SNAP's 2025 error rate hit 10.62%. 43 states now owe a share.
Summary
USDA's annual measure of how accurately states determine SNAP eligibility and benefit amounts found a 10.62% national payment error rate for fiscal year 2025 -- $10.1 billion in combined over- and under-payments, nearly double the 6% threshold Congress set in law. Under the One Big Beautiful Bill Act, FY2025 is the first year states can use to calculate a new benefit cost-share, ranging from 0% to 15%, that takes effect as soon as fiscal year 2028. Forty-three of the 53 SNAP jurisdictions came in at or above that 6% line.
What the error rate measures
The rate isn't a fraud measure. 's own explainer states plainly that 'SNAP error rates are not fraud rates⧉' -- they capture largely unintentional mistakes: an applicant approved who shouldn't have been, or an eligible household paid more or less than it was owed. Every year, states review a sample of roughly 50,000 cases nationwide, and re-reviews about half of those to validate the results before calculating each state's rate.
View data as table
| Below 6% -- 0% state share | 10 |
|---|---|
| 6% to 8% -- 5% state share | 6 |
| 8% to 10% -- 10% state share | 16 |
| 10%+ -- 15% state share | 21 |
A law that turns the error rate into a bill
Section 10105 of the One Big Beautiful Bill Act⧉ sets four tiers: under 6% error, states pay nothing; 6% to 8% costs a state 5% of its benefit spending; 8% to 10% costs 10%; and 10% or higher costs 15%. The requirement generally starts as soon as fiscal year 2028, and states get to choose whether their FY2025 or FY2026 error rate determines that first bill. Applying FY2025's numbers alone, 21 of the 53 jurisdictions would land in the top, 15% tier.
View data as table
| Alaska | 23.2% |
|---|---|
| District of Columbia | 18.7% |
| New Mexico | 16.8% |
| Delaware | 16% |
| Georgia | 15.2% |
| Iowa | 5.3% |
| Kentucky | 4.7% |
| Wyoming | 4% |
| Idaho | 3.9% |
| South Dakota | 2.5% |
The spread between states is enormous
Alaska's 23.15% error rate was the nation's highest -- more than nine times South Dakota's 2.47%, the lowest. The law includes a safety valve for the most extreme cases: if a state's error rate times 1.5 reaches 20% (roughly a 13.3% error rate or higher), that state can push its cost-share start date back a year, to FY2029. Seven jurisdictions -- Alaska, Washington DC, Delaware, Georgia, Illinois, New Mexico, and Oregon -- cleared that bar using FY2025 data alone.
The takeaway
- The national rate missed the legal threshold by a wide margin. 10.62% against a 6% ceiling, $10.1 billion in combined errors.
- Most jurisdictions are over the line. 43 of 53 states and territories reported FY2025 error rates at or above 6%, the point where a state cost-share begins to apply.
- The gap between states is the real story. Alaska's rate was more than nine times South Dakota's -- and states above roughly 13.3% get an extra year before the new cost-share takes effect.
States with a FY2025 or FY2026 error rate at or above 6% must also submit a Corrective Action Plan to and may face a separate quality-control financial penalty, independent of the new benefit cost-share described here. Because states may elect to use either their FY2025 or FY2026 error rate to calculate their first required cost-share, the tier assignments in this article -- based on FY2025 data alone -- are illustrative of what the law would produce this year, not a final determination of what any state will actually owe starting in FY2028.
Sources(4) ▾
- U.S. Department of Agriculture, USDA Announces FY 2025 State Payment Error Rates in SNAP (2026-06-24) — 's own announcement of the annual payment error rate results, including the national rate, the $10.1 billion total, the congressional threshold, and a summary of the new state cost-share consequences under H.R. 1 (the One Big Beautiful Bill Act). Fetched via the Food and Nutrition Administration's mirrored newsroom page after the primary usda.gov URL timed out; text confirmed against the official FNA press release. usda.gov · original document
- U.S. Department of Agriculture, Food and Nutrition Administration, Supplemental Nutrition Assistance Program: Payment Error Rates, Fiscal Year 2025 (2026-06-24) — 's official data table: over-payment, under-payment, and combined payment error rates for all 50 states, DC, Guam, and the Virgin Islands (53 jurisdictions), plus the national weighted rate. Fetched directly from fna.usda.gov and converted with pdftotext -layout; read in full and used as the basis for the tier-count and ranking analysis in this article. fna.usda.gov · original document
- U.S. Department of Agriculture, Food and Nutrition Administration, SNAP Quality Control (program and FAQ page, including 'What is the new SNAP matching funds requirement?') (2026-07-13) — 's own explainer of the quality control process and the new OBBBA matching-funds requirement, used to confirm the general 0%-15% match range and effective-date mechanics independent of secondary-source characterizations. fna.usda.gov · original document
- U.S. Government Publishing Office (govinfo.gov), One Big Beautiful Bill Act, Public Law 119-21, Sec. 10105 (Matching Funds Requirements) (2025-07-04) — The actual enacted statutory text creating the state cost-share tiers, amending Section 4(a) of the Food and Nutrition Act of 2008 (7 U.S.C. 2013(a)). Fetched directly from govinfo.gov and read in full to confirm the exact tier percentages, effective date, and the delayed-implementation ('20 percent') provision independent of secondary summaries. govinfo.gov · original document
Comments
Always open. Logged-in readers can annotate paragraphs in place.
's annual SNAP payment error rate determination for fiscal year 2025 found a national rate of 10.62% -- nearly double the 6% threshold set in law -- representing $10.1 billion in combined over- and under-payments nationwide. The measurement carries new weight this year: under the One Big Beautiful Bill Act, FY2025 is the first year whose data a state could use to calculate a benefit cost-share of up to 15%, a financial obligation that did not exist before. It is not the only option -- states may instead elect to use FY2026 data, not yet public, whichever produces a more favorable result for them.