A working paper from the National Bureau of Economic Research reports that after 10 states implemented restrictions on using Supplemental Nutrition Assistance Program (SNAP) benefits to buy soda and candy, households receiving SNAP bought about 12% less soda. The authors say this change corresponds to roughly 34 fewer 12‑ounce cans of soda per person per year. The paper is a working paper and has not yet been peer-reviewed.
The study’s lead commentators described the reduction as not huge but meaningful. “If the goal is to reduce obesity, reduce the cases of diabetes caused by people being overweight and overconsuming sugary beverages, this is probably one part of a broader set of policies,” said Matt Notowidigdo, an economics professor at the University of Chicago Booth School of Business and a co-author of the study.
The analysis used grocery purchase records from the first six months of 2026 for about 15,000 households using SNAP benefits. Of those households, 3,291 were located in the 10 states that had implemented new restrictions on using SNAP dollars to buy soda, candy, and similar items. The paper’s funding included a grant from Bloomberg Philanthropies.
Experts quoted in the article noted limitations in representativeness. Benjamin Chrisinger, an assistant professor of community health at Tufts University, said that while the data are useful, they may not reflect the broader population of SNAP recipients.
Using the observed drop in soda purchases, the study’s authors estimated a modest long-term public health effect: a 2.6% reduction in the risk of developing type 2 diabetes over the next 10 years, which they translate into about 34,000 fewer new diabetes cases in the United States. The authors also estimated the change would yield about $1 billion in annual savings to the health care system.
Notowidigdo framed that saving in context, noting that while $1 billion is meaningful, it is small relative to total U.S. health spending. The article cites overall U.S. health care spending of about $5.3 trillion in 2024 to illustrate scale.
The paper reports that SNAP recipients diverted a portion of money not used on soda toward other beverages not covered by the bans. Up to 39% of the funds saved from reduced soda purchases were spent on other sugary drinks and fruit juices that were not restricted in those states. The authors highlight that if the objective is to reduce overall sugar consumption, narrower bans that exempt other sugar-sweetened beverages will blunt the policy’s impact.
The study’s findings underline how the specific items included in state restrictions affect outcomes: partial restrictions can produce substitution toward other high-sugar beverages instead of net decreases in sugar intake.
A survey of SNAP recipients conducted as part of the study found respondents were more likely to report feeling judged or disrespected after the restrictions. Outside experts called this finding concerning. Benjamin Chrisinger noted that stigma is real and difficult to weigh against other health indicators, complicating assessments of trade-offs.
Robert Paarlberg, a food-policy scholar quoted in the piece, questioned whether the modest health gains among the portion of Americans who use SNAP justify potential stigma. He suggested that alternatives that do not single out SNAP recipients might avoid these harms.
At the time of the story, 23 states had received waivers from the U.S. Department of Agriculture to implement varying restrictions on purchases with SNAP benefits, though federal judges had ordered suspensions of bans in five of those states. State policies differ in which items are prohibited.
Critics and some policy analysts raised alternatives. Paarlberg pointed to sugary-beverage taxes as one option that reduced consumption across whole populations in prior implementations and generated revenue for public projects. The article cites Philadelphia’s experience as an example where a tax achieved larger reductions in consumption across all groups, according to the commentator.
The authors and outside experts caution about constraints on interpreting the findings. The data cover only six months in 2026 and a subset of SNAP households; results may not generalize to all recipients or persist long term. The study is a working paper that has not undergone peer review. Funding from Bloomberg Philanthropies is disclosed; the outlet noted that Bloomberg provides funding to STAT but does not influence editorial decisions.
Experts also flagged behavioral substitution and stigma as important considerations when weighing the net benefits of the policy. The study’s estimates of diabetes cases averted and cost savings are model-based projections tied to the observed short-term reduction in soda purchases, and the authors characterize the health gains as modest components of a broader set of potential interventions.
The NBER working paper presents evidence that bans on buying soda and candy with SNAP benefits in 10 states were associated with a 12% reduction in soda purchases among SNAP households observed in the data, equating to about 34 fewer 12‑ounce sodas per person per year. The authors estimate modest downstream reductions in diabetes incidence and health care spending, but substitution toward other sugary beverages and reported increases in feelings of stigma are important caveats. The findings are based on a limited sample and time window and the study has not yet been peer-reviewed.