UC Davis

With obesity rates at epidemic levels in the United States, some public health advocates and policy makers have turned to soda taxes as a way to curb soda consumption. A number of U.S. cities, motivated by health-related or fiscal considerations, have implemented or are considering soda taxes that vary by enactment process, scope, and magnitude. These excise taxes are imposed on sugar-sweetened sodas and non-carbonated drinks such as sport drinks, fruit drinks, and iced teas. In some jurisdictions beverages sweetened with artificial sweeteners are also taxed. This study examines how current soda taxes affect the price and volume sales of soda products and related products. In particular, this study explores not only the short-run effects, but also the long-run effects, an important aspect missing in the literature. I also explore the heterogeneity in the impact across product category and package size, and how responses vary among retailers based on their types. Regressivity is a hotly debated topic surrounding an excise tax. This study investigates whether the tax burdens fall more heavily on the relatively poor and ethnic diverse neighborhoods. I additionally explore whether geographic proximity to neighboring untaxed stores alters the pricing behavior of retailers, as well as their soda sales. Finally, I compare the results among current tax policies and further evaluate their effectiveness. In the empirical work, I mainly use a large national IRI scanner data and employ both a Difference-in-Differences approach and a Triple-Difference approach to estimate the effects of soda taxes. Results suggest that soda taxes in all taxed jurisdictions had a significantly positive impact on the price of taxed beverages, with pass-through rates less than 100%, and retailers raised prices gradually over time, rather than pass on the tax all at once. These results indicate that the estimates of the impact of soda taxes based on short-term data might be under stated. The pass-through rates of soda taxes that were passed through a council budget vote are larger than those of taxes that were passed through public referendum. This result is consistent with the hypothesis that the massive media campaigns before an election can draw consumers’ attention to the tax and to avoid losing sales, retailers choose not to raise the price significantly in the immediate aftermath. Moreover, the volume sales of soda drinks in Oakland, San Francisco, and Boulder did not decline due to soda taxes, and Berkeley’s soda sales did not drop until the fourth year. The main goal of soda taxes enacted by these jurisdictions is to curb soda consumption, and the insignificant consumption effects suggest that establishing soda taxes by public vote may not succeed in achieving the goal of improving public health. Results also suggest that consumers did not switch to untaxed and healthier beverages in response to soda taxes; they may drive across the city border and purchase sodas in neighboring untaxed stores. In Philadelphia, there is evidence that the sales reduction of taxable beverages was larger in stores near the city border than in those farther away from the border. Given the possibility that the nearby stores of a taxed city can be indirectly affected, future research on soda taxes should be cautious about using them as controls in the empirical analysis. The impact of soda taxes exhibited huge heterogeneity across store characteristics and product characteristics. Results show that large retailers passed on a smaller proportion of the tax to consumers than small retailers as they may risk losing all profits from consumers’ shopping baskets if they raise the price by a large percentage and cause consumers to shop elsewhere, e.g., in untaxed jurisdictions. The pattern of the heterogeneous impact across products also varies by store type. For example, the pass-through rate for large bottles of sodas was greater than that for small bottles at large retailers, while small retailers saw the opposite result. These findings suggest that different types of stores adopt different pricing strategies and consider cross-product relationships when responding to the soda tax. Therefore, it is necessary for future research to analyze at a disaggregated level, or at least distinguish between store types. Soda taxes might be regressive in relatively large jurisdictions and the results for soda sales suggest that the soda taxes might be less effective at reducing soda consumption among low-income households who tend to purchase more sodas and are more adversely affected by obesity. The tax burdens are also shown to fall more heavily on racially diverse neighborhoods. This study provides insights on how retailers’ responses to soda taxes can vary across different dimensions, and how soda taxes of varying scope and objective can differ in terms of the effects and effectiveness. With more jurisdictions considering an introduction of soda tax, the findings of this study provide important implications for future tax design.