The Role of Functional Form in Estimating the Effect of a Cash-Only Food Stamp Program

Larger marginal impacts on household food spending come from food stamps than from equivalent cash income, according to previous studies. These studies have been limited, however, in using food expenditure equations that are linear in the coefficients, placing prior constraints on the estimated marginal propensities to spend (MPS) from cash and from stamps. This article re-examines the earlier MPS estimates in light of a more general and flexible food expenditure equation, comparing estimates from a common data set under alternative functional forms. The article estimates changes in food spending that would result under each functional form from "cashing out" food stamps, and replacing stamps with equivalent cash benefits. Results show that MPS estimates vary widely depending on functional form, past estimates tended to substantially exaggerate the cash-out effect in reducing food spending, and the most general, consistent, and flexible forms show a 10-cent reduction in food spending for each dollar of food-stamp benefit shifted to a cash payment.