05/13/2026 | Press release | Distributed by Public on 05/13/2026 11:17
The Supplemental Nutrition Assistance Program (SNAP) assists millions of vulnerable Americans each year. It is also a massive taxpayer investment, costing nearly $100 billion per year. Congress tasks the U.S. Department of Agriculture (USDA) with managing that investment and statutorily demands integrity and accountability practices be implemented by both state agencies and USDA. To fulfill that responsibility, USDA will leverage every available tool to prevent fraud and protect the generosity of the American taxpayer.
To that end, in May 2025, Secretary Brooke L. Rollins directed states to share their SNAP eligibility data with USDA, launching a historic and collaborative effort to root out fraud, waste, and abuse. The Department established the first ever SNAP Program Integrity Data Team to analyze the data provided by states and compare it to readily available federal databases. What they found was alarming; in the data provided by 29 state agencies, initial estimates indicate the team identified at least $3 billion a year of potential fraud, waste, and abuse. This report provides an overview of the team's methodologies and findings.
Today, USDA is working together with these state agencies to verify and, where appropriate, take action on the flags the data analyses revealed. The Trump Administration firmly believes this federal-state partnership is vital to strengthening SNAP integrity. Together, USDA and states will safeguard SNAP - and American taxpayers - from fraud, waste, and abuse, preserving benefits for those most in need.
Patrick A. Penn
Deputy Under Secretary
Food, Nutrition, and Consumer Services
U.S. Department of Agriculture
U.S. Department of Agriculture and the Food and Nutrition Service (FNS) use SNAP eligibility data to ensure the integrity of government programs, including by cross referencing SNAP recipient eligibility against federally maintained databases, identifying and eliminating duplicate enrollments, and performing additional eligibility and program integrity checks specified herein. This Standard Operating Procedure (SOP) establishes a framework for state SNAP agencies to share data with the USDA/FNS for fraud, waste, and abuse detection and program integrity purposes. The SOP balances federal data access requirements under Executive Order 14243 and the Food and Nutrition Act with privacy protections, minimizes unnecessary data collection, and incorporates security standards while maintaining public trust and compliance with applicable law.
Data Requested from StatesData requested from states can be found in Appendix A. Only data elements necessary to achieve specific, legally permissible goals, such as fraud detection, duplicate enrollment prevention, and program integrity checks, were collected and used. The scope of data collection was limited to excluding sensitive personally identifying information (PII) unless directly relevant to these goals (i.e., the collection of data elements spelled out in the Privacy Impact Assessment, found in Appendix B). Data derived from third-party sources (employment verification databases, financial institution records, and property records) that states use for verification but are not part of applicant-reported information are excluded and were not requested. The applicable system of records notice for this protocol is USDA/FNS-15, "National Supplemental Nutrition Assistance Program (SNAP) Information Database," as found in Appendix C.
Data Transfer and SecurityThe SNAP Information Database's infrastructure and operational procedures were designed to ensure the highest levels of security and compliance. The primary storage location is the Amazon Web Services (AWS) GovCloud (US) region, with backup and disaster recovery systems in a secondary AWS region. No storage, processing, or backup activities occur outside of GovCloud or other federally secure environments. All infrastructure adheres to Federal Risk and Authorization Management Program (FedRAMP) High Baseline standards. Data are encrypted to prevent AWS personnel from accessing unencrypted information, with encryption keys managed exclusively by federal USDA personnel, ensuring no access by AWS or third-party vendors.
Access controls are robust, requiring multi-factor authentication for all access points. USDA implements role-based access control (RBAC) following the principle of least privilege, with access limited to specifically designated employees within the FNS Program Integrity Division. Comprehensive access logging and audit trails are maintained for a period equal to the data retention period, with quarterly access reviews conducted by the USDA. Access is immediately revoked upon employee termination or assignment changes.
Encryption standards ensure that all data is encrypted both at rest and in transit, with encryption keys managed solely by USDA federal personnel, and no keys are held by commercial vendors. Data segregation and compartmentalization practices are in place, with SNAP data isolated in a separate, dedicated AWS GovCloud environment and not directly integrated with other USDA systems. Access is restricted to individuals with specialized oversight, who are required to take annual training on procedures for handling sensitive data, sign and attest to confidentiality requirements, and submit personal financial disclosures.
Finally, all data transfers are conducted using secure tools that meet or exceed FedRAMP High Baseline standards, ensuring the integrity and security of the data throughout the transfer process.
States were provided individual multi-factor authenticated access to the AWS location to securely pass data to specific federal USDA personnel who then moved data into secure analysis environment. Access to state data was limited to USDA personnel who meet the requirements outlined above and access was not granted to outside agencies.
ProcessingStates store and organize their data very differently, so state-specific processing procedures were developed to rapidly read in and synchronize data across all states when received. Special attention was paid to coordinating date fields, ensuring true recipients' data were in fact active participants in SNAP, and removing duplication. Some states provided multiple instances of a given recipient when their data were updated (i.e. address change), and care was taken to ensure current information was being considered. Data were arranged into tidy format, such that a single row per active recipient and a single column per category of data were created. States were consulted during this process to ensure proper interpretation of data formats and column definitions.
Flagging State DataUSDA employed a foundational fraud, waste, and abuse verification that focused on identity verification, income and eligibility verification, immigration status, reported residential and mailing addresses, verification of disqualified recipients, as well as non-recipient household member data. USDA cross referenced state supplied data against auxiliary data from other sources including Social Security Administration (SSA), Systematic Alien Verification for Entitlements (SAVE), and other internal FNS data sources to establish data integrity. Date of birth, household size, and spending patterns were used to glean insight on income and eligibility verification. No official flag or determination was made with regard to income, but spending patterns were linked to eligibility data from internal FNS data sources. SSA data coupled with SAVE data were used to flag possible immigration status discrepancies. The FNS Electronic Disqualification Recipient System was used to flag recipient disqualification status.
In addition to the core fraud detection functions, USDA employed additional techniques including flagging intrastate and interstate duplication, deceased individuals, synthetic identity patterns (new social security numbers [SSNs] with inconsistent biographical data), and geographic anomalies (such as high out-of-state transactions). Intrastate and interstate duplication flags were raised on individual social security numbers that were duplicated within (intrastate) or between states (interstate). Deceased individuals and synthetic identity patterns were flagged based on links with SSA data and consistency within state active eligibility data (i.e. same names and birth dates assigned to different SSNs, heads of household with unrealistic birthdates, or data that fit typical dummy-data patterns, such as SSNs of 123-45-6789 or similar configurations).
Reporting to StatesAfter analysis, identified data discrepancy flags were returned to some of the state agencies with a summary of all flag counts by zip code and an identified category tab for each of the flags described above. States were asked to review a sample of each flag in each of the respective categories and provide feedback on the efficacy of the flagging procedure. States were encouraged to provide final disposition of each case review. USDA personnel will use feedback to improve initial data read in or state-specific flagging or output procedures.
ExclusionsData was not used for tax administration or tax compliance, immigration enforcement, law enforcement investigations beyond coordination regarding criminal and administrative SNAP violations, administration of non-SNAP federal assistance programs (e.g., Medicaid, Temporary Assistance for Needy Families, housing assistance), sharing with foreign governments or international organizations, commercial use or transfer to private entities, or determining federal eligibility for benefits other than SNAP.
Table 1. SNAP Participation and Data Supplied by State Agency| State/Territory | SNAP Participation June 20251 | Responded to Data Request? |
| Alabama | 732,974 | Yes |
| Alaska | 66,572 | Yes |
| Arizona | 886,806 | No |
| Arkansas | 241,210 | Yes |
| California | 5,477,070 | No |
| Colorado | 614,911 | No |
| Connecticut | 361,655 | No |
| Delaware | 118,766 | No |
| District of Columbia | 140,716 | No |
| Florida | 2,928,850 | Yes |
| Georgia | 1,385,834 | Yes |
| Guam | 39,285 | Yes2 |
| Hawaii | 149,284 | No |
| Idaho | 132,534 | Yes |
| Illinois | 1,869,744 | No |
| Indiana | 580,902 | Yes |
| Iowa | 266,947 | Yes |
| Kansas | 186,560 | No |
| Kentucky | 593,934 | No |
| Louisiana | 791,032 | Yes |
| Maine | 163,056 | No |
| Maryland | 665,084 | No |
| Massachusetts | 1,079,234 | No |
| Michigan | 1,474,701 | No |
| Minnesota | 448,841 | No |
| Mississippi | 356,756 | Yes |
| Missouri | 660,033 | Yes |
| Montana | 80,066 | Yes |
| Nebraska | 149,699 | Yes |
| Nevada | 492,270 | Yes |
| New Hampshire | 75,489 | Yes |
| New Jersey | 821,038 | No |
| New Mexico | 458,019 | No |
| New York | 2,955,731 | No |
| North Carolina | 1,350,026 | Yes |
| North Dakota | 51,179 | Yes |
| Ohio | 1,437,112 | Yes |
| Oklahoma | 691,754 | Yes |
| Oregon | 772,310 | No |
| Pennsylvania | 1,941,067 | No |
| Rhode Island | 141,801 | No |
| South Carolina | 570,687 | Yes |
| South Dakota | 75,329 | Yes |
| Tennessee | 682,128 | Yes |
| Texas | 3,457,259 | Yes |
| Utah | 174,386 | Yes |
| Vermont | 64,094 | Yes |
| Virginia | 813,228 | Yes |
| Virgin Islands | 20,610 | Not required |
| Washington | 903,442 | No |
| West Virginia | 269,687 | Yes |
| Wisconsin | 687,133 | No |
| Wyoming | 26,927 | Yes |
| TOTAL | 41,575,762 |
1 Preliminary data as of September 2025.
2 Guam was not required to provide data but voluntarily responded to the request.
| Issue |
Total Count 29 State Agencies |
Median State Count1 | Median State Percent2 | Annual Implied Dollars3 |
| Intrastate Duplication | 247,575 | 467 | 0.10% | 558,529,200 |
| Interstate Duplication | 108,670 | 2,057 | 0.50% | 245,159,520 |
| Deceased (per SSA record) | 185,986 | 2,005 | 0.60% | 419,584,416 |
| Dummy SSN | 441,572 | 2,462 | 1.20% | 996,186,432 |
| No SSN Found | 26,333 | 254 | 0.10% | 59,407,248 |
| Disqualified but Active | 4,442 | 56 | 0.03% | 10,021,152 |
1 Median count of each item for 29 state agencies supplying data.
2 Median percent of records per state for each item in 29 state agencies supplying data.
3 Calculated as $2256 per recipient annually ($188 per month).
Intrastate Duplication: Recipients with same identifiers (SSN) that show up multiple times within a state.
Interstate Duplication: Recipients with same identifiers (SSN) that show up in multiple states.
Deceased (per SSA Record): Based on SSA lookup, the recipient is deceased.
Dummy SSN: Recipient has a non-normal SSN (i.e. 111-11-1111 or 999-99-9999).
No SSN Found: Recipient does not have an SSN in data provided.
Disqualified but Active: Recipient has been identified as a disqualified participant but is still active on state agency rolls.
The review of SNAP eligibility data across 29 State agencies identified several categories of discrepancies that could signal improper or erroneous benefit issuance. The largest issues by volume were cases with dummy or missing SSNs and intrastate duplication, together representing hundreds of thousands of records. Interstate duplication and deceased individuals appearing as active participants also showed substantial counts. Although the median percentage of each issue within States appear small, general under 1.2%, the implied annual financial exposure is significant. Dummy SSNs and intrastate duplication together account for more than $1.5 billion in estimated annual risk. Overall, the findings indicate that even small error rates can translate into substantial fiscal impact when applied across large caseloads, underscoring the importance of strengthened data-matching and eligibility verification controls.
These results highlight the need for targeted oversight, improved data integrity efforts, and sustained collaboration with states to reduce vulnerabilities and ensure program accuracy.