Public health advocates and policymakers frequently debate how to discourage the consumption of sugary beverages and candy. One common approach is a Pigouvian tax, often called a “sin tax,” which is a fee levied on goods that create negative societal effects. This raises the price of these items for all consumers. Recently, some governments have tried a different tactic within the Supplemental Nutrition Assistance Program (SNAP), the largest social safety net program in the United States. Rather than taxing the items, states requested waivers to remove the eligibility of sugar-sweetened beverages and candy so they cannot be purchased using government benefits.
A recent NBER working paper examines the rollout of these eligibility restrictions across several U.S. states in 2026. The researchers sought to understand if this policy actually changed household purchasing behavior and what mechanisms drove those changes.
Tracking Shopping Habits Across Retailers
The research was conducted by Crossan Cooper and Katja Seim of Yale University, along with Jintaek Song of the University of Florida. They looked at 10 states that implemented SNAP restrictions on soda, along with several states that restricted candy, energy drinks, and sports drinks.
To measure how the policy affected purchasing behavior, the authors utilized a large consumer panel dataset from the market research firm Numerator. Participating households upload photographs of their shopping receipts across all types of stores, including supermarkets, club stores, dollar stores, and convenience stores. The dataset allowed the researchers to identify which households used SNAP benefits, what specific items they purchased, and how they paid for their orders.
The researchers also collected their own dataset of retail prices. They scraped item-level price data for beverages at major chains like Kroger, Walmart, Publix, and Dollar General across implementing states and non-implementing states. This allowed them to monitor whether retailers adjusted their pricing or promotional strategies in response to the changing SNAP rules.
Changes in Checkout Behavior and Diet
The analysis showed that spending on restricted items by SNAP households fell by 11 percent following the policy changes. Specifically, soda purchases declined by 12 percent, candy by 8.8 percent, energy drinks by 13.4 percent, and sports drinks by 21 percent. The researchers found no evidence that households simply took the money they would have spent on soda at the grocery store and spent it at restaurants instead.
The drop in spending varied depending on the type of store. At large grocery and mass-merchandise stores, soda spending fell by 16.2 percent. At smaller convenience, drug, and dollar stores, the spending change was a statistically insignificant increase of 1.9 percent. The authors note that while SNAP usage is historically lower at small stores, those locations still account for nearly a quarter of SNAP households’ soda purchases.
Because households bought fewer of these targeted items, the nutritional profile of their shopping baskets changed. Across all beverage purchases, calories fell by 7.3 percent, total sugar fell by 8.1 percent, and sodium fell by 8.9 percent.
The researchers also looked at whether shoppers substituted restricted items for similar unrestricted items. For beverages, the response was slightly asymmetrical. In states that allowed fruit juices to remain eligible, calories from fruit juice purchases rose by 5.6 percent, offsetting about a quarter of the calorie decline from soda. For candy, the results were different. Restrictions on candy did not lead to increased purchases of eligible packaged snacks like chips, crackers, or ice cream. Instead, spending on those still-eligible snack items actually fell by 4.2 percent. The authors suggest this might be due to shoppers misunderstanding the exact boundaries of the new eligibility rules.
The Fungibility Puzzle
A central question in the study is why consumption dropped at all. In economics, there is a principle that money is interchangeable. Because SNAP benefits act as a cash equivalent for eligible groceries, households can redirect their government benefits toward items like vegetables and bread, freeing up their actual cash to buy restricted items like soda. Since they still have the same total purchasing power, standard economic theory suggests the restriction should just cause a reshuffling of payment methods rather than a drop in consumption.
The researchers confirmed that over 95 percent of the SNAP households in their sample spent enough money on eligible groceries that they could have simply swapped their payment methods without changing their overall shopping basket.
To understand why households reduced their purchases instead of swapping payment methods, the researchers investigated several potential mechanisms. First, they looked at the supply side. The scraped price data revealed that retailers did not systematically raise or lower the shelf prices of restricted items to account for shifting demand. No retailer-specific price change estimate exceeded 0.4 percent in either direction, and retailers did not significantly alter their promotional discounts.
Next, they looked at sales taxes. Purchases made with SNAP are exempt from state and local sales taxes. Once an item loses eligibility, a shopper using cash must pay the tax. The researchers calculated that losing this exemption raised the tax-inclusive price of the goods by an average of only 1.9 percent. This is largely because only 35 to 42 percent of pre-policy spending on these products was actually paid for using SNAP in the first place. The researchers conclude that a 1.9 percent effective price increase is too small to explain a 12 percent drop in sales.
Mental Budgets and Checkout Stigma
The researchers also tested behavioral explanations. One common concern with restricting SNAP eligibility is checkout stigma. If a shopper brings a restricted item to the register and is told they cannot use their benefits, they might feel embarrassed and leave the item behind. To test this, the authors compared in-store purchases with online grocery orders, where there is no face-to-face interaction with a cashier. They found that the proportional decline in soda spending was actually slightly larger online (19.1 percent) than in stores (11.2 percent), indicating that public embarrassment at the register was not the primary driver.
The policy changes and local news coverage might have acted as a public health signal, warning shoppers that the state considered the items unhealthy. The researchers tested this by looking at non-SNAP households. Soda spending did fall among non-SNAP households by about 3 to 4 percent, showing some broader market reaction, but this was much smaller than the 12 percent drop among SNAP participants.
The authors interpret the findings as evidence of a psychological departure from standard economic models. They argue that consumers do not treat their SNAP benefits and cash as a single, pooled budget. Instead, households engage in mental accounting, associating certain shopping trips or retail chains with their SNAP benefits and others with cash. When soda lost its SNAP eligibility, shoppers appeared to simply stop adding it to their SNAP-associated grocery trips rather than undertaking the mental effort of reorganizing their payment methods across different accounts.
Comparing Bans to Taxes
The researchers evaluated the financial and societal impact of this policy by comparing it to an alternative intervention: an excise tax on the targeted products calibrated to achieve the exact same purchase reduction. Past studies have noted that traditional taxes reduce the purchasing power of lower-income households by raising shelf prices for everyone.
Assuming a nationwide rollout, the authors calculated that an eligibility restriction would result in SNAP households paying an additional $176 million in newly applicable sales taxes. An equivalent excise tax designed to lower consumption by the same amount would cost those same households roughly $1.35 billion. The out-of-pocket financial burden of the eligibility restriction is roughly 13 percent of the burden created by an equivalent tax.
The researchers used a standard economic framework to estimate the efficiency gains of applying these restrictions to SNAP. By valuing the estimated reductions in future healthcare costs and correcting for consumer overconsumption, the authors estimate that the restrictions on soda, candy, and energy drinks generate between $1.07 billion and $1.16 billion in annual welfare gains.
