Evaluating the Barbados SSB tax: a mixed-method natural experimental study

Consuming sugar-sweetened beverages (SSBs) has been associated with increased rates of obesity and type 2 diabetes, making SSBs an increasingly popular target for policy interventions such as taxation. SSB taxes are intended to reduce SSB consumption, and several national evaluations have suggested that they may be effective. However, SSB taxes have been implemented in diverse settings (e.g. in terms of baseline SSB consumption) and designed in a variety of ways (e.g. in terms of tax structure), with important implications for effectiveness. The Government of Barbados introduced a 10% excise tax on SSBs in 2015. In this thesis, I evaluated the Barbados SSB tax with the aims of 1) evaluating the impact of Barbados SSB tax on price and sales change and 2) contributing to the development of broader SSB taxation theory. My research objectives were to assess 1) price change, 2) sales change, 3) how the tax design compared with pre-tax SSB consumption patterns, and 4) whether the introduction of the tax conveyed information about the health risks of SSBs (i.e. through a risk signalling effect). I used a mixed methods approach and analysed data from a major grocery store chain, a nationally representative cross-sectional survey, interviews with members of the public and archived media data. My findings suggest that the tax was associated with an increase in the price of SSBs and a decrease in SSB sales, after controlling for pre-existing time trends, seasonality and other time-varying confounders. Prices of non-SSBs did not change and sales of non-SSBs were estimated to have increased. However, the Barbados SSB tax only addressed around 60% of SSB-related free sugars and did not clearly differentiate between high- and low-sugar SSBs. Finally, the Barbados SSB tax may have conveyed additional information about the health risks of sodas but not sugar-sweetened juice drinks. In terms of theoretical contributions, I proposed three criteria to improve the design of SSB taxes using pre-existing nutritional survey data and operationalised legal theory around an SSB taxation risk signaling effect. In addition, I found suggestive evidence that the introduction of an SSB tax may encourage industry to 1) increase SSB advertising (diminishing any potential risk signal) and/or 2) introduce low-cost SSBs (diminishing any potential price effect). Overall, I suggest that the Barbados SSB tax may have reduced consumption of taxed SSBs but had several unexpected effects, with important implications from a health perspective. After considering these findings in the context of the broader SSB taxation evidence base, I identify a number of implications for policymakers, civil society organisations and future SSB tax evaluation teams. Finally, I suggest that evaluations of SSB taxes and other population-level health interventions may benefit from the integration of multiple types of evidence, a focus on eliminating alternative explanations and an emphasis on contributing to theory development.Gates Cambridge Trus