Introduction Directory UMM :Data Elmu:jurnal:E:Economics Letters:Vol67.Issue1.Apr2000:

Economics Letters 67 2000 75–85 www.elsevier.com locate econbase The effect of food stamp cashout on undernutrition a b c , John A. Bishop , John P. Formby , Lester A. Zeager a Department of Economics , East Carolina University, Greenville, NC 27858-4353, USA b Department of Economics , Finance, and Legal Studies, University of Alabama, Tuscaloosa, AL 35487-0224, USA c Department of Economics , East Carolina University, Greenville, NC 27858-4353, USA Received 25 January 1999; accepted 9 September 1999 Abstract We investigate the effect of food stamp cashout on undernutrition using household data from cashout experiments in Alabama and San Diego, California. We apply dominance methods to test for differences in nutrient distributions at or below the recommended dietary allowances.  2000 Elsevier Science S.A. All rights reserved. Keywords : Food stamp program; UnderNutrition; Cashout; Stochastic dominance JEL classification : I38 Provision and effects of welfare programs

1. Introduction

The architects of the US Food Stamp Program FSP aimed to help low-income households eat a nutritious diet by supplementing their food purchasing power. Since its inception, the FSP has been providing benefits in the form of food coupons rather than cash, reflecting the popular notion that coupons are more likely than cash to raise food consumption and improve nutrition. The effect on recipient behavior of in-kind versus cash benefits has been a longstanding interest of economists Southworth, 1945, but direct comparisons of consumer responses to cash and coupon benefits were not possible until recently. To investigate the issue, the US Department of Agriculture USDA – the agency responsible for oversight of the FSP – sponsored cashout demonstration projects in 1989. Two of these projects were designed as classic experiments. In cooperation with state and local social service agencies, USDA randomly assigned large numbers of households to treatment cash and control food coupon groups Corresponding author. Tel.: 11-252-328-6408; fax: 11-252-328-6743. E-mail address : zeagerlmail.ecu.edu L.A. Zeager 0165-1765 00 – see front matter  2000 Elsevier Science S.A. All rights reserved. P I I : S 0 1 6 5 - 1 7 6 5 9 9 0 0 2 4 4 - X 76 J .A. Bishop et al. Economics Letters 67 2000 75 –85 in Alabama and San Diego, California, which provides an ideal opportunity for comparing the effects of cash and coupons on actual behavior. In both experiments, substantial proportions of households have nutrient deficits, especially for folacin, vitamin E, and calcium. Other households are well above requirement thresholds and are not undernourished. To compare relative undernutrition in the treatment and control groups, we apply dominance methods as proposed by Kakwani 1989. Comparing the nutrient distributions rather than 1 summary statistics e.g. the distribution means permits us to focus on undernourished households. Kakwani 1989 has shown that cumulative distributions of nutrients can be used to evaluate relative undernutrition in the same way that cumulative distributions of incomes have been used to evaluate relative poverty Foster and Shorrocks, 1988. Kakwani’s approach has been applied in India Kakwani, 1989, Indonesia Ravallion, 1990, 1992, Puerto Rico Bishop et al., 1996, and the United States Bishop et al., 1992. Statistical inference procedures have also been developed for identifying significant differences between two cumulative distributions. We apply Kakwani’s approach and the statistical inference procedures to data from the cashout experiments. A dominance approach is well suited to the cashout data for two reasons. First, the classic experimental design of the cashout demonstrations ensures that any differences between the nutrient distributions of cash and coupon recipients are attributable to the treatment variable. Second, the specific form of the distribution of nutrient requirements in the population does not need to be known; we need only assume the populations have the same requirement distributions — a reasonable assumption for comparisons of the treatment and control groups in the cashout experiments. The remainder of the paper is organized as follows. The next section provides a brief description of the methods used in the paper. After describing the data from the experiments, the empirical results are presented and discussed. A final section offers concluding comments.

2. The dominance approach to comparing relative undernutrition