Five minutes to apply, 3.7 million flags, 937 sentencings a year.
A stolen identity sold for about eight dollars; a fintech platform advertised a PPP application in about five minutes. GAO flagged 3.7 million PPP and EIDL recipients for fraud indicators; in fiscal 2024 federal courts sentenced 937 offenders in government-benefits-fraud cases, counting every such program. Dividing one by the other runs the horizon into thousands of years — a measure of scale, not a forecast; GAO's flagged set already includes recipients it says were prosecuted, never pursued, or flagged for reasons other than fraud.
A Two clocks
Pick a scheme. The left clock is what it takes to file; the right clock is what it takes to answer a case.
The dials are illustrative — the sourced figure sits at the center of each. The left clock counts minutes and dollars; the right counts years.
B The money
What the fraud costs to run, what a prosecuted case turns out to have lost, and what the government spends to answer it.
C The backlog
Prosecution scaleSet the size of the flagged population and the speed of enforcement. The horizon recomputes as you drag.
D Built in vs. bolted on
Share of program dollars already committed before the fraud controls were running. Each bar keeps its own source’s basis: GAO reports the PPP share as approved and the COVID-EIDL share as disbursed, and the UK figure is loan value paid out.
The controls existed; they were added after the money moved.
Sources
SBA Office of Inspector General, Report 23-09 (original: sba.gov · stored capture) (more than $200 billion in potentially fraudulent PPP and EIDL loans; SBA management disputed the figure with roughly $36 billion across its pandemic programs). Government Accountability Office: GAO-23-105331 (original: gao.gov · stored capture) (3.7 million of 13.4 million PPP and COVID-EIDL recipients carried fraud indicators; GAO states that further review, investigation and adjudication is needed to determine whether fraud exists); GAO-25-107746 (original: gao.gov · stored capture) (3,096 defendants publicly charged through the end of 2024); GAO-25-107267 (original: gao.gov · stored capture) (PPP dollars approved and COVID-EIDL dollars disbursed before the full control stack was running). U.S. Sentencing Commission, QuickFacts, fiscal 2024 (median loss of $137,600 in sentenced government-benefits-fraud cases; 937 such sentencings). That publication is not mirrored in this archive, so the 937 denominator behind the headline ratio could not be checked against a local copy. Transactional Records Access Clearinghouse (TRAC), Syracuse University, Federal Prosecution of White-Collar Crimes Receiving Less and Less Attention local copy (May 2025 — 4,332 white-collar prosecutions filed in fiscal 2024, fiscal 2025 projected at 3,862). House Select Subcommittee on the Coronavirus Crisis, December 2022 staff report (original: coronavirus-democrats-oversight.house.gov · stored capture) (thirty-second application reviews and processor accounting). Treasury Inspector General for Tax Administration, Report 2022-46-059 (original: oversight.gov · stored capture) (amended returns carrying Employee Retention Credit claims that met the criteria were not referred to Examination as required). National Audit Office, The Bounce Back Loan Scheme: an update (original: nao.org.uk · stored capture) (December 2021 — 93% of UK scheme loan value issued before HMRC’s turnover cross-check began). Internal Revenue Service (CP2000 / Automated Underreporter document-matching letters; volume here is an order-of-magnitude estimate with no document mirrored in this archive). Ten-year statute of limitations for PPP and EIDL fraud: Public Laws 117-166 and 117-165, August 5 2022; neither reaches unemployment-insurance fraud or the Employee Retention Credit. Application-speed claims: Womply’s Fast Lane launch and the Federal Trade Commission’s case against Womply (original: ftc.gov · stored capture); the Kabbage “applications in minutes” line comes from the company’s own announcement of April 7, 2020.
Indicators, estimates and criminal charges measure different things and do not convert into one another. Figures are recomposed from the cited public records; several are exposure or order-of-magnitude estimates rather than audited totals.