work Opportunity Reconciliation Act (PRWORA) in 1996. Those legislative changes are collectively referred to as welfare reform.

Before those changes were implemented, the government provided cash payments to qualifying recipients through Aid to Families With Dependent Children (AFDC). Into the early 1990s, that program had only minimal work requirements. From 1993 through 1996, most states imposed their own work requirements on AFDC participants by using waivers provided by the Bush and Clinton Administrations. In 1997, as a result of PRWORA, TANF replaced AFDC. For a family to receive cash payments from TANF, all able-bodied parents in it generally must participate in work-related activities. PRWORA also imposed a work requirement on able-bodied adults without dependents in SNAP.

Welfare reform substantially increased the amount of work supports available to participants in means-tested programs. PRWORA consolidated four child care assistance programs into the CCDF and added $1 billion in funding. (All figures are in 2019 dollars.) It also allowed states to reallocate TANF funding from cash payments to work supports and other services. The states have used that discretion to drastically change the types of assistance TANF provides. In 1996, for example, about 84 percent (or $31 billion) of the federal and state funding for the programs that preceded TANF was spent on recurring cash payments, whereas about 5 percent (or $2 billion) was spent on work supports (see Figure 1-1). By 2019, states had allocated about 51 percent (or $14 billion) of TANF funding to work supports. In addition, states reallocated funding for recurring cash payments to a wide array of other services, leaving only about 23 percent (or $7 billion) of TANF funding for those recurring cash payments in 2019.6

Figure 1-1.

Spending for TANF and the Programs That Preceded It, by Type of Assistance

Percentage of Total Spending

States used the flexibility provided by TANF to shift funding from cash assistance to work supports, including subsidized child care.

Data source: Department of Health and Human Services. See www.cbo.gov/publication/57702#data.

Before the creation of TANF in 1997, AFDC distributed recurring cash assistance, the Job Opportunities and Basic Skills Training program provided work support, and the Emergency Assistance program supplied other services for low-income families.

Costs to administer the program and monitor compliance with eligibility rules are distributed proportionally among the three types of assistance.

AFDC = Aid to Families With Dependent Children; TANF = Temporary Assistance for Needy Families.

Welfare reform coincided with a large increase in the EITC, which also supports work. The Omnibus Budget Reconciliation Act of 1993 roughly tripled the maximum credit for families with two or more children and roughly doubled the maximum credit for families with one child. Those increases were phased in from 1994 through 1996. In addition, policymakers created the child tax credit in 1998. That credit was available to few working families with low income before 2001, however.

Changes in Federal Spending and Enrollment

Over the past few decades, federal spending on SNAP and Medicaid has increased substantially, whereas spending on AFDC/TANF has fallen drastically (see Figure 1-2). In 1989, the federal government spent $63 billion on Medicaid, compared with roughly $20 billion each on SNAP (which was known as the Food Stamp program at that time) and AFDC. (Again, all figures are in 2019 dollars.)

Figure 1-2.

Federal Spending for Selected Means-Tested and Work Support Programs

Billions of 2019 Dollars

Over the past few decades, federal spending on SNAP and Medicaid has increased substantially, whereas spending on AFDC/TANF has fallen from $20 billion to $16 billion. The CCDF and WIOA have each provided about $5 billion per year in work supports.

Data source: Congressional Budget Office. See www.cbo.gov/publication/57702#data.

Work supports are funded by the CCDF (subsidized child care), WIOA (workforce development services), and TANF (subsidized child care and workforce development services). The transfers that states make from TANF to the CCDF are only included as spending on TANF. Amounts for WIOA are appropriations because data on spending were not available.

CCDF = Child Care Development Fund; SNAP = Supplemental Nutrition Assistance Program; TANF = Temporary Assistance for Needy Families; WIOA = Workforce Innovation and Opportunity Act.

a. Includes spending on predecessors: the Child Care Development Block Grant for the Child Care Development Fund, the Workforce Innovation Act and the Job Training Partnership Act for the Workforce Innovation and Opportunity Act, and Aid to Families With Dependent Children for Temporary Assistance for Needy Families. The WIOA and its predecessors funded a wide range of programs.

Medicaid spending exceeded spending on the other programs in 1989 largely because health care coverage was much more costly than other benefits, a trend that has persisted. Spending per enrollee and the number of enrollees in Medicaid have increased substantially since 1989, bringing the program’s total federal spending to $409 billion in 2019 (in addition to the $221 billion states spent on the program in that year). Federal spending on SNAP has also increased (to $60 billion in 2019), driven mostly by a rise in the number of recipients.7 For Medicaid and SNAP, funding has been adjusted over the years so that benefits are available to everyone who applies for and qualifies for the programs (see Box 1-1). Spending on SNAP has also fluctuated because of changes in food prices and how those changes affect the benefits provided.

Box 1-1.

How Are Means-Tested Programs Funded?

Means-tested programs are funded in various ways. Lawmakers furnish a specific amount of money for Temporary Assistance for Needy Families (TANF), which is primarily provided in the form of a block grant. The state family assistance grant, which totals about $16.5 billion, has accounted for 95 percent of TANF’s federal funding in most years. In recent years, the contingency fund has provided around $600 million in additional funding.1

Funding for the Supplemental Nutrition Assistance Program (SNAP) is handled differently. The amount of money appropriated for SNAP each year is intended to cover the cost of providing benefits to all people who apply for and are eligible for the program. If the appropriated amount does not cover those costs, lawmakers would need to appropriate additional funds, or the Administration would have to cut benefits. There has not been any need to use supplemental appropriations or to implement any reduction in benefits in recent years. (Supplemental appropriations were last provided about 30 years ago.)

Medicaid is an entitlement, which means that federal funding for medical services provided to eligible individuals is open-ended. Funding for the program adjusts automatically as enrollment or costs per enrollee change.

All three programs are jointly financed by the federal government and state governments. To varying degrees, states fund a portion of the services provided through TANF, SNAP, and Medicaid. TANF has a maintenance-of-effort requirement, which is designed to limit the extent to which federal funding displaces money that state governments would otherwise have spent on the program. In contrast, states cover a share of the administrative costs for SNAP, which include employment and training services. For Medicaid, states cover a share of the costs of the services provided by the program in addition to a share of the administrative costs.


1. The contingency fund is a mechanism that can increase the amount of TANF funding available to states that are experiencing economic downturns. For additional details on the funding of TANF, see Congressional Budget Office, Temporary Assistance for Needy Families: Spending and Policy Options (January 2015), www.cbo.gov/publication/49887.

Funding for TANF, CCDF, and WIOA is handled differently. Lawmakers furnish a specific amount of money for those programs each year, which might not be sufficient to provide benefits to all qualified applicants. Since TANF was implemented in 1997, lawmakers have not increased the nominal amount of the state family assistance grant, which provides states with most of their funding for the program; that value has been and remains at about $16.5 billion. Because the amount has not been raised to account for inflation, in effect the value of the program’s funding has fallen by about 30 percent since 1997.8 (To receive all that funding, states are required to spend a certain minimum amount of their own funds on the program. That amount is also not adjusted for inflation; it has been and remains at about $10 billion.) Lawmakers have increased funding for CCDF and WIOA on multiple occasions. Those nominal increases have elevated the inflation-adjusted value of CCDF and approximately maintained it for WIOA.

Enrollment in Medicaid and SNAP has risen substantially over the past three decades, but enrollment in AFDC/TANF has plummeted (see Figure 1-3). In 1989, Medicaid and SNAP had about 20 million recipients each in an average month, and about 10 million people received cash payments through AFDC. Over the subsequent 30 years, enrollment has changed for each program.

Figure 1-3.

Participation in Selected Means-Tested and Work Support Programs

Millions of People

Even though enrollment in Medicaid and SNAP has risen greatly over the past three decades, the number of people receiving recurring cash payments has plummeted.

Data source: Congressional Budget Office. See www.cbo.gov/publication/57702#data.

Data for the CCDF were not available before 1998. CBO did not compile participation data for the Workforce Innovation and Opportunity Act because participation in the programs it funds can range from a brief self-directed job search to a full year of intensive job training with income support.

CCDF = Child Care Development Fund; SNAP = Supplemental Nutrition Assistance Program; TANF = Temporary Assistance for Needy Families.

a. Consists of recipients of recurring cash assistance and includes participants in its predecessor, Aid to Families With Dependent Children.

  • Medicaid. Enrollment has skyrocketed as lawmakers have repeatedly expanded eligibility for the program, including 2014’s expansion under the Affordable Care Act. The number of Medicaid recipients has roughly tripled since 1989, reaching 75 million (or about one-quarter of the U.S. population) during an average month in 2019.
  • SNAP. Participation rose substantially in the 2000s, primarily as a result of outreach initiatives. In general, participation in the program tends to fluctuate, rising during and after recessions and diminishing when the unemployment rate is low.9
  • TANF. Participation has dropped drastically since the onset of welfare reform, having fallen by 85 percent from 1993 to 2019. Some of that decrease can be attributed to rising earnings among single mothers, but it primarily has been driven by declining participation among families whose income is low enough for them to qualify for the program. PRWORA contributed to the decline in the participation rate through the changes it made to AFDC/TANF.10

1. The maximum SNAP benefit increased by slightly more than 20 percent in October 2021. In contrast, TANF benefits have not risen recently in most states and are unlikely to substantially increase in the future because federal funding for that program is not scheduled to rise.

2. For more details, see Congressional Budget Office, Effective Marginal Tax Rates for Low- and Moderate-Income Workers in 2016 (November 2015), www.cbo.gov/publication/50923.

3. The Trump Administration encouraged states to apply for waivers that would allow them to impose work requirements on Medicaid recipients and then approved the applications of 10 states. However, the Biden Administration rescinded those approvals before most of the states had implemented the proposed requirements.

4. The Congress temporarily increased CCDF funding in 2020 and 2021 in response to the coronavirus pandemic.

5. Head Start provides free care and education for the young children of parents with low income. The EITC was designed to encourage people to work by providing a refundable tax credit that initially increases as earnings rise. The child tax credit also encourages work in the same way (although it temporarily stopped functioning as a work support in 2021 because the American Rescue Plan Act of 2021 made it available to all families with qualifying children regardless of their earnings).

6. States spent the remaining 26 percent of TANF funding on a broad array of other services, including initiatives to reduce out-of-wedlock pregnancies, encourage two-parent families, and support the foster care system.

7. In 2019, states spent $4 billion on administrative costs for SNAP.

8. The decline in the inflation-adjusted value of funding for TANF exceeds the decline in the value of spending for the program because several states did not spend all the funding that they received in the program’s first two years.

9. For more details on how participation in SNAP has changed over the years, see Congressional Budget Office, The Supplemental Nutrition Assistance Program (April 2012), www.cbo.gov/publication/43173.

10. Since before 1997, many states have not increased the nominal value of the maximum payment available in TANF, even though the rising cost of living has eroded the value of that payment by about 30 percent. A growing number of income-eligible families appears to find those payments an inadequate incentive to go through the process of demonstrating eligibility for the program—a process made more onerous by the imposition of the work requirements. See Zachary Parolin, “Decomposing the Decline of Cash Assistance in the United States, 1993 to 2016,” Demography, vol. 58, no. 3 (April 2021), pp. 1119–1141, https://doi.org/10.1215/00703370-9157471. PRWORA has contributed to states’ not maintaining the purchasing power of benefits by converting the funding mechanism into a grant of constant nominal value and by allowing states to divert most of that money from cash payments to a wide array of other services.

Chapter 2Effects of Work Requirements on People’s Employment and Income

The effects of work requirements vary among participants in different means-tested programs. Most of the research conducted on those effects has focused on work requirements in the Temporary Assistance for Needy Families program and its predecessor, Aid to Families With Dependent Children, in the 1990s, when welfare reform greatly altered those programs. Evidence of the effects of work requirements in the Supplemental Nutrition Assistance Program and Medicaid is more limited.

Overall, available evidence indicates that the effects of work requirements on employment and income probably differ among the three programs.1 In addition, the gain in employment among program participants who are subject to work requirements is probably offset in part by temporary reductions in employment among people who are not subject to the requirements.2 The analysis in this chapter is mostly based on data from the annual supplement to the Census Bureau’s Current Population Survey, so the numbers provided are for calendar years.

TANF

Nationwide, most adults in TANF are required to participate in work-related activities. Specifically, the federal government requires a certain percentage of able-bodied parents of children receiving monthly cash payments—50 percent, before adjustments are made—to either have sufficient hours of employment or participate in approved activities that could lead to employment, such as vocational training. Any state that does not meet that work standard risks having its federal funding for the program reduced, so states typically decrease or terminate a family’s benefits if the parent’s work hours are not sufficient. Many states have chosen to impose more stringent work requirements even when they do not appear to be at risk of violating the work standard.

In general, those work requirements have increased employment, which has substantially boosted the income of some single mothers.3 (Most TANF recipients are in families headed by single mothers.) The number of people receiving cash payments in TANF has declined by about 85 percent since the mid-1990s, however, and many single mothers have very low income because they are not working or receiving assistance. (This analysis focuses on parents who do not receive disability benefits, because they are most likely to be subject to the work requirements.)

Employment

Most TANF families are headed by single mothers with no education beyond high school. Employment among that group rose substantially in the 1990s, in part because of the widespread imposition of work requirements. In the years leading up to welfare reform, single women without education beyond high school were far less likely to be employed if they had a child (see Figure 2-1).4 From 1993 to 2000, though, the employment rate for less-educated single mothers increased by 16 percentage points (rising from 52 percent to 68 percent), whereas the employment rate of their childless counterparts increased by only 4 percentage points. During those years, many states added work requirements to AFDC (from 1993 to 1996), all states transitioned from AFDC to TANF (in 1997), and states that did not have work requirements already in place started imposing them in TANF (from 1997 to 1999).

Figure 2-1.

Employment Rates for Single Women With No Education Beyond High School, by Presence of Children

Percent

The employment rate for less-educated single mothers—the primary group served by Aid to Families With Dependent Children and Temporary Assistance for Needy Families—rose substantially in the 1990s after work requirements were added to the cash assistance programs and work supports were increased.

Data source: Census Bureau, Annual Social and Economic Supplement to the Current Population Survey, from IPUMS-USA. See www.cbo.gov/publication/57702#data.

The data are by calendar year and are limited to unmarried women between the ages of 18 and 61 who have no postsecondary education, are not students, and are not receiving disability benefits. The employment rates are for March of the specified year.

Although work requirements accounted for some of the 16 percentage-point increase in employment from 1993 to 2000, most of the increase was probably the result of other policy changes (see Box 2-1). Experimental evaluations conducted in AFDC waiver programs during the 1990s indicate that work requirements similar to those in TANF, when combined with increases in work supports, increased the employment rate by about 5 percentage points over five years.5

Box 2-1.

How Policy Changes Made to Means-Tested Programs During the 1990s Affected Single Mothers’ Employment

To encourage employment and limit cash payments, policymakers in the mid-1990s began making changes to the program that would become Temporary Assistance for Needy Families (TANF). Probably the most well-known change was the imposition of work requirements, which made the receipt of benefits contingent on working or engaging in work-related activities. Those work requirements are one of several components of TANF that contributed to the rise in employment among single mothers in the late 1990s. Other contributing changes include the following:

  • A five-year lifetime limit on TANF’s cash payments. Researchers have found evidence that the five-year lifetime limit on cash payments contributed to the rise in employment, even though no families could have reached that limit by 2000. Those researchers argue that the limit caused some families to leave TANF for employment earlier so that they could save some of their potential time in the program in case of future hardship.1
  • A weakened disincentive to work. Benefits drop more slowly as earnings increase under TANF than under its predecessor program, Aid to Families With Dependent Children. Many states weakened that disincentive to work by reducing the benefits available to families without earnings and by allowing families to keep more of their benefits once their earnings rose.2
  • Additional funding for work supports. States shifted some TANF funds from cash payments to work supports, including subsidized child care and workforce development services.

Policy changes not directly related to TANF also boosted the employment of single mothers. Of those changes, increases in the Child Care Development Fund (CCDF) and the earned income tax credit (EITC) probably contributed the most to their rise in employment from 1993 to 2000.3 Lawmakers substantially increased the EITC in phases from 1994 to 1996, which elevated the compensation that parents received for working and thus boosted their incentive to work. Furthermore, lawmakers doubled spending for the CCDF from 1993 to 2000; the additional subsidized child care provided by that increase made employment more feasible and profitable for some single mothers.


1. Francesca Mazzolari, “Welfare Use When Approaching the Time Limit,” Journal of Human Resources, vol. 42, no. 3 (Summer 2007), pp. 596–618, https://doi.org/10.3368/jhr.XLII.3.596; and Jeffrey Grogger, “The Effects of Time Limits, the EITC, and Other Policy Changes on Welfare Use, Work, and Income Among Female-Headed Families,” Review of Economics and Statistics, vol. 85, no. 2 (May 2003), pp. 394–408, https://doi.org/10.1162/003465303765299891.

2. Rebecca M. Blank, “Evaluating Welfare Reform in the United States,” Journal of Economic Literature, vol. 40, no. 4 (December 2002), pp. 1105–1166, https://doi.org/10.1257/002205102762203576. That study suggests that the rapid growth of the economy also contributed to the rise in employment.

3. For a description of the research on the effects of the EITC on the employment of single mothers, see Diane Whitmore Schanzenbach and Michael R. Strain, Employment Effects of the Earned Income Tax Credit: Taking the Long View, Working Paper 28041 (National Bureau of Economic Research, November 2020), www.nber.org/papers/w28041.

TANF’s work requirements have continued to boost employment for the single parents who have entered the program in recent years, according to the limited evidence that is available. Since 2000, the employment rate of single mothers without education beyond high school has remained relatively steady, moving in tandem with the rate of their childless counterparts. But the changes to work requirements have been gradual enough that their effects on employment are obscured by 20 years of changes in the economy and other labor market policies.6

To bolster the available evidence, the Congressional Budget Office analyzed Alabama’s extension of its work requirement to parents of young children and found that it increased their employment rate by 11 percentage points.7 That increase is similar in size to the increase of 10 percentage points that occurred during the first year after parents entered the AFDC waiver experiments.8 However, the overall size of the increase in employment was modest because TANF serves far fewer families than AFDC did.9 Research on the effects of recent changes to work requirements made in other states is scant.

Income

Work requirements in TANF have had a small effect on average income for single mothers. According to studies from the 1990s, that is because increases in their earnings and in the earned income tax credit roughly equaled reductions in cash payments and food assistance.10 Those findings are consistent with trends in the data showing that the average cash income of single mothers with no education beyond high school has generally moved in parallel with the average for their childless counterparts over the past 30 years (see Figure 2-2).

Figure 2-2.

Average Cash Income for Single Women With No Education Beyond High School, by Presence of Children

Thousands of 2019 Dollars

Pretax cash income of less-educated single mothers has changed little, on average, despite the addition of work requirements in the 1990s.

Data source: Census Bureau, Annual Social and Economic Supplement to the Current Population Survey, from IPUMS-USA. See www.cbo.gov/publication/57702#data.

The data are by calendar year and are limited to unmarried women between the ages of 18 and 61 who have no postsecondary education, are not students, and are not receiving disability benefits.

The effect on the income of single mothers has probably been uneven, though. In recent years, single mothers who found work while in TANF saw their cash income rise to an average of about $1,300 per month at the point they left the program, but mothers who left the program without a job had very little cash income.11

Income of Former TANF Recipients Who Found Work. About half of single mothers are employed when they leave TANF, enabling some of their families to rise out of poverty. Studies of families that left TANF in the late 1990s documented their income in subsequent years. About half of the families with a working adult had monthly cash income (in 2019 dollars) above $1,716, which was the poverty threshold for a single parent with two children in 2019. The primary source of that income was the parent’s earnings, and some parents had a partner with substantial earnings as well. Earnings for those families grew modestly in later years, on average.12

Work requirements have raised the income of employed single mothers by increasing how quickly they found a job, which reduced the amount of time they spent in AFDC/TANF. Between the early stages of welfare reform in 1993 and PRWORA’s implementation in 1997, the percentage of mothers who had earnings 24 months after entering the program rose from about 31 percent to about 46 percent (see Figure 2-3).13 That increase in earnings boosted their income, but some of the increase was offset by a decline in the cash payments they received. From 1993 to 1997, the share of families headed by single mothers who were still receiving cash payments from AFDC/TANF 24 months after entering the program fell from about 65 percent to about 35 percent. Over the following years, most single mothers continued to leave the program within 24 months. About half of those mothers had earnings. For most of them, the increase in earnings was larger than the reduction in benefits from leaving TANF. In addition, about 25 percent of single mothers who had earnings 24 months after entering the program were still receiving cash payments.

Figure 2-3.

Sources of Income for Families of Single Women With Children 24 Months After They Started Receiving Cash Assistance, by Year

Percentage of Families Receiving Income From the Specified Source

After work requirements were added to cash assistance programs, recipients tended to find employment more quickly and stop receiving benefits sooner.

Data source: Survey of Income and Program Participation. See www.cbo.gov/publication/57702#data.

Data are limited to unmarried mothers between the ages of 18 and 61 who received cash assistance from AFDC or TANF, are not students, and are not receiving disability benefits.

Data are presented only for the years in which the sample size provided sufficient precision.

AFDC = Aid to Families With Dependent Children; TANF = Temporary Assistance for Needy Families.

Income of Families Without Earnings or Cash Assistance. About half of single mothers are not employed when they stop receiving cash assistance, and the share of single mothers without earnings or cash assistance from AFDC/TANF has risen substantially since welfare reform began.14 Work requirements have probably contributed to that increase in two ways: Families that do not meet those requirements are removed from TANF, and other families are deterred from entering the program. However, the low levels of participation in TANF appear to be primarily driven by other changes made to AFDC/TANF by PRWORA.

Trends in the Data. In the early 1990s, nearly all less-educated single mothers received recurring cash payments if they had not worked for an extended period. As the number of families receiving cash assistance plummeted, though, the share of less-educated single mothers who did not have income from work or AFDC/TANF rose from close to zero in 1993 to 9 percent in 2000, CBO estimates (see Figure 2-4). Over that period, participation in AFDC/TANF declined among those who had earnings, so the portion of less-educated single mothers who received cash payments fell by about 34 percentage points in total. In comparison, the portion of less-educated single mothers who had earnings (whether they participated in AFDC/TANF or not) increased by about 14 percentage points from 1993 to 2000.

Figure 2-4.

Sources of Cash Income for Families of Single Women With Children and With No Education Beyond High School

Percentage of Families

The portion of single mothers without income from cash assistance or work rose substantially in the late 1990s and early 2000s and has stayed elevated since then.

Data sources: Census Bureau, Annual Social and Economic Supplement to the Current Population Survey, from IPUMS-USA; Department of Health and Human Services. See www.cbo.gov/publication/57702#data.

Data are by calendar year and are limited to unmarried mothers between the ages of 18 and 61 who have no postsecondary education, are not students, and are not receiving disability benefits. The only sources of income considered are the mother’s employment and recurring cash assistance through TANF or its predecessor, AFDC. Sources of income have been adjusted for errors in reporting.

AFDC = Aid to Families With Dependent Children; TANF = Temporary Assistance for Needy Families.

Over the next two decades, the share of less-educated single mothers who did not have income from work or TANF continued growing, albeit more gradually. Reductions in TANF receipt were no longer countered by substantial increases in employment, leaving one-in-seven less-educated single mothers without income from work or TANF in 2019. (That estimate is highly uncertain because of imprecision in the underlying data.)15

Sources of Income. In recent years, most single mothers without earnings or cash assistance reported being in deep poverty. They had little cash income, although they typically received benefits through SNAP to purchase food. About 39 percent of those mothers lived with an adult who reported working, typically the children’s (unwed) father. Other mothers reported receiving cash from people who they did not live with or were not related to; the most common forms of assistance were child support (11 percent) and help from friends or relatives (11 percent). The mothers might understate cash income from those sources, which could help explain why single mothers tend to report spending substantially more income than they report receiving.16

The Role of Work Requirements. Work requirements probably increase the number of single mothers without earnings or cash assistance by removing them from TANF when they are not employed. Work requirements can affect that number by changing the length of time nonworking mothers receive assistance and by changing whether those mothers are employed when they stop receiving assistance. The expansion of Alabama’s work requirement in 2018 appears to have had little effect on the latter variable, but it did reduce the length of time families received assistance by about a month during the first year after they entered the program, on average. Thus, the work requirement probably caused the number of families without earnings or cash assistance to increase modestly. Other research finds evidence of a larger increase in the number of such families stemming from a more stringent work requirement implemented when the labor market was weaker.17 In addition, the AFDC waiver experiments showed that work requirements often increase the number of families in deep poverty, and deep poverty is common among families without earnings or cash assistance.18

Although some single mothers end up disconnected from employment and cash payments because they are removed from TANF for not complying with its work requirements, most are probably disconnected because they do not apply for TANF or because their application is rejected. Before welfare reform, most single mothers whose income and assets were low enough to qualify for recurring cash payments received that assistance. But TANF participation among that group has fallen over the past three decades, from about 80 percent to roughly 25 percent.19 Many single mothers with sufficiently low income and assets are not entering the program because they are ineligible for other reasons or because they do not complete the application process.

Work requirements may have increased the number of single mothers without earnings or cash assistance by stopping some of them from enrolling in TANF. Work requirements can reduce entry into TANF by discouraging parents from applying or by causing them to be ineligible. The expansion of Alabama’s work requirement does not appear to have had an immediate effect on the number of families entering the program. Repeated violation of the work requirement can lead to families’ being ineligible to enter the program for a year, however. (Ten other states impose ineligibility periods of a year or more for repeated violations of the work requirement; six of those states permanently bar families from reentering the program.) Thus, TANF’s work requirements probably reduce entry into the program by mothers without earnings by removing them from the program when they are not working and then barring their reentry. However, the low levels of participation in TANF appear to be primarily driven by other changes made to AFDC/TANF by PRWORA.20

SNAP

The federal government places two work requirements on adults in SNAP. First, most able-bodied nonelderly adults who are not working at least 30 hours a week or caring for a child under age 6 must register for work—that is, notify their state’s employment office that they are available to work—and accept a suitable job if one is offered. Second, able-bodied adults under the age of 50 who do not live with dependent children (so-called ABAWDs, or able-bodied adults without dependents) are generally limited to 3 months of benefits in any 36-month period unless they are working at least 80 hours per month or in job training.

ABAWDs accounted for 7 percent of SNAP recipients in 2019 and a larger percentage in years when many localities waived the work requirement.21 Waiving of the work requirement—mainly because of adverse economic conditions—has occurred frequently over the past 14 years.22 This report focuses on the work requirement for ABAWDs because of the lack of research on the income and employment of other SNAP recipients who must register for work.

Employment

The work requirement for ABAWDs slightly increases the employment of the oldest of the workers who are subject to it. Most studies of SNAP’s work requirement compare employment of ABAWDs in their late 40s with that of able-bodied adults without dependents who are just over 49 because those two groups are similar in their age and education but differ in whether they are subject to the work requirement. Using data for four states (Colorado, Pennsylvania, Missouri, and Virginia), those studies estimate that the employment rate is 1 to 4 percentage points higher among the group of ABAWDs subject to the requirement.23 SNAP participants between the ages of 45 and 49 account for only about 15 percent of workers who are subject to the work requirement, though, so those findings might not apply more broadly.

SNAP’s work requirement might induce younger ABAWDs to work more, but it probably boosts employment less than TANF’s work requirements do. Younger SNAP recipients might be more likely to seek employment because they will be ineligible for benefits for more years if they do not (unlike, say, a recipient who is 48 and thus subject to the work requirement for only two more years). SNAP’s work requirement is probably less effective than TANF’s requirement, however, because most SNAP recipients do not receive intensive case-management services. Those services, typically provided one on one by a caseworker, can boost the effectiveness of work requirements by monitoring progress, assigning work supports, and providing job referrals. One study of unemployment insurance found that a 40-minute meeting with a caseworker reduced the duration of unemployment by 5 percent, on average.24

Income

SNAP’s work requirement probably reduces the average income of ABAWDs by causing many of them to stop receiving benefits. Studies of recipients who are near the age of 50 found that the work requirement reduced the number of ABAWDs receiving benefits by 22 percent, 31 percent, and 53 percent in Pennsylvania, Missouri, and Virginia, respectively. (The work requirement did not appear to change the number of ABAWDs receiving benefits in Colorado.) Most of those ABAWDs lost SNAP benefits because they did not demonstrate sufficient participation in work-related activities. That reduction in benefits appears larger than the increase in earnings among the smaller percentage of ABAWDs who worked more because of the work requirement.

SNAP’s work requirement results in many ABAWDs having very low income because SNAP benefits had been their primary source of income. Most ABAWDs who stop receiving benefits from SNAP have few or no other sources of income, and many of them are homeless.25 Only about 30 percent of ABAWDs have any earnings—and none of those people are eligible to receive cash payments through TANF.26 (In contrast, most TANF recipients who lose cash payments because of noncompliance with the program’s work requirements still receive full SNAP benefits because they are not ABAWDs.)

Medicaid

Although federal law does not impose any work requirements on Medicaid recipients, states have sought waivers to do so. In 2018, the Trump Administration began approving requests from states to waive the entitlement to Medicaid in order to experiment with work requirements. Most of the waivers required ABAWDs to perform work-related activities for at least 80 hours a month to retain their health insurance coverage. The Administration approved waivers for 11 states, but implementation has been halted for various reasons: The courts set aside four of the approvals, six waivers have not been implemented, and the only state that implemented a work requirement in 2020 suspended it after two months because of the pandemic. (In 2021, the Biden Administration rescinded all the approvals.)

Employment

States’ limited experience with work requirements in Medicaid indicates that they cause a substantial portion of adults who are not exempt from them to lose coverage, and they appear to have little effect on employment. Only Arkansas had a work requirement in place for more than a few months—from June 2018 to March 2019, when it was set aside by the courts. Under that requirement, ABAWDs lost Medicaid coverage if they did not document sufficient hours spent in work-related activities for three months in the same year. Over the seven months of 2018 that the requirement was in effect, about 23 percent of Medicaid recipients who were subject to it lost coverage for failure to comply.

Researchers who designed and conducted a phone survey targeted at measuring employment among Medicaid recipients in Arkansas after the work requirement took effect found mixed results. The employment rate decreased slightly using one statistical approach and increased a bit using another approach; in both cases, though, the sample size was too small to rule out no change or a sizable change.27

Income and Expenses

Arkansas’ work requirement probably reduced families’ resources overall because it caused a large reduction in Medicaid coverage. In addition, the number of hours that recipients worked did not appear to increase, which suggests that their earnings did not rise. The targeted phone survey used by researchers to measure employment indicates that the work requirement substantially increased the number of adults who were uninsured by terminating their Medicaid coverage. Instead of having most of their medical care covered by Medicaid, many of the affected adults had to pay for most of it themselves. The work requirement appears to have roughly doubled the portion of those adults who reported having serious problems paying their medical bills, according to the researchers.28

The combination of a large reduction in Medicaid receipt and a small or negligible effect on employment among adults subject to the work requirement might have been driven by two factors. First, many employed participants might have been unaware of the requirement, or they might have found it onerous to demonstrate compliance. The targeted survey indicates that about 35 percent of Medicaid recipients in the age range subject to the work requirement were not aware of it. Second, other studies have shown that even modest increases in the effort required to demonstrate eligibility for means-tested benefits can cause large declines in the number of recipients.29 A lack of awareness and burdensome reporting could explain why only about 9 percent of Medicaid recipients who needed to demonstrate compliance filed a report. (In most cases, the administrators of the program were able to verify that recipients were in compliance without contacting them.)


1. This report focuses on changes in employment and income within the standard period used for the Congressional budget process, which extended through 2031 when this report was prepared. Evidence of the longer-term effects of work requirements and supports is limited, although research indicates that access to SNAP benefits during childhood tends to increase earnings in adulthood. See Marianne P. Bitler and Theodore F. Figinski, Long-Run Effects of Food Assistance: Evidence From the Food Stamp Program (August 2019), https://tinyurl.com/2p974kdj (PDF, 9.2 MB). Thus, work requirements that terminate benefits for families with children might reduce their earnings later in life.

2. Most of the research on unemployment insurance indicates that policies that increase employment by encouraging people to search for a job decrease employment among people not in the program. Such spillovers are probably rarer when jobs are plentiful, and those studies were conducted during periods of high unemployment. For example, see Ioana Marinescu, “The General Equilibrium Impacts of Unemployment Insurance: Evidence From a Large Online Job Board,” Journal of Public Economics, vol. 150 (June 2017), pp. 14–29, https://doi.org/10.1016/j.jpubeco.2017.02.012.

3. Initially, AFDC only provided benefits to single mothers or the wives of men who were unable to work. Eligibility was extended to single fathers and families with two able-bodied parents before the transition to TANF. However, participation in the programs has remained low among those groups.

4. Families of women without dependents are not eligible for cash payments from TANF. Because such childless women are largely unaffected by TANF, they are used as proxies for how the employment and income of single mothers without education beyond high school would have changed in the absence of welfare reform.

5. The experimental evaluations compared recipients who were randomly assigned to a combination of work requirements, intensive job-search assistance, and additional subsidized child care with recipients of the standard array of assistance. For a summary of the findings of those evaluations, see Gayle Hamilton and others, National Evaluation of Welfare-to-Work Strategies (submitted by Manpower Demonstration Research Corporation to the Department of Health and Human Services, December 2001), p. 86, https://tinyurl.com/2p89s5ad. The Congressional Budget Office estimated the employment effects from evaluations of programs that focused on labor force attachment because that is TANF’s focus.

6. The Urban Institute’s welfare rules database indicates that more states have tightened work requirements than have loosened them. See Urban Institute, “Welfare Rules Databook,” https://wrd.urban.org/wrd/databook.cfm (accessed May 31, 2022).

7. Alabama expanded the number of people subject to TANF’s work requirement when it stopped providing exemptions to parents with a child between the ages of 6 months and 11 months. In CBO’s estimation, that expansion increased the employment rate among participants who were newly subject to the work requirement by about 4 percentage points in the year following their entry into TANF. Because the requirement only applied to parents when their child was between the ages of 6 months and 11 months, though, participants were newly subject to the work requirement for only about half of the first year. Had those participants never been subject to the work requirement before the expansion and became subject to it for the full year after the expansion, their employment rate would have increased by about 9 percentage points in the first year after entering the program, CBO estimates. For more details about the methods CBO used for this analysis, see Appendix B.

8. See Gayle Hamilton and others, National Evaluation of Welfare-to-Work Strategies (submitted by Manpower Demonstration Research Corporation to the Department of Health and Human Services, December 2001), p. 352, https://tinyurl.com/2p89s5ad. The employment effects shrank over time. CBO did not have sufficient data to estimate the effect of Alabama’s TANF expansion over a longer period.

9. In Alabama, TANF served only about 4 percent of single parents in poverty during an average month in 2018. Nationally, TANF served about 10 percent of that group, whereas AFDC served about 77 percent of that group during the years leading up to enactment of the Personal Responsibility and Work Opportunity Reconciliation Act.

10. See Gayle Hamilton and others, National Evaluation of Welfare-to-Work Strategies (submitted by Manpower Demonstration Research Corporation to the Department of Health and Human Services, December 2001), https://tinyurl.com/2p89s5ad.

11. For details on CBO’s analysis, see Appendix B.

12. Those findings come from a summary of studies conducted in 12 states, although data generally were not available for all 12 states. See Gregory Acs and Pamela Loprest, Leaving Welfare: Employment and Well-Being of Families That Left Welfare in the Post-Entitlement Era (W. E. Upjohn Institute for Employment Research, 2004), https://doi.org/10.17848/9781417550012.

13. Federal law requires states to impose work requirements on families within their first 24 months of receiving assistance. For details, see Appendix A, which compares the changes in employment and benefits for single mothers in SNAP who were not subject to work requirements.

14. One study found that TANF’s work requirements led to some nonworking single mothers’ receiving more cash assistance through Supplemental Security Income than they would have through TANF. The prospect of losing TANF benefits because of the work requirements appears to have led some mothers to undertake the more strenuous application process for SSI, which generally provides larger monthly payments than TANF. See Lucie Schmidt and Purvi Sevak, “AFDC, SSI, and Welfare Reform Aggressiveness: Caseload Reductions Versus Caseload Shifting,” Journal of Human Resources, vol. 39, no. 3 (Summer 2004), pp. 792–812, https://doi.org/10.3368/jhr.XXXIX.3.792.

15. The underreporting of earnings and cash assistance in household surveys is well established. See Bruce D. Meyer and others, “The Use and Misuse of Income Data and Extreme Poverty in the United States,” Journal of Labor Economics, vol. 39, no. S1 (January 2021), pp. S5–S58, https://doi.org/10.1086/711227. CBO adjusted downward the number of single mothers without earnings on the basis of inconsistencies between multiple interviews covering the same years. That adjustment decreased the estimated number of families without income from work or AFDC/TANF by 18 percent in 1993; by 2019, the size of that adjustment had grown to 29 percent. In addition, CBO adjusted the percentage of single mothers who receive recurring cash payments so that it matched the administrative data.

16. Bruce D. Meyer and James X. Sullivan, “Changes in the Consumption, Income, and Well-Being of Single Mother Headed Families,” American Economic Review, vol. 98, no. 5 (December 2008), https://doi.org/10.1257/aer.98.5.2221.

17. Tazra Mitchell, LaDonna Pavetti, and Yixuan Huang, Life After TANF in Kansas: For Most, Unsteady Work and Earnings Below Half the Poverty Line (Center on Budget and Policy Priorities, February 2018), https://tinyurl.com/mpxsxkbj.

18. Stephen Freedman and others, National Evaluation of Welfare-to-Work Strategies (submitted by Manpower Demonstration Research Corporation to the Department of Health and Human Services, June 2000), p. 199, https://tinyurl.com/38k9exee.

19. Linda Giannarelli, What Was the TANF Participation Rate in 2016? (Urban Institute, July 2019), https://tinyurl.com/3tny44sc.

20. For a summary of the other channels through which PRWORA has reduced participation, see Zachary Parolin, “Decomposing the Decline of Cash Assistance in the United States, 1993 to 2016,” Demography, vol. 58, no. 3 (April 2021), pp. 1119–1141, https://doi.org/10.1215/00703370-9157471.

21. Kathryn Cronquist, Characteristics of Supplemental Nutrition Assistance Program Households: Fiscal Year 2019, SNAP-20-CHAR (submitted by Mathematica Policy Research to the Department of Labor, March 2021), www.fns.usda.gov/snap/characteristics-snap-households-fy-2019.

22. The federal government waived the work requirement from April 2009 through September 2010, and waivers remained in effect in most of the country for a few years after that because of high unemployment rates. Starting in April 2020, the Families First Coronavirus Response Act compelled all states to waive the work requirement by suspending it; that suspension will continue through the month after the public health emergency declaration is lifted.

23. Colin Gray and others, Employed in a SNAP? The Impact of Work Requirements on Program Participation and Labor Supply, Social Science Research Network working paper (August 2020), http://dx.doi.org/10.2139/ssrn.3676722; and Laura Wheaton and others, The Impact of SNAP Able-Bodied Adults Without Dependents (ABAWD) Time Limit Reinstatement in Nine States (Urban Institute, June 2021), https://go.usa.gov/xtYxE.

24. Amelie Schiprowski, “The Role of Caseworkers in Unemployment Insurance: Evidence From Unplanned Absences,” Journal of Labor Economics, vol. 38, no. 4 (August 2020), pp. 1189–1225, https://doi.org/10.1086/706092.

25. Laura Wheaton and others, The Impact of SNAP Able-Bodied Adults Without Dependents (ABAWD) Time Limit Reinstatement in Nine States (Urban Institute, June 2021), https://go.usa.gov/xtYxE.

26. Kathryn Cronquist, Characteristics of Supplemental Nutrition Assistance Program Households: Fiscal Year 2019, SNAP-20-CHAR (submitted by Mathematica Policy Research to the Department of Labor, March 2021), www.fns.usda.gov/snap/characteristics-snap-households-fy-2019.

27. Benjamin D. Sommers and others, “Medicaid Work Requirements—Results From the First Year in Arkansas,” New England Journal of Medicine, vol. 381, no. 11 (September 2019), pp. 1073–1082, http://dx.doi.org/10.1056/NEJMsr1901772.

28. Benjamin D. Sommers and others, “Medicaid Work Requirements in Arkansas: Two-Year Impacts on Coverage, Employment, and Affordability of Care,” Health Affairs, vol. 39, no. 9 (September 2020), pp. 1522–1530, https://doi.org/10.1377/hlthaff.2020.00538.

29. Manasi Deshpande and Yue Li, “Who Is Screened Out? Application Costs and the Targeting of Disability Programs,” American Economic Journal: Economic Policy, vol. 11, no. 4 (November 2019), pp. 213–248, https://doi.org/10.1257/pol.20180076.

Chapter 3Effects of Work Supports on People’s Employment and Income

Two of the most common ways that the federal government supports work among participants in Temporary Assistance for Needy Families, the Supplemental Nutrition Assistance Program, and Medicaid are by subsidizing child care and funding workforce development services. Those work supports can boost the employment of recipients and thus increase their income. Employment gains among program participants might hold down employment among people who do not have access to those work supports, though, because they compete for the same types of jobs.

Subsidized Child Care

The federal government primarily funds subsidized child care through the Child Care Development Fund, which the states administer. The CCDF provides subsidies to low-income families who are working or preparing to work. Recipients pay about 5 percent of their income for child care, on average, and the subsidy covers the remainder.

The most compelling research on the effects of subsidized child care on employment and income comes from a study of Child Care Development Block Grant recipients. That study’s findings can be extrapolated to recipients in TANF, SNAP, and Medicaid. Subsidized child care confers two main benefits: It increases earnings by enabling recipients to work more, and it boosts consumption by reducing the amount recipients spend on child care.

Employment

Subsidized child care has boosted the employment of low-income single parents, who represent about 80 percent of CCDF recipients. Most data do not indicate the direction of causality—that is, whether people are working because they are receiving a subsidy or receiving a subsidy because they are working. The latter explanation probably prevails often in states that have employment as an eligibility requirement, but the subsidies boost employment in other instances.

One study addresses that issue by recording whether single mothers were working before they applied for subsidized child care. The researchers found that the subsidies provided by the Child Care Development Block Grant—now a major component of the CCDF—increased the employment rate of single mothers with very low income by 8 to 25 percentage points.1 That research is based on a 1988 survey of single mothers in Kentucky who did not have to work to receive cash payments.

The Congress boosted CCDF funding in March 2018, allowing many states to substantially decrease the length of their waitlists for subsidized child care.2 By February 2020, there was one wait-listed child for every 12 children who received a CCDF subsidy.3 Additional funding for subsidized child care would probably benefit participants in Medicaid and SNAP and increase their employment more than it would assist participants in TANF. That is because most states already give priority for child care subsidies to parents who are subject to TANF’s work requirements.

Recent boosts to CCDF funding probably have increased employment less than might be suggested by the Kentucky study. As that study demonstrates, applicants are more likely to work once they get through the waitlist for subsidized child care, but the potential to increase employment by reducing waitlists has diminished because the lists have gotten shorter. Shorter waitlists indicate that fewer parents have been delaying work because child care subsidies are not available. However, because some states continue to use increases in CCDF funding to reduce their waitlists, such funding continues to boost employment, although the effect is probably small.

In addition to reducing waitlists, states have responded to the recent boost in CCDF funding by expanding eligibility to families with higher income and by reducing the amount recipients pay out of pocket for child care. Research indicates that such policies do not increase employment as much as subsidies that are provided to single parents with very low income.4

Income and Expenses

Subsidized child care substantially increases the resources available to some low-income families. It boosts earnings for some single parents by enabling them to work, and it bolsters consumption for some parents who would have worked anyway by reducing the amount they spend on child care.

The average subsidy funded by CCDF is about $500 per month, which is about one-third of the average monthly income for a single parent and child at the poverty threshold. The size of the average increase in resources from subsidized child care is much harder to determine and highly uncertain because researchers have provided little indication of how many recipients would have purchased child care in the absence of the subsidy. In many instances, the subsidy probably had little effect on a family’s resources because it replaced care that would have been provided by a relative for little or no charge. About 68 percent of working single parents who were in TANF, SNAP, or Medicaid received free informal care in an average week from 2013 through 2016, typically from a child’s grandparent.5 In contrast, only about 23 percent of families in those programs were paying for child care without any subsidy during those years, and that share has probably decreased as a result of the CCDF’s recent expansions.

Workforce Development Services

This report focuses on three common types of workforce development services: job-search assistance, job training, and subsidized employment. Researchers have typically analyzed how those services affect the employment and income of TANF recipients and participants in two of the largest workforce development programs, the Adult and Dislocated Worker programs. Participants in those two programs are broadly similar to the able-bodied nonelderly adults in SNAP and Medicaid in that many of them have low income, little education, and sporadic employment.6

Job-Search Assistance

Intensive job-search assistance increases employment and income. In the 2010s, the Department of Labor (DOL) conducted a national experimental evaluation of the intensive job-search assistance provided through the Adult and Dislocated Worker programs.7 In the year following the receipt of assistance, people who had been assigned to intensive job-search assistance and counseling earned about $2,200 (or 18 percent) more, on average, than people who had only been given access to online tools and literature with little direct assistance from staff. (The experiment did not evaluate the effects of that less intensive assistance.)

That substantial increase in earnings resulted from a combination of a higher employment rate and better paying jobs. The gains from intensive job-search assistance could be smaller among participants in TANF, SNAP, and Medicaid because they face more barriers to employment, such as not having graduated from high school. But intensive job-search assistance appears to also have boosted the employment of less-educated recipients in the Aid to Families With Dependent Children’s experimental evaluations. Intensive job-search assistance was a component of all the experimental evaluations of AFDC work requirements that were found to boost employment.8

Job Training

Job training provided by the Adult and Dislocated Worker programs has not been shown to increase participants’ employment or earnings. DOL’s national experimental evaluation found that recipients of such training tended to earn less during the training period than people who were provided job-search assistance immediately. In addition, after the training and job-search assistance were completed, the earnings of those two groups were similar, on average. Likewise, experimental evaluations of AFDC recipients indicate that programs focusing on immediate job-search assistance generally boost employment and earnings more than programs that first provide vocational training or education.9

One possible explanation is that the training did not align well with the demands of local employers. In DOL’s experiment, only about 40 percent of participants in occupation-specific training found a job in that occupation.10 Since DOL conducted its evaluation, the Congress has made several modifications to those programs that are intended to better align job training with the needs of local employers.11

Recent research demonstrates that job training provided by smaller programs can increase employment and income when it focuses on occupations that are in high demand locally and that have the potential for career advancement.12 Such programs differ from the Adult and Dislocated Worker programs in that they are administered by community-based organizations. Whether those programs would increase employment and income if implemented on a national scale is unclear. But the findings suggest that the Congress’s recent attempts to focus job training on the needs of local employers may have led to federally provided job training that increases participants’ employment and income.

Subsidized Employment

The federal government provides funding for subsidized employment to the states, which use it to temporarily cover some or all of the costs of compensation paid by a participant’s employer (usually a private firm). That strategy has been used to boost employment during recessions and to help participants with few skills progress toward unsubsidized employment.

Subsidized employment increases workers’ employment and earnings, though the gains are usually temporary. Many recipients of subsidized employment would not have been employed otherwise, either because jobs were scarce or because they lacked job skills. The subsidies boosted their income by raising their earnings. Most studies have found that those gains disappeared once the subsidies ended.13 However, an evaluation of 13 recent programs found that 6 of them generated more persistent and substantial increases in earnings for the affected workers.14


1. Mark C. Berger and Dan A. Black, “Child Care Subsidies, Quality of Care, and the Labor Supply of Low-Income, Single Mothers,” Review of Economics and Statistics, vol. 74, no. 4 (November 1992), pp. 635–642, https://doi.org/10.2307/2109377.

2. Consolidated Appropriations Act of 2018, Public Law 115-141.

3. Those findings are CBO’s analysis of data from Karen Schulman, On the Precipice: State Child Care Assistance Policies 2020 (National Women’s Law Center, May 2021), https://tinyurl.com/ycksszu2 (PDF, 2.4 MB).

4. The cost of unsubsidized child care is less likely to deter higher-income parents from working. A randomized controlled trial conducted on families near the poverty line found that a substantial reduction in the percentage of income recipients had to pay for child care had no effect on their employment or earnings. See Charles Michalopoulos, Effects of Reducing Child Care Subsidy Copayments in Washington State, OPRE 2011-2 (submitted by MDRC to the Department of Health and Human Services, June 2010), https://tinyurl.com/mumukvfk.

5. In the survey, respondents were asked about their child care usage in a typical week in the fall each year. The amounts cited are based on CBO’s analysis of data from the Survey of Income and Program Participation. The analysis was limited to recipients with a child under age 13; those families account for almost all CCDF recipients.

6. For a description of the characteristics of participants in the Adult and Dislocated Worker programs, see Kenneth Fortson and others, Providing Public Workforce Services to Job Seekers: 30-Month Impact Findings on the WIA Adult and Dislocated Worker Programs (submitted by Mathematica Policy Research to the Department of Labor, May 2017), https://tinyurl.com/2p8kkbvs.

7. Ibid.

8. Gayle Hamilton and others, National Evaluation of Welfare-to-Work Strategies (submitted by Manpower Demonstration Research Corporation to the Department of Health and Human Services, December 2001), https://tinyurl.com/2p89s5ad.

9. For a summary of the findings of many studies on job training, see Gayle Hamilton, Improving Employment and Earnings for TANF Recipients, OPRE Brief 6 (Urban Institute, March 2012), https://tinyurl.com/yc346ddv. For a study that found evidence of long-term benefits from job training, see V. Joseph Hotz, Guido W. Imbens, and Jacob A. Klerman, “Evaluating the Differential Effects of Alternative Welfare-to-Work Training Components: A Reanalysis of the California GAIN Program,” Journal of Labor Economics, vol. 24, no. 3 (July 2006), pp. 521–566, https://doi.org/10.1086/505050.

10. Kenneth Fortson and others, Providing Public Workforce Services to Job Seekers: 30-Month Impact Findings on the WIA Adult and Dislocated Worker Programs (report submitted by Mathematica Policy Research to the Department of Labor, May 2017), https://tinyurl.com/2p8kkbvs.

11. Workforce Innovation and Opportunity Act of 2014, Public Law 113-128, www.congress.gov/bill/113th-congress/house-bill/803.

12. Lawrence F. Katz and others, Why Do Sectoral Employment Programs Work? Lessons From WorkAdvance, Working Paper 28248 (National Bureau of Economic Research, December 2020), www.nber.org/papers/w28248.

13. David Butler and others, What Strategies Work for the Hard-to-Employ? Final Results of the Hard-to-Employ Demonstration and Evaluation Project and Selected Sites From the Employment Retention and Advancement Project, OPRE Report 2012-08 (submitted by MDRC to the Department of Health and Human Services, March 2012), https://tinyurl.com/49d46rxh; and Erin Jacobs and Dan Bloom, Alternative Employment Strategies for Hard-to-Employ TANF Recipients: Final Results From a Test of Transitional Jobs and Preemployment Services in Philadelphia, OPRE Report 2011-19 (submitted by MDRC to the Department of Health and Human Services, December 2011), https://tinyurl.com/bxyypzte.

14. Danielle Cummings and Dan Bloom, Can Subsidized Employment Programs Help Disadvantaged Job Seekers? A Synthesis of Findings From Evaluations of 13 Programs, OPRE Report 2020-23 (submitted by MDRC to the Department of Health and Human Services, February 2020), https://tinyurl.com/bdf2bfcf.

Chapter 4Factors That Affect How Work Requirements and Work Supports Change People’s Employment and Income

The extent to which work requirements and work supports boost employment and income probably depends on many factors. When policymakers consider proposals that aim to increase employment and income among participants of means-tested programs through work requirements or work supports, it is important to consider three factors—the extent to which participants are already engaged in work-related activities, barriers to work, and economic conditions—and how those factors could change over time.

Employment, training, and job-search assistance are already widespread among some groups of participants in means-tested programs, which limits the extent to which work requirements can raise employment. Many program participants who are not engaged in work-related activities report having multiple conditions that make employment difficult, such as health issues or child care responsibilities.1 The ability of low-income adults to find and keep employment also depends on the health of the economy.

Participation in Work-Related Activities in the Absence of Work Requirements and Work Supports

The ability of work requirements and work supports to boost the employment of recipients is limited by the extent to which the recipients would participate in work-related activities in the absence of those provisions. Because of that constraint, expansions of work requirements and work supports in the Supplemental Nutrition Assistance Program or Medicaid could be less effective than that expansion was decades earlier in the Aid to Families With Dependent Children program. In 1993, about 55 percent of AFDC recipients reported not working, searching for work, training for a job, or being a student (see Figure 4-1). In contrast, from 2013 through 2016, about 30 percent of SNAP and Medicaid recipients reported not participating in at least one of those activities. Because the employment rate among SNAP and Medicaid recipients is much higher than it was for AFDC recipients, it has less room to increase further.

Figure 4-1.

Work-Related Activities of Adult Recipients of Assistance, by Program

Percent

The ability of work requirements to boost employment is limited by the extent to which recipients would have participated in work-related activities in the absence of those provisions. Most SNAP and Medicaid recipients are working or preparing to work even though few of them are subject to work requirements.

Data source: Survey of Income and Program Participation. See www.cbo.gov/publication/57702#data.

Data are limited to recipients between the ages of 18 and 61 who were not receiving disability benefits.

Some recipients are engaged in multiple activities. They are included in the lowest applicable segment of the bar and not in higher segments.

AFDC = Aid to Families With Dependent Children; SNAP = Supplemental Nutrition Assistance Program; TANF = Temporary Assistance for Needy Families.

a. Recipients are defined as not working or preparing to work if they report not working, searching for work, being in a job-training program, or being in school.

When work requirements are added to programs in which many recipients are already working enough hours to meet the requirements, they are likely to reduce recipients’ total income. That is because those requirements are unlikely to increase their earnings but can reduce the benefits they receive. Employed adults must document their compliance with the work requirements (typically each month or every few months), a process that can be arduous and time-consuming. Some employed adults lose their benefits because they do not submit the required information, either because they find doing so too onerous or because they are not aware of the requirement.2 Thus, when work requirements are imposed on recipients among whom employment is more common—such as those in SNAP and Medicaid—the reduction in benefits is more likely to exceed any increase in earnings.

Barriers to Work

Work requirements are less likely to lead to employment and more likely to reduce income, including benefits, when applied to people who have conditions that make it difficult to find and keep a job. Participants in Temporary Assistance for Needy Families, SNAP, and Medicaid who are not working or preparing to work often report having such conditions.3 Many of them report having a disability that limits their ability to work, such as physical or mental health issues. Other common barriers to employment include caring for young children or not having a high school diploma or equivalent degree.4

In a 2014 survey, about 70 percent of recipients in TANF and SNAP who were not working or preparing to work reported having at least one barrier to employment, and about 20 percent reported having more than one (see Figure 4-2). Such barriers were also common among their counterparts in Medicaid, but less so. Under the stringent work requirements in TANF, recipients who report having those barriers are much less likely to be employed when they leave the program. The employment rate for former TANF recipients who report having multiple barriers to employment is about 35 percent, compared with about 70 percent for former recipients who report having none of those employment barriers.

Figure 4-2.

Barriers to Employment Among Adult Recipients of Assistance Who Are Not Engaged in Work-Related Activities, by Program

Percent

Most recipients of assistance who are not working or preparing to work report facing at least one of three common barriers that hinder employment.

Data source: Survey of Income and Program Participation, 2014. See www.cbo.gov/publication/57702#data.

Data are limited to recipients between the ages of 18 and 61 who were not receiving disability benefits, working, searching for work, or in school.

SNAP = Supplemental Nutrition Assistance Program; TANF = Temporary Assistance for Needy Families.

a. The barriers to employment reflected in the data are having a work-limiting disability, not having a high school diploma or equivalent degree, and having a child younger than 6 while being a single parent or having a working spouse.

Work supports can have a larger effect on employment when they are directed at people who face barriers to finding and keeping jobs instead of all participants. For example, subsidized child care can boost employment more when it is targeted at people who are not working because they cannot afford adequate care. In contrast, extending child care subsidies to people with higher income would probably have a smaller effect on employment because those people are more likely to be able to afford child care without a subsidy. Similarly, job training can boost employment more when it is directed at people who are unlikely to find a job because they lack marketable skills. Of the 11 approaches to getting AFDC recipients working that were evaluated in the 1990s, the largest increase in employment came from assigning the most disadvantaged participants to training before they began searching for a job.5 However, even in that program many people remained unemployed, which is indicative of the limits of both work requirements and work supports.

A 2015 survey of SNAP recipients provides detailed data on the barriers to employment faced by recipients of means-tested benefits.6 As in the 2014 survey, physical and mental health issues are pervasive among those recipients. They constitute the most common barrier to employment, even though recipients of Supplemental Security Income and Social Security were excluded from the survey (see Figure 4-3). The 2015 survey also excluded recipients who were caring for an incapacitated adult or a child under age 6 or who were participating in a drug or alcohol treatment program; still, 11 percent of respondents reported a lack of child care as a barrier to employment, indicating that not having access to before- and after-school care can limit employment. Other reported barriers, such as a lack of education or transportation, could be addressed by workforce development programs. Overall, about 79 percent of the surveyed SNAP recipients reported facing at least one barrier to employment, and 48 percent reported more than one.

Figure 4-3.

Barriers to Employment for Selected Recipients of SNAP

Percent

A 2015 survey of SNAP recipients indicates that they face multiple barriers to employment that are not measured by the Survey of Income and Program Participation.

Data source: Department of Agriculture. See www.cbo.gov/publication/57702#data.

Data are limited to SNAP recipients who had to register for work. Thus, the data probably exclude most adults who are receiving disability benefits, caring for an incapacitated adult or a dependent child under age 6, or participating in a drug or alcohol treatment program.

SNAP = Supplemental Nutrition Assistance Program.

Economic Conditions

When economic conditions are poor, work requirements can have adverse effects on participants in means-tested programs. In particular, those requirements can substantially reduce program participation while boosting employment only a little, and they can reduce average income because the lost benefits are not offset by additional earnings. When jobs become harder to find, work requirements are difficult to meet through employment. At those times, if the requirements are not eased, they put families at greater risk of losing benefits before they can find work—and they are less likely to spur recipients to find work.

In recent years, TANF’s work requirements have probably reduced the average income of recipients during periods of high unemployment. To examine the effects of work requirements during such periods, the Congressional Budget Office used data from the Department of Health and Human Services that cover 2012 through 2018.7 Like other job seekers, parents in TANF find jobs far less frequently when the local unemployment rate is elevated. Yet the rate at which those parents lose cash payments does not decline substantially at such times.8 Instead, states terminate more families’ benefits for violations of the work requirements (see Figure 4-4). Those terminations probably contribute to fewer families’ being employed when they stop receiving cash payments, which leads to lower income. Between 2012 and 2018, the average cash income of families in the month they stopped receiving assistance was $485 (in 2019 dollars) when the unemployment rate was 6 percent or higher, compared with $697 during periods of lower unemployment.

Figure 4-4.

Employment and Access to Benefits for Families of Single Adults With Children Who Left TANF, by Local Unemployment Rate

When the unemployment rate is higher, families are more likely to stop receiving cash payments because they do not meet the work requirements, are less likely to be employed when they leave, and tend to have lower income when they leave.

Data source: Department of Health and Human Services, 2012 through 2018. See www.cbo.gov/publication/57022#data.

The analysis is generally limited to the families of single women with children who are subject to the federal work standard and who live in one of the nine states for which CBO has adequate data.

During economic downturns, policymakers are sometimes able to waive or reduce work requirements to lessen the negative effects on recipients’ employment and income. Federal law allows states to waive SNAP’s work requirement for able-bodied adults without dependents, for example, in areas where the unemployment rate exceeds 10 percent or is more than 20 percent above the national average. In addition, policymakers temporarily reduced the work requirements in TANF and SNAP during the 2020 recession.9 But TANF has fixed funding that constrains states’ ability to provide benefits to families for a longer period by relaxing its work requirements during economic downturns. Ninety-five percent of federal funding for TANF is provided through the State Family Assistance Grant, which remains at about $16.5 billion regardless of economic conditions. Thus, it is more difficult for states to fund extended TANF benefits in response to an increase in economic hardship than it is for states to expand assistance through SNAP or Medicaid, both of which generally receive enough federal funding to provide benefits to all eligible people.

When economic conditions deteriorate, some work supports probably become more effective than others at boosting employment. Subsidized employment and job training, for instance, can build skills at times when participants are unlikely to gain experience through an unsubsidized job because so few are available. (When jobs are plentiful, those work supports are probably more likely to replace unsubsidized employment.) In contrast, subsidized child care might boost employment less when jobs are scarce. That is because fewer of the people who would work more if they received subsidized child care would be able to find additional work.


1. Elderly people and people who receive disability benefits through Supplemental Security Income or Social Security are excluded from this analysis because they are exempt from work requirements.

2. Research has shown that many families whose income appears low enough for them to qualify for SNAP do not receive benefits because they do not adhere to the requirements for demonstrating that their income is low enough. However, that research does not examine whether the work requirement for able-bodied adults without dependents creates additional reporting requirements that lead to more eligible adults losing their benefits. See Colin Gray, “Leaving Benefits on the Table: Evidence From SNAP,” Journal of Public Economics, vol. 179 (November 2019), https://doi.org/10.1016/j.jpubeco.2019.104054; and Tatiana Homonoff and Jason Somerville, “Program Recertification Costs: Evidence From SNAP,” American Economic Journal: Economic Policy, vol. 13, no. 4 (November 2021), pp. 271–298, https://doi.org/10.1257/pol.20190272. When Arkansas instituted a work requirement for Medicaid recipients, it substantially increased the frequency with which some adult recipients had to demonstrate continued eligibility, which probably led to a loss of Medicaid coverage among working adults.

3. CBO used data from the 2014 Survey of Income and Program Participation to examine barriers to employment among recipients.

4. Parents were classified as caring for a young child if they had a child under the age of 6 and were a single parent or had a working spouse.

5. Gayle Hamilton and others, National Evaluation of Welfare-to-Work Strategies (submitted by Manpower Demonstration Research Corporation to the Department of Health and Human Services, December 2001), https://tinyurl.com/2p89s5ad.

6. Gretchen Rowe, Elizabeth Brown, and Brian Estes, SNAP Employment and Training (E&T) Characteristics Study: Final Report (submitted by Mathematica Policy Research to the Department of Agriculture, October 2017), https://go.usa.gov/xtgYQ.

7. For details on CBO’s analysis, see Appendix B.

8. Before 1993 (when the work requirement was imposed), AFDC recipients tended to stay in the program longer when the unemployment rate was high. See Hilary Hoynes, “Local Labor Markets and Welfare Spells: Do Demand Conditions Matter?” Review of Economics and Statistics, vol. 82, no. 3 (August 2000), pp. 351–368, https://doi.org/10.1162/003465300558812.

9. HHS notified state administrators that it would “grant reasonable cause exceptions… to the maximum extent possible” for 2020; see Department of Health and Human Services, Administration for Children and Families, “TANF-ACF-PI-2020-01 (Questions and answers About TANF and the Coronavirus Disease 2019 [COVID-19] Pandemic),” March 2020, https://go.usa.gov/xtggc. In addition, the Families First Coronavirus Response Act suspended the work requirement for ABAWDs through the month after the public health emergency declaration is lifted.

Chapter 5Options for Changing Work Requirements and Work Supports

Changing the stringency of work requirements and funding for work supports would affect the employment and income of the families served by Temporary Assistance for Needy Families, the Supplemental Nutrition Assistance Program, and Medicaid. By altering program participation and average benefit amounts, policy changes could also affect federal spending.1

The Congressional Budget Office has examined seven options that policymakers could pursue to expand or reduce work requirements or work supports. The options are assumed to take effect at the beginning of fiscal year 2023. This report gives CBO’s conclusions about the options’ broad effects on participation, employment, income, and spending. In many cases, the magnitude of the effects is highly uncertain.

Two Options That Would Expand Work Requirements

CBO examined two of the many ways in which lawmakers could expand work requirements to promote employment and reduce federal spending (see Table 5-1, top panel). In general, expanding work requirements would cause some newly covered recipients to work more, and most of those recipients would see their earnings rise by more than their benefits fell. But other recipients would be left without earnings or benefits because they did not comply with the new requirements. The reduction in benefits would lower federal spending if the work requirements were imposed on recipients of SNAP or Medicaid because the federal government funds a fixed percentage of those benefits.

Table 5-1.

Effects of Particular Options That Would Change Work Requirements

Source: Congressional Budget Office.

The options would be implemented at the beginning of 2023.

The hours-worked criteria limits ABAWDs to 3 months of benefits in any 36-month period if they are not working at least 20 hours per week or in job training.

ABAWD = able-bodied adult without dependents; SNAP = Supplemental Nutrition Assistance Program; TANF = Temporary Assistance for Needy Families.

a. Many of those people would be exempt from the expanded work requirements because they fall into certain categories (for example, caring for an incapacitated person) or, for SNAP, because they live in an area with high unemployment. For similar reasons, many of the adults potentially subject to the reduced work requirements are exempt from the work requirement under current law.

Require Able-Bodied Medicaid Recipients Without Dependents to Work

The first option that CBO examined would require all states to impose work requirements on their Medicaid recipients. Under this policy, which would be similar to the requirements imposed in Arkansas, able-bodied adults who have no dependents and who are between the ages of 19 and 49 would lose Medicaid coverage if they did not participate in work-related activities for at least 80 hours per month for three or more months over the course of a year. Adults would be exempt from the requirements if they have dependents, receive benefits from Supplemental Security Income or Disability Insurance, are pregnant, or have certain mental or physical health conditions. In addition to employment, hours spent in school, training for a job, searching for a job, or performing community service would count toward the 80-hour minimum. About 30 million people would be potentially subject to the work requirements each year, though many of them would qualify for an exemption.

This option would substantially increase the amount that people who lost Medicaid coverage would pay out of pocket for medical services, and it would increase employment very slightly. By CBO’s estimate, a very small portion of adults who were subject to the new requirements would work more hours to avoid losing insurance through Medicaid. Employment would not increase much because many Medicaid recipients have limited access to work supports and do not receive intensive case-management services. For every recipient who worked more, several recipients would lose Medicaid coverage because they failed to demonstrate compliance, and many of those adults would become uninsured. They would forgo some health care and pay more out of pocket for the care they did use, leaving them with less money to spend on nonmedical goods and services, on average. The increase in medical expenses from the loss of insurance coverage under this option is far more certain than the change in employment.

This option would decrease federal spending by about $135 billion from 2023 to 2031, CBO estimates. Spending would fall by about $15 billion per year, on average, because about 2.2 million adults would lose Medicaid coverage. That reduction in enrollment represents a substantial portion of the adults who would be subject to the work requirements. (Many other adults would be exempt.)

Expand SNAP’s Work Requirement While Providing More Employment and Training Services

Under current law, only able-bodied adults who do not have dependents and who are under the age of 50 (ABAWDs) risk losing SNAP benefits if they do not spend 20 hours a week, on average, participating in work-related activities. This second option that CBO examined would potentially extend that requirement to about 10 million able-bodied adults between the ages of 18 and 49 who have dependents, though many of them would qualify for an exemption.2 (Other members of those adults’ households would not be at risk of losing benefits because of the requirement.) Participants would receive additional employment and training services to boost their chances of finding a job.

This option would reduce federal spending if the decrease in spending on SNAP benefits exceeded the increase in the program’s administrative costs, which include the cost of providing additional employment and training services. If this option increased that spending substantially, it could push up spending overall. Spending on SNAP benefits would fall primarily because of reductions in benefits for adults who did not comply with the work requirement. (Spending on benefits would also fall for adults whose earnings rose because they worked more to meet the requirement.) To help adults comply with the work requirement, the option would substantially increase federal funding for SNAP’s employment and training programs and would require states to offer case-management services to all adults subject to it. The states would not be required to match any of the additional federal funding for such programs. In 2019, the federal government provided $124 million in such unmatched funding for employment and training programs.

In all likelihood, this option would modestly increase employment among the additional recipients who would become subject to the expanded work requirement. Its effect on employment would probably exceed that of the option that would impose a work requirement on Medicaid recipients because SNAP recipients would be offered more work supports. (Some states would probably not be prepared to offer those supports for several years.) But the evidence of employment increases from the work requirement for ABAWDs suggests that this option would boost employment only a little if recipients were not provided with intensive case-management services, such as assistance provided in person by a trained caseworker. The effects of this option on employment and earnings are also highly uncertain because of the lack of research on SNAP’s employment and training programs.

Under this option, income would decline, on average, for households with adults who were not exempt from the work requirement. Many of those households have few other resources. In particular, only about 53 percent of SNAP households with children had earnings in 2019. Many of the nonworking households whose benefits would decrease under this option would have very low income. For example, a single parent with two children who was not working would have his or her maximum monthly SNAP benefit reduced from about $658 to around $459 in 2022 (excluding the emergency allotments that will expire once the current public health emergency declaration is lifted); in many cases, that payment would still be the family’s primary source of income because few families receive cash payments from TANF. Other adults would probably work more under this option, but the size of the increase in their earnings from the combination of the expanded work requirement and work supports is very uncertain. According to CBO’s assessment, that increase in earnings would be smaller than the reduction in benefits, thus decreasing families’ income, on average.

Two Options That Would Reduce Work Requirements

Lawmakers could reduce work requirements in many ways to boost the income of the poorest families. This analysis examines two of them (see Table 5-1, bottom panel).

In general, if work requirements were lessened, many newly exempt recipients would receive benefits for a longer period, and fewer of those recipients would work.3 During periods of high unemployment, the boost in income from extended benefits would probably be larger than the reduction in earnings from less work. Generally, the increase in benefits would lead to higher federal spending for SNAP and Medicaid but not for TANF.

Eliminate Hours-Worked Criteria for ABAWDs’ Continued SNAP Eligibility

Under this option, able-bodied adults without dependents would no longer be limited to 3 months of benefits in any 36-month period if they were not working at least 20 hours per week or in job training. Eliminating that hours-worked criteria would increase spending on SNAP by boosting the number of people receiving benefits. About 5 million fewer people would be potentially subject to the work requirement each year, though many of them are exempt from it under current law. This option would increase federal spending by providing benefits to more people.

Among the people who would no longer be subject to SNAP’s work requirement, this option would probably decrease employment slightly—but it would substantially increase income for the many SNAP recipients who would no longer lose benefits because of the work requirement. A much smaller number of people would have lower income from working less because they would no longer be at risk of losing SNAP benefits. But the specific size of the decrease in employment (and the resulting effect on earnings) is highly uncertain. In contrast, this option would clearly reduce the number of nonworking ABAWDs with very low income by providing them with more assistance to purchase food, and it would probably increase ABAWDs’ income, on average.

Prohibit States From Imposing TANF’s Work Requirements on Single Parents With a Child Younger Than One

Although states are allowed to exclude from TANF’s work requirements single parents with a child who has yet to turn one, about half of states choose to apply those requirements to some or all of those parents. Under this option, states would not be allowed to reduce the benefits given to those parents for insufficient participation in work-related activities. That change would reduce the number of TANF recipients whose families were potentially subject to the work requirements by about 50,000 each year, CBO estimates, though many of those families would have been exempt under current law. The resulting increase in the number of TANF recipients would not change federal funding for the program, which is set at a fixed amount.

To better understand the potential implications of this option on employment and income, CBO examined the experience of TANF recipients in Alabama. In October 2017, that state stopped exempting from TANF’s work requirements the single parents of children between the ages of 6 months and 11 months.4 Even though this option would make a change in the opposite direction (eliminating instead of imposing the work requirements on parents with a young child), the effects are probably comparable in size. CBO used data provided by the Department of Health and Human Services for this particular analysis.

This option would decrease employment among the parents subject to it by slowing the rate at which they found jobs. Alabama’s application of its work requirement to single parents of children between the ages of 6 months and 11 months substantially increased their employment.5 Although raising the age threshold back to 12 months in Alabama might lower employment to its previous level, this option would be unlikely to have an effect that large on employment nationwide. The unemployment rate in Alabama was about 4 percent when it implemented the change in work requirements, whereas this option would probably exempt mothers in areas of the country where jobs were scarce as well as in areas where they were plentiful.

Alabama’s application of its work requirement to single parents of children under age one appears to have increased their average cash income during the year after they entered TANF. Average cash income increased because the additional earnings for some families were larger than the loss of cash payments for families who did not meet the work requirement. This option, which would reverse such a policy in a number of states, would have effects in the opposite direction.

Nevertheless, CBO expects that this option would not change recipients’ income much, on average. Although raising the age threshold back to 12 months in Alabama might lower average cash income in that state to its previous level, the net effect of such a change on earnings nationwide would probably be much smaller, for two reasons. First, jobs would not be as plentiful in some of the states that would be forced to raise the threshold, which would decrease the number of recipients that could find work under current law and thus reduce the income loss under this policy. Second, Alabama provides smaller cash payments than most other states, so families who, because of this option, stayed in the program longer in those other states would see a larger increase in payments.

This option would reduce the number of families with very low income, in CBO’s estimation. Many of the parents who would not work regardless of the requirement would be spared from having very low income because they would remain eligible for cash payments. Other parents would have lower income because they would choose to work less in the absence of the work requirement, but eligibility for cash payments would spare many of them from having very low income.

Two Options That Would Expand Work Supports

To promote employment and increase the income of poor families, lawmakers could expand work supports for participants in means-tested programs. In general, expanding work supports would increase employment and boost consumption because it would reduce recipients’ spending on child care and job preparation as well as increase their earnings. (The effect on consumption would be greater than that from expanding work requirements.) Expanding work supports would push up federal spending if the funding was not diverted from other services, however. Although increases in earnings from expanded work supports might reduce spending on means-tested benefits, those reductions would be unlikely to offset a large portion of the spending on the work supports. Among the many ways that lawmakers could expand work supports, CBO examined two options (see Table 5-2, top panel).

Table 5-2.

Effects of Particular Options That Would Change Work Supports

Source: Congressional Budget Office.

The options would be implemented at the beginning of 2023.

SNAP = Supplemental Nutrition Assistance Program; TANF = Temporary Assistance for Needy Families.

a. For most programs, reducing work supports would decrease people’s income and reduce federal spending.

Increase Funding for Subsidized Child Care to Support More TANF, SNAP, and Medicaid Recipients

To make subsidized child care available to more parents in TANF, SNAP, and Medicaid, this option would increase federal funding for the discretionary component of the Child Care Development Fund by $3 billion per year (which represents 50 percent of nonsupplemental funding).6 In response to that increase in federal funding for child care, states could reduce the amount paid by the many single parents in those three programs who already receive subsidies. In addition, states could use a portion of the additional federal funding to assist some of the single parents in those programs who pay for unsubsidized child care despite meeting the federal requirements for the CCDF. By expanding subsidized child care for hundreds of thousands of parents each year, this option would increase federal spending by about $26 billion from 2023 to 2031, CBO estimates.

Under this option, states would provide subsidies to more program participants and increase the size of the subsidies, which would reduce participants’ share of the cost of child care and probably help them obtain higher-quality care. Beginning in 2018, lawmakers increased funding for the CCDF by 41 percent, and by 2019 the states had used that funding to raise spending about 34 percent above its level in 2017. As a result, the number of subsidy recipients rose by about 7 percent, and the number of children on waitlists for subsidized child care fell by about 53 percent.

States primarily increased their spending in 2019 by boosting child care subsidies. States increased the subsidies by decreasing the amount that parents have to contribute and by raising the maximum rate a child care provider can charge and still be eligible for a subsidy.7 Raising the maximum rate allowed parents to select more expensive providers and thus probably obtain higher-quality care.

This option would increase employment slightly among the parents who received more benefits through the CCDF. The increase in employment would be concentrated among recipients who would not have received a subsidy under current law, and the 2018 expansion indicates that they would constitute a modest portion of the recipients affected by the expansion. In addition to enabling more parents to work immediately, this option would allow more parents to undertake job training and job searching by making more subsidies available to parents who are preparing to work.

Some recipients of TANF, SNAP, and Medicaid would have substantially more resources to spend on goods and services other than child care under this option. In addition to the boost in earnings from more employment, parents would tend to spend less on child care. The number of parents paying for unsubsidized care would fall, and some states would use the additional funding to reduce copayments tied to subsidized care, which would benefit the parents who were already receiving subsidies.

Increase Funding for TANF’s Work Supports When Unemployment Is High

Under current law, federal funding for TANF changes little in response to economic conditions.8 To develop the workforce more during periods of weak economic growth—when jobs are scarce—this option would increase a state’s TANF funding when its unemployment rate was high. Specifically, funding would rise by 10 percent when a state’s unemployment rate was above 6 percent, 20 percent when it was above 7 percent, and 30 percent when the rate exceeded 8 percent.

The increase in funding would be focused on periods when unemployment was widespread. For example, if this option had been in effect during 2009 and 2010, when the national unemployment rate peaked at 9.9 percent, it would have increased funding in 47 states by a total of $9 billion. In contrast, the option would have increased funding in only one state in 2018 and 2019, when the national unemployment rate fell to a low of 3.6 percent.

The additional funding would be used to cover 80 percent of increases in spending on job training, subsidized employment, and recurring cash payments that occur during the period of high unemployment. Other uses of the funds would be prohibited. Constraints on how the funding is spent would be similar to those that accompanied the emergency contingency fund used in the wake of the 2007–2009 recession. This option would affect tens of thousands of program participants each year, and it would increase federal spending by $5 billion from 2023 to 2031, CBO estimates.9

States would cover the remaining 20 percent of additional spending on job training, subsidized employment, and recurring cash payments, but they would not have to increase their own total spending on TANF. Instead, they could offset their additional spending on those services by spending less on other TANF services.10 Thus, even if states found it difficult to allocate more of their own revenues to TANF, they could still receive the additional funding. Program requirements would stipulate that the funding be spent promptly on services that increase participants’ job skills or income.

This option would probably raise employment eventually by bolstering participants’ job skills. But the initial effect on employment in each state would depend on which services it chose to expand. On the one hand, employment could fall because some recipients of job training would have worked otherwise and because additional spending on cash payments would make some people work less. But those two effects would probably be modest because few recipients would be able to find work during periods of high unemployment. On the other hand, employment could rise initially because some recipients of subsidized employment would not have been hired had the employer had to bear all the cost.11 Eventually, the skills that recipients developed through job training and subsidized employment would probably help them find slightly more work, on average, across the various state programs.12

This option would increase recipients’ income in the short term and most likely in the long term as well. Additional spending on subsidized employment and cash payments would boost their income immediately, and the skills they learned through subsidized employment and job training would probably lead to higher future earnings, on average.

One Option That Would Decrease Work Supports

One way to reduce federal spending would be to decrease the work supports provided through means-tested programs. In general, decreasing work supports would reduce recipients’ employment, lower their consumption (by causing them to spend more on child care and job preparation), and decrease their earnings. If, instead, funding was shifted from work supports to cash payments for nonworking families, the number of families in deep poverty could be reduced. Federal spending would be unchanged in that case, however (see Table 5-2, bottom panel). CBO analyzed that option—even though it is not representative of the typical effects of decreasing work supports—because it is the only recent proposal for reducing work supports that the agency identified.

Require States to Spend at Least 20 Percent of TANF’s Funding on Cash Payments

This option would require states to spend at least 20 percent of TANF’s total funding, excluding administrative and systems costs, on recurring cash payments for recipients—reversing a modest portion of the downward shift in that portion of funding that has occurred over the past 25 years. Hundreds of thousands of TANF recipients would be affected each year by that change, and their income would increase substantially as a result. This option would not affect federal spending.

To cover the increased cost of the cash payments, some states would probably reduce spending on other TANF-funded services; other states might spend more of their own funds on TANF as a whole. States would probably offset most of the increase in spending on cash payments by reducing spending on work supports and social services. In 2019, states that spent less than 20 percent of their TANF funding on recurring cash payments spent about 40 percent on subsidized child care and an additional 15 percent on other work supports. Furthermore, they allocated about 30 percent toward social services for children, including initiatives aimed at boosting educational attainment of parents, promoting family preservation and reunification, and providing guidance for parents on child health and development.

This option would expand the resources of low-income families by increasing the amount of cash assistance they received. If this option had been in effect in 2019, the 27 states that apportioned less than 20 percent of their TANF funding to cash payments would have had to increase spending on those payments by $1.1 billion, which constitutes an 18 percent increase in the total cash payments provided by all states—and 10 of them would have had to more than double their spending on those payments. States could provide more cash assistance in two ways: by paying families more or by making payments to more families (a task they could accomplish by loosening eligibility standards and conducting outreach to increase the percentage of eligible families that apply for benefits).

The increase in cash payments would substantially boost the income of some of the poorest families in the affected states. In the 27 states that spent less than 20 percent of their TANF funding on cash assistance, about 35 percent of the families who receive those payments report being in deep poverty. One reason income is so low in those states is that the states provide little cash assistance to recipients—families in those states receive $321 per month, on average, compared with $508 per month in other states. In addition to making larger payments to some recipients, the 27 states could provide benefits to more families in poverty. Currently, those states provide cash assistance to only about 6 percent of the families of single parents who are in poverty.

This option would decrease employment by reducing work supports and lessening recipients’ incentive to work. The number of people who would no longer work would probably be small compared with the number of people who would receive larger cash payments. The size of the change in employment is highly uncertain, though, because it would depend on how states chose to increase their spending on cash assistance. Increasing the amount of the payments to nonworking recipients and boosting the number of nonworking recipients who got those payments would decrease employment. But some states might choose to make cash payments available to working recipients for longer periods, which could increase employment.


1. In some cases, changes to a program can affect spending on other programs in addition to spending on the changed program. For example, policies that increased the cash payments provided through TANF would reduce the benefits provided through SNAP. Changes in spending on other programs are not included in the estimates of costs provided in this analysis.

2. States could still waive the work requirement if their unemployment rate was high. Certain categories of people (pregnant women, the parents of children younger than 6, and people who care for an incapacitated person) would be exempt from the requirement. States would have some discretion to exempt individuals otherwise subject to the requirements.

3. In addition to possibly boosting income, on average, these options would allow nonworking people to continue receiving benefits, thus affording them a steady income stream. Compelling those people to try to find and keep employment, by contrast, could result in their periodically having no income.

4. The only time one of the 14 states in CBO’s data from HHS increased the exemption to 11 months appears to be in 2014. That change is difficult to evaluate because during the months leading up to it, the state rarely enforced the work requirement on families with young children.

5. For a detailed description of CBO’s results and methods, see Appendix B.

6. Increases in funding larger than $3 billion would probably be less targeted at parents in TANF, SNAP, and Medicaid because a greater portion of the additional subsidies would go to parents whose income is too high for them to qualify for those programs.

7. Karen Schulman, Early Progress: State Child Care Assistance Policies 2019 (National Women’s Law Center, October 2019), https://tinyurl.com/2p8nj6n6 (PDF, 2.3 MB).

8. Marianne P. Bitler and Hilary W. Hoynes, “The State of the Social Safety Net in the Post-Welfare Reform Era,” Brookings Papers on Economic Activity, no. 2 (Fall 2010), https://tinyurl.com/2p8bpv7v.

9. For a description of how CBO forecasted a distribution of national unemployment rates, see Michael McGrane, A Markov-Switching Model of the Unemployment Rate, Working Paper 2022-05 (Congressional Budget Office, March 2022), www.cbo.gov/publication/57582. Using historic tendencies in states’ unemployment rates, CBO estimated a distribution of those rates around each national unemployment rate. CBO’s models are designed to forecast the number of states that would have high and low unemployment rates but not which states those would be.

10. Some of the increase in employment and income from more spending on job training and subsidized employment could be offset by the states’ spending less on other work supports.

11. Experience indicates this option might boost employment among people who are not in TANF. In the wake of the 2007–2009 recession, states appear to have used the emergency contingency fund to provide subsidized employment to many people who were not in TANF. See Mary Farrell and others, Subsidizing Employment Opportunities for Low-Income Families: A Review of State Employment Programs Created Through the TANF Emergency Fund (submitted by MDRC to the Department of Health and Human Services, December 2011), https://tinyurl.com/2p86bdjr.

12. Some job training and subsidized employment programs have achieved persistent increases in employment and earnings, but others have not (see Chapter 3 for details). That evidence indicates that this option would be more likely to increase employment and earnings if job training is focused on occupations that have the potential for career advancement.

Appendix ASources of Income for Families 24 Months After They Start Receiving Cash Assistance

The Temporary Assistance for Needy Families (TANF) program emphasizes reducing recipients’ dependence on government benefits by promoting self-sufficiency through employment. Its work requirements, which were enacted into law in 1996 as part of the Personal Responsibility and Work Opportunity Reconciliation Act (PRWORA), are intended to help families more quickly replace cash payments with earnings. By law, states are required to ensure that most recipients are working within 24 months of first receiving cash assistance. To meet that requirement, states impose work requirements on most parents before they have been in the program for 24 months.

This appendix describes how the Congressional Budget Office used data from the Survey of Income and Program Participation (SIPP) to evaluate the extent to which TANF’s work requirements increased the portion of families that were able to replace cash payments with earnings in two years.1

The SIPP Data That CBO Used

The SIPP is a nationally representative survey of U.S. households that is designed as a series of panels. For each panel, the Census Bureau continually interviews survey respondents in the same households—every four months in most panels—for periods ranging from 32 months to 64 months. CBO analyzed the SIPP panels that began in 1990, 1991, 1992, 1993, 1996, 2001, 2004, 2008, and 2014. The largest panel includes around 50,000 households, and the smallest one includes about 15,000 households.

In each interview, survey respondents were asked about their employment, income, and participation in means-tested programs, including Aid to Families With Dependent Children (AFDC, the program that preceded TANF and was in place until 1997), TANF, and the Supplemental Nutrition Assistance Program (SNAP). Because far fewer respondents participate in TANF than SNAP or AFDC, estimates for that program are less certain.

Researchers have shown that participation in AFDC/TANF is sometimes misreported in the SIPP.2 To improve the accuracy of its analysis, CBO’s sample focused on single mothers—the group that constitutes the vast majority of adults who receive benefits from TANF, according to administrative data. Single fathers and married parents who report being in TANF are probably more likely to have misreported their participation.

CBO’s Approach to Analyzing the SIPP Data

CBO estimated the effects of work requirements on families’ sources of income by comparing changes in those income sources for single mothers in TANF and for single mothers in SNAP in selected years in the 1990s and 2000s. That period encompasses the years during which welfare reform was most active (the mid-1990s). In those years, policymakers undertook a series of actions—including a large expansion of work requirements in AFDC/TANF—to encourage employment. Several states expanded work requirements in AFDC starting in 1994; by 1997, many states had implemented TANF’s broad work requirements. Thus, CBO compared sources of income in 1993 (the final year before those changes occurred) with those in 1997 and in later years for which data were available.

For sources of income, CBO focused on cash payments and income from employment 24 months after single mothers first started receiving cash payments from either AFDC or TANF. CBO relied on self-reported data to mark the start of assistance. Many single mothers were already receiving payments from AFDC/TANF or SNAP in the first month they responded to the survey; for those recipients, the date of their first response was considered their start of assistance.

CBO aggregated responses about sources of income 24 months after assistance began across all participants in a particular program within a panel. Aggregating fewer months of data (by taking an average for each year, for example) would result in imprecise estimates because few respondents entered TANF in some years. To determine which date to associate with each data panel, CBO took the average of the dates at which respondents entered the program and added 12 months to get the midpoint between the start of assistance and the period in which the agency evaluated sources of income. That date determined whether the estimates from that panel would be included in the period before or after welfare reform. For the 1992 panel, that date was at the end of January 1993; for the 1996 panel, it fell in the middle of May 1997. Thus, the 1992 panel was the last one in the prepolicy period, and the 1996 panel was the first one in the postpolicy period.

Families’ sources of income could have changed from 1993 to 1997 as a result of factors other than work requirements, such as the expansion of the earned income tax credit (EITC) or the strengthening economy. To evaluate whether the changes in sources of income were caused by the work requirements, CBO compared those changes for TANF recipients and SNAP recipients. (SNAP did not impose work requirements on single mothers, but those women were eligible for the EITC just like their counterparts in TANF were.) Specifically, CBO focused on single mothers in SNAP who were not in TANF. Single mothers enrolled in both programs were excluded from the SNAP sample because they would have been subject to TANF’s work requirements. That approach also enabled CBO to determine whether the strengthening economy caused the changes in the sources of income: Single mothers in both TANF and SNAP would probably have responded to economic conditions similarly because those participants did not differ much in their age, educational attainment, and other characteristics measured in the data.

Families’ Sources of Income 24 Months After Receiving Cash Assistance

The AFDC/TANF work requirements substantially reduced the number of single mothers who continued to receive cash payments two years after entering the program. The prevalence of those longer spells in AFDC/TANF fell by half from 1993 to 1997 and remained low in subsequent years (see Figure A-1). Although the prevalence of longer spells fell by 30 percentage points among AFDC/TANF participants, it fell by only 8 percentage points for SNAP recipients who were not in TANF. That contrast suggests that the drop in the prevalence of longer periods of cash assistance was mostly the result of changes to AFDC/TANF policies, not the EITC or strengthening economy.

Figure A-1.

Share of Families Continuing to Receive Assistance After 24 Months, by Program and Year

Percent

Single mothers in TANF, who were subject to work requirements starting in the mid-1990s, tended to stop receiving benefits much sooner than they had before. In contrast, the share of single mothers in SNAP who continued to receive benefits remained about the same.

Data source: Survey of Income and Program Participation. See www.cbo.gov/publication/57702#data.

The analysis is limited to unmarried mothers between the ages of 18 and 61 who are not students and are not receiving disability benefits.

Data are presented only for the years in which the sample size provides sufficient precision.

AFDC = Aid to Families With Dependent Children; SNAP = Supplemental Nutrition Assistance Program; TANF = Temporary Assistance for Needy Families.

a. Nearly all TANF recipients also receive benefits from SNAP.

The AFDC/TANF work requirements appear to have temporarily decreased the number of single mothers who continued to receive nutrition assistance two years after entering SNAP. Mothers who entered both AFDC/TANF and SNAP 24 months earlier were 19 percentage points less likely to still be receiving benefits through SNAP in 1997 than they were in 1993, even though SNAP did not impose work requirements on families with children (see Figure A-2). One likely reason for the decline is that the work requirements added to AFDC increased the earnings of some participants enough that they no longer qualified for SNAP. But many of the parents who stopped receiving benefits from both programs appeared to still be eligible for SNAP because they were not employed. In 1993, about 58 percent of nonworking single mothers who left AFDC/TANF also left SNAP at the same time; that number fell to only 41 percent in 1997.3 In subsequent years, about 17 percent of nonworking single mothers who left TANF also left SNAP at the same time, and the prevalence of longer spells in SNAP rose back to its pre-PRWORA level, according to CBO’s analysis of SIPP data.

Figure A-2.

Share of Families Continuing to Receive SNAP Benefits After 24 Months, by Program and Year

Percent

After the imposition of work requirements, single mothers who received benefits through both SNAP and TANF tended to stop receiving SNAP benefits much sooner than they had before—even though continued receipt of SNAP benefits was not subject to work requirements.

Data source: Survey of Income and Program Participation. See www.cbo.gov/publication/57702#data.

The analysis is limited to unmarried mothers between the ages of 18 and 61 who are not students and are not receiving disability benefits.

Data are presented only for the years in which the sample size provides sufficient precision.

AFDC = Aid to Families With Dependent Children; SNAP = Supplemental Nutrition Assistance Program; TANF = Temporary Assistance for Needy Families.

TANF’s work requirements led to a higher employment rate for single mothers two years after they entered the program. The employment rate for mothers who had entered TANF two years earlier rose by about 14 percentage points from 1993 to 1997 (see Figure A-3). The strong economic conditions of the late 1990s do not appear to have had a large direct effect on employment because the employment rate among single mothers in SNAP who were not subject to TANF’s work requirements changed little. In subsequent years, employment among single parents who had participated in TANF remained elevated, whereas the employment rate for their counterparts in SNAP remained relatively low.

Figure A-3.

Share of Mothers Employed 24 Months After Family Began Receiving Assistance, by Program and Year

Percent

After work requirements were added to cash assistance programs, the single mothers in those programs tended to find employment more quickly. That was not the case for single mothers in SNAP, who were not subject to work requirements.

Data source: Survey of Income and Program Participation. See www.cbo.gov/publication/57702#data.

The analysis is limited to unmarried mothers between the ages of 18 and 61 who are not students and are not receiving disability benefits.

Data are presented only for the years in which the sample size provides sufficient precision.

AFDC = Aid to Families With Dependent Children; SNAP = Supplemental Nutrition Assistance Program; TANF = Temporary Assistance for Needy Families.

a. Nearly all TANF recipients also receive benefits from SNAP.

The increase in employment for single parents two years after entering TANF was smaller than the decrease in their continued receipt of cash assistance, which indicates a rise in the number of families with neither earnings nor cash payments. Even though employment rates rose by 14 percentage points from 1993 to 1997, the percentage of single parents still receiving cash payments two years after entering the program fell by 30 percentage points. Those results are consistent with other evidence indicating that TANF’s work requirements contributed to the increase in the number of families with neither earnings nor cash assistance.


1. CBO also used the SIPP to evaluate the characteristics of recent participants in TANF, the Supplemental Nutrition Assistance Program, and Medicaid as well as the types of child care used by single parents in those programs.

2. Bruce D. Meyer, Wallace K. C. Mok, and James X. Sullivan, “Household Surveys in Crisis,” Journal of Economic Perspectives, vol. 29, no. 4 (Fall 2015), pp. 199–226, http://doi.org/10.1257/jep.29.4.199.

3. Another study also found that the transition from AFDC to TANF led to large reductions in SNAP participation. See Janet Currie and Jeffrey Grogger, “Explaining Recent Declines in Food Stamp Program Participation,” Brookings-Wharton Papers on Urban Affairs, vol. 2 (2001), pp. 203–244, www.jstor.org/stable/25058786.

Appendix BThe Effects of Work Requirements on People’s Employment and Income in Recent Years

Within the past few years, states have imposed work requirements more widely, including during periods of high unemployment and on participants in the Temporary Assistance for Needy Families (TANF) program who have children under the age of one. To continue receiving benefits, program participants must work or participate in work-related activities.

The Congressional Budget Office used administrative data from the Department of Health and Human Services (HHS) to estimate the effects of work requirements in TANF on recipients’ employment and income in recent years and during periods of high unemployment.1 The analysis presented here supplements the limited evidence on how work requirements have affected employment and income since 2000.

The HHS Data That CBO Used

The administrative data that HHS collects give a detailed description of the economic situations for many families in TANF and track changes in those situations. State agencies that administer the TANF programs report the data to HHS to demonstrate compliance with federal requirements, including the work standard and the five-year limit on cash payments.

This analysis is based on data from 2012 to 2018 for TANF recipients in 14 states.2 The states that CBO included in the sample provided records for all their TANF recipients (which no states did before 2012), and family identifiers enabled CBO to observe how the circumstances of those families changed over time while they were in TANF. CBO used the fine details in the data, such as recipients’ dates of birth, to confirm that family identifiers were used consistently in the 14 states.

CBO could not use data for the other 36 states because they were subject to at least one of three limitations.

  • First, only 28 states and the District of Columbia provided data for all their recipients. The other 22 states provided data for a small portion of recipients. Because those samples are redrawn each month, CBO cannot track changes in the circumstances of those recipients and the length of their spells—or months of consecutive assistance—in TANF.
  • Second, of the 28 states that provided data for all their recipients, several frequently transferred recipients to other cash assistance programs for which CBO does not have data.
  • Third, CBO could not confirm that the family identifiers were unique and used consistently for several states; use of data from those states might have obscured changes in the circumstances of recipients in those states.

The main data file from HHS includes monthly records of families’ participation in work-related activities, earnings, other sources of income, and demographic traits. The sources of cash income included are earnings, child support, and benefits from TANF, Social Security, Supplemental Security Income, unemployment insurance, and workers’ compensation. Also included are benefits received from the Supplemental Nutrition Assistance Program (SNAP) and subsidies for child care.

HHS also provided data files that record the circumstances of TANF recipients when they leave the program, including their reason for leaving and their income in the month they stop receiving cash payments. Those data allowed CBO to confirm that the TANF spell ended in the month following the last month that the family appeared in the main file.

Effects of Work Requirements When the Unemployment Rate Is High

During periods of adverse economic conditions, work requirements can have negative effects on program participants. Those requirements can substantially lessen program participation while boosting employment only a little and reducing recipients’ average income. When work becomes harder to find, it can be difficult for recipients to meet those requirements through employment. If the work requirements are not eased at those times, more families are at risk of losing benefits before they can find work.

To examine how the effects of work requirements on employment vary with economic conditions, CBO looked at recipients’ employment, income, and program participation over a certain period. Specifically, CBO estimated how those factors varied with the unemployment rate across 708 counties in 9 selected states between 2012 and 2018.3 To explore whether that variation was the result of TANF’s work requirements, CBO examined whether families were more likely when the unemployment rate was high to have their TANF spells end because they violated the work requirements.4

CBO found that TANF’s work requirements probably boosted recipients’ employment less when the unemployment rate for their county was high. In particular, only 4.8 percent of job seekers in TANF found employment each month when their county’s unemployment rate was 6 percent or higher, compared with 8.6 percent when the rate was below 6 percent. That reduction in employment occurred even though states continued to impose work requirements on most adult TANF recipients. About 36 percent of those recipients were exempted from the work requirements when the unemployment rate was 6 percent or higher. (When the rate was below 6 percent, about 31 percent of recipients were exempted.)

In addition, CBO found that TANF’s work requirements probably reduced the average income of recipients during periods of high unemployment by halting cash payments to more families with little or no earnings. The percentage of spells ending because of work-requirement violations rose from 6 percent when the unemployment rate was below 6 percent to 18 percent when the unemployment rate was 6 percent or higher. Those violations reduced the extent to which additional cash payments offset the reduction in earnings from less employment. Families only received cash assistance for about one month longer, on average, when the unemployment rate was 6 percent or higher than when the unemployment rate was below 6 percent, even though job seekers in TANF took nine more months to find a job, on average, when the unemployment rate was higher, CBO estimates.

During periods of high unemployment, TANF’s work requirements caused a large portion of the families leaving the program to have very low income. When the unemployment rate was 6 percent or higher, 18 percent of families leaving TANF were expelled for violating the work requirements, CBO found, and their average cash income was $153 in the month after they lost benefits, which is less than 10 percent of the poverty threshold for a family of three. If SNAP benefits are included, average monthly income for those families rose to $569. In contrast, when the unemployment rate was lower than 6 percent, the share of families leaving TANF that were expelled for violating the work requirement was 5 percent.

Effects of Work Requirements on Families With Young Children

Although states are allowed to exclude from the work standard single parents with a child under the age of one, about half of states choose to apply work requirements to some or all of those parents. CBO examined how expanding work requirements to those parents affected their employment and income by comparing trends in outcomes for families who were subject to the expansion with families who were not. CBO found that expanding work requirements to parents who entered TANF with a child younger than one increased their employment but had little effect on their income, on average.

How CBO Analyzed the Effects of Work Requirements on Families With Young Children

To estimate the effects of work requirements on the employment and income of families with children younger than one year of age, CBO examined the recent experiences of TANF recipients in Alabama. In 2017, Alabama was one of 25 states that provided work-requirement exemptions to single parents whose youngest child was less than one year old. At the beginning of 2018, though, Alabama stopped providing exemptions to nearly all single parents who had a child between the ages of 6 months and 11 months. CBO used two approaches to examine the effects of that policy change on the employment and income of TANF recipients.

First, CBO examined the effects on families while they were in TANF and in the month after they left the program. To do that, the agency compared the employment and income of single parents who had a child between the ages of 6 months and 11 months in 2018 with the employment and income of single parents who had a child in that age range in 2017. The agency compared those changes in employment and income with changes in those outcomes among single parents whose youngest child was older to adjust for differences that were driven by other factors (such as economic conditions) instead of the expansion of the work requirement. That approach illustrates the effects of the expansion, but it does not fully capture the reduction in income from lost benefits because families remained in the sample for only one month after they stopped receiving cash payments. CBO thus used a second, more comprehensive, approach, which examined employment and income in the months after families left TANF as well as while they were in it.

Using that second approach, CBO estimated the effects of Alabama’s work-requirement expansion over the year following families’ entry into TANF in 2018. Whether a family was subject to the expansion of the work requirement was determined by the age of its youngest child when it entered TANF. Nearly all single parents with a child between the ages of 6 months and 11 months in 2018 were subject to the expansion, as were single parents with a child between the ages of zero and 5 months if they remained in the program after that child reached 6 months of age.

CBO compared the employment and income in 2017 and 2018 of families whose youngest child was not older than 11 months when the family entered TANF. For families that were in TANF for less than a year, employment and income for the remainder of the year were set equal to their value in the month after those families left the program.5 As with the first approach, CBO compared the changes in employment and income among the parents of young children with those changes among the parents of older children in TANF. By examining changes in employment and income among single parents of older children, CBO intended to capture the effects of changes in economic conditions or policy changes other than the expansion of the work requirement.6

What CBO’s Analysis Found

Using the first approach, CBO determined that Alabama’s expansion of TANF’s work requirements in 2018 increased the employment rate of single parents whose youngest child was between the ages of 6 months and 11 months by 11 percentage points during the months they were in the program (see Figure B-1). Employment rose because families started searching for work sooner and found jobs faster when subject to the work requirement.

Figure B-1.

Employment for Single Parents in Alabama’s TANF Program, by Age of Youngest Child in Months and Whether the Family Could be Exempt From the Work Requirement

Percent

When Alabama began imposing work requirements on parents whose youngest child was between the ages of 6 months and 11 months, the employment rate of parents in that group rose by about 11 percentage points. Employment also rose among parents of children between the ages of 12 months and 16 months, apparently because some of those parents were subjected to the work requirements for more months.

The employment rate at program exit stayed roughly the same from 2017 to 2018. That shows that the employment increase was not driven by nonworking parents’ being expelled from the program for violating the work requirement.

Data source: Department of Health and Human Services. See www.cbo.gov/publication/57702#data.

The analysis is limited to the single-parent families that Alabama could include in the work standard.

TANF = Temporary Assistance for Needy Families.

Fitted values give the trends in employment rates by the age of the youngest child.

Parents whose youngest child was 4 months or 5 months old are excluded from this analysis because they were occasionally subject to the work requirement in 2018. Parents whose youngest child was between 12 months and 16 months of age were likely to have been subjected to the work requirement for more months because of its expansion. Affected data are shown as stand-alone dots in this figure.

Employment also rose among parents of children between the ages of 12 months and 16 months. That change appears to be the result of the expansion’s having subjected some of those parents to work requirements for more months, because the employment rates differed only among parents who entered TANF before their youngest child turned 12 months old.

The rise in employment was probably the result of the work requirement and not other possibilities, such as changes in the sample studied, the presence of other policies (subsidized child care, for instance), or economic conditions. Specifically:

  • The employment rate at program exit stayed roughly the same from 2017 to 2018. That shows that the employment increase was not driven by nonworking parents’ being expelled from the program for violating the work requirement.
  • The percentage of parents with a child between the ages of 6 months and 11 months who received subsidized child care increased by 10 percentage points, but that rise was the result of more employment, not the cause of it. In both 2017 and 2018, parents generally only received subsidized child care if they were working or searching for a job. The expansion of the work requirement does not appear to have changed their access to subsidies: Alabama gave subsidies to about 45 percent of parents who were in the labor force and had a child between the ages of 6 months and 11 months both before and after the expansion. That indicates that the employment increase was not driven by an increase in subsidized child care.
  • If the employment increase had been driven by changes in other policies or economic conditions, then employment would probably also have increased among single parents who were not subject to the expansion because their children were slightly older when they entered TANF. That did not happen, which shows that the employment increase was probably not driven by changes in other policies or economic conditions.

The expansion of the work requirement had countervailing effects on the average cash income of families with children between the ages of 6 months and 11 months. On the one hand, average cash income for families with young children rose while they were in the program because of additional earnings from more employment (see Figure B-2). On the other hand, earnings at program exit appear to have fallen slightly, and the expansion reduced the amount of cash assistance families received. TANF benefits decreased primarily because violations of the work requirement reduced the number of months that families remained in the program (see Figure B-3).7

Figure B-2.

Cash Income for Single-Parent Families in Alabama’s TANF Program, by Age of Youngest Child in Months and Whether the Family Could be Exempt From the Work Requirement

2019 Dollars

When Alabama began imposing work requirements on parents whose youngest child was between the ages of 6 months and 11 months, the average cash income of those families rose substantially while they were in the program because of additional earnings from more employment.

Average cash income at program exit appears to have fallen slightly from 2017 to 2018 for families whose youngest child was between the ages of 6 months and 11 months.

Data source: Department of Health and Human Services. See www.cbo.gov/publication/57702#data.

The analysis is limited to the single-parent families that Alabama could include in the work standard.

TANF = Temporary Assistance for Needy Families.

Fitted values give the trends in cash income by the age of the youngest child.

Parents whose youngest child was 4 months or 5 months old are excluded from this analysis because they were occasionally subject to the work requirement in 2018. Parents whose youngest child was between 12 months and 16 months of age were likely to have been subjected to the work requirement for more months because of its expansion. Affected data are shown as stand-alone dots in this figure.

Figure B-3.

Cash Assistance for Single-Parent Families in Alabama’s TANF Program, by Age of Youngest Child in Months and Whether the Family Could be Exempt From the Work Requirement

Percent

When Alabama began imposing work requirements on parents whose youngest child was between the ages of 6 months and 11 months, violations of the work requirement became a far more common reason for those families to stop receiving cash assistance. The percentage of families that stopped receiving cash assistance in an average month also increased, which offset some of the increase in average cash income from more employment.

Data source: Department of Health and Human Services. See www.cbo.gov/publication/57702#data.

The analysis is limited to the single-parent families that Alabama could include in the work standard.

TANF = Temporary Assistance for Needy Families.

Fitted values give the trends in the termination of cash assistance by the age of the youngest child.

Parents whose youngest child was 4 months or 5 months old are excluded from this analysis because they were occasionally subject to the work requirement in 2018. Parents whose youngest child was between 12 months and 16 months of age were likely to have been subjected to the work requirement for more months because of its expansion. Affected data are shown as stand-alone dots in this figure.

The increases in employment among participants do not appear to stem from a reduction in the number of people (or a decrease in the number of less-skilled people) entering the program. If entry into the program had declined substantially, as it did after enactment of the Personal Responsibility and Work Opportunity Reconciliation Act, then unmeasured changes in employment and income among families who were deterred from entry might offset the increases experienced by participants. But the parents of children between the ages of 6 months and 11 months continued to account for about 21 percent of entrants in 2018 (see Figure B-4). Even if entry did not decline for that group overall, a reduction in entry among less-skilled parents could have inflated the employment rate and average income. However, the expansion of the work requirement does not appear to have substantially decreased the frequency with which parents who have little education entered the program.

Figure B-4.

Entry Into Alabama’s TANF Program Among Single-Parent Families, by Age of Youngest Child in Months and Whether the Family Could be Exempt From the Work Requirement

Work requirements can boost employment among TANF recipients by encouraging work or by deterring participation by people who are unlikely to find employment. The expansion of the work requirement in Alabama does not appear to have reduced the number of families entering TANF or their average level of education.

Data source: Department of Health and Human Services. See www.cbo.gov/publication/57702#data.

The analysis is limited to the single-parent families that Alabama could include in the work standard.

TANF = Temporary Assistance for Needy Families.

Parents whose youngest child was 4 months or 5 months old are excluded from this analysis because they were occasionally subject to the work requirement in 2018. Parents whose youngest child was between 12 months and 16 months of age were likely to have been subjected to the work requirement for more months because of its expansion. Affected data are shown as stand-alone dots in this figure.

Using the second approach, CBO found that TANF’s work requirements increased employment over the year after families with young children entered the program. For single parents of children who were under one year of age upon entry into TANF, CBO estimates that the expansion of Alabama’s work requirement increased their employment rate by 6 percentage points while they remained in TANF. Over the year after they entered the program, some of that initial increase in employment was offset by rates of employment that were similar for families after they left TANF; averaging those rates brings the estimated effect on the employment rate down to an increase of about 4 percentage points.

Also using the second approach, CBO found that TANF’s work requirements might have increased the average cash income of families with young children over the year after they entered the program. But higher average earnings while in the program were partly offset by a reduction in cash payments because some people were in the program for a shorter time.

A Comparison of CBO’s Findings Under the Two Approaches

The estimated effects on employment tend to be smaller under the second approach than under the first because the second approach concentrates less on the months in which families were subject to the expansion of the work requirement.

The first approach focuses on the period when families were directly subject to the expansion because they had a child between the ages of 6 months and 11 months. The second approach expands the examination to all months of assistance for any families that could be subject to the expansion for at least one month during their spell in the program. As a result, the number of months examined using the second approach was 137 percent higher than the number under the first approach. Although the increase in the employment rate was smaller for that broader group, the increase in the months of employment was larger—primarily because the employment rate rose among parents who had a child between the ages of 12 months and 16 months and who had been subject to the work requirement earlier as a result of the expansion.

A Comparison of CBO’s Findings With Results of the AFDC Waiver Experiments

In the 1990s, a series of evaluations were conducted on waiver programs in Aid to Families With Dependent Children (AFDC). The results from those AFDC waiver experiments have remained a primary source of evidence on the effects of work requirements.8 CBO’s estimates of the effect of Alabama’s work-requirement expansion on employment and income are similar to the results from the experimental evaluations once adjustments are made for differences in the way the requirements were expanded.

The experimental evaluations generally measured the effect of fully implemented work requirements by comparing income and employment for a group in which almost everyone was subject to the requirement and a group in which no one was. In contrast, the single parents who entered Alabama’s TANF program with a child under the age of 6 months in 2018 were rarely subject to the requirement until their child reached 6 months of age. In addition, the single parents who entered the program with a child under the age of 12 months in 2017 were typically subject to the requirement once their child reached 12 months of age. Among single parents who entered the program with a child under 12 months of age, the portion of months within a year that they were subject to the work requirement increased by 0.41 (or 5 months). That increase would have been more than twice as large (almost 12 months) if almost everyone was subject to the requirement in 2018 and no one was in 2017.

The estimate of the effect on employment from Alabama’s work-requirement expansion is consistent with the estimate from the AFDC waiver experiments. CBO estimates that the employment rate increased by about 4 percentage points over the first year after single parents entered TANF in Alabama. Thus, full implementation of a work requirement in 2018 would have boosted employment by about 9 percentage points if the increase in employment had been proportional to the change in the percentage of single parents subject to the work requirement. The average of estimates from the AFDC waiver experiments that focused on labor force attachment was about 10 percentage points.

CBO found more evidence of an increase in families’ average cash income from Alabama’s work-requirement expansion than the evaluations of the AFDC waiver experiments found. Most of the difference in those estimates can be explained by the larger gains in earnings for participants in Alabama’s work-requirement expansion, and some can be explained by the smaller reduction in cash payments for those same participants. The reduction in cash assistance was smaller because Alabama provides smaller monthly cash payments than the programs in the AFDC waiver experiments provided. In contrast, the reduction in the number of months that cash assistance was received was larger in CBO’s analysis than in the AFDC waiver experiments.


1. The analysis focuses on the families receiving TANF that are headed by single parents who could be counted by the states toward the work standard. Families headed by single parents are excluded from that standard if the parents are disabled or are caring for a disabled family member.

2. The 14 states are Alabama, Alaska, Georgia, Hawaii, Idaho, Indiana, Iowa, Kentucky, Louisiana, Minnesota, Oklahoma, Oregon, Washington, and Wyoming.

3. Five of the 14 states that generally provided consistent measures of spell length have inconsistencies in the data on employment. Thus, this analysis is based on data from Alabama, Alaska, Georgia, Indiana, Iowa, Louisiana, Minnesota, Oklahoma, and Oregon.

4. The data probably understate the number of families that stop receiving benefits because of the work requirement. Some families probably have their spell end because of the work requirement without the states knowing whether they are in violation of it. For example, families may not submit the documentation necessary to determine continued eligibility because they know that they do not meet the work requirement.

5. The analysis is unlikely to be sensitive to that imputation because a family’s employment and earnings when it left the program tended to be similar in 2017 and 2018. Thus, allowing for reversion to the average employment rate or for earnings to grow with job experience would not substantially change the estimates of the effects on employment or earnings.

6. That approach appears to capture differences in economic conditions because directly adding measures of economic conditions, such as the local unemployment rate, did not substantially change the estimated effects.

7. Most families who were expelled because of the work requirement were not employed at that time. That decrease in employment at exit was offset by higher employment among families that left the program for other reasons.

8. Gayle Hamilton and others, National Evaluation of Welfare-to-Work Strategies (submitted by Manpower Demonstration Research Corporation to the Department of Health and Human Services, December 2001), https://tinyurl.com/2p89s5ad.

About This Document

This report was prepared at the request of the Chairman of the House Committee on the Budget. In keeping with the Congressional Budget Office’s mandate to provide objective, impartial analysis, the report makes no recommendations.

Justin Falk prepared the report with assistance from Julia Heinzel (formerly of CBO) and guidance from Joseph Kile, Xiaotong Niu, and Julie Topoleski. William Carrington (formerly of CBO), Molly Dahl, Elizabeth Cove Delisle, Kathleen FitzGerald, Jennifer Gray, Rebecca Heller, Nadia Karamcheva, Justin Latus, Susanne Mehlman, Alexandra Minicozzi, Shannon Mok, Robert Stewart, and Sree Yeluri offered comments on the draft. Elizabeth Ash fact-checked the report. Marianne Bitler of the University of California‒Davis and Bruce Meyer of the University of Chicago commented on an earlier draft. The assistance of external reviewers implies no responsibility for the final product; that responsibility rests solely with CBO.

Jeffrey Kling and Robert Sunshine reviewed the report. Christine Bogusz edited it, and Jorge Salazar created the graphics and designed the cover. R. L. Rebach and Jorge Salazar prepared the report for publication. The report is available at www.cbo.gov/publication/57702.

CBO seeks feedback to make its work as useful as possible. Please send any comments to communications@cbo.gov.

Phillip L. Swagel

Director

June 2022