A matter of desing in Soda Taxes : Tax sugar instead of volume

Soda taxes are meant to improve public health outcomes insofar as they curb consumption of sugary beverages and thus sugar intake. This research is the first to get closer to the aim of the policy by evaluating the effect of the soda tax introduced in Berkeley in 2015, focusing on its effect on sugar consumption. I conduct this analysis on a unique dataset that merges retailer point-of-sale (POS) data with matched Nutrition Facts information for sugar content. Due to the vast range of sugar densities across beverages, this setting provides information about how a tax levied at volume, rather than sugar content, affects demand for taxable items. I develop a theoretical model that accounts for how a tax such as the one in Berkeley, incentivizes consumers to substitute toward drinks with higher sugar concentration, as unitary sugar prices among taxed drinks are affected differently. I evaluate these hypotheses for Berkeley relative to comparable locations and assess the effects of soda taxes on consumption and price by volume and sugar content. I find that sugar intake fell as a response to the tax, but at a lower rate compared to volume consumption. Following the introduction of the tax, in other words, the unitary sugar price exhibited a lower impact in more concentrated sugar beverages. This indicates that the policy incentivized consumers to buy drinks with higher sugar concentrations among tax beverages and that a levy at the sugar content level would be more effective