If a tax could be considered trendy, the label would aptly apply to the soda tax. In 2013, no U.S. city levied a tax on sugary drinks. Today, seven major cities do so (see also here). Such taxes have also been enacted at the national level in a diverse group of countries including Mexico, France, and Sri Lanka. The tax’s proliferation heightens the need for data on its efficacy, especially when implemented at the local level. Stephan Seiler, Anna Tuchman, and Song Yao’s recent work on Philadelphia’s tax on sugar-sweetened beverages responds to such a need. Their research offers valuable insights for public health advocates and policymakers considering a soda tax, whether as a source of revenue or as a response to increasing obesity rates.
Philadelphia’s tax is structured as a 1.5 cent-per-ounce levy on sugar-sweetened and artificially-sweetened beverages. Although advocates typically present the tax as a public-health measure, Philadelphia implemented the tax primarily to raise revenue. This revenue goal explains the inclusion of artificially-sweetened beverages, which are typically excluded from taxation as a healthier, zero-calorie alternative to sugary drinks.
To measure how Philadelphia vendors and shoppers responded to the tax, the authors examine sales data from chain stores operating both inside and outside of the taxed area. They enhance these data by including local demographic and nutritional information. Using a difference-in-differences framework, they compare changes to beverage prices and sales quantities before and after the tax took effect relative to a control group of stores outside the taxed area. Although the data appears to exclude independent stores, I suspect these have a relatively low market share.
As the authors explain, the effect of any sin tax depends on its incidence as well as how consumers respond to increased prices. If consumers substitute away from sugary drinks toward healthier beverages, such as water, the tax will have successfully altered behavior. However, if they substitute toward equally unhealthy beverages, or travel outside the taxed area to purchase sugary drinks, the tax will have failed to achieve its goals.
Unfortunately, the authors’ findings suggest that the latter has occurred in Philadelphia. Although purchases of sugar-sweetened beverages in Philadelphia declined 42%, this reduction is almost entirely offset by an increase in purchases outside the taxed area. Due to such “cross-shopping,” the authors found no significant change in calorie or sugar intake. Further, consumers in low-income/high-obesity neighborhoods changed behavior the least. That is, they are least likely to reduce sugar consumption as well as to travel outside the taxed area. Thus, not only has Philadelphia’s tax failed to achieve both of its goals, but it appears regressive in both structure and effect.
The authors hypothesize that low-income households are less likely to cross-shop due to a lack of transportation options. Considering transportation realities sheds light on how true this must be. Imagine a shopper who lives within walking distance of a grocery store in the taxed area, but who must take a bus in order to shop outside the tax zone. Philadelphia bus fare is $2.50 per trip (according to the SEPTA website), which makes $5.00 round trip. Assuming that the tax is passed through 100% (which is a conservative estimate given the pass-through rates found in the paper), our shopper must buy at least 333 ounces of soda in order for the tax savings to exceed the bus fare. That is about five 2-liter bottles. While not impossible, it is a hefty parcel to carry home on the bus.
The authors also find that stores pass the tax through to consumers at a rate of 75-115%, which causes a 30-40% price increase. While this pass-through rate is relatively steady across products and stores, there are a few interesting exceptions that suggest a need for qualitative research to add texture to the data. For example, one merchant increased the price of soda only, but did not increase prices for other taxed beverages. Perhaps this store felt that demand for soda is uniquely inelastic. Or perhaps they believed that consumers would be willing to bear the tax on soda because they would understand the new tax as a “soda tax,” but would balk at price increases on other products, like sugary juices. The data also reveal higher pass-through rates in low-income/high-obesity areas. This could be due to lower competitive pressures—perhaps due to the lack of transportation mentioned above—or some other force. Qualitative data could shed light on sellers’ pricing decisions, and inform future tax design to ensure the desired incidence.
Seiler et al.’s findings suggest that sin taxes will be more effective if enacted at the state or national level or in larger localities where cross-shopping is more difficult. While perhaps unsurprising, such insights should likely play a larger role in policy design to help avoid potentially ineffective or regressive tax policies.