Heterogeneous response and spillover effects of sugar-sweetened beverage taxes

In this paper we evaluate the tax on sugar-sweetened and artificially sweetened beverages by exploiting a natural quasi-experiment in Spain: in 2021 an increase in Value Added Tax was introduced in all regions of Spain except a few that do not have VAT. These regions serve as a control group, offering a unique opportunity in the literature because, in addition, unlike most previous work, which uses sales data, we use microdata from a rich household survey with information on their spending and quantities purchased. We find a pass-through of the tax to prices of more than 90% and a 12% drop in soft drink consumption among the poorest tercile of households, especially among those with children aged between 5 and 16. As a novel finding, we find an almost equivalent reduction in spending on unhealthy complementary goods among the same households in the first tercile. However, the rest of households did not react to the tax by reducing their consumption of soft drinks or of complementary goods. Our results show the importance of considering household structure, as well as the response in the consumption of complementary goods, when evaluating the effect of this type of tax on consumption.