State and local governments invest millions of dollars in social programs, expecting to generate some community level economic benefit from these expenditures. Most base estimates of economic activity on multipliers developed for national scale economies, and few ever generate the amount of revenue projected or jobs gained predicted by national multiplier models. Nuances in local and state conditions may be lost in macroeconomic forecasts. This study employs a well-used predictive methodology, input-output modeling, at the state and county level to compare estimates generated with models of the federal hunger relief program known as the Supplemental Nutrition Assistance Program. The results show that more localized economies cannot achieve the results predicted by national models, likely because of the leakage of economic activity outside the boundaries of a single state or county.

This study may shed light on methodological tools that local, regional, and state forecasters and policymakers can use in making decisions about social safety net programs and in understanding what sorts of economic impact may be associated with them. This paper assists in providing a methodology for states and localities to use in assessing the economic benefits of SNAP spending in communities by generating three types of multipliers. The first, a Type I multiplier, focused on direct and indirect effects of the number of jobs directly created by increased SNAP spending, perhaps including grocery store clerks or farm hands. The Type II multipliers produced induced effects arising from labor income (e.g, increased SNAP spending resulting in hiring truck drivers to transport goods from distribution centers to food retailers).  The third, a Type SAM multiplier, used a social account matrix to estimate induced effects generated by changes in household spending.

Exogenous changes within a geographic region were modeled by specifying the desired multipliers (Type I, II, or SAM), including government expenditures and payments. Using the industry standard methodology for economic impact analyses and I/O modeling, this research utilizes data compiled from a variety of federal, state and local data sets to derive estimates of direct, indirect, and induced economic impacts of SNAP spending at the local level.  The data for this study came from a variety of sources, including 2009 SNAP participation reports obtained from the North Carolina Department of Health and Human Services (NCDHHS, 2012), the 2010 Decennial U.S. Census (Census, 2010), and IMPLAN. Exogenous demand changes were modeled using Sector 324, Retail Food and Beverage Stores in IMPLAN 3.0.

The study is important for the policy guidance it can offer. It is common for state and local officials to pursue additional federal dollars as part of community economic development strategies. These strategies typically include attracting, retaining, or expanding military bases, attracting or retaining federal transportation dollars for transit or highway projects, attempting to win federal contracts or grants, or even attracting federal government facilities or regional offices. And this paper argues that modeling transfer program participation at the appropriate unit of analysis is an important step in the generation of reliable estimates for programmatic and policy decisions.