No Evidence of Food or Alcohol Substitution in Response to a Sweetened Beverage Tax

Abstract

Evidence suggests real-world beverage taxes reduce sweetened beverage purchases, but it is unknown if consumers consequently increase food or alcohol purchases. This study examines whether Philadelphia's 1.5 cents/ounce beverage tax was associated with substitution to 3 kinds of hypothesized substitutes: snacks, nontaxed beverage concentrates, and alcohol. Using commercial retail sales data and a difference-in-differences approach, analyses compared logged volume and dollar sales of snacks and beverage concentrates between 2016 (pretax) and 2017 (post-tax) at chain food retail stores in Philadelphia (n=180) and Baltimore (nontaxed control city; n=60), and logged volume and dollar sales of wine and spirits at liquor stores in Philadelphia (n=44) and nearby Pennsylvania counties (alternate control; n=66). Additional food analyses examined change in logged volume sales of hypothesized products compared to control products (other foods). Analyses were conducted in 2020. Across store types, analyses showed no statistically significant increases in logged volume or dollar sales of snacks or spirits in Philadelphia stores compared to control sites (decreased, ranging from -10% to 0%). Supermarket analyses showed substitution to nontaxed beverage concentrates (27% increase in volume, 36% increase relative to other food) but remained a relatively small percentage of overall beverage dollar sales (12% at baseline, 15% at post). At the population level, there is no evidence that Philadelphia's decline in taxed beverage purchases is offset by increases in snacks or spirits purchasing, but there is evidence of substitution to beverage concentrates in supermarkets. Future studies should explore individual-level purchasing changes.