More than 4 million people have dropped off the Supplemental Nutrition Assistance Program since the Fiscal Responsibility Act’s tougher work rules went into effect, a decline that has outpaced some federal projections and raised questions about whether administrative tightening by states is amplifying the policy’s reach. The participation drop, tracked through USDA monthly caseload data through March 2026, coincides with expanded time limits for able-bodied adults without dependents and narrower state waiver authority. For the roughly 42 million Americans who relied on SNAP before the changes, the shrinking rolls signal a shift in who qualifies, who gets recertified, and how aggressively states enforce eligibility checks.
Why the post-FRA participation drop matters right now
The Fiscal Responsibility Act of 2023, signed into law as the debt-limit compromise, changed the age ceiling for ABAWD (able-bodied adults without dependents) time limits and curtailed the discretionary exemptions states had used to shield recipients from work requirements. A December 2024 final rule published by the Food and Nutrition Service, available in the agency’s regulatory notice, formalized those statutory changes, setting the framework that state agencies now follow when making eligibility determinations and defining how age-based time limits, exemptions, and good-cause provisions must be applied.
The decline in SNAP participation is not happening in a vacuum. USDA’s FY 2025 payment error rate results put several states under heightened scrutiny, creating a parallel pressure on local agencies to tighten verification and recertification procedures. States with above-average error rates face financial penalties, and the resulting administrative response, including shorter certification periods, more frequent document requests, and stricter income verification, can push eligible households off the rolls before they fail to meet any work requirement. That dynamic makes it difficult to separate how much of the 4-million-person drop stems from people actually gaining employment versus how much reflects paperwork-driven attrition in states scrambling to lower their error rates.
Georgia offers a concrete example. The state publishes monthly caseload reports showing active SNAP cases during each reporting period. Georgia’s caseload reductions have tracked with both the federal work-rule rollout and the state’s own administrative adjustments, but the available data do not isolate the cause. No primary state-level enforcement records currently break out how many individuals lost eligibility specifically because they failed to meet work-rule requirements versus those who were dropped during routine recertification tightening or missed paperwork deadlines.
Federal data and CBO projections versus the observed decline
USDA’s Keydata reports, the canonical monthly series for SNAP participation measured in persons, households, and total benefits, document the net change from the month the FRA provisions became enforceable through March 2026. The agency’s March 2026 Keydata release shows a cumulative loss of more than 4 million participants compared with the baseline month, even as total benefits have not fallen as sharply because remaining households often have children, disabilities, or other characteristics that qualify them for higher benefit amounts. The series records both person-level and household-level counts, and the gap between the two metrics matters: a household that loses one member to a work-rule disqualification may remain on SNAP at a reduced benefit, meaning the person-level drop can exceed the household-level drop by a wide margin.
The Congressional Budget Office, in its analyses of the Fiscal Responsibility Act and related proposals, estimated how many ABAWDs would cycle off SNAP as time limits expanded and state waivers narrowed. Those projections assumed that some affected adults would increase their earnings or hours to meet the new work standards, while others would lose benefits after exhausting their allowable months. The observed 4-million-person decline, however, appears to be running ahead of the narrow subgroup CBO expected to be directly touched by ABAWD rules, suggesting that the policy change is interacting with broader administrative and economic forces.
One explanation is that state agencies, under simultaneous pressure to reduce payment errors and implement new work rules, have adopted more conservative practices when evaluating borderline cases. If a worker’s hours fluctuate around the minimum threshold, caseworkers facing tight performance metrics may be more likely to close the case at recertification rather than risk an overpayment finding. In states that have shortened certification periods, these borderline households now face more frequent rechecks, increasing the chances that a missed notice, an incomplete form, or a late employer verification can trigger a cutoff.
Another factor is the complexity of the new age tiers and exemptions. The FRA raised the upper age limit for ABAWD time limits in stages, with different birth years phasing into the requirement over time. The December 2024 rule spells out those details, but for recipients and even some frontline staff, keeping track of who is subject to which rule at any given moment can be challenging. Confusion can lead to eligible older adults being coded as ABAWDs and timed out, or younger adults incorrectly told they no longer qualify, further inflating the participation decline beyond what lawmakers or modelers anticipated.
Looking ahead, the key policy question is whether the current caseload contraction represents a one-time adjustment to new rules or the start of a longer trend of tightened access. If the drop is largely driven by administrative barriers, future USDA guidance or corrective action plans could restore some participation without changing the underlying statute. If, instead, the decline reflects sustained earnings gains and higher employment among former recipients, policymakers may point to the FRA as evidence that stricter work rules can reduce dependency. For now, the data through March 2026 show a system in flux, with millions fewer people receiving food assistance and only partial visibility into why they left the program.
Free tool for readers: Curious where your retirement stands on a 0–100 scale? You can get your free Retirement Safety Score in about five minutes — no account, no bank details, just your number and a few steps to improve it.
