The Dynamics of Food and Housing Assistance Before and After the Great Recession

Year: 2015

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

Investigator: Wilmarth, Melissa J., Martin C. Seay, and Robert B. Nielsen

Institution: University of Alabama

Project Contact:
Melissa J. Wilmarth
University of Alabama
Department of Consumer Sciences
305 Adams Hall
Box 870158
Tuscaloosa, AL 35487
Phone: 205-348-7954


Despite extensive literature on people who receive food and housing assistance separately, a relatively small literature investigates the characteristics of households that experience hardships related to food and housing simultaneously. Further, little research exists on potential cross-programmatic benefits of assistance programs that target each of these hardships individually. Where cross-program benefits exist, hardships related to one basic need may be reduced through programs that address other basic needs. To inform both researchers and policymakers about these possibilities, this study investigated the following research questions:

  1. What are the characteristics of households that receive food assistance and/or housing assistance?
  2. What are the characteristics of households that are, and are not, housing-cost burdened?
  3. Does the receipt of food assistance covary over time with the receipt of housing assistance?

Data from the 2004 and 2008 panels of the Survey of Income and Program Participation (SIPP) were used to investigate these questions. The SIPP is a nationally representative survey of the noninstitutionalized, civilian population conducted by the U.S. Census Bureau. These two separate panels combine to provide data corresponding to calendar years 2004 to 2012.

The first two research questions were investigated using parallel methods. First, bivariate analyses identified the characteristics of households that received housing assistance, food assistance, and were housing-cost burdened either moderately (housing-related expenses comprise 30–50 percent of annual household income) or severely (housing-related expenses comprise more than 50 percent of annual household income). Next, logistic and ordinary lest squares regressions predicted the amount of time that households received food and/or housing assistance and the duration of four-year panels in which households were housing-cost burdened.

The third research question exploited each panel’s 12 waves (4 years) of data to understand how the receipt of assistance in a previous time period was associated with receipt of assistance in a later time period. These structural equation models with reciprocal and lagged assistance indicators for both food and housing assistance included control variables for time-varying measures of household composition, sociodemographic characteristics, and key indicators of material hardship and disability. Finally, a lagged fixed-effects logistic regression model investigated whether the receipt of food or housing assistance reduced the incidence of later housing-cost burden.

Results informing the first two research questions indicate that household characteristics vary by the receipt of food and housing assistance, as well as by housing-cost burden status. Households that received food assistance were more likely to be housing-cost burdened, and they also spent a greater percentage of time being housing-cost burdened. The opposite was true for the receipt of housing assistance; the receipt of housing assistance was associated with a reduction in the amount of time a household was housing-cost burdened. Despite apparent evidence of the efficacy of housing assistance programs on housing-cost burden, no evidence was found to suggest that food assistance lowered households’ odds of being housing-cost burdened.

Results from the third research question provided no evidence to support the hypothesis that cross-programmatic benefits exist in terms of food and housing assistance. Rather than reducing the need for assistance, consistent results from before and after the recession showed that prior receipt of one type of assistance was associated with higher odds of receiving the other assistance type. The positive relationship between the two types of assistance held true when estimating single-year models, longitudinal fixed-effects models predicting each type of assistance individually, and generalized structural equation models with lagged reciprocal predictors that allowed for the two types of assistance to be predicted simultaneously. Moreover, no evidence was found to suggest that the prior receipt of food assistance lowered the likelihood of being housing-cost burdened. Indeed, the opposite was found. The receipt of food assistance was, with just one exception in the 2004 panel, positively associated with the odds of being housing-cost burdened in a later period. This result may suggest that, given relatively easy accessibility, households first seek food assistance while waiting on housing assistance to alleviate housing-cost burden, though additional research is needed to assess this possibility.

Overall, the results suggest that there is a positive and potentially complementary relationship between housing and food assistance, but the exact dynamics are complex. Clearly, the best predictor of receipt of assistance is earlier receipt of that same assistance. This consistent, positive relationship was accompanied by no evidence that households seek housing assistance or food assistance to alleviate overall economic hardship instead of the specific category of hardship. Receipt of food assistance may come easier than housing assistance and is perhaps the point where households first receive help, possibly because of program-related obstacles encountered when seeking housing assistance.