GUBMENTPlain talk · policy frontier
Filings / Disability / Sources / §3 Findings: Fraud / improper-payment
GBMT-12 · Research record · No. 12

§3 Findings: Fraud / improper-payment number (KC1 workstream)

disability/research/ws03-integrity.md
This is a working research document from the disability filing, published as written — including the parts later corrected. It is the underlying record for Whitepaper No. 12, not a summary of it.

Date: 2026-08-11. Extends Phase 0 anchor 4 / KC1; fills anchor 10 directionally; adjudicates H3. Programs: SSI and OASDI improper payments are separate series. Medical CDR cessations are a third object. Intentional fraud / CDI prosecutions are a fourth. Do not collapse them.

Search log (PRISMA-lite)

Field Record
Date 2026-08-11
Strings SSA AFR payment integrity FY2023 SSI 10.62; PaymentAccuracy.gov Social Security improper; OIG PIIA SSI improper payments; CDR cessations full medical review; CDI judicial actions SSA OIG; improper payment not fraud stewardship review
Sources searched SSA FY2024 AFR Other Reporting Requirements (Payment Integrity); PaymentAccuracy.gov (via AFR pointers); SSA OIG PIIA / progress releases; SSA CDI Expansion Progress Report 2022; SSA BBA §845(a) FY2020 program-integrity narrative; CRR/Mathematica WP 2022-11 (CDR volume cite); SSA OIG SAR (prosecution/conviction metrics); Phase 0 register S12–S13
Inclusion Agency primary definitions and cause tables; OIG CDI judicial-action person counts; CDR cessation person counts from SSA-cited research or agency tables
Exclusion Media “X% fraudulent” claims without root; equating IP dollars to fraud dollars; treating CDR medical improvement as original-award fraud

1. Primary quotes — improper ≠ fraud (AFR / OIG)

SSA FY2024 AFR, Payment Integrity background (emphasis added by filing for construct clarity; wording is the agency’s):

Each year, we report IP findings, both overpayments (OP) and underpayments (UP), from our stewardship reviews of the nonmedical aspects of the OASDI and SSI programs.

We conduct Medical Continuing Disability Reviews (CDR) to determine whether disability beneficiaries meet the programs’ medical criteria. Terminating disability benefits after a CDR does not necessarily mean that the original determination was incorrect; it means the beneficiary’s medical condition has improved and the beneficiary no longer meets our definition of disability. Therefore, we consider the benefits received before improvement to be proper.

OMB/PIIA categories SSA reports as improper include agency computation mistakes, failure to obtain or act on available information, beneficiary failure to report, or incorrect beneficiary reports — not a labeled intentional-fraud rate.

OIG PIIA compliance work critiques SSA’s progress on reducing IPs; it does not redefine the stewardship IP percentage as intentional fraud (Phase 0 S13).

Confidence: strong — primary AFR definitional text.

2. Magnitudes and cause roots (PaymentAccuracy lineage)

FY2023 stewardship estimates (AFR):

Program Outlays (approx.) Improper payments IP rate Of which OP / UP
SSI $61.0B $6.5B 10.62% OP 9.18% / UP 1.44%
OASDI $1.35T ~$4.1B ~0.30% OP ~0.24% / UP ~0.06%

Leading cause categories (AFR, last-5-year stewardship findings):

PaymentAccuracy.gov is the public root dashboard the AFR points to for program-level cause detail. No stewardship table in the AFR equates these percentages to intentional fraud at the circulating magnitude.

KC1 remains fired: the integrity debate’s headline percentage is an improper-payment construct.

3. CDR continuance / cessation vs prosecution person-counts (anchor 10)

Object Approximate person-count scale Source / vintage
Disabled-worker medical CDR volume FY2019: 215,720 full medical reviews + 766,913 mailers CRR WP 2022-11 citing SSA
Initial cessations + FO FTC terminations (disabled workers) FY2019: 39,056 Same
DDS CDR cessations (pre-COVID comparison year in OIG DDS workload audit) 136,481 cessations (vs 78,284 in COVID comparison window) SSA OIG A-01-21-51038
CDI claim cessations/denials FY2020: 1,729 claims ceased or denied after CDI SSA BBA §845(a) FY2020 narrative
CDI judicial actions (sentencing, diversion, civil settlement, CMP) FY2020: 77; FY2021: 74 Same BBA narrative; CDI Expansion Progress Report 2022
OIG Office of Investigations convictions (all program areas, half-year SAR window) 221 convictions + 7 pretrial diversions (Oct 2025–Mar 2026 SAR) SSA OIG Spring 2026 SAR — not disability-only

Reading: even using the narrower disabled-worker FY2019 cessation figure (~39k), CDR exits dwarf CDI judicial actions (~70–80/year) by roughly two orders of magnitude. Using the larger DDS cessation tally (~100k+) widens the gap. OIG’s broader conviction count across all investigative categories remains in the low hundreds per half-year — still far below CDR cessation person-counts.

Construct caution: CDR cessation ≠ finding that the original award was fraudulent (AFR quote above). CDI judicial actions ≠ all overpayment cases. Anchor 10’s claim is ordinal (cessations ≫ prosecutions in person terms), not a claim that CDRs measure fraud.

CSV note: SSA’s open-data Periodic CDR CSV returned access-denied from this environment (2026-08-11); person-counts above use published secondary cites of SSA tables and OIG/agency narrative reports. Full table archive remains a §7 polish item if the CSV unlocks.

Anchor 10 status: Directionally verified (cessations substantially exceed criminal/CDI judicial person-counts). Exact same-year paired extract still owed for whitepaper table polish.

4. H3 adjudication

H3 — The fraud number is not a fraud number.

Criterion Result
Primary SSA/OIG IP publication attributes a majority of dollars to error/complexity/status change rather than intentional fraud Met. SSI causes = financial accounts, wages, ISM; OASDI = relationship, SGA reporting, computations. Stewardship = nonmedical.
Or every high-circulation “X% fraudulent” claim root-traces to one non-independent source Not required once the first leg holds; media citogenesis audit still owed in §8.

Verdict: SUPPORTED (definitional / primary-publication form). Matches Phase 0 KC1. No administrative series located that measures intentional fraud at the circulating IP magnitude with transparent methods → KC4 does not fire.

Confidence: strong.

Implications

← All Disability research documents Sources digest Read the whitepaper