October 27, 2020
In a forthcoming paper in American Economic Journal: Economic Policy, PSTC economists Jesse Shapiro and Justine Hastings, along with Ryan Kessler, a recent Brown Economics Ph.D. graduate, used data from a retail panel to understand how participation in the Supplemental Nutrition Assistance Program (SNAP) affects the composition of household food purchases. They conclude that policymakers aiming to close socioeconomic gaps in diet quality should consider measures beyond the current program design to effectively serve low-income populations.
Shapiro and Hastings, who frequently study the intersection of public policy and economics, say that this research evolved from their 2018 study, which used some of the same data to evaluate the effect of SNAP on household spending patterns. Of the 2018 findings, Shapiro noted, “This led us to wonder how SNAP affected the composition of foods purchased, and our forthcoming paper grew out of our interest in that question.”
The researchers utilized detailed transaction data from a large US grocery retailer in conjunction with nutritional information to track the composition of purchased foods for each household, which included thousands of households that transitioned to SNAP benefits during the course of the panel – nearly seven years. “We study how the composition of foods purchased by these households changes when they enter the program, using statistical tools to account for simultaneous changes in factors like income,” Shapiro explained. “As in our earlier study, we can also measure how total food spending changes when households enter the program. This allows us to relate changes in food spending to changes in food composition.”
Shapiro added, “Our estimates imply that helping low-income households to increase their food spending may not help to narrow socioeconomic differences in the healthfulness of purchased foods. Policymakers interested in pursuing that goal may therefore want to explore other policy tools.”
A pre-print version of the article is available here.