71st International Atlantic Economic Conference

March 16 - 19, 2011 | Athens, Greece

Food Stamp Participation and Obesity: Matching Estimators and Their Sensitivity

Thursday, 17 March 2011: 10:20
Achilleas Vassilopoulos, M.B.A. , Agricultural Economics and Rural Development, Agricultural University of Athens, Athens, Greece
Rodolfo M. Nayga Jr., Ph.D , Department of Agricultural Economics and Agribusiness, University of Arkansas, Fayetteville, AR
Panagiotis Lazaridis, Ph.D , Agricultural Economics and Rural Development, Agricultural University of Athens, Athens, Greece
Over the last few decades, the prevalence of obesity among US citizens has grown rapidly. Oddly, lower income individuals and households exhibit higher prevalence rates. This has led to questioning the effectiveness of nutritional assistance programs such as the Supplemental Nutrition Assistance Program (SNAP), formerly known as the Food Stamps Program (FSP). The FSP is by far the largest nutrition assistance program in the US. As implemented in 1964, it was designed as an entitlement program that targets low income people to help them acquire the necessary amount of food that is needed to maintain good health, thus to alleviating hunger by distributing coupons that could only be used to purchase food at grocery stores.  Currently, electronic benefit transfers that operate essentially as debit cards have replaced food stamp coupons. At least 35 million individuals nation-wide participate in this program. Eligibility and benefits are based on household size, household assets, and income.  FSP benefits are given to a single person or family who meets the program’s requirements pertaining to income, assets, work and immigration status. Most benefit periods last for 6 months but some can be as short as 1 month or as long as 3 years. The FSP is also one of the most thoroughly surveyed programs in the literature with respect to its relevance with various measures of well-being.

The results from previous studies in the literature are often contradictory. What separates our analysis from previous studies is that we attempt to take into account the complex endogeneity issues inherent in these types of analysis and test the robustness of our results to deviations from functional form assumptions. In this study we examine the effect of FSP participation on the likelihood of being obese among adult individuals using data from the National Health And Nutrition Examination Survey (NHANES) of 2005-06. To address causality we apply a propensity score matching estimator and scrutinize our results through an extensive sensitivity analysis procedure taking into account the misclassification errors as well as the weaknesses of our econometric technique.