Food Stamps, Unemployment Insurance, and the Safety Net

Year: 2006

Research Center: The Harris School of Public Policy Studies, University of Chicago

Investigator: Schroeder, Daniel

Institution: University of Texas at Austin

Project Contact:
Daniel Schroeder
Ray Marshall Center for the Study of Human Resources
LBJ School of Public Affairs
University of Texas at Austin
3001 Lake Austin Boulevard, Suite 3200
Austin, TX 78703
Phone: 512-471-2196
Fax: 512-471-0585


The Food Stamp Program (FSP) and cash assistance were reformed in 1996 and later to emphasize work as a route out of poverty. When employment opportunities were plentiful, as they were in the late 1990s, many families were able to transition off program rolls and into jobs. However, when the employment situation reversed starting in 2000, social supports were needed. This study attempts to determine whether Unemployment Insurance (UI) was a significant source of support for these families, as might be expected because many former welfare recipients should have developed work histories that would have made them eligible for UI benefits. In particular, the study asks whether UI was able to replace or complement food stamps for unemployed, welfare-eligible families.

This study uses a rich data set of electronic case records from Texas on food stamp receipt, UI earnings, and UI claim benefits to examine patterns of employment, unemployment, and program participation over the period 1996-2004. Research using administrative records is advantageous due to large sample sizes and accurate accounts of the beginning and ending dates of assistance spells. The primary disadvantage of this approach is lack of depth in the constructs measured.

Data were assembled into longitudinal data sets for analysis. Descriptive analysis of UI earnings data over time indicates that monetary eligibility for UI among FSP recipients was moderate in the beginning (40-41 percent). However, as expected with the increased emphasis on work, eligibility increased by several percentage points throughout the study period, peaking at 46 percent around 2002. Concurrent receipt of UI benefits among FSP recipients also rose to a peak around 2002, at the height of the recession, at 7 percent, several times higher than its level at the beginning of the study.

Receiving unemployment insurance benefits is associated with increased exit rates from the Food Stamp Program.

Cox proportional hazards regressions were used to identify factors associated with exit from an FSP spell and found that both currently and recently receiving UI benefits was associated with increased rates of exit from the FSP. This effect held even when controlling for earning levels and monetary eligibility for UI benefits, suggesting that UI can act as a substitute for the FSP among families that qualify for and receive this benefit.

Similar attempts to model return to the FSP among former recipients also found strong positive associations with currently and recently receiving UI benefits. Contrary to the deterrent effect that had been expected, this finding suggests that UI might be acting as a first-tier safety net, with the FSP being the second tier. Although there was some concern that receiving UI benefits in this regression may have served as a marker for job loss, inclusion of UI earnings and employment measures in this regression did not diminish the UI benefit effects.

A final attempt to ascertain the role of UI benefits in FSP participation was estimated by modeling takeup, or entry into the FSP, among the entire population of workers with earnings reported to the UI system. Among individuals with a recent decline in earnings (20 percent or greater decline in quarterly earnings within a year), entry into the FSP was found once again to have a strongly positive association with both current and recent UI benefit receipt.

Whereas most prior analysis suggested that the UI role in FSP dynamics did not vary in the economic expansion and recession/post-recession periods, this last model did suggest some variation. Before 2001, concurrent receipt of UI (in the same quarter) strongly predicted entry into the FSP, while recent UI receipt (in the prior quarter) only weakly predicted FSP entry. But in 2001 and later, recent UI receipt is a much stronger predictor of FSP entry, while the estimated effect of concurrent receipt is substantially weaker. This situation can also be observed in descriptive statistics, which show that, in the early years of the study, 5-10 percent of FSP spells were preceded by receiving UI compared with 16-20 percent of FSP spells in 2002 and later. As with the findings regarding re-entry to the FSP, this finding suggests that UI benefits act as a first-tier safety net, with the FSP being the second tier. This trend appears to be increasing over time.

The pattern of the findings suggests that food stamps were not the causal factor in unemployment insurance dynamics.

Similar methods were also used to identify factors associated with exit from and re-entry and entry into spells of receiving UI benefits, but FSP participation was found to play little or no role in UI benefit dynamics. Although indicators for concurrently receiving food stamps were sometimes associated with measures of UI entry and re-entry, prior FSP receipt did not show such associations.

In summary, findings from this study suggested that receiving UI benefits plays a large role in FSP dynamics but that the FSP plays little role in the dynamics of receiving UI benefits. FSP recipients who qualified for and received UI benefits tended to have shorter spells in the FSP, suggesting a substitution effect. Although the analysis did not reveal the expected deterrent effect of receiving UI benefits on FSP participation, UI benefits did seem to increasingly provide a first-tier safety net, whereas the FSP constituted the second tier. These findings suggest that the increased emphasis on work has led to a greater role for UI in the dynamics of the FSP but that room for improvement remains.