PHILADELPHIA: Taxed-beverage volume at large chain stores in Philadelphia fell 51 per cent after the city’s beverage tax began in 2017, a study published in JAMA in 2019 found. Chain-store sales in bordering Pennsylvania ZIP codes rose enough to offset 24.4 per cent of that drop.

The study, led by Christina Roberto of the University of Pennsylvania, tracked sales at large chain retailers. Volume inside the city fell from 2.475 billion ounces to 1.214 billion ounces. Chain stores in bordering ZIP codes in Bucks, Montgomery and Delaware counties sold 308.2 million ounces more.

A city can cut soda volume on paper and still leave drinking habits mostly unchanged if shoppers just drive across the line. Philadelphia’s beverage tax is a clean case of that gap: chain-store ounces inside the city halved, about a quarter of the drop showed up over the border, and a separate study found adult consumption barely moved.

The Tension

The City of Philadelphia’s beverage tax page still lists the rate at $0.015, or 1.5 cents, an ounce. The tax has been in place since 2017.

Inside the city, chain-store volume halved. Over the city line, sales at chain stores rose by an amount equal to about a quarter of the ounces lost inside Philadelphia.

Net of that border shift, the drop comes to about 39 per cent. The paper does not state that figure. It is a DMNews calculation from the study’s own numbers: 51.0 per cent multiplied by one minus 0.244.

The data have limits. They cover large chain stores only, which account for about 25 per cent of taxed ounces sold in the city. Sales at every other kind of shop fall outside the count.

New Jersey was also left out. Only Bucks, Montgomery and Delaware counties in Pennsylvania were counted as border areas, so any shift in purchases to New Jersey is not part of the 24.4 per cent.

The study measures what chain stores sold. It does not measure what people drank. The same paper also found that pharmacies passed far more of the tax into prices than supermarkets did.

Roberto’s paper appeared in JAMA volume 321, issue 18, on pages 1799 to 1810. Its DOI is 10.1001/jama.2019.4249 and its PubMed ID is 31087022. She is at Penn’s Perelman School of Medicine.

Two later corrections, one in 2019 and one in 2020, were housekeeping.

The Noise

The 51 per cent figure describes taxed-beverage volume at one category of store in one city. It says nothing about how much anyone drank.

A separate paper by John Cawley of Cornell and colleagues, published by the National Bureau of Economic Research as working paper 25052, measured consumption as well as purchases. It also found that purchases moved across the border.

Adult consumption changed only modestly, the NBER paper found. It found no detectable effect on children.

The City published a post on where the tax money goes on 3 April 2017.

On 27 October 2025, Chalkbeat reported that City Council members criticised the use of the tax to fund free pre-K. In November 2025, Penn LDI published a call to keep the tax. Both positions date from late 2025.

Sources: