How do federal audits measure overpayments versus intentional SNAP fraud?
Executive summary
Federal audits measure SNAP overpayments primarily through an official improper payment estimate that found an 11.7% error rate (about $10.5 billion) for FY2023 and is produced using a two‑tiered quality control process overseen by USDA’s Food and Nutrition Service (FNS) and reviewed by GAO [1]. Separate investigations and state fraud units pursue intentional recipient or retailer fraud—activities GAO and OIG describe as distinct from the broad “improper payment” estimate and often measured by different tools like trafficking studies, prosecutions, and card‑theft incident tracking [1] [2] [3].
1. How federal audits define and count “improper payments”
Auditors use the FNS improper payment estimation system, a statistically driven two‑tiered quality control process that produces a national improper payment rate—11.7% for FY2023, or roughly $10.5 billion in payments deemed wrong by FNS’s methodology—which covers payments that are the wrong amount or should not have been made [1]. GAO’s reports describe this estimate as a broad umbrella metric designed to capture all overpayments, regardless of whether they result from error, administrative fault, or potential wrongdoing [1].
2. Why “improper payment” is not the same as intentional fraud
GAO and FNS make a technical distinction: the improper payment estimate aggregates all overpayments, while intentional fraud (for example, retailer trafficking or deliberate recipient deception) is a narrower category that requires case‑level proof and different measurement methods [1]. GAO has repeatedly urged FNS to improve how trafficking and other intentional misuse are estimated because assumptions in current methods (such as the share of benefits trafficked) are imperfect and can misattribute causes inside the broader improper‑payment number [1].
3. How auditors try to separate mistakes from malfeasance
Audits and oversight mix statistical sampling with targeted investigations. The national improper payment process relies on quality control reviews and sampling to estimate the total; by contrast, fraud is identified through investigative work by state fraud units, OIGs and law enforcement, which verify intent via documentary and transactional evidence before pursuing penalties or prosecution [1] [2]. GAO has recommended FNS refine trafficking estimates and better validate the assumptions behind how much of the improper‑payment pool stems from theft or intentional abuse [1].
4. The limits and gaps GAO highlights in current practice
GAO notes persistent weaknesses: several recommendations from 2018 about estimating retailer trafficking remain open, and FNS needs to better evaluate its assumptions about trafficking percentages [1]. GAO also found that FNS lacks comprehensive data on state‑level theft‑prevention measures and their effectiveness, which hampers the agency’s ability to target assistance and to disentangle fraud from other causes of improper payments [3].
5. State investigative units and recovered overpayments — ground truth on fraud
State SNAP fraud units and Benefits Program Integrity teams run case investigations that produce recoveries and referrals for prosecution; for example, Texas BPI reported nearly $44.3 million recovered and hundreds of investigations/referrals in part‑year reports, demonstrating how proven fraud and recoveries are tracked separately from the statistical improper‑payment estimate [2]. These units document intentional concealment of income, household composition fraud, and other deliberate schemes that are actionable in ways statistical audits are not [2].
6. What this means for interpreting headlines about “massive fraud”
High‑level numbers in political statements about “massive fraud” risk conflating three different things GAO separately measures: overall improper payments (a statistical error/overpayment estimate), documented theft or trafficking incidents (investigative outcomes), and the dollar value of reported stolen benefit claims (reported incidents of card theft or trafficking). Journalists and policymakers should check whether claims cite the improper‑payment rate, trafficking estimates, or recovered fraud figures—GAO cautions that current methods can overstate or misattribute the share tied to intentional misuse [1] [3] [2].
7. How measurement can improve and why it matters
GAO recommends FNS refine trafficking estimates, validate assumptions, and comprehensively assess state theft‑prevention measures so that federal audits can more accurately separate inadvertent overpayments from intentional fraud; doing so would let resources be targeted to prevention, prosecutions, or administrative fixes as appropriate [1] [3]. Until those recommendations are fully implemented, the improper‑payment figure will remain useful for tracking scale but limited for assigning causation between error and intentional fraud [1] [3].
Limitations: available sources do not provide the full technical sampling protocol or the exact statistical model FNS uses, nor do they include raw underlying data USDA may have used in recent public claims (not found in current reporting) [1] [4].