SNAP Participation and Employment Among Childless Adults: The Effects of Restriction Waivers After the American Recovery and Reinvestment Act of 2009

Year: 2013

Research Center: Institute for Research on Poverty, University of Wisconsin-Madison

Investigator: Mykerezi, Elton, Timothy Beatty, and Joel Cuffey

Institution: University of Minnesota

Project Contact:
Elton Mykerezi
University of Minnesota Extension
218-F Ruttan Hall
1994 Buford Avenue
St. Paul, MN 55108
Phone: 612-265-2749


The Supplemental Nutrition Assistance Program (SNAP) is one of the cornerstones of the U.S. social safety net. As such, changes to SNAP were included in the set of tools designed to mediate the effects of the Great Recession as part of the 2009 American Recovery and Reinvestment Act (ARRA). In particular, one of the main changes was a 1-year waiver of work requirements that limits participation of Able Bodied Adults Without Dependents (ABAWDs) to 3 out of 36 months. SNAP participation during and after the recession hit record highs, with substantial increases in atypical participants such as younger adults with no children and with relatively high education levels. Many of these participants were first-time program users. One of the main scopes of this study was to investigate the impact of ABAWD restriction waivers on this increase.

Additionally, one of the primary concerns with operating a social safety net is that it may generate unproductive incentives, especially for means-tested programs, since eligibility and benefits are affected by earnings. ABAWD work restrictions were introduced as a way to limit such incentives among young adults, without affecting the well-being of children.

The Personal Responsibility and Work Opportunity Reconciliation Act (PRWORA) of 1996 changed the welfare system. SNAP participation of childless adults between the ages of 18 and 49 was limited to 3 months out of every 3 years, unless they worked a minimum of 20 hours per week (or earn 20 hours’ worth of minimum wage pay), or satisfied State employment and training or workfare requirements. States were given an option to waive these restrictions in areas (cities, counties, or regions) where adults would face substantial difficulty in finding work. These “labor surplus areas” are defined in a variety of ways, including unemployment of 10 percent or more over the last 3 months or year; unemployment 20 percent higher than the national average for 24 months; and labor surplus area designation by the Department of Labor. USDA’s Food and Nutrition Service (FNS) considered State requests for waivers periodically (usually each year, with 2-year waivers allowed after 2006), but localities covered by FNS-approved waivers still could choose not to waive restrictions. In addition, the Balanced Budget Act (BBA) of 1997 gave States the option to waive restrictions for up to 15 percent of their ABAWDs at State discretion. Finally, States were also eligible for statewide waivers if the Department of Labor determined that they were eligible for extended unemployment benefits. In 2009, a 12-month nationwide waiver was implemented as part of the ARRA.

This study obtained all State requests and FNS responses related to ABAWD waivers since 2003. Also, an email survey of State human services representatives was conducted to collect information on how the 15-percent exemptions of ABAWDs were used and instances in which localities had chosen not to implement FNS-approved waivers. A database of county-level program status between 2004 and 2011 was compiled based on FNS and State documents for more than half of all States.

Data from the American Community Survey (ACS) were used to examine the impact of ABAWD waivers on SNAP participation, employment and various other workforce outcomes. The ACS was chosen because of its large sample size (because ABAWDs are only a small fraction of SNAP participants) and because it provides geographic designations for sub-State regions.

Empirically, the challenge is to separate the effects of SNAP ABAWD policy from other plausible explanations of changes in program participation and employment, especially since waivers were granted primarily based on local unemployment rates. A generalized triple- differences approach was used to identify the causal effects of ABAWD waivers.

Identification relies on two primary features of the program. First, while most areas changed status at least once, others held waivers throughout the first decade of the 2000s, and some chose not to implement them at any point before or after ARRA. This allows comparisons of trends in outcomes of 18- to 49-year-olds across areas. Trends for those in areas experiencing a program change were compared to those of adults in areas that had no program change throughout the study period.

Second, adults are no longer subject to restrictions at age 50. This allows comparisons of 50- to 65-year-old adults to slightly younger counterparts within areas that experience at least one program change. In general, program effects can be separated from year-specific shocks because waivers were granted in different years across areas. Also, comparisons of areas that never changed status (before or after ARRA) to areas that only changed status once in 2009 because of the ARRA are particularly meaningful. That is because this was a nationwide waiver that did not involve effort from Sates to apply, it was not based on local area unemployment rates (unlike in previous years), and it was fairly well-advertised.

A summary of policy data indicated that States varied widely in how they approached ABAWD policy in terms of both requesting waivers and using the 15-percent exemptions afforded by the BBA. In terms of exemptions, some States put forward their best efforts to exempt as many areas as possible, efforts which sometimes led to FNS rejections. Other States chose to apply only for areas that were clearly eligible (for example, designated labor-surplus areas), while others failed to apply at all in certain years. Some of these behaviors were dictated by politics. In response, at least two States implemented legislation that compels their human services agencies to apply for waivers every year, removing the choice from annual politics. They also used their 15-percent exemptions differently; some waived specific areas (often those included in an application that was rejected by FNS), others waived certain adults only (for example, 47 years or older in certain areas in Maryland), and others distributed them to nonwaived counties in proportion to historical ABAWD caseloads.

The study also found some evidence that ABAWD waivers increased SNAP participation among young adults, but the effect of waivers was far smaller than that of worsening economic conditions due to the recession. Participation across all areas had already started to increase well before the 12-month nationwide waiver implemented as part of ARRA was in place.

The study considered three outcomes related to employment: (1) workforce participation, to account for discouraged workers; (2) incidence of unemployment; and (3) working for at least 20 hours per week (the restriction cutoff). Theory would predict the last outcome to be the most sensitive to this policy because the marginal wage rate spike associated with working just below 20 hours to just above is highest. No evidence was found that ABAWD restriction waivers had any adverse effects on any labor force outcome.