The overwhelming objective of the Food Stamp Program, the largest government food assistance program, with annual expenditures of over $32 billion, is to decrease food insecurity among low-income households. Yet, many U.S. households continue to report food insecurity and hunger. USDA estimates that in 2005, 12.6 million households, composing 11 percent of all U.S. households, reported being food insecure with over a third of food-insecure households (4.4 million) experiencing severe food insecurity.
In recent years, considerable attention has been devoted to measuring the impact of food stamp use on food insecurity, but most studies find either no significant relationship between food stamp use and food insecurity or, in some cases, a paradoxical positive correlation. Identifying the effect of food stamp participation on the severity of food insecurity with nonexperimental data has proven particularly difficult, as households with higher severity of food insecurity are often more likely to choose to participate in the Food Stamp Program.
This study uses an instrumental variable model to estimate the causal effect of food stamp participation on food insecurity. State-level food stamp policy variables are used as instruments. The study also explicitly examines the impact that important and often previously omitted factors, like income variability and inelastic expenditures on addictive substances, have on food insecurity. Nationally representative data from the Panel Study of Income Dynamics (PSID) are used for the analysis.
The study finds that food stamp participation is associated with a 27-percent reduction in the severity of food insecurity for the average low-income food-insecure household. Several different model specifications produce estimates of the Food Stamp Program impact that range between a 22-percent and a 29-percent reduction in the severity of food insecurity. The study also finds evidence that smoking habits increase food stamp participation significantly. Income volatility increases food insecurity in moderate-income households but not low-income households. Accumulated wealth, current income, and average income reduce household food insecurity, as do education and age of the household head. Households headed by Blacks show higher food insecurity, as do households that experienced unemployment.
Direct inquiries about this study to the Project Contact listed above.