The impact of taxing sugar-sweetened beverages on cola purchasing in Catalonia: an approach to causal inference with time series cross-sectional data

Sugar-sweetened beverage (SSB) taxes related to the quantity of sugar have appeared as a popular regulatory tool around the world during the last decade showing important variations in their implementation and impact. We evaluated the impact of a new SSB tax implemented in Catalonia since May 1, 2017 on the purchased quantities and penetration rates of taxed and untaxed cola beverages. We use aggregate time series of cola beverages purchases in all 17 Spanish Autonomous Communities before and after the implementation of the SBB tax in Catalonia, from January 2013 to June 2019. A comparison between two different types of causal inference methods was conducted: a two-way fixed effects difference in differences model and a modified synthetic control model. Regular cola purchases decreased 12.1% and their penetration rate decreased by 1.27 points during the two post-intervention years using the preferred model. Diet cola purchases increased 17.0% and their penetration rate also increased by 1.65 points. Only regular cola results were robust to all placebo test checks. The SSB tax implemented in Catalonia in 2017 significantly reduced the volume and penetration rates of regular colas with no robust evidence for the substitution effect on diet colas.