Food stamp caseloads over the business cycle.

• Author: Ziliak, James P.
• Position: Illustration

1. Introduction

   The onset of the current economic downturn following the longest expansion in U.S. history has renewed interest in the role of automatic stabilizers designed to insure households against negative cyclical income shocks. This system of income and consumption stabilizers includes, among others, the individual income tax and the panoply of social insurance programs, such as unemployment insurance and the Food Stamp Program. While the income tax and unemployment insurance have received recent attention (Gruber 1997; Auerbach and Feenberg 2000; Kniesner and Ziliak 2002a, b), little is known about the role of food stamps as an automatic stabilizer even though the program served over 27 million Americans at its peak in 1994. This gap in the literature is particularly acute in light of passage of the Personal Responsibility and Work Opportunity Reconciliation Act of 1996 (Welfare Reform Act), which introduced new rules on cash-assistance recipients and thus the nearly half of food stamp recipients who also receive ca sh welfare. Moreover, the 1996 welfare reform also had a direct administrative effect on the Food Stamp Program by ending the eligibility of some recipients, reducing average benefit levels, and requiring states to replace paper coupons with Electronic Benefit Transfer cards. Importantly, the Welfare Reform Act also eliminated the entitlement status of the main cash-welfare program, Aid to Families with Dependent Children (AFDC), thus positioning food stamps as a more prominent countercyclical consumption stabilizer.

   A first task in assessing the possible stabilization role of food stamps in the era of welfare reform is to identify the cyclical sensitivity of food stamp caseloads. If food stamp caseloads are acyclical, then we would expect no role for food stamps in smoothing consumption over the business cycle. If, however, food stamp caseloads are highly cyclical, then food stamps may function as an important antirecessionary tool. By the same token, the new welfare policies may independently affect food stamp caseload movements and interact with the business cycle, thereby altering the possible stabilization role of the program. In this paper, we specify a dynamic model of food stamp caseloads to estimate the responsiveness of the caseload to the business cycle, to welfare policies, and to interactions between the business cycle and welfare policies.

   Unlike cash welfare, there has been scant research on the cyclicality of food stamp caseloads. (1) Wallace and Blank (1999) are a recent exception in their use of both static annual and dynamic monthly food stamp caseload models based on state-level panel data for the 1980 to 1998 federal fiscal years. With annual data, food stamp caseloads were strongly countercyclical, and reform of AFDC led to weak declines in total caseloads. (2) Specifically, they attribute up to 44% of the 1994 to 1998 food stamp caseload decline to economic conditions and about 6% of the decline to welfare reform. They reach broadly similar conclusions with monthly data. However, Wallace and Blank calculate that upward of 85% of the post-1996 decline in food stamp caseloads can be attributed to welfare reform if one is willing to ascribe all the unexplained residual to welfare reform, which as the authors note is undoubtedly an overestimate of welfare reform's effect on food stamp caseloads (p. 85). In their static models, Figlio, Gund ersen, and Ziliak (2000) reached conclusions similar to Wallace and Blank's estimates with annual-level data. In their preferred dynamic models, they attribute about 35% of the 1994 to 1998 caseload decline to the macroeconomy and virtually nothing to welfare reform.

   We improve on this previous research on food stamp caseloads along several dimensions. First, we not only follow other research in estimating the impact of AFDC policy changes on food stamp caseloads but also examine the impact of policies that are focused directly on food stamps. Specifically, we consider how food stamp caseloads respond to state decisions regarding the introduction of Electronic Benefit Transfer cards, waivers from the work requirement for unemployed able-bodied adults without dependents (ABAWDs), and administrative error rates. Additionally, we consider an alternative method of modeling welfare waiver variables because changes implemented after the Welfare Reform Act may be quite different in effect from those introduced beforehand. Instead of specifying the AFDC policy reform as an aggregate "any waiver" variable, we distinguish policies that were implemented before the Welfare Reform Act from those implemented after the act. Third, in contradistinction to previous work by Wallace and Bla nk (1999), our dynamic model directly admits aggregate macroeconomic conditions that more fully capture national reforms such as the expansions in the Earned Income Tax Credit in the mid-1990s. This is likely to be important, as Wallace and Blank (1999) report that the magnitude of the estimated impact of welfare reform on AFDC caseloads declines by about half with even rudimentary macroeconomic controls. Fourth, because most AFDC recipients receive food stamps and nearly half of food stamp recipients receive AFDC, we examine the links between food stamp and AFDC caseloads and for possible changes in these links after welfare reform. Finally, as states with robust economies may foster a more hospitable environment to implement welfare reform (e.g., transitions from welfare to work are more rapid), we test for interactions between welfare reform and macroeconomic performance. (3)

   Using state-level panel data for federal fiscal years 1980 to 1999, our results demonstrate a strong countercyclical relationship between macroeconomic activity and food stamp caseloads. A one-percentage-point decrease in the unemployment rate leads to a 2.3% decrease in food stamp caseloads after one year and upward of an 8% decrease in the long run. These results suggest an important role to be played by the Food Stamp Program to counteract the impact of negative income shocks. We also find evidence, after the Welfare Reform Act, of a breakdown in the traditionally strong link between TANF and the Food Stamp Program.

2. Empirical Model

   To gauge the potential countercyclical insurance role of food stamps, one could employ aggregate time-series data. However, time-series data mask important heterogeneity in caseload movements across states because of heterogeneity in state economic conditions and state variation in the types and timing of adoption of welfare policies. This masked heterogeneity in macroeconomic conditions is readily apparent in a comparison of Figure 1, which contains aggregate food stamp caseloads and unemployment rates, and Figure 2, which contains state-specific changes in caseloads between 1984 and 1989. In Figure 1, it is clear that aggregate unemployment fell substantially in the 1980s expansion but that food stamp caseloads declined only slightly. In Figure 2, however, we observe the reasons for this modest aggregate decline because while many states experienced caseload declines, others experienced caseload increases.

   A likely source of the varied caseload experience is due to cross-state differences in business cycle conditions. For example, in the 1980s, while many states experienced robust growth (e.g., New England), other areas suffered recessions (e.g., the oil bust in Texas). Indeed, in Figure 3, across the two aggregate business cycles over the past 20 years, we see the substantial heterogeneity in state unemployment rates, ranging from a low of 3.7% in Nebraska to a high of 10.4% in West Virginia. For modeling purposes, then, one should admit this state heterogeneity to more accurately identify the cyclicality of food stamp caseloads.

   More recently, movements in food stamp caseloads have likely been affected by the radical changes in the administration of welfare programs, changes that affected both cash assistance and in-kind programs. The momentum toward passage of the Welfare Reform Act began in the early 1990s when the U.S. Department of Health and Human Services (DHHS) selectively granted states' requests for waivers from federal AFDC requirements. These waivers included policies such as terminal time limits, work requirements, and personal responsibility measures. Thus, in addition to meeting the usual sequence of income and asset tests in order to qualify for program benefits, recipients in the 35 states with statewide waivers had to satisfy many new rules. Passage of the Welfare Reform Act codified these state-specific waivers into federal law.

   The benefits distributed by the Food Stamp Program are federally funded, and it is an entitlement program. States are, however, responsible for the administration of the Food Stamp Program. While many of the rules for food stamps are the same across all states (e.g., the benefit levels are established in the same manner), states still have substantial autonomy with the construction of the Food Stamp Program. There are three primary ways this autonomy has manifested itself in recent years. First, even though all states have the same benefit calculation formula, the ways and frequency with which this information is garnered can differ widely. For example, states can have shorter recertification periods if they believe this will allow them to more accurately calculate recipient benefit levels and eligibility status. Second, states had control over when they implemented the new Electronic Benefit Transfer card. This card is operationally similar to an ATM card and is replacing the previous method of dispensing fo od stamp benefits: paper coupons. The Electronic Benefit Transfer program is designed to reduce the stigma associated with food stamp use in stores, to prevent theft and loss of benefits, to impede misuse and illegal resale of benefits, and to improve the distribution of benefits. Third, the...