Consumer Demand and Firm Advertising Responses to a Local Beverage Tax: Evidence from Philadelphia

• In Event: Food and Beverage Policies

Abstract

Local beverage taxes have become a prominent policy tool for reducing unhealthy consumption, yet their full market effects remain incompletely understood. Existing evaluations focus primarily on whether tax-induced price increases reduce purchases, leaving a parallel margin largely unexplored: firms operating in taxed jurisdictions may also adjust local advertising, and these non-price responses may reinforce or offset the intended demand effect. This paper jointly studies consumer purchasing and firm advertising responses to Philadelphia’s 1.5-cent-per-ounce beverage tax, implemented in 2017.

I combine two Nielsen datasets covering 2013–2019. On the consumer side, I use Homescan Consumer Panel data to estimate a two-part difference-in-differences model on a balanced household-by-beverage-category-by-week panel, allowing separate analysis of purchase incidence and purchase intensity. On the firm side, I use Nielsen Ad Intel Spot TV data to construct a brand-by-DMA-by-week panel and estimate changes in local advertising in Philadelphia relative to matched comparison markets.

I find that the tax reduced household purchases of caloric taxed beverages, lowering both the probability of purchase and unconditional weekly ounces purchased.
The probability of purchasing caloric taxed beverages fell by 2.7 percentage points, and unconditional weekly ounces purchased declined by 6.3 ounces. I find no comparable demand response for diet beverages, which faced the same statutory tax rate, consistent with substitution toward zero-calorie alternatives. Event-study estimates support parallel pre-trends and indicate that the consumer response is largest immediately after implementation and attenuates thereafter. These results are robust to alternative control groups and border-based comparisons.

On the firm side, local television advertising in Philadelphia also declined after the tax. Both the probability that a brand aired any local advertisement and local advertising expenditures fell relative to comparison markets, with larger reductions among brands with stronger pre-tax Philadelphia market presence. Placebo tests indicate that this decline is specific to local advertising in the taxed market rather than part of a broader industry-wide trend.

A complementary analysis links the two margins: brands with greater pre-tax national advertising intensity experienced larger post-tax purchase declines in Philadelphia relative to control markets, consistent with a role for advertising in amplifying the tax’s demand-reducing effect. Taken together, the results indicate that local beverage taxes affect both consumer purchase responses and firm-side marketing adjustments. Evaluations focused only on prices and purchases may therefore understate the full demand-reducing effect of these policies. More broadly, the findings suggest that firm advertising responses should be incorporated into projections of beverage taxes and related health excise policies.

Author

• Qianhui Xu, Georgia State University
  Presenting Author