The effects of food sales taxes on household food spending: An application of a censored cluster model

This study proposes a model of household food spending that accounts for zero censoring and can be applied to data collected through a clustered survey design to investigate the impact of food sales taxes on three groups: households who are eligible for and participate in the Supplemental Nutrition Assistance Program (SNAP), households who are eligible but do not participate in the program, and households who are not eligible for the program. We find that SNAP participating households are largely insensitive to grocery taxes and respond to restaurant taxes by shifting more of their food dollars towards at-home foods. Among households who are eligible for SNAP but do not participate in the program, grocery taxes reduce spending on foods purchased for at-home consumption, and thereby increase the amount of the total food budget allocated to away from home foods. This is concerning from a nutrition and health standpoint since away from home foods tend to be more calorie dense and nutritionally poorer than at home foods.