SNAP Purchase Restrictions Are Creating Checkout Confusion, Not Clearer Nutrition

SNAP purchase restrictions are being implemented through a patchwork of state waivers that differ enough to make the word “restriction” misleadingly singular. Justin King, policy director at Propel, said 23 states had approved waivers as of the discussion, with Montana the latest approval. Ten states already had live restrictions: Florida, Idaho, Indiana, Iowa, Louisiana, Nebraska, Oklahoma, Texas, Utah, and West Virginia. Ten more were scheduled to go live in 2026, with a few others set for 2027 and 2028.

Propel’s position in this debate is practical as well as policy-oriented. King said the company’s app is used by one in four SNAP recipients nationally and by more than 5 million households every month to manage benefits. That gives Propel both a reason to track state rules closely and a channel for hearing from recipients trying to use those rules in stores.

What the waivers prohibit varies widely. King described state policies that range from soda bans to restrictions on energy drinks, juice, flavored and sweetened carbonated beverages, desserts in some places, and candy. Candy itself, he said, is “a surprisingly difficult thing to define.” Some state rules ban what most shoppers would intuitively call candy. In other places, a product may be allowed if it contains flour, making a Twix bar a recurring example of the policy’s definitional edge cases.

That complexity moves directly to the register. As King and Margaret Mannion described it, restrictions are state-specific, retailer-mediated, and product-level. A product can be treated differently depending on the state, the store’s classification decisions, and what the checkout system tells the clerk or shopper. In many households’ experience, the policy is not learned from a clear posted list or a nutrition lesson; it is learned when an item is rejected at checkout.

King said Propel surveyed almost 3,500 people in early May across seven states where waivers had been in place as of March 1: Idaho, Indiana, Louisiana, Nebraska, Oklahoma, Utah, and West Virginia. Awareness was high: 93% of respondents said they knew about the restrictions. Understanding was more limited. About half said they had a good handle on what was in and out; the other half said either that they had only a general idea or that the rules were very confusing.

The most common way people said they found out whether a product was eligible was at checkout. Almost 60% said they learned there. Nearly two-thirds said they had had an item declined because of the new rules. Among those who had experienced a declined transaction, 80% said it had happened more than once, and 20% said it had happened more than five times.

King said nearly half of survey respondents also reported getting different answers on the same products at different stores. He connected that inconsistency to a basic design problem: stores are being asked to decide gray areas, and the consequences of those decisions fall on shoppers and frontline staff.

Some large retailers have built eligibility information into store-specific apps. King said about a third of respondents identified those apps as a source of information. But he also noted their limits. A tool that works inside Walmart or Target does not necessarily help at a convenience store, a small independent grocer, or a community market. The informational capacity is unevenly distributed, just as the restrictions are unevenly defined.

Celia Cole, CEO of Feeding Texas, framed the debate as one that begins with a shared stated goal and then splits quickly over values, evidence, and implementation. Feeding Texas supports 20 food banks around the state through food distribution, SNAP connection work, and policy advocacy for a stronger federal nutrition safety net. Cole said she has worked on SNAP policy for about 30 years and has been involved in the restrictions debate for at least two decades, consistently on the critic side.

She still laid out the argument for restrictions before criticizing it. SNAP is a taxpayer-funded benefit, supporters argue, and the public has an interest in ensuring that the benefit supports good nutrition rather than subsidizing purchases that may worsen health outcomes. Sweetened drinks and candy are frequent targets because they are linked, in that argument, to obesity, diabetes, and other conditions that disproportionately affect low-income households. Restrictions are also presented as a nudge: if SNAP cannot be used for less healthy products, households may shift toward more nutritious foods, with downstream benefits for health and health-care costs. A final argument is comparative: WIC restricts eligible items according to nutritional value, so supporters ask why SNAP should not be held to similar expectations.

Cole’s critique begins with whether SNAP restrictions are aimed at the real problem. She said research shows that SNAP participants’ purchasing patterns are not dramatically different from those of non-recipients with similar income levels. To her, that suggests the issue is not a SNAP-specific behavioral problem but affordability and access to healthy food. SNAP is also supplemental and rarely covers a household’s full food budget. If soda or candy cannot be bought with SNAP dollars, a household may buy the same item with cash.

That distinction matters because the restriction may change payment method without changing consumption. It may also reduce the effective value of the benefit. Cole described that as a de facto cut: if people are pushed toward more expensive alternatives—protein, fruits and vegetables, less processed foods—then the same benefit buys less. If those healthier foods are not available in nearby stores, the cut becomes more practical than theoretical.

Cole also emphasized implementation costs for retailers, particularly smaller grocers in rural areas, who may leave the program rather than absorb additional compliance burdens. For recipients, she said, the rules create confusion and potentially stigma.

Her most acute concern was medical. People with hypoglycemia or insulin-dependent diabetes may need fast-acting sugar sources in emergencies. If restricted items are the affordable or available source of sugar in a small store, the restriction can become a safety risk. Glucose tablets may be more expensive or unavailable.

Cole’s policy position has been shaped by years of watching different coalitions form and break. In Texas, she said, past efforts to restrict SNAP could often be defeated by an unusual alliance: anti-hunger advocates worried about food access, retailers and grocers warning about implementation, and conservatives arguing that government should not tell people what to eat. That coalition no longer held. Cole said changes in the Texas Legislature, the new federal administration, and the rhetoric around the Make America Healthy Again movement created a “perfect storm” for passage.

Feeding Texas worked with bill authors during the session and, in Cole’s telling, succeeded in narrowing the proposal. It began with a broader list of prohibited items, including cookies, and was reduced to soda, sugar-sweetened beverages, artificially sweetened beverages, candy, and sugar-coated candy. But narrowing the list did not solve the definitional problem. The moment flour, coatings, beverages, and sweeteners become eligibility triggers, products that look similar to shoppers can be treated differently by the program.

Cole said she does not doubt there will be another effort in the next session to expand the list of restricted items. Interim work is underway to monitor the legislation. But she was skeptical that Texas will generate meaningful evidence on whether the policy works, especially if “works” means improved health.

At Restore OKC in Northeast Oklahoma City, the restrictions are visible in purchase patterns and in the work of a store built around community voice. Rachel Newman, Restore OKC’s community health manager, said the organization works in zip codes where life expectancy is, on average, 18 years lower than the state average. Its work focuses on social determinants of health, but Newman emphasized that the organization’s model begins with community-based participatory research and listening.

The grocery store itself came out of a practical crisis and a community preference. Newman said the neighborhood’s final grocery store closed in 2019. When COVID hit the next year, Restore OKC began a food pantry. But community members said they did not want a food pantry; they wanted a place to spend their money on things they wanted. The result was a local grocery store operated as a social enterprise through a public-private partnership, with profit less central than community access.

The store incentivizes healthy purchases and dedicates more space to produce than would typically be recommended for a store of its size. But Newman was clear that it also carries items people want. If it does not, it goes out of business.

That makes the store a useful window into what happens when restrictions encounter actual household preferences. Newman’s main theme was “innovation coming out of resilience.” Communities that have experienced poverty over long periods, including across generations, have also had to develop creativity in response to changing rules. SNAP restrictions are the latest iteration, not a wholly new condition.

In Oklahoma, Newman said, powdered drinks and baked items were not restricted. Restore OKC saw soda sales decrease, but powdered drink sales increase. Those products remained eligible and familiar. The restriction did not eliminate the desire for sweet drinks; it redirected some purchases to a permitted substitute.

The same pattern appeared with sweets. If candy could not be bought with SNAP, shoppers might buy a single donut across the aisle or a Hostess snack. Those items “scratch that same itch,” Newman said, without falling under the restriction.

Household dynamics complicate the idea that the individual shopper will respond to a restriction by choosing the healthier option policymakers might prefer. Newman said 83% of Oklahoma SNAP households include a young child, an elderly person, or a person with disabilities. Caregivers are often buying for people with strong preferences. What a child, elder, or disabled family member will eat can matter as much as abstract nutrition guidance.

Restore OKC’s data, as Newman described it, does not show a major transformation in what people bring home. People are still buying many of the items they bought before, sometimes with their own money and sometimes in substitute forms. SNAP’s supplemental nature matters again: households may use SNAP for what remains eligible and reserve cash for restricted products. That may reduce the quantity of soda or candy coming into the household, but it does not mean those items disappear.

Newman contrasted restrictions with incentives. Restore OKC participates in healthy incentive programs, including Oklahoma’s Double Up program. Newman described Oklahoma’s program as giving shoppers additional value when they spend SNAP dollars, saying they get “double the amount back.” Because 60% of the store’s shoppers use SNAP, she said, the store also has a large share of shoppers receiving a healthy incentive when they spend SNAP dollars.

Her recommendation was to move away from restriction as the primary model and toward incentives that make healthy food more accessible and economically attractive. The goal, as she put it, should be to make people feel they are leaving money on the table if they do not choose healthier foods, rather than taking choices away in a way that adds stress.

For convenience stores, SNAP restrictions arrive alongside new stocking requirements, and Margaret Mannion described that combination as a threat to access. Mannion, speaking for the National Association of Convenience Stores, said the association represents more than 150,000 convenience stores nationally. Of those, 118,000 are SNAP retailers, more than 75% of the industry.

She also said that, in USDA’s SNAP retail universe, convenience stores make up the largest portion of authorized SNAP retailers, at 45%. That matters because convenience stores do not play the same role as supermarkets. Mannion described them as fill-in grocery locations: places to buy milk late at night, food during a second or third shift, items accessible by walking or public transit, or groceries in rural areas where the gas station or convenience store may be the only option for miles.

Many of those stores have limited administrative capacity. Mannion said 60% of the convenience-store industry consists of single-store operators. The owner may work in the store every day with a small team, handling checkout, HR, accounting, and stocking. Asking that operator to interpret product definitions across hundreds or thousands of items is a significant burden.

The definitions vary by state. Mannion said the 23 states approved by USDA were, in effect, working with different definitions. A soft drink in Iowa may not be treated the same way as a soft drink in South Dakota. For multi-state operators, that can mean managing compliance across many state-specific systems with limited time and resources.

New stocking requirements add another layer. Mannion said USDA issued a final rule earlier in the month increasing what stores must stock to remain SNAP retailers. Under the current rule, a retailer qualifying under the staple-food stocking pathway must stock three varieties of food in each of four staple food groups. A change rooted in the 2014 farm bill raises that requirement to seven varieties, with implementation tied to USDA’s definition of “variety.”

For small-format stores, the definition is decisive. Mannion said NACS is concerned that USDA is pushing stores to devote scarce shelf space to items customers are unlikely to buy in convenience stores, such as raw grains, grain-based flour, barley, or quinoa. A supermarket or dollar store may have room for those products. A convenience store that depends on quick, grab-and-go turnover may not.

The business question becomes whether participation is still worth it. If restrictions reduce customer purchases and stocking rules require stores to carry unsold inventory, some operators may decide that SNAP authorization no longer makes economic sense. Mannion said that concern had moved from theoretical to real among small-format and single-store operators. Some are saying they do not know if they can comply because the rules are too complex and burdensome. The new stocking requirements were set to go into effect on November 4.

Compliance risk is not limited to lost sales. Mannion later described what she called a “two-strike” concern: if a retailer is found out of compliance twice, it can face an involuntary withdrawal from the program, meaning the SNAP license for that location may be pulled. State governments have driven the restrictions, but enforcement ultimately sits with USDA. For retailers, that means state-level variation plus federal enforcement exposure.

At checkout, the human burden falls on workers who did not make the policy. Mannion said retailers had expected customer anger, but what they have heard most is confusion. The person explaining why one candy is rejected and another is allowed is not the governor or the USDA secretary; it may be an 18- or 19-year-old clerk behind the counter reading what the system says. Customers ask why the store down the street can sell a fountain beverage while this one cannot. Retailers, Mannion said, want ease of transaction and ease of access. The restrictions have made that harder.

The stated public-health rationale for SNAP restrictions depends on an empirical claim: that limiting purchases of certain items with SNAP dollars will improve nutrition and health. Cole argued that the evaluation structure, at least in Texas as she described it, is not designed to answer that question.

Celia Cole said the waiver submitted to USDA frames success in terms of healthy choices and healthy outcomes. But the primary evaluation instrument is a survey that does not include actual health outcome measures. It does not ask about diet quality, health conditions, or food security. Instead, it focuses on behavioral proxies: what people buy, whether they know about the restrictions, whether they spend out of pocket on the same items, and whether the policy changed purchasing behavior.

That can answer a narrower question: did SNAP purchases of restricted items decline, and did some spending shift to non-SNAP dollars? It cannot show whether health improved. Cole said the design also relies heavily on self-reporting, a bias risk the state itself acknowledged in its waiver request.

Margaret Mannion called the evaluation component critical because it could shape the length of the pilots. Her concern was that USDA might equate a lack of soda and candy sales with healthier people. She said NACS is trying to emphasize in Washington that the relationship is more nuanced.

Cole added an important caveat: Texas Health and Human Services was not funded to conduct a formal health evaluation. Waivers like this typically require evaluation, but the state did not receive money from the Legislature to do more robust measurement. As a result, she said, the agency is doing what it can with limited resources while facing rapid policy changes elsewhere. There is no funding for pre- and post-testing A1C levels or similar health markers. Food security is not included either, so the evaluation will not directly answer whether the restriction functions as a de facto benefit cut.

Rachel Newman said Restore OKC’s approach to evaluation is shaped by a different concern: research fatigue and extraction in high-poverty communities. Communities like hers are often repeatedly surveyed or studied. Restore OKC therefore tries to structure research around benefits to participants. The organization looks at purchase patterns and runs opt-in pilots, including food-as-medicine programming that provides produce and access to healthier options.

Newman said that, in her reading of existing research and in Restore OKC’s work, incentives have shown results: when people have better access to healthy items, they buy more healthy items. But when they are restricted from buying things they already purchase, they do not necessarily replace those items with healthier options. Her store’s purchasing data, she said, does not show a huge difference so far. It shows continued purchasing of familiar items, sometimes with cash, and shifts toward unrestricted substitutes.

Justin King said Propel does not aim to replace formal evaluations, but it can provide rapid insight and trend data about recipient experience over time. One of the survey findings he highlighted cuts directly against the assumption that restrictions change consumption: 62% of respondents said they were buying the same items with their own dollars rather than SNAP. Some also reported spending less overall so they could preserve cash for the things they wanted.

The effect, King said, is not only physical-health-related but financial. Households may rearrange constrained budgets to maintain autonomy over purchases. Propel also saw a strong emotional response: frustration, confusion, and a sense that dignity and autonomy had been violated.

Kummer showed a printed Propel chart of open-ended survey responses titled “What people say about their experiences with SNAP purchase restrictions.” The visible portion of the chart showed “Frustration” at 24.1%, accompanied by the quote “A living hell,” and “Supportive” at 22.0%, accompanied by the quote “Better choices & feel amazing.”

King cautioned against treating those two categories as equivalent. If the various forms of frustration are grouped together, he said, frustration, confusion, and loss of autonomy or dignity vastly outweigh supportive comments. He also said the most severe comments should not be dismissed as hyperbole, especially because some respondents described restrictions interfering with their ability to buy products they need to manage serious health conditions.

That point brought the debate back to Cole’s warning about medical equity. A restriction designed around population-level nutrition goals can have very different effects for a household managing diabetes, hypoglycemia, disability, time scarcity, transportation limits, and a narrow cash budget.

When asked what should happen over the next 6, 12, or 18 months, the panelists’ answers converged around resources, access, and incentives rather than prohibition.

Celia Cole said the best solutions come from the people affected by the problem. When people with limited incomes are asked what prevents healthy eating, she said, they name lack of resources, limited access to places that sell healthy food, time constraints, inability to cook fresh meals from scratch every day, and, for people with complex diet-related health needs, sometimes lack of understanding about what foods would best serve those needs. Policies that respond to that input would increase people’s resources and their ability to find and afford healthy food. They would also include broader economic policies that make healthy food cheaper for everyone.

Rachel Newman pointed to the economics of her store. Restore OKC’s grocery is part of a workforce development program, and workforce funding helps offset overhead. That matters because grocery margins are thin, which Newman put at five to ten percent, and because 60% of the store’s shoppers use SNAP. If SNAP changes reduce revenue, the store feels it. Public-private partnership and workforce development funding give the store a way to absorb some of that pressure while serving the community.

Healthy incentives are the other part of her answer. Oklahoma’s Double Up program, as Newman described it, gives shoppers additional value when they spend SNAP dollars. From Restore OKC’s perspective, she said incentives work by making healthy purchases more accessible and appealing. Her “gentle suggestion” was to move away from a restrictive model that reduces autonomy and toward an incentive model that makes healthier eating feel like an economic opportunity.

Margaret Mannion agreed with the emphasis on choice. NACS supports the SNAP Choice Coalition in Washington and argues that customers can make the best choices for themselves without government dictating them. But Mannion’s central warning was about access: if retailers leave SNAP because restrictions and stocking rules make participation too burdensome, that harms the program itself. A SNAP benefit requires a place to spend it. If the local convenience store is the only practical option and it loses authorization or exits the program, the household’s access to food is reduced.

No panelist identified states that were pairing new restrictions with nutrition incentives as a matter of policy. Mannion’s answer was careful rather than definitive: she said, “I don’t think that any state” is pairing the nutrition-incentive piece with the restrictions. “It’s just the restrictions.” Kummer treated that as an action step for people working within states.

The question of recipient organizing produced a more constrained answer. Cole said Feeding Texas does not have capacity for grassroots organizing, though its food-bank network has contact with five or six thousand people each month applying for or renewing SNAP benefits through application assistance. The organization is trying to document experiences and help people lift them up through media or legislators. But she said most of its current organizing energy is focused on a larger impending threat: a potential SNAP cost share that could shift hundreds of millions of dollars in benefit costs to Texas.

Mannion said NACS had not heard from USDA asking how implementation was going, and she had not heard of members being contacted either. Retailers and recipients are generating operational knowledge quickly, but Mannion’s account did not indicate a federal feedback process collecting it in real time.

Artificial intelligence could make implementation easier in theory, but the panelists described a more basic obstacle: the rules do not resolve to a single source of truth across stores.

Margaret Mannion said there is likely a world in which AI could help identify what qualifies as soda or candy. But she emphasized that many definitions being used were written for tax policy, not food policy. That makes the classification problem difficult even before a software tool tries to automate it.

Justin King said Propel has experimented in this area. He compared the potential to barcode-scanning tools in diet apps, then described a more advanced possibility: a user points a camera at a product, the tool identifies it, references ingredients, and checks the product against the rules. Technically, he said, what is possible is “mind-blowing.”

But the barrier is not only image recognition or ingredient parsing. If the store’s point-of-sale system is where the eligibility decision is ultimately made, and if that decision differs across stores, then a statewide or national tool cannot reliably tell a shopper what will happen at checkout. It may be wrong even without AI hallucination because it cannot know how the particular store has designated the product.

That is why large retailers’ proprietary tools can work better: they have their own source of truth. The tool is connected to the retailer’s internal classification. A cross-store tool serving people across a state or the country does not have that same anchor.

King’s conclusion was that the deeper fragmentation affects everyone. It makes it harder to equip consumers with accurate information, harder to evaluate policy impact, harder for retailers to comply, and harder to assess whether the policy improves health. Until that reference-point problem is solved, even powerful tools risk becoming another layer of confusion.