Can Food Taxes Improve Health?

The Lempert Report
June 28, 2022

On today’s Bullseye we ask the question once again – can food taxes and subsidies help us eat healthier – this time with the help of researchers from the University of Connecticut and the University of Illinois who were recruited by The World Health Organization. There are precedents that have been successful – taxes on tobacco and alcohol for example. Around the globe there are other success stories – as in Mexico where there is a 1 peso per liter tax on sugar-sweetened drinks – and over 3 years we have seen the results – the probability of being a non-consumer of soft drinks has increased by 4.7 percent, probability of being a medium or high consumer of these beverages decreased by 6.8 percentage points and 6.1 percent respectively.

So what did the researchers find? In the first paper published in the Journal of the American Medical Association they reviewed 54 articles of which 15 studies were included in their meta-analysis. The result? Fruit and vegetable subsidies like doubling the value when using SNAP benefits were associated with just a moderate increase in fruit and vegetable sales. In the second paper, also published in JAMA they focused on the outcome of taxes on sugar sweetened beverages. They reviewed 86 studies from around the world with a meta-analysis of 62 of them and found that indeed these taxes increased prices and reduced sales for these drinks. More than 45 countries and several local jurisdictions they report have implemented sugar-sweetened beverage taxes with the intention to improve nutrition and population health. The researchers also found that these taxes have prompted many of the brands producing these drinks to reformulate and reduce their sugar content. They also point out that the study does have limitations and further studies should be conducted to understand the association with diet and health outcomes as well as assessing the variability of the consumer populations by demography to improve policy reach and effectiveness.