SNAP payment error rates by state

A SNAP payment error rate measures the percentage of Supplemental Nutrition Assistance Program (SNAP) benefits that are issued incorrectly, either overpayments to households or underpayments to eligible ones. These rates are determined through the USDA's Quality Control (QC) review process, which audits a statistically representative sample of cases in each state. The FY 2024 combined error rate includes both over-issuance (9.26% nationally) and under-issuance (1.67% nationally) improper payments.

Based on the most recent USDA FNS FY 2024 Quality Control data, Alaska has the highest SNAP payment error rate at 24.66%. This is significantly above the national average of 10.93%. High error rates can result from complex eligibility rules, staffing shortages, outdated technology systems, and high caseload volumes.

The national average SNAP payment error rate for FY 2024 is 10.93%, comprising a 9.26% overpayment rate and a 1.67% underpayment rate. This means roughly 1 in 9 SNAP benefit calculations contain some form of improper payment. South Dakota has the lowest rate at 3.28%, while Alaska has the highest at 24.66%. The national average is calculated using a dollar-weighted methodology by USDA FNS.

H.R. 1, signed into law in July 2025, establishes a tiered cost-sharing framework for states based on their SNAP payment error rates, taking effect in FY 2028. States with error rates below 6% owe no cost share. States at 6–8% share 5% of benefit costs, those at 8–10% share 10%, and states at or above 10% share 15% of their total benefit allotments. These provisions are designed to incentivize states to invest in payment accuracy and process improvements.

SNAP error rates are calculated through the USDA Food and Nutrition Service (FNS) Quality Control (QC) system. Each state conducts monthly reviews of a random sample of active SNAP cases and negative actions (denials and terminations). Federal reviewers then re-review a subsample for accuracy. The error rate is computed as the total dollar value of over-issuances and under-issuances divided by the total dollar value of benefits issued, expressed as a percentage.

Under H.R. 1, states are subject to tiered cost-sharing starting in FY 2028 based on their error rate. States at 6–8% share 5% of benefit costs, those at 8–10% share 10%, and states at or above 10% share 15%. However, H.R. 1 includes a temporary delay for states with the highest error rates. If a state's FY 2025 payment error rate exceeds 13.33%, it is exempt from cost-sharing until FY 2029. If its FY 2026 rate still exceeds 13.33%, the exemption extends to FY 2030. This means the highest-error states may not face cost-sharing obligations for up to two additional years; if their error rate falls from above 13.33% to between 6% and 13.33% during that period, cost-sharing is triggered at the applicable tier rate. The cost-share estimates shown on this page reflect each state's current tier obligation based on FY 2024 error rates and do not account for this temporary delay provision.

States can strengthen SNAP payment accuracy through several strategies: investing in modern eligibility verification technology, implementing improper-payment detection systems powered by machine learning, improving caseworker training and support, simplifying application and recertification processes, and conducting targeted quality improvement reviews. Savor Snap partners with state human services agencies on applied research and human-centered software, including AI tools, that help eligibility staff and supervisors catch improper payments before they reach families.

The USDA FNS publishes official SNAP error rates annually, typically releasing data from the prior fiscal year. The QC review process runs continuously throughout the year, but the compiled and verified state-level error rates are released on an annual cycle. States receive preliminary data before the official publication, allowing them to begin corrective actions. This map is updated when new official data becomes available from USDA FNS.