REPORT Food InsecurityHungerPovertySNAP

The Dismal Economics of SNAP’s Work-Hours Test and Time Limit



The Supplemental Nutrition Assistance Program, or SNAP, reduces food insecurity, which in turn improves people’s health and economic well-being.1 Yet, SNAP’s effectiveness is limited by restrictions that target people deemed undeserving by some policymakers — undeserving, essentially, of a healthy but basic diet.2 This brief discusses one of the most onerous of these restrictions: a complex work-hours test and time limit, applied on top of SNAP’s other work-related requirements. The test requires adults under age 50 to clock at least 80 hours of countable work each month unless they qualify for an exemption from the test or live in an area where the test has been waived due to insufficient job opportunities. If they fail the test or cannot show they passed it to the satisfaction of local officials, they are excluded from SNAP. The federal government has waived the 80-hour work test nationwide while the COVID-19 public health emergency declaration is in place, but this declaration will be lifted in May 2023. Once this happens, people subject to the work test will start to lose SNAP food vouchers in October 2023. States and local SNAP agencies have some discretion in implementing the test, so their decisions will partly determine how many people are harmed.

This paper reviews relevant empirical research on the work-hours test. According to conventional economic theory, SNAP and other means-tested programs provide a minimum guaranteed income to very low-income households or people. That minimum income is then reduced as earnings or other income increases. In theory, this can lead people to remain out of the labor market or reduce their work hours, depending on how they value their time. This theory may hold when the guaranteed income is a monetary benefit that is large enough to afford housing, utilities, food, clothing, transportation, and other basic needs, especially if any earnings received reduce the benefit on a dollar-for-dollar basis. But SNAP only provides a modest, in-kind benefit — a food “value voucher” that people can only use to buy food to prepare meals at home — and also has benefit calculation rules that make it financially advantageous to have a job. It seems unlikely that many people will opt not to work or reduce their work hours — and forgo the money income that they receive for it — simply to receive a small, in-kind benefit like SNAP.3

Taken as a whole, the empirical studies reviewed in this report provide strong evidence that the test is counterproductive — it has no or little positive impact on employment while excluding a substantial number of vulnerable people from SNAP. The test’s complexity places administrative burdens on both people receiving SNAP and officials in local SNAP offices. The collateral consequences include people who should be exempt from the test losing access to SNAP and the federal government wasting money to administer the test that could be put to better use elsewhere. The administrative burdens partly explain why the test has a negative spillover effect on homeless people and disabled people, even though disabled people are legally exempt from the test, and homeless people generally should be exempted by local officials.

Rep. Barbara Lee (D-CA) has introduced legislation in the current session of Congress that would repeal the work-hours test. Given the well-documented harm caused by the test, Congress should pass repeal legislation when it reauthorizes the Farm Bill later this year.4 Despite the strong evidence showing that the test is counterproductive, some conservative policymakers have called for subjecting more people to it. Sen. Rick Scott (R-FL) and Rep. Dusty Johnson (R-SD) have introduced legislation that would extend the test to more groups of people receiving SNAP vouchers, including many older people — 50- to 59-year-olds in Scott’s bill and 50- to 65-year-olds in Johnson’s — and parents and other adults living with school-age children.5  Both bills would also take away states’ ability to waive the test in areas with insufficient job opportunities. If legislation like this is enacted, experts estimate that more than 10 million people could lose SNAP food vouchers.6 Based on the evidence compiled here, these massive losses would not be offset by increased employment and earnings and would have other adverse impacts on health and well-being.

Background on SNAP and the 80-Hours Work Test

SNAP provides vouchers — in the form of an electronic benefit transfer, or EBT, card — that eligible people can use to purchase groceries. Over 42 million people — about 1 in 8 people in the United States — currently buy some or most of their groceries with SNAP vouchers. The maximum value of SNAP vouchers varies by household size. The average value of the food voucher that households receive depends on income and other factors. A person living alone and receiving the average monthly amount ($195) can use it to buy groceries amounting to $2.17 per meal. For a household of three, the average monthly amount ($577) can be used to purchase food for meals costing about $2.14 for each person in the household.7 

People subject to the work-hours test are conventionally called ABAWDs — short for able-bodied adults without dependents — and the test itself is called the ABAWD work requirement and time limit. The ABAWD label is problematic for a number of reasons. “Without dependents” implies that no people harmed by the test are parents. Yet, the test applies to parents who do not live regularly with their children, even if these parents provide care and support that their children depend on. “Able-bodied” is disfavored by many people in the disability rights movement.8 Moreover, the research discussed below suggests that many people with disabilities still end up being excluded from SNAP because of the test. Finally, there is no common ABAWD identity in the world outside of the SNAP program — it is hard to imagine anyone self-identifying as an able-bodied adult without dependents or an ABAWD.9 For these reasons, this brief generally uses the terms “80-hour work-test,” “work-hours test,” or just “test” instead of ABAWD requirement.

SNAP’s Standard Employment-Related Requirements

SNAP requires most unemployed 18- to 59-year-olds receiving benefits to register for work and accept almost any job they are offered.10 In addition, employed workers receiving SNAP may not voluntarily quit a job or reduce their work hours to less than 30 hours a week without good cause.11 States can also require adult household members to participate in the SNAP Employment and Training Program, except for members who are elderly, disabled, caring for young children, or who fall into certain other exempt categories. States must operate a SNAP Employment and Training Program, but they have considerable flexibility when it comes to deciding to whom to provide services and how to structure their programs. Programs typically involve various basic employment and training services, including employment assessments and job search assistance, and may also include vocational training, subsidized employment, and supportive services.

The 80-Hours Work-Test

The 80-hours work test is an additional employment-related requirement. The test is convoluted and difficult to describe in plain language, so what follows is a simplification.  Over a 36-month period, adults subject to the test can receive up to three months of SNAP during months in which they work less than 80 hours. Once they have received SNAP for three months in which they did not work at least 80 hours, they are barred from receiving SNAP in any of the remaining months in the 36-month period. They can requalify by showing that they have worked at least 80 hours in a subsequent month. In addition to hours of paid work, hours spent in some types of unpaid labor and other activities count toward the test, including hours of participation in a SNAP Employment and Training Program or other approved work program. The three months that trigger the test do not need to be consecutive. Thus, if a person subject to the test works 140 hours monthly in January, March, and May but only 79 hours monthly in February, April, and June, they are barred from receiving SNAP in any of the subsequent 27 months in which they work 79 or fewer hours.

Who Has to Pass the Test

The test applies to 18- to 49-year-olds unless they are pregnant, live with a minor child, care for an incapacitated person, are determined to be physically or mentally unfit for employment or meet one of a few other narrow exemptions. If a person is not receiving temporary or permanent disability benefits, like Supplemental Security Income, they will typically need to provide certification from a doctor that they are physically or medically “unfit” for work. Depending on where a person subject to the test lives, there may be other state-specific exemptions.12 States can also ask the USDA to waive the work-hours test in geographic areas with insufficient jobs. The convoluted nature of the test places considerable administrative burdens — beyond those due to SNAP’s income and asset tests and standard work and training requirements — on SNAP agencies and the people who are subject to the test or trying to obtain exemptions from it. As a result, some people give up on obtaining SNAP benefits for which they are eligible, and that would improve their health and well-being, while others lose eligibility and then return a few months later.

Waivers in Areas with Insufficient Jobs

As noted above, the federal government suspended the test nationwide during the COVID-19 public health emergency. In 2009, as part of the American Recovery and Reinvestment Act, Congress also waived the work-hours test in response to the Great Recession.13 The national suspension ended in 2010. Some states reinstated the test statewide, regardless of job availability; others obtained permission from USDA not to apply the test in areas where jobs were insufficient.14

Administrative Burdens

State and local SNAP officials have long noted the burdensome nature of the work-hours test. In a 2016 report, the USDA’s Office of Inspector General noted that state SNAP officials used terms like “administrative nightmare” and “operational nightmare” to describe the work-hours test. Officials also expressed concerns about the costs of administering the test — both in terms of time and resources — and how the convoluted nature of the test increases the likelihood that people are erroneously denied benefits for which they are eligible.15

Estimated Number of People Subject to Test and Demographics

As part of a major federally funded study that examined the effects of the work-hours test in nine states after it was reinstated in the mid-2010s, researchers at the Urban Institute and Westat estimated that 4 to 9 percent of people receiving SNAP food vouchers in each of the states were potentially subject to the test.16 If the same share of people currently receiving SNAP are potentially subject to the test when it is reinstated later this year, about 1.6 million to 3.7 million people are at risk of losing SNAP benefits. In a 2018 analysis, Lauen Bauer and her colleagues at the Hamilton Project estimated that about 2.2 million people who reported receiving SNAP benefits nationwide in 2017 were potentially subject to the work-hours test.17

The Urban Institute’s nine-state report used SNAP administrative data to examine the characteristics of 18- to 49-year-olds receiving SNAP at the time of reinstatement of the work-hours test in each state in the mid-2010s. About 50 to 60 percent of people subject to the work-hours test had a high school degree, about 20 to 30 percent had no high school degree, and very few had education beyond high school.18 Most people subject to the test live in single-person SNAP households and have lower total incomes than other 18- to 49-year-olds receiving SNAP.  Although adult women are more likely to participate in SNAP than adult men, men are more likely to be subject to the work-hours test than women in large part because men are more likely to live in households without children.19 Five states in the Urban study had SNAP administrative data on homelessness. In all of these states, the percentage of homeless people subject to the work-hours test was much higher than the percentage of other 18- to 49-year-old homeless people receiving SNAP.

Literature Review

We identified 19 studies published or made publicly available since 2010 that use empirical methods to estimate the impact of the work-hours test on various outcomes, most commonly labor supply, SNAP participation, and income. Eleven of the studies are peer-reviewed journal articles. The remaining studies comprise one book chapter, one major federally funded report, one research brief, and five working papers. We did not include research published before 2010 both because it is well-covered elsewhere and generally much weaker methodologically than more recent research.

Researchers seeking to estimate the effects of the SNAP work-hours test face a number of challenges. The test has never been evaluated using a randomized controlled trial in which people are randomly assigned to a group subject to the test and a control group not subject to the test. Instead, most of the studies summarized here use quasi-experimental methods — typically to estimate the impact of the reinstatement of the work-hours test in the wake of the Great Recession. As noted above, after the national suspension of the test was lifted in 2010, there was considerable variation between states and areas within states in the applicability of the work-hours test. This variation makes it possible for researchers to control for, to some extent, measurable characteristics and other factors that influence labor supply, SNAP receipt, and health and well-being. Still, the lack of randomization means that results may be caused by factors that are not controlled for in the research.

Estimating the effect of the work-hours test is also challenging because it impacts a relatively isolated group of people who are typically unemployed or precariously employed and often have unstable housing arrangements. As a result, they may not be well-captured in household surveys. Moreover, participation in SNAP is underreported in household surveys, which increases the likelihood of selection bias and sampling errors.20  Many of the studies discussed in this paper rely in part on administrative data, including information collected from people as part of the SNAP eligibility process and employee wage and hour information that employers provide to Unemployment Insurance officials on a quarterly basis. Administrative data has some advantages over survey data, especially when addressing underreporting benefits in household surveys. But survey data has some advantages over administrative data, especially regarding demographic detail. Researchers have taken different approaches to address these challenges. The Appendix table summarizes the methodology and key findings for each study.

Summary of Impacts

Table 1 summarizes the findings of the 19 studies reviewed (for a full description of methods and findings, see Appendix Table A1). The growing body of research on the SNAP work test tells a relatively consistent story about its impacts. The test has little or no positive impact on employment but does reduce access to SNAP food vouchers among vulnerable people with few resources. These findings generally hold across the different methodological approaches used by researchers. There is some evidence of harmful spillover effects, including on homeless and disabled people, who generally should be exempted from the test. There is also some evidence of negative impacts on individual and community well-being indicators, including health, past-due debt, property crime, and food-pantry utilization.

Table 1


Impact on Receipt of SNAP Food Vouchers

Fifteen of the studies examined the effect of the work-hours test on SNAP participation. Not surprisingly, all of these studies found that the test reduced participation. Among the notable published journal articles, Ku et al. (2019) use administrative data from 33 states and estimate that after the reinstatement of the work-hours test in the mid-2010s, more than one-third of people subject to the test lost benefits.21 If representative of all US states, this would mean that about 600,000 people lost about $2.5 billion in food vouchers between 2013 and 2017 due to the test. Brantly et al. (2020) use national data from the American Communities Survey (ACS) and find a similarly large reduction in SNAP participation due to the test. They also find that the test reduced SNAP participation among disabled 18- to 49-year-olds, a group that should be exempted from the test, suggesting that the disability exemption is not consistently applied, perhaps in part due to administrative burdens involved in obtaining it.

Impact on Employment and Earnings

Of the 10 studies that examined the impact of the test on employment, seven found no positive employment impacts, and three found positive impacts. Among the notable published journal articles, Han (2022) uses national household survey data (ACS) and finds that increasing the share of people in a substate area who are exempt from the test —due to waivers or additional exemptions — does not increase employment. Similarly, Gray and colleagues (2023) use administrative data (SNAP and UI wage records) from Virginia and find no significant change in employment after the reinstatement of the test in the wake of the Great Recession.

Two published journal articles and one working paper find employment increases. Harris (2021), using the ACS, finds a small (1.3 percent) increase in employment that was limited to urban areas.22 Cuffey et al. (2022) also find a positive employment impact, but their methodology is among the weakest of those used in the published studies. Their sample is limited to adults without a high school diploma — even though most people subject to the test have a high school diploma — and the survey they rely on (CPS) has a much smaller sample size and much more limited geographic detail than the ACS.23 In fact, the authors acknowledge the “likelihood that this group [low-income adults who do not have a high school diploma] is sufficiently small and that SNAP work restrictions would not have a meaningful effect on [the larger overall group of adult SNAP participants who do not live with children] or labor market outcomes of low-income adults in general.”

Impacts on Income and Other Wellbeing Indicators

If the test reduces SNAP participation without increasing employment, people impacted by it will have lower net incomes until they can offset the benefit loss by finding other sources of income. These could include disability benefits, transfers from family members or friends, using credit cards, spending down liquid assets, or borrowing money. Using administrative data (SNAP and UI wage data) from three states, Veriker et al. (2023) estimate that the work-hours test reduces combined income from SNAP and earnings by 8 to 21 percent, but they are not able to estimate total net income. The only other published journal article to find suggestive evidence of net income loss is Coffey et al. (2023). Using food pantry data from three states, they find that food pantries in urban areas where the work-hour test was reinstated served 34 percent more households in the eight months after the reinstatement than food pantries in urban areas that did not reinstate the test.

Some of the working papers also include suggestive, if preliminary, evidence of reduced income and consumption. Using CPS data and SSA administrative data, Stith (2019) finds that the work-hours test may increase applications for Supplemental Security Income, a means-tested disability benefit, particularly among people who self-report activity limitations other than blindness. Using national consumer credit data, Dodini, Larrimore, and Trafaglia (2022) find that reinstatement of the work-hours test led to increased credit balances, including past-due balances. In theory, this could be due to increased employment, which the authors could not measure in their research. But after considering the possibility, they conclude that the results reflect “financial vulnerability” and “strongly point to consumers seeking out and using new credit in order to make up for lost SNAP benefits.” In addition to reductions in SNAP participation, Lippold and Remy (2021) find increases in property crime and homelessness.

Finally, two published journal articles document health-related impacts. Allen et al. (2023) use Medicaid and SNAP administrative data from West Virginia and finds that reinstatement of the test in nine counties increased the probability of visiting a mental health provider for a mood disorder or anxiety for both men and women. They note that the increase could be due to the loss of SNAP directly impacting mental health, increasing the likelihood of visiting a provider to address pre-existing issues, and increasing the likelihood that people with pre-existing issues visit a provider to obtain the certification needed for a disability exemption from the test. Feng (2022) analyzes data from the monthly state data from the CDC’s Behavioral Risk Factor Surveillance System and finds that, besides having no impact on employment, the test increases the likelihood of experiencing a physically unhealthy day by 14 percent.

Implications for Future Research

Taken as a whole, the research summarized here provides conclusive evidence that the work-hours test reduces SNAP participation and is very strong evidence that it does not increase employment. Even if the test has a positive effect on employment and earnings, it appears to be very small. Still, it would be helpful to have more research that examines whether the effects of the test on employment and SNAP participation vary by class, ethnicity and race, gender, age, and other categories. Researchers also shouldn’t assume that disabled people are automatically exempted from the test, especially if they are not receiving benefits like SSI and SSDI. In addition, as Dong and Feng (2021) note, the work-hours test is likely to negatively impact formerly incarcerated individuals given their high unemployment rates, but beyond suggestive small-n qualitative studies, researchers have not focused on this group.24 Finally, additional research is needed to estimate the impacts of the work-hour test on net income, consumption, housing stability, health, credit usage, and other outcomes.


Congress added the work-hours test to SNAP in August 1996 during a presidential election year when President Clinton was in the last year of his first term, and Congress was controlled by Republicans, including Rep. Newt Gingrich, then the Speaker of the House, and Sen. Bob Dole, the Senate Majority leader who was running against Clinton. As political scientists have documented, anti-Black racism and paternalism played a central role in passing legislation — the Personal Responsibility and Work Opportunity Reconciliation Act, or PRWORA — that included the work-hours test and other benefit cuts. Both crime and welfare had become “racially coded” issues that activated White people’s racist beliefs that Black people are lazy and immoral.25 When Congress passed PRWORA, political scientist Martin Gilens published research finding that negative views of Black people receiving “welfare” had become more politically potent than ones of White people and helped to generate opposition to means-tested programs.26 Other major pieces of legislation enacted during this period — including the 1994 Crime Bill and the U.S.-China Relations Act of 2000, which normalized long-term trade relations with China — are now widely viewed as missteps that harmed working-class and vulnerable people on a mass scale. The work-hours test has gotten much less attention over the years, but it should also be viewed as a counterproductive misstep that has harmed the working class.

When the work-hours test became law in 1996, the economy was strong, and unemployment was relatively low — the average annual unemployment rate was 5.4 percent in 1996 and continued downward until reaching an average annual rate of 4 percent in 2000, which was, at that point, the lowest annual rate since 1969.27 At the same time, these headline numbers masked underlying weaknesses. Among 25- to 54-year-old men without college degrees, labor force participation declined sharply in the early 1990s — while it stabilized for a few years in the latter half of the 1990s, it has continued to trend downward over time. In contrast to other countries with higher levels of labor force participation and employment among 25- to 54-year-old men, the United States has an extremely limited welfare state for people in this age range. SNAP is one of the few meaningful benefits they may qualify for, but there is no good reason to think that a program providing a modest food voucher plays any significant role in this long decline. Punitive policies like the work-hours test have likely intensified working-class men’s struggles in recent decades.

The empirical research summarized here suggests that the labor market and working-class people’s employment decisions are more complex than conventional theory and the thinking behind the work-hours test assumes. The test has done more harm than good. Congress should repeal it this year and expand social insurance and employment and training services for working-class people, including non-disabled people who do not live with children. Finally, the fact that so many working-class people need basic food vouchers has little to do with the people themselves and everything to do with the rules that structure the economy. Those rules need to be changed in ways that produce inclusive prosperity.



Author and Year

Type of Publication

Impact on Labor Supply

Impact on SNAP Participation

Impact on Other Indicators



Allen et al. (2023)

Journal Article



Increased visits to mental health providers

Summary: Event-study design using administrative data (Medicaid and SNAP) to estimate impact of work-hour tests on visits for depression and anxiety in West Virginia in 2016-2018.

Sample: Balanced panel sample of 18- to 49-year-olds enrolled in both SNAP and Medicaid in 2015, excluding those not continuously enrolled in Medicaid during study time period, and those who turned 18 or 50. Data does not allow for exclusion of adults living with children under age 18, which authors note biases the study towards null results.

Reinstatement of work-hours test in nine counties in 2016:: 1) increased probability of visiting a mental health provider for a mood disorder for both men and women (by 13-14%); 2) increased number of mental health care visits made by women, and 3) increased probability of visiting a mental health provider for anxiety (17.8% for women and 24.3% for men).

Authors note that increase in mental health visits could be due to: 1) loss of SNAP negatively impacting mental health; 2) potential loss of SNAP pushing people with pre-existing mental health issues to visit a provider to obtain proof needed for an exemption from the test, and/or address issues that are a barrier to employment

Brantley et al. (2020)

Journal Article


Negative, including for disabled adults who should be exempt from test

Spillover effect on disabled adults

Summary: Difference-in-differences and triple-difference methods using ACS data to estimate effects of work-hours test reinstatement between 2012 and 2017 on annual SNAP participation, including by disability and race/ethnicity.

Sample: Low-income adults (money income below 200% of the federal poverty line) under age 50 stratified by disability status. Excludes people living with a minor child and certain others who are federally exempt from general SNAP work requirements or who face certain other specific SNAP eligibility requirements (e.g., non-citizens). N = (866,000)

Reinstatement of the work-hour test reduces SNAP participation by 4-percentage points in both the sample that excludes disabled people and the one that is limited to disabled people. Negative spillover impact on disabled adults (who are supposed to be exempt from work-hours test) is equivalent to a 7.8 percent reduction in SNAP participation among disabled beneficiaries. Larger declines among Black and Hispanic adults (7.2 and 5.5 pp decrease respectively) than white adults (2.6 pp decrease).

Cuffey et al. (2022)

Journal Article




Summary: Regression discontinuity (age 50 cutoff) design using CPS (Basic Monthly) to estimate effect of work-hours test on employment.

Sample: Adults who (a) do not live with anyone under 18, (b) are U.S. citizens, (c) have less than a high school education, and (d) live in a low-income household (money income under 250% of the poverty line).

Work-hours test increases employment among low-income adults with no high-school diploma. However, authors acknowledge “likelihood that this group [low-income adults who do not have a high school diploma] is sufficiently small and that SNAP work restrictions would not have a meaningful effect on [the larger overall group of adult SNAP participants who not live with children] or labor market outcomes of low-income adults in general.” Authors also note that waivers of the work-hours have a smaller effect on adults living in high-minimum wage states (suggesting that general labor standards are more important in decisions to work than SNAP policy).

Cuffey et al. (2023)

Journal Article



Increase in food pantry visits.

Summary: This paper uses data from a cohort of food pantries in AL, FL, and MS (Source: Feeding the Gulf Coast). Localities in all 3 states reinstated work-hour tests in 2016. Authors run event study models using 2022 data to predict the effect of work-hour test reinstatement on the number of households served by food pantries. Data spans January - December 2016 and covers 3 months before the reinstatement and the 9 months following.

Sample: All 257 FTGC food pantries that operated between January - December 2016. 24 counties of MS, FL, and AL reinstated work-hour test requirements in that time period. Not all food pantries operated every month, the final pantry-month sample size is N = 2,812. Pantry addresses were used to obtain information that controlled for local factors influencing food pantry participation and also to define exposure to the SNAP work requirement. They also controlled for time-varying economic conditions via county unemployment rates. primary limitation of this study is that the results might be driven by unobserved time-varying factors

Work-hour test reinstatement is associated with an increase in the number of households served by food pantries. This impact is concentrated in urban areas.

On average, a food pantry in an urban area with greater exposure to work-hour test requirements reinstatement served 34% more households in the 8 months after the reinstatement than food pantries in urban areas that did not reinstate work-hour test requirements.

Dodini, Larrimore, Trafaglia (2022)

Federal Reserve Discussion Series Paper



Increase in credit applications, credit balances, and past due balances

Summary: Difference-in-differences and event study designs using Federal Reserve Bank of New York/Equifax Consumer Credit Panel (ECCP) quarterly data to examine impact of work-hours reinstatement on credit and debt between 2010 and 2017.

Sample: Authors draw a random sample from the ECCP which contains 0.5 percent of individuals in the US with a credit report. They then limit the sample to those aged 18-49 and those whose lowest Equifax Risk score was lower than 700 at any point between 2010-2017. They then limit to those who have never had a mortgage between 2010-2017 and those who are residing in an area designated as a Commuting Zone based on 2000 definitions. Final sample covers 3,108 counties in 707 Commuting Zones with ~8.67 million observations.

For people in sample, reinstatement of work-hours test led to (in years 1 to 6 following the reinstatement): 1) a 17 percent increase in credit inquiries in the past six months; 2) an 18 percent increase in total credit accounts; 3) a 36-percent increase in total combined credit limits ($1,500); 4) a 29 percent increase in total card balances increased (approximately $500), and a 1.3 percentage point (5 percent) increase in prevalence of past due balances. Effects are significantly stronger in states that had asset limits, which suggests that reinstatement imposes additional financial pressure for people who lack meaningful liquid assets to withstand loss of SNAP.

Authors interpret results as “strongly pointing to consumers seeking out and using new credit in order to make up for lost SNAP benefits” and “reflecting financial vulnerability of people likely to lose SNAP as a result of test reinstatement. Authors consider and reject alternative explanations that increase in credit is driven by increased employment.

Feng (2022)

Journal Article



Increase in physically unhealthy days

Summary and Sample: Triple-difference approach using Behavioral Risk Factor Surveillance System (BRFSS) data, which is collected monthly and representative at state level, to examine health and employment impacts of reinstating work-hours test waivers in 2015 and 2016 on non-disabled 18- to 50-year-olds who are not living with children.

Among low-income adults potentially subject to the work-hours test, reinstating the test statewide: 1) increased incidence of reporting physically unhealthy days by 14%, and 2) had no impact on employment. BRSFF does not directly measure SNAP participation, but based on supplemental analysis, Feng infers that it decreased in states that reinstated the test partial or statewide basis.

Gray et al. (2023)

Journal Article




Summary: Longitudinal study using administrative data (SNAP and UI earnings) from Virginia between 2007-2015. Regression discontinuity design uses age 50 cutoff to investigate the impact of the work-hours test reinstatement beginning in October 2013.

Sample: Adults enrolled in SNAP who have no known exemptions or disabilities and do not live in a household with a child <18.

Eighteen months after reinstatement, the work-hours test reduced SNAP participation by 64% among incumbent participants and by 53% among all adults subject to the test, but had no effect on employment. In some specifications, some suggest increased earnings among a limited subset of participants.

Largest declines in SNAP participation were among homeless people. Initial cost-benefit calculations suggest that eliminating the test may transfer more resources per dollar of public expenditure to low-income adults subject to the test than certain other programs targeting a similar group, including the EITC and housing assistance.

Hall (2022)

Research Brief




Summary: Difference-in-differences and other methods using state administrative data (SNAP and UI) to estimate impact of work-hours test reinstatement on employment, earnings, and SNAP participation in Maryland counties.

Sample: 18- to 50-year olds not living with children and not receiving disability benefits who received SNAP in MD for at least one month between January 2015 and December 2017. (N=235,545 individuals, 911,591 observations). N control = 662,316 observations, N treatment = 249,275).

Reinstatement of the work-hours test had: 1) no impact on employment or earnings, and 2) reduced SNAP participation.

Han (2022)

Journal Article

Mostly none



Summary: Triple-difference approach using American Communities Survey (2005-2017) to compare labor supply differences between low-income people just below 50 and those just above 50 by share of each PUMA’s population that is not subject to the work-hours test because of USDA waivers or state-specific 15% exemptions.

Sample: Adults nationwide who are below 300% of the federal poverty line, except for adults receiving disability benefits, adults living with minor children, and adults living with an elderly person >70 or another adult who receives disability benefits.

Increasing the share of a PUMA’s population not subject to the work-hours test (because of USDA waivers or state-specific exemptions) has: 1) a small positive effect on likelihood of receiving SNAP in last 12 months, and a likely larger positive effect on duration of SNAP participation among people who receive it; 2) no significant effect on likelihood of employment in employment in last 12 months; and 3) in some specifications, a small negative effect on usual hours worked conditional on working.

Based in part on supplemental analysis, Han suggests that the work-hours test has little to no impact on employment because: 1) SNAP’s benefit calculation rules ensure that moving from non-employment to employment increases overall income, and 2) most people who are able to enroll in SNAP because of USDA waivers or state-specific exemptions have not worked during the prior year (so they’re not moving from employment to non-employment in order to access SNAP).

Han also systematically compares his findings to other studies and concludes that studies finding that the work-hours test increases employment have limitations that suggest they are outliers. For example, he notes that Cuffey et al. (2022) uses the Current Population Survey-ASEC, which has a much smaller sample size and less geographic detail than the ACS.

Harris (2021)

Journal Article




Summary: Triple differences (age 50 cutoff) and difference-in-differences (time and geographic variation) methods using 2010-2017 ACS data to estimate the impact of work-hours reinstatement on employment and SNAP participation.

Sample: 25- to 54-year-olds, excluding those with disabilities, living with children <18, students, noncitizens, and those who are institutionalized (N=42,699). Author reweights the ACS sample to align with the QC sample more closely on age, gender, race, ethnicity, educational attainment, state and year.

Reinstatement of work-hours test increased employment by 1.3 percentage points. Somewhat larger effect on Black adults than whites adults subject to test. No significant effect on employment of adults living in rural areas.

Reinstatement also reduced SNAP participation among adults subject to the test with the largest decline among adults without a high school diploma.

Ku et al. (2019)

Journal Article




Summary: Longitudinal study (county-level 2-way fixed effect model) using FNS administrative data to estimate effect of work-hours test reinstatement on SNAP participation and amount of benefits received.

Sample: SNAP participants in for 33 states and DC (2,410 counties) across 10 semiannual periods with 24,100 initial observations reduced to 21,690 in final sample for methodological reasons). About 7/8ths of all SNAP participants live in counties in the main sample.

Reinstatement of work-hours test reduces total number of households receiving SNAP by 4.5% in quarter after test reinstatement and by 3% in following quarter. Authors estimate that these overall declines mean that more than one-third of adults subject to the test lost benefits after it was reinstated. Authors estimate that if sample counties are representative of the US as a whole, about 600,000 participants lost about $2.5 billion SNAP food vouchers due to work-hours test between 2013-2017.

Lippold and Remy (2021)

Working Paper

Positive (hours worked)


positive (increased property crime and homelessness)

Summary: Regression-discontinuity design (using unemployment rate cutoffs for USDA waivers of work-hours test) using BLS LAUS, SNAP QC data, and other data to estimate impact of test on employment, SNAP participation and several other outcomes.

Sample: Counties in the US observed in July of each year from 2004-2018, excluding: 1) April 2009 to September 2010 (work-hours test waived nationwide), and 2) all counties in TX, DE, and SD (no work-test waivers except when waived nationwide).

Work-hour test reduces SNAP enrollment, increases hours worked, increases homelessness and property crime.

Mueser et al. (2019)

Book Chapter




Summary: Event-history analysis using SNAP administrative data to estimate differences in SNAP exit probabilities by month in areas with and without work-hours test in Georgia, Missouri, and South Carolina in 2001-2007.

Sample: Nonelderly SNAP households with no children, and two pseudo-control groups: households with children, and households with only elderly members.

Work-hours test reduced SNAP participation among nonelderly households with no children.

Ribar, Edelhoch, Liu (2010)

Journal Article




Summary: Event-history study using administrative data (SNAP and UI wage records) to examine differences in SNAP participation among adult-only households in South Carolina between 1996 and 2005.

Sample: Random sample of adult-only households in SC who began receiving SNAP between October 1996 and December 2005. N = 28,628 participants.

Adult-only households potentially subject to work-hours test had lower rates of SNAP participation than other households (about 20% lower), were more likely to have earnings-related exists, but also higher rates of exits without earnings, suggesting mixed employment effects.

Ritter (2018)

Working Paper




Summary: Regression discontinuity (age 50 cutoff) design using Current Population Survey (Basic Monthly, not ASEC) and SNAP Quality Control data from 2003 to 2017 to estimate effect of work-hours test on employment.

Sample: Main estimation sample is 45- to 55-year olds without high school diplomas who are US citizens and do not live with children under age 18 (from CPS monthly). Additional estimation samples include adults in the age range who are not disabled and do not live with children <18 (from SNAP QC data, which consists of cases that are randomly selected from each state’s SNAP administrative data).

Work-hours test did not increase employment.

Stacy et al. (2018)

Working Paper




Summary: Regression discontinuity (age 50 cutoff) design using SNAP administrative data and ACS data from 2005 to 2015 to estimate effect of work-hours test on employment, SNAP participation, and receipt of disability benefits in nine states.

Sample: Low-income (money income under 250 percent of poverty line) 25- to 54-year olds who do not live with children under age 18 and did not report being disabled.

Work-hours test reduced SNAP participation but did not increase employment or hours worked. No impact on disability benefit receipt, although authors note they did not have administrative data on disability benefit receipt, which may bias estimates.

Negative impact of the work-hours test is larger for two of three subgroups with “worse job prospects” (people living in high unemployment counties and people participating in SNAP in the aftermath of the Great Recession).

Stith (2019)

Working Paper



positive (SSI); none (SSDI)

Summary: Difference-in-differences and event study analysis using administrative data (SSA) and Current Population Survey (both monthly and ASEC) data to estimate effect of work-hours test on applications for and receipt of disability benefits (Social Security Disability Income and Supplemental Security Income, or SSDI and SSI).

Sample: Low-income (money income under 150 percent of federal poverty line) 18- to 49-year-olds who do not live with children. Sample further limited to observations with county identifiers. N=29.539.

Reinstatement of work-hours test has no effect on SSDI or combined SSDI-SSI applications and receipt, but may increase SSI-only applications, especially among sample members with self-reported disabilities, with effects concentrated among women, white people, the less educated and those with disabilities other than blindness.

Vericker et al. (2023)

Journal Article

Mixed, negative or none


reduced net income

Summary: Quasi-experimental study using administrative data (SNAP and UI earnings) from Colorado, Missouri, and Pennsylvania to investigate the impact of work-hours test reinstatement in the mid-2010s on SNAP participation, employment, and earnings.

Sample: 18- to 47-year-olds subject to general SNAP work requirements who do not live in a household with a child <18.

Twelve months after reinstatement, work-hours test reduced SNAP participation by 7 to 32 percentage points, employment by 2 to 7 percentage points, earnings by $247 to $1,230, and combined income from earnings and SNAP by 8 to 21 percent. An alternative analysis found no employment effect in CO and PA and a small positive effect in MO.

Wheaton et al. (2021)

Urban Institute Report




Summary: Mixed methods study funded by the USDA Food and Nutrition Service conducted by the Urban Institute. Qualitative methodology includes interviews with seven Food and Nutrition Service Regional SNAP Directors in 2016. Quantitative methodology includes descriptive and inferential analysis of the demographic and economic characteristics of people potentially subject to the work-hour test including longitudinal trends in SNAP participation in nine states, and estimating the impact of reinstating work-hours test in those states. Also estimates the impact of the work-hours test on employment in three of the nine states.

Sample: Uses administrative SNAP caseload data from VT, PA, MD, TN, AL, MO, CO, OR, MN to look at trends in two groups: 1) “potentially subject” to the work-hours test (SNAP participants 18-49 who are subject to SNAP’s general work requirements and do not live with children under 18); 2) “subject” to the work-hours test (live in an area where the time limit is in effect, are not pregnant, and have not been determined unfit for work). The study uses the broader group (1) to examine results across areas and time periods with and without the test.

Administering the work-hours test is complex and challenging for both local SNAP officials and SNAP recipients. Implementing rule changes takes time and resources.

People subject to the test are among the most likely to need SNAP food vouchers and simultaneously the hardest to reach. Compared to other people receiving SNAP, they are more likely to be disconnected from the workforce, less likely to have traditional means of communication (phone, email, mailing address), more likely to have extremely low incomes, and have lower levels of educational attainment.

Most SNAP E&T programs focus on job search and job search training, but only 39 hours a month in these activities is countable toward the 80-hour work test. States face challenges in funding, providing, and increasing E&T participation among people subject to the test.

Between 4 and 9% of SNAP participants in the study states were potentially subject to the test. More likely to be men and have lower incomes than the broader group of SNAP recipients 18-49. In 5 of the 9 states, people subject to the test were disproportionately Black.

Only about 5 to 12 percent of people subject to the test meet it in the month it becomes binding.

Homeless people are less likely to meet the test than housed people, and men are less likely to meet the test than women.

Reinstatement of the test reduces SNAP participation in 8 of 9 states; reduces employment; and reduces combined value of earnings plus SNAP food vouchers. Vericker et al. (2023), the subsequently published version of these results for three states, is summarized above.


  1.  On SNAP and food insecurity, see, e.g., https://www.journals.uchicago.edu/doi/abs/10.1086/673999 and Gundersen, Craig, Brent Kreider, and John V. Pepper. 2017. “Partial Identification Methods for Evaluating Food Assistance Programs: A Case Study of the Causal Impact of SNAP on Food Insecurity.” American Journal of Agricultural Economics 99 (4): 875–93. https://doi.org/10.1093/ajae/aax026.
  2. Dickinson, Maggie. 2022. “SNAP, Campus Food Insecurity, and the Politics of Deservingness.” Agriculture and Human Values 39 (2): 605–16. https://doi.org/10.1007/s10460-021-10273-3.; Fox, Ashley, Wenhui Feng, and Megan Reynolds. 2023. “The Effect of Administrative Burden on State Safety-Net Participation: Evidence from Food Assistance, Cash Assistance, and Medicaid.” Public Administration Review 83 (2): 367–84. https://doi.org/10.1111/puar.13497.; Craig Gundersen, Brent Kreider, and John V. Pepper. 2018. “Reconstructing the Supplemental Nutrition Assistance Program to More Effectively Alleviate Food Insecurity in the United States.” RSF: The Russell Sage Foundation Journal of the Social Sciences 4 (2): 113. https://doi.org/10.7758/rsf.2018.4.2.06.; Ettinger de Cuba, Stephanie, Mariana Chilton, Allison Bovell-Ammon, Molly Knowles, Sharon M. Coleman, Maureen M. Black, John T. Cook, et al. 2019. “Loss Of SNAP Is Associated With Food Insecurity And Poor Health In Working Families With Young Children.” Health Affairs 38 (5): 765–73. https://doi.org/10.1377/hlthaff.2018.05265
  3. In conventional economic discourse, the assumption is usually that reductions in labor supply are bad “distortions” that policymakers should always avoid. But this is an oversimplification.
  4. H.R. 1510 (“Improving Access to Nutrition Act of 2023”) https://www.congress.gov/bill/118th-congress/house-bill/1510/text?s=1&r=6
  5. H.R. 1581 (“America Works Act of 2023”)  https://www.congress.gov/bill/118th-congress/house-bill/1581 and S. 39 (“Let’s Get to Work Act of 2023”) https://www.congress.gov/bill/118th-congress/senate-bill/39?s=1&r=8
  6. Bergh , K. and Rosenbaum, D. (2023) House Republicans’ proposals could take food away from millions of low-income individuals and families, Center on Budget and Policy Priorities. Available at: https://www.cbpp.org/research/food-assistance/house-republicans-proposals-could-take-food-away-from-millions-of-low (Accessed: April 9, 2023).
  7. “Policy Basics: The Supplemental Nutrition Assistance Program (SNAP).” 2022. Center on Budget and Policy Priorities. https://www.cbpp.org/research/food-assistance/a-quick-guide-to-snap-eligibility-and-benefits; https://fns-prod.azureedge.us/sites/default/files/resource-files/29SNAPcurrPP-3.pdf
  8. Authors’ discussion with Kings Floyd and Kimberly Knackstedt of the Disability Economic Justice Collaborative. The National Center on Disability and Journalism includes “able-bodied” on its list of terms to avoid because they are outdated, imply inferiority or have other negative connotations. See https://ncdj.org/2015/09/terms-to-avoid-when-writing-about-disability/
  9. See, e.g., Leopold et al. (2014)  (“Describing a control group as consisting of “normal” or “able-bodied” individuals conveys specific implications about the treatment groups that seldom are correct, and in some instances, can be both misleading and insulting.”); Badger & Sanger-Katz (2018)
  10. SNAP’s general employment and training rules, as set out in the Code of Federal Regulations (7 CFR 273.7), are as follows. Household members who are elderly, disabled, or caring for young children, or meet other criteria, are exempt. Nonexempt participants must accept suitable employment. SNAP considers all employment suitable unless it meets at least one of the following criteria: pays less than the minimum wage; requires the employee to join, resign from, or refrain from joining any legitimate labor organization; is at a site subject to a strike or lockout; or fails to meet state-established suitability criteria. In a few other limited situations, a participant can decline a job, including if she or he can show it entails an unreasonable risk to health and safety or that it interferes with his or her religious convictions or observances.
  11. “Supplemental Nutrition Assistance Program (SNAP) – COVID-19 Voluntary Quit and Good Cause Policy Clarification.” 2022. Food and Nutrition Service. U.S. Department of Agriculture. https://fns-prod.azureedge.us/sites/default/files/resource-files/COVID-19-voluntary-quit-good-cause-policy-clarification.pdf#page=4
  12. Each year, USDA provides each state with a specific number of “discretionary exemptions”; the number of exemptions equals 12 percent of the number of people ineligible for program benefits because of the ABAWD time limit. FNS (2022) SNAP – Able-Bodied Adults Without Dependents Percentage Exemption Totals for FY 23 – Not Adjusted for Carryover. U.S. Department of Agriculture. Food and Nutrition Service. Available at: https://www.fns.usda.gov/snap/abawds-percentage-exemption-totals-fy23-not-adjusted
  13. “A Short History of SNAP.”  2018. Food and Nutrition Service | U.S. Department of Agriculture. 2018. https://www.fns.usda.gov/snap/short-history-snap#2009.
  14. Wheaton et al. (2021); Brian Stacy, Laura Tiehen, and David Marquardt. Using a Policy Index To Capture Trends and Differences in State Administration of USDA’s Supplemental Nutrition Assistance Program, ERR-244, U.S. Department of Agriculture, Economic Research Service, February 2018. https://ageconsearch.umn.edu/record/276250
  15. Office of the Inspector General and USDA (2016) FNS Controls Over SNAP Benefits For Able-Bodied Adults Without Dependents. Available at: https://www.usda.gov/sites/default/files/27601-0002-31.pdf
  16. Wheaton, L, Vericker, T., Schwabish, J., Anderson, T., Gasper, J., Sick, N., Werner, K., and Baier, K., 2021. The Impact of SNAP Able Bodied Adults Without Dependents (ABAWD) Time Limit Reinstatement in Nine States. https://www.urban.org/sites/default/files/publication/104451/the-impact-of-snap-able-bodied-adults-without-dependents-abawd-time-limit-reinstatement-in-nin_0.pdf
  17. Bauer, L., Schanzenbach, D.W. and Shambaugh, J. (2018) Work Requirements and Safety Net Programs. rep. The Hamilton Project. Available at: https://www.hamiltonproject.org/assets/files/WorkRequirements_EA_web_1010_2.pdf (Accessed: April 3, 2023).
  18. See Figure 10
  19. Wheaton et al. (2021) found that “men make up between 31 and 43 percent of SNAP participants ages 18 to 49 in the study States but account for 52 to 64 percent of participants potentially subject to the time limit and 50 to 65 percent of ABAWDs subject to the time limit.”
  20. Meyer, Bruce D., Nikolas Mittag, and Robert M. Goerge. “Errors in survey reporting and imputation and their effects on estimates of food stamp program participation.” Journal of Human Resources 57, no. 5 (2022): 1605-1644. https://doi.org/10.3368/jhr.58.1.0818-9704R2.
  21. Ku, Leighton, Erin Brantley, and Drishti Pillai. "The effects of SNAP work requirements in reducing participation and benefits from 2013 to 2017." American Journal of Public Health 109, no. 10 (2019): 1446-1451. https://doi.org/10.2105%2FAJPH.2019.305232
  22. Harris, Timothy F. "Do SNAP work requirements work?." Economic Inquiry 59, no. 1 (2021): 72-94. https://doi.org/10.1111/ecin.12948
  23. Cuffey, J., Beatty, T. K., & Mykerezi, E. (2022). Work effort and work requirements for food assistance among US adults. American Journal of Agricultural Economics, 104(1), 294-317. https://ageconsearch.umn.edu/record/205821
  24. Dong, Kimberly R., and Wenhui Feng. "SNAP restrictions that punish people formerly and currently under correctional supervision." Journal of Health Care for the Poor and Underserved 32, no. 2 (2021): 654-663. https://muse.jhu.edu/pub/1/article/794637/summary; Al Abosy, J., Grossman, A. and Dong, K.R., 2022. Determinants and Consequences of Food and Nutrition Insecurity in Justice-Impacted Populations. Current Nutrition Reports, 11(3), pp.407-415. https://link.springer.com/article/10.1007/s13668-022-00421-4
  25. See e.g., Martin Gilens, Why Americans Hate Welfare: Race, Media, and the Politics of Antipoverty Policy (University of Chicago Press, 1999); Jill Quadrango, The Color of Welfare: How Racism Undermined the War on Poverty (New York: Oxford University Press, 1996). Sanford Schram, Joe Soss, and Richard Fording, eds., Race and the Politics of Welfare Reform (University of Michigan Press, 2003); Joe Soss, Richard Fording, and Sanford Schram, Discipling the Poor: Neoliberal Paternalism and the Power of Race (University of Chicago Press, 2011).
  26. Martin Gilens, “‘Race Coding’ and White Opposition to Welfare,” American Political Science Review, 90:3, September 1996.
  27. U.S. Bureau of Labor Statistics, Unemployment Rate [UNRATE], retrieved from FRED, Federal Reserve Bank of St. Louis; https://fred.stlouisfed.org/series/UNRATE, April 05, 2023.

    Support Cepr

    If you value CEPR's work, support us by making a financial contribution.