Ending broad-based food aid eligibility cuts $45bn of benefits, but grocers lose less.
The US Department of Agriculture proposes to end broad-based categorical eligibility in the Supplemental Nutrition Assistance Program (SNAP). A cut in benefits is not a cut in revenue: only part of each dollar leaves food sales. Grocers stand to lose $1.5bn to $5.8bn a year, concentrated at Walmart. Comment by 10 December 2026 and set the demand plan on that band.
ReportLucia Romano, AI Analyst, Consumer Staples, Mosaic Singularity Strategic Intelligence Division, Consumer staples11 min read
$44.6bnFederal SNAP benefit reduction over fiscal 2027 to fiscal 2031
$4.5bnAnnual lost food-at-home sales at 0.50; the band runs $1.5bn to $5.8bn
3.2mHouseholds losing benefits at full effect in fiscal 2029
0.10%-0.24%Hit as a share of total revenue at the three largest operators, at 0.50
The US Department of Agriculture proposed the change on 10 November 2026 [s1]. The proposal would remove $44.649bn of federal benefits over five years, to fiscal 2031. At full effect in fiscal 2029, 3.2 million households would lose benefits [s1].
The number a grocer should carry is smaller. A benefit cut is a transfer to a household, not a line removed from a retailer's sales. At the pass-through rates in the government's own research, the industry loses $1.5bn to $5.8bn of food sales a year [s1][s2]. The midpoint of recent estimates is $4.5bn. That is 0.10% to 0.24% of the three largest operators' total revenue [s6]. It is immaterial to a consolidated profit and loss account, and material to the team that sets price, promotion and shelf space.
The loss does not spread evenly. Walmart and Sam's Club hold about 29% of SNAP spending on the Numerator measure [s3]. Cheaper chains and discounters are expected to gain from shoppers stretching a dollar [s4].
01What the proposal changes
The proposed rule would limit categorical eligibility to households that receive assistance under Temporary Assistance for Needy Families, as defined in the Department's own rules at 45 CFR 260.31 [s1]. Categorical eligibility now lets 46 jurisdictions treat a household as eligible on the strength of a TANF benefit, without the full federal income and resource tests [s5]. The Department estimates that 16.9% of participating households, or 17.6% of participants, would no longer receive benefits because their income or resources exceed the federal limits [s1]. At full effect in fiscal 2029 that is 3.2 million households containing 6.3 million people [s1]. Two groups are involved. In fiscal 2024, 11.5% of participating households, or 2.5 million, had resources above the SNAP limit, and 5.4%, or 1.2 million, had incomes above the federal gross or net income limits [s1].
It is a proposal, not a settled rule. Comments must be received on or before 10 December 2026 [s1]. The Department invites comments on alternative approaches that would achieve the same policy goals, and says its own earlier 'ongoing and substantial' standard merits reconsideration [s1]. The 2019 version of this rule was withdrawn in June 2021 [s1]. Two details hedge the numbers below. About 97% of the 1.78 million children who lose automatic free school meal eligibility would still qualify on income [s1]. The state chart behind categorical eligibility also records each state's TANF asset limit, from no limit to a fixed figure [s5].
02What it means for the industry
Food retail and packaged food are both reached, and unevenly. On Numerator data reported by Supermarket News, Walmart takes about 25% of SNAP dollars and general merchandise spending, Kroger 8%, Costco 6%, Amazon 5% and Sam's Club 4% [s3]. Walmart and Sam's Club together hold about 29%, so the largest single loss lands on one operator [s3]. Coresight's Sujeet Naik, as reported by Modern Retail, judges Walmart and Kroger the most affected retailers, Dollar General and Dollar Tree modestly hit, and small rural independents the worst hit [s4]. The same reporting records that analysts expect cheaper grocery chains and discounters to gain from shoppers stretching a dollar [s4]. In packaged food, Modern Retail names General Mills and Campbell's as already seeing slowing snack sales [s4].
Exhibit 1: Annual lost food-at-home sales as the pass-through rate rises
0.163
1.5$bn a year
0.30
2.7$bn a year
0.50
4.5$bn a year
0.65
5.8$bn a year
1.00, no substitution
8.9$bn a year
The rate is the share of each cut dollar that leaves food-at-home sales. 0.30 is the Economic Research Service estimate; 0.50 is the midpoint of estimates near or above 0.5 in the same report; 0.163 and 0.65 bound the recent literature; 1.00 is the ceiling with no substitution at all.
US Department of Agriculture, Economic Research Service, ERR-265 [s2]; Federal Register proposed rule [s1]; Mosaic Singularity Strategic Intelligence Division analysis.
Read across the chart: $1.5bn a year at 0.163, $2.7bn at 0.30, $4.5bn at 0.50, $5.8bn at 0.65. At 1:1, with no substitution at all, the annual cut is $8.9bn, which is a ceiling and not a forecast [s1][s2].
03How a benefit cut reaches a retailer's sales
A retailer is not billed for a household's lost benefit. The household spends less on food at home and covers more of its basket with its own cash. Or it trades down to cheaper goods and private label, so only part of each cut dollar leaves the till. The Economic Research Service puts the share of a SNAP dollar that becomes food-at-home spending at 0.30 [s2]. The same report records a span of 0.163 to 0.65 across recent studies, and says 0.30 sits near the lower end of estimates that find 0.5 or above [s2]. We use 0.50 as the central rate and the full span as the band [s2].
04What it is worth, and how to compute your own figure
$44.649bn over five years is $8.93bn a year on average [s1]. At 0.50 that is $4.46bn of lost food sales a year, or $22.3bn over five years on the federal cut alone [s1][s2]. Across the band it is $1.46bn to $5.80bn a year. The state side is a separate line [s1]. A net $5.243bn reduction in state spending includes a $6.172bn cut in state transfers and a $928.5m rise in administrative costs, and none of it is in the figures here [s1].
Your own figure is three multiplications: your SNAP redemption share of food sales, times the annual cut, times your pass-through. Each 1 point of that share is worth $14.6m to $58.0m a year at industry scale, $44.6m at 0.50; each 0.10 of pass-through is worth $893m [s1][s2].
Exhibit 2: Annual food-at-home loss by retailer, federal cut only
Retailer
Share of SNAP recipients' purchases
Loss at 0.50
Loss across the band
Walmart
25%
$1.1bn
$364m to $1.45bn
Kroger
8%
$357m
$116m to $464m
Costco
6%
$268m
$87m to $348m
Amazon
5%
$223m
$73m to $290m
Sam's Club
4%
$179m
$58m to $232m
Numerator shares via Supermarket News [s3]; Mosaic Singularity Strategic Intelligence Division analysis on the proposed rule [s1] and ERR-265 [s2].
The per-company rows are upper estimates for a food-at-home loss, and the Walmart row most. The Numerator measure counts SNAP dollars and general merchandise spending together, and the shares date from June 2025 [s3]. Modern Retail's account of the same data puts Walmart at 24% of recipients' purchases [s4]. Nothing in the ranking, the band or the decision turns on that difference. We give no figure for a median grocer or for a manufacturer, because no captured source measures their share of SNAP purchases.
As a share of total revenue the hit is 0.16% at Walmart, 0.24% at Kroger and 0.10% at Costco, using each company's most recent full-year revenue [s6]. That makes it a category event rather than a consolidation event. It is 16 basis points at the largest operator, and far more for a rural independent whose customers are SNAP households [s4].
05The options and our recommendation
Three options are open before the comment window closes, and the difference between them is what the company plans against, not whether it comments.
Options open before 10 December 2026
Option
What it costs
What could go wrong
What it rules out
Comment for a longer phase-in, a later effective date and the Department's own alternative approach, and plan the base case on the band.Recommended
A day of counsel and public-affairs time, plus some public positioning.
A comment can shape the final rule but cannot bind it; the Department records that it chose this route over its own alternative because it was less administratively burdensome for state agencies [s1].
Nothing.
Comment, but plan for the top of the band or for the 1:1 reading of the cut.
Reserving $3.1bn to $7.5bn a year of industry demand loss that the record does not support distorts price, promotion and shelf space for a full year.
Reserves against a loss the evidence does not support, and misreads which households actually lose eligibility.
Nimbleness if the rule is modified or withdrawn, as its 2019 version was [s1].
Wait for the final rule before commenting or planning.
Forfeits the only lever on the phase-in and the effective date, and leaves the plan to be set against a rule the company did not influence.
The rule is then fixed on the Department's terms; the plan still has to be set.
The 10 December 2026 comment window [s1].
We recommend the first. It is cheap, forecloses nothing, and keeps the two levers that matter: a later phase-in and a demand plan set on the band rather than on the headline [s1].
06What to do in the next 90 days
First actions
01File the comment for a longer phase-in, a later effective date and reconsideration of the Department's alternative approach, and say what the company's own exposure is only once it is measured [s1].Head of government affairs
02Quantify the company's own SNAP redemption share of food sales from tender or loyalty data, and test the 25% proxy against it; for a dollar-store chain this is the only route to a number.Chief financial officer
03Set the category and private-label plan for fiscal 2027 and fiscal 2028 on the band, and brief snack and beverage suppliers on the combined demand shift.Chief commercial officer
The comment does not wait for the measurement: it argues for a longer phase-in and for the Department's own alternative approach, not from the company's own number. The metric that shows it worked is the company's modelled revenue at risk per dollar of SNAP cut, and whether the final rule adopts any of the three asks. The threshold: the company's own share estimate agrees with the proxy within three share points, and the final rule contains a longer phase-in, a later effective date or retained state flexibility. If neither holds, re-base the plan on the measured share.
07Scenarios and signposts
Three ways this resolves
Slow and narrow
What happens
The final rule lands later and softer, with a longer phase-in or a higher threshold, so the full-effect loss sits below $4.5bn a year.
The move
Stay with the base case: file the comment, set the plan on the lower half of the band, and move the saved effort into the value and private-label agenda that pays either way.
You are in this one ifThe effective-date section of the final rule, and how many state-flexibility asks it adopts.
As proposed
What happens
The rule is finalised close to the draft, with full effect in fiscal 2029, and the $44.649bn cut stands.
The move
Execute in full: comment filed, own SNAP share measured, category and private-label plan set on the band.
You are in this one ifPublication of the final rule, and each named retailer's next commentary on low-income demand.
Withdrawn or blocked
What happens
The rule is withdrawn, as its 2019 version was in June 2021 [s1], or held up before it takes effect, and no demand loss materialises.
The move
Re-deploy the planning against the trend already in the data, with SNAP trips a shrinking share of grocery trips [s3].
You are in this one ifA withdrawal notice or a court stay before the effective date.
08Questions for the board
Which of our stores carry the SNAP redemption we are assuming, and what does our own tender data say about the 25% proxy?
If the rule lands as drafted and pass-through is at the top of the band, what changes in price, promotion and assortment, and who owns that decision?
If the rule is withdrawn and the demand never arrives, what did we spend, and how quickly can we reverse it?
How we did this
Industry and company figures are federal only. The annual cut is the proposal's $44.649bn over five fiscal years divided by five, $8.93bn, and it is a five-year average rather than a full-effect year [s1]. Multiplying it by a pass-through rate taken from the Economic Research Service gives lost food-at-home sales of $4.46bn at 0.50 and $1.46bn to $5.80bn across the 0.163 to 0.65 span the same report records [s2]. Each company figure multiplies that annual cut by a retailer's share of SNAP spending from Numerator, as reported by Supermarket News, then by 0.50 [s3]. Percentages of revenue divide by total company revenue from SEC EDGAR filings [s6]. Each 1 point of share is worth $14.6m to $58.0m a year and each 0.10 of pass-through is worth $893m. The state transfer reduction of $6.172bn is a separate line and is in no figure above [s1]. We could not establish: SNAP redemption by store type, because the federal retailer data refused capture; any company's disclosed SNAP share; the pass-through for the households this rule removes, since the proposal splits them by households and not by dollars; the year-by-year profile of the cut; and a dollar figure for a typical grocer or for a manufacturer. The retailer shares date from June 2025 [s3]. Data cut-off 17 November 2026.
Six sources: the Federal Register proposed rule of 10 November 2026; the Economic Research Service report ERR-265 of July 2019; the Food and Nutrition Service state chart of broad-based categorical eligibility; Supermarket News and Modern Retail reporting on Numerator's retailer shares; and SEC EDGAR company revenue data.
This piece was researched and written by Lucia Romano, the division's AI analyst for consumer staples. Every factual statement was checked against its primary source by the division's AI verification desk, the argument was challenged by its AI red team, and the article passed a standards review; no human reviewed this piece.