Academic journal article Journal of Regional Analysis & Policy

A Bayesian Spatial Econometric Analysis of SNAP Participation Rates in Appalachia

Academic journal article Journal of Regional Analysis & Policy

A Bayesian Spatial Econometric Analysis of SNAP Participation Rates in Appalachia

Article excerpt

Abstract. The Supplemental Nutrition Assistance Program (SNAP) is a federal program that provides assistance to low- and no-income people in the United States. Its aim is to increase individuals' food-purchasing power and improve the nutritional content of their diet. We employed recent advances in Bayesian spatial econometric modeling to determine the appropriate model for drawing inferences about the percentage of SNAP recipients in Appalachia. We found that there is significant spatial dependence justifying the use of spatial econometric methods. We also examined how changes in an independent variable affect the dependent variable for orders of neighbors over space.

(ProQuest: ... denotes formulae omitted.)

1. Introduction

The Supplemental Nutritional Analysis Program (SNAP) is the largest food assistance program for low-income households in the United States. The United States Department of Agriculture administers SNAP at the federal level, while state agencies administer the program at the state and local levels (USDA, 2010). States determine eligibility based on income and asset requirements before allocating and distributing benefits using an electronic benefit transfer system. SNAP is currently the cornerstone of the federal food assistance programs and catered to an average of 28.2 million people every month over the last 5 fiscal years (USDA, 2010).

The SNAP program can considerably improve a poor working family's ability to purchase food. However, reports by the United States Department of Agriculture (USDA, 2011) show that not all eligible people take part in the program, recording a 60- 70% participation rate over the last four years (Cunnyngham and Castner, 2009). This has led to an increasing interest among researchers and governmental agencies in determining what influences individuals or households to participate in the SNAP program. Earlier studies on SNAP participation rates and caseloads were carried out at state and national levels using data from 1980 to 2004, with a majority employing Ordinary Least Square (OLS) and Feasible Generalized Least Square (FGLS) methods to conduct their analyses (Figlio, Gundersen, and Ziliak, 2000; Currie and Grogger, 2001; Kornfeld and Wilde, 2002; Kabbani and Wilde, 2003). A study by Goetz, Rupasingha, and Zimmerman (2002) utilized spatial econometric methods to analyze participation in the food stamp program across the United States.

There is a paucity of studies analyzing the effects of social conditions on the SNAP program at the regional level. Furthermore, the use of Bayesian techniques in such studies has not been fully explored in the literature. This study examines the SNAP program at the regional level and controls for spatial autocorrelation. Spatial autocorrelation exists where the dependent variable or the error terms are correlated in a systematic manner over space. Ignoring this condition may lead to improper inferences because the coefficients estimated and standard errors may be biased, inconsistent, or both (LeSage, 1997). We formulated our model based on past studies and used Bayesian spatial econometric techniques to determine which variables affect SNAP usage in the 417 Appalachian counties. We also addressed the model comparison issue in spatial econometric studies before selecting the preferred model and reporting results.

In the rest of the paper which follows, Section 2 gives a brief overview of the past literature, while section 3 develops the empirical model for analysis. Section 4 lays out the Bayesian econometric models. Section 5 presents the Bayesian spatial econometric results, while section 6 describes the impacts over space. Concluding remarks are given in section 7.

2. Literature review

Some studies have analyzed the supplemental Nutritional Analysis Program (SNAP) with regard to caseloads and participation. Studies by Figlio, Gundersen and Ziliak (2000), Goetz, Rupasingha, and Zimmerman (2002), Kornfeld and Wilde (2002), Kabbani and Wilde (2003), and Klerman and Danielson (2009) examined participation in the SNAP program by modeling recipients of the program, while Currie and Grogger (2001) and Ratcliffe et al. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.