Oh, SNAP!

Part 1

It started in the candy aisle

I was looking for gum in the candy aisle of my local HEB in Austin in early April when I noticed a cadre of shelf talkers yelling “ALL CANDY AND SOME SWEETENED DRINKS ARE NO LONGER SNAP ELIGIBLE.” The signs were loud. My first thought was blunt: Hershey's stock is going to take a hit. I pulled up the ticker on my phone. It looked wholly unaffected.

A few questions I wanted to answer:

• How much money are these companies about to lose?
• How much of the revenue sitting in Hershey's or Mondelez's 10-K actually rode in on a SNAP card, and what happens to it now?
• Which hedge funds and big institutional holders are exposed, by how much, and how on earth are they modeling for it?

If I were tasked with helping a hedge fund understand its exposure, how would I build the expert-network diligence to answer it?

So I did what a research analyst does in a grocery store, which is start asking questions. The store manager was filling in on a checkout lane, so I put it to him plainly. I have been looking at the SNAP signs, I said; have you actually seen a drop in how much candy and junk food people are buying? He did not hedge. "The change is massive."

Massive. Not "a little soft this month." That one word, from a person who watches these registers every day, is what sent me down the path that became this report. Because if the change is "massive" at a single HEB, basic logic would lead one to believe that the repercussions at scale are colossal.

The Author's Bias

Before reading further, you should know my personal bias: I lift or run every day, have finished a full Ironman, run marathons, played college baseball, think about food as fuel, and cook all of my own meals. So when state after state started pulling soda and candy out of what taxpayers fund, my instinct was to nod along. Should the government spend $3 of our taxpayer dollars on a pound of chicken breast, frozen peas and carrots, an energy drink, candy bars, or a bag of chips? That is a values question, and reasonable people land all over the spectrum, from dignity and autonomy for recipients on one side to public-health and fiscal-stewardship arguments on the other. I am not here to litigate it.

When the federal government changes what tens of billions of dollars are allowed to buy, that money moves, and somewhere downstream a publicly traded company's volume moves with it.

Part 2

Executive thesis

For the PM who has ninety seconds. The rest of the report is the proof.

A regulatory wave is moving through the American grocery aisle, seemingly unpriced. Under the MAHA (Make America Healthy Again) banner, the USDA has approved waivers in 22 states that pull candy and sweetened beverages out of SNAP eligibility. Even the word "waiver" is confusingly named here: it is federal sign-off for a state to start restricting purchases, not an exemption from doing so. The restrictions are real, and they are expanding. What I could not find, though, was a number. I went looking through 10-Ks and financial disclosure statements, and on none of them could I see exactly how many dollars a company collects through SNAP. That revenue is obfuscated, blended into the top line and one step removed from the manufacturer, because the benefits are spent at the retailer rather than the brand. It makes the shift genuinely hard to measure, and because it is fragmented state-by-state, harder still.

What the tape did on each rollout date

If the restrictions were a repricing event, you would see it on these dates: a gap down, a volume spike, something. Hover the chart to inspect any single day. Every name is rebased to 100 at mid-December 2025, meaning each line tracks percentage change from that starting point rather than its dollar price, so moves can be compared proportionately across names. The grey dashed line is the consumer-staples ETF (XLP) as a market benchmark, and the red dashed verticals mark the rollout dates. Use the legend to toggle any company on or off, and the dropdown to switch the volume series.

Where the exposure actually concentrates

The intuitive "big junk-food names get hit" framing is half wrong. The hit is concentrated, not broad, and it lands on the pure-sugar players, not the diversified snackers.

And the bigger exposure isn't the brands at all

Tracing only the candy and soda makers misses the larger pool. Those companies are Channel 1, hit by a shift in what SNAP can buy. The bigger dollar exposure is Channel 2: the retailers and payment plumbing whose revenue tracks the total SNAP pool itself, which is shrinking through a $6.2B funding cut, a cost-shift onto the states, and enrollment already down roughly 4.2 million people.

The four non-obvious calls

These are the points that separate this read from the headlines, and each is a place a 10-K can't help you:

That is the uncomfortable part for anyone who thinks a question like this can be Googled or handed to a model. The number that settles it is published nowhere. It lives with the category manager watching the registers and the former state director who knows how the rule is really enforced. You cannot scrape it. You have to ask a human who has seen it.

This is what an expert network is for, and it is part of why I wrote this. The claim that AI or a search bar makes this work obsolete has it backwards: the harder something is to find in public data, the more it takes real people to answer. This is one of those problems.

Part 3

The public equities framework

Picture two calls landing at our desk. One is a public equities investor holding long positions in these names, asking what its real exposure is. The other is a short seller circling the same names, looking for a drop. Both say a version of the same thing: "You run an expert network, so use it. Help me understand how to build a case strategy to price this in."

This is how I would scope it, the same way we have seen large-scale CDDs run on CPG assets. The market breaks into three layers, each one a question, and each ending in the same place, where the public filings go dark.

Part 4 · The three layers

The policy shift

Before the company math, the policy. What SNAP is, where the money comes from, and what is actually changing.

SNAP is federally funded and large: roughly $100 billion a year, about $107.5 billion appropriated for 2026, reaching some 42 million people. The federal government pays the benefit; the recipient spends it at the register like a debit card. That is the baseline everything else moves against.

Two distinct things are happening at once, and they are easy to conflate. The first is what can be bought, the purchase restrictions imposed through the state waivers, which is the next section. The second is the money itself. Under the 2025 reconciliation bill, mandatory SNAP funding falls by roughly $6.2 billion from 2026 to 2027, the federal share of administrative costs drops from 50% to 25% starting in FY2027 (pushing the other 75% onto the states), and enrollment is already down about 4.2 million people year over year.

Net: the pool is shrinking and the states are absorbing more of the cost, even before you count the categories being pulled out of eligibility. For the exposed companies that is two pressures at once, a smaller pot and a narrower slice of it.

Part 5 · The three layers

The state rollout

The restrictions are not one federal rule. They are 22 separate state decisions, rolling out on different dates with different definitions.

Twenty-two states hold USDA-approved waivers, and ten are already live as of mid-2026. The first wave, Indiana, Iowa, Nebraska, Utah, and West Virginia, went live on January 1; then Idaho and Oklahoma, Louisiana, Texas on April 1, and Florida on April 20. The rest roll out through the back half of 2026, with Kansas and Wyoming in early 2027 and Nevada not until 2028. It cascades by design: the early states test the mechanics, and more follow.

No two waivers are identical. Most restrict “soft drinks” and “candy,” several add energy drinks, and a few add prepared desserts. The definitions are inconsistent enough that the same product can be eligible in one state and not the next, which is the compliance friction the National Grocers Association is pushing back on.

The carve-outs matter as much as the bans. Salty snacks, chips, pretzels, cheese puffs, stay eligible in 21 of the 22 states; only Iowa, which restricts all taxable food, sweeps them in. Sweetened cereal, baked goods, and ice cream mostly stay eligible too. The bite falls on sugar-sweetened drinks and candy, not the broader ultra-processed aisle, which is precisely why the exposure is concentrated rather than industry-wide. The full state-by-state matrix, with effective dates, restricted categories, diet treatment, and a source for each, is in the companion file.

Part 6 · The three layers

Company exposure: two channels and a 54-name map

SNAP exposure runs through two channels, and the second is where the real dollars sit.

Channel 1: the products that can no longer be bought

This is the mix-shift channel: the companies that make the candy, soda, energy drinks and sweetened desserts now ineligible in restricting states. The exposure is concentrated, not broad: Hershey and Coca-Cola screen highest, the energy pure-plays Monster and Celsius are hit wherever energy drinks are banned, and Mondelez and PepsiCo are more insulated than their reputations suggest (biscuits and salty snacks, neither restricted). The broader packaged-snack names carry only a thin slice, mostly in the expansive states.

Channel 2: the revenue that tracks the shrinking pool

This is the larger channel by dollars. These companies do not make restricted products. Their revenue tracks the total SNAP pool, which is contracting on the three fronts described earlier: the funding cut, the admin cost-shift, and falling enrollment. The most reliant names are the dollar and discount stores, where SNAP runs around 13% of sales at Grocery Outlet and roughly 9% at Dollar General and BJ's, and where Dollar General has already reported SNAP-driven declines. The mass grocers carry a smaller percentage (Walmart around 5.5%, Kroger around 5%) but enormous absolute dollars, with Walmart alone capturing about a quarter of all SNAP spending in the country. Convenience stores sit in both channels, since candy, soda and energy drinks are their core basket. And one step further back is the cleanest infrastructure play of all: the EBT processors, FIS and Conduent, which hold exclusive state contracts and whose transaction volume is the benefit pool itself.

The same test, applied to Channel 2

The retailers and processors trade around more day-to-day than the food brands, and a couple (Conduent is a sub-$2 stock) are simply volatile. But run the identical test and the rollout dates are non-events here too. The one broadly red column, Feb 17, was a market-wide staples selloff, not a SNAP signal. The more important caveat: the category-restriction dates are not really the right catalyst for these names at all. Their exposure is to the shrinking benefit pool, which plays out on a slower timeline (the FY2027 funding cut, the admin cost-shift, falling enrollment), so the decisive test for Channel 2 is still ahead.

The other side: who gains

A credible thesis names the winners too - or does it? If SNAP dollars redirect toward real food, the beneficiaries should be the protein and better-for-you names such as Simply Good Foods and BellRing, bottled water like Primo Brands, or the fresh-skewed grocers such as Sprouts and Natural Grocers.

One step further upstream sit the sweetener, cocoa, and packaging suppliers (ADM, Ingredion, Ball). That exposure is real but diffuse, and the most-affected nodes, the cane and beet sugar refiners, are largely private, so for equity purposes it is a watch-list rather than a thesis.

Representative names shown. The full 54-name index with every ticker, exposure tag and note is in the companion file (SNAP Exposure Index). “Read” is directional, not a price target.

Part 7 · The diligence

Why you can't get this from a 10-K

Everything to this point has pointed at the same wall: the number that would settle the thesis is not in any public document. This is where the research has to leave the filing cabinet.

I went looking for the obvious version of this story, a line item, and found nothing. No issuer breaks out how much of its revenue is paid for with SNAP benefits, because none of them have to. SNAP is not a material disclosure requirement, and the dollars are structurally hidden: a shopper's benefits are spent at the retailer, blended into the store's top line, one step removed from the brand that made the product.

The state side is no clearer. There is no national list of which products lost eligibility. The waivers restrict categories, not SKUs, and “candy” is defined off each state's sales-tax code, which is why a flour-containing Kit Kat stays eligible while a Hershey's bar does not. A few states sent retailers UPC files, but even those are incomplete. You cannot reconcile the policy to a product list from the outside, let alone to a company's revenue.

The absence is itself information. Either these companies do not track SNAP exposure with precision, which would be its own story, or they do and have chosen not to volunteer it. Either way, the gap is the opportunity.

If the decision-useful number is not in a filing, the edge belongs to whoever builds it first.

Which leaves one question: who can actually see it? A specific set of people, none of them quoted in a 10-K and all of them reachable, which is exactly what the next section maps out.

Part 8 · The diligence

How we would run this diligence

The exposure number does not exist in public data, so you build it from primary sources. Here is the exact sourcing plan: four angles, the specific people behind each, and the screens that keep it clean.

Each angle captures a vantage point the filings cannot. Run together, they triangulate a bottom-up SNAP exposure map, name by name and state by state. This is the part a fund cannot replicate from a terminal, and it is what TCE is built to execute.

A fifth thread sits one step out: the EBT processors (FIS, Conduent) whose transaction volume is the benefit pool itself, useful for sizing the Channel 2 picture rather than any single brand.

The state-by-state targeting

Because the regulatory and retail angles are state-specific, sourcing keys off each state's dominant chain. In Texas that is H-E-B; in Florida, Publix and Winn-Dixie; in Iowa, Hy-Vee; and across the Mountain West and Midwest, Kroger's regional banners (King Soopers, Smith's, Dillons, Fred Meyer), which is who you actually source rather than “Kroger.” Walmart, the single largest SNAP retailer, is a target in every state.

Part 9 · The diligence

The pre-scoped brief, and how to engage

Everything above is the thinking. This is the thing you can actually greenlight: a pre-scoped expert-network project, ready to run as-is or plug into your own process.

We have already done the scoping work that normally eats the first week of an engagement, the angles, the target profiles, the screens, the state-by-state targeting. What is left is execution.

Sequencing. We start with the regulatory and retail angles, the fastest leading indicators, then layer in distributors and CPG. First conversations begin within 48 to 72 hours of greenlight.

A living project. As each new state goes live and the benefit pool keeps shrinking, the same framework re-runs. Every wave is a fresh read before the print. Expert honoraria run the standard $150 to $750 per hour by seniority; total scope and cost are sized to the mandate.

Part 10

Appendix & sources

The data behind the piece, and where it came from.

Companion data assets (available on request): the 22-state rollout matrix (effective dates, restricted categories, diet treatment, sources); the company exposure model; the SNAP Exposure Index of 54 public names by sector and channel; the state-by-state retailer sourcing map; and the full expert-network project brief.

Primary sources: USDA Food and Nutrition Service, SNAP Food Restriction Waivers (state list, effective dates, category definitions); USDA Economic Research Service, SNAP participation and spending; Numerator, estimated category sales loss; Credit Suisse, retailer SNAP-exposure estimates; company filings and earnings commentary (Hershey, PepsiCo, and peers); Financial Modeling Prep, end-of-day price and volume data; and reporting from Civil Eats, the National Grocers Association, and Propel on definitions and rollout.