111,620 Borrowers, One Sweep

Here is a number to sit with: 111,620. That is how many California borrowers the SBA suspended in a single announcement, tied to 118,489 loans and 8.6 billion dollars in suspected fraud. Now ask the only question that matters. How many of those 111,620 files did a human being open, read, and judge before the suspension went out? You already know. The honest answer is the whole story.

Published June 16, 2026 • Filed under: The Machine Decides Now

A wall of dense electronic circuitry and data nodes, representing the automated SBA flagging system that suspended 111,620 California borrowers in a single sweep

The SBA wants the California announcement to read as muscle. Administrator Kelly Loeffler stepped to the microphone in early February 2026 and rolled out the figures like a trophy. The agency had suspended 111,620 California borrowers who collectively received 118,489 Paycheck Protection Program and Economic Injury Disaster Loans worth 8.6 billion dollars in suspected fraud, with the cases being referred to the Department of Justice. The headline practically writes itself: tough agency, staggering theft, accountability at last. And buried inside that triumphant number is the part nobody at the podium wants you to do, which is the arithmetic.

So let us do it. One hundred and eleven thousand, six hundred and twenty borrowers. If a trained reviewer spent a generous thirty minutes on each file, reading the application, checking the business, verifying the documents, and confirming the fraud, a single person working without breaks would need more than 55,000 hours to get through the pile. That is over six years of continuous, round-the-clock labor for one reviewer, or a small army of them working for months. The SBA did not do that. The SBA cannot do that. The suspensions went out in a sweep, which means the deciding was done by a system, not a person, and the flag came first while the judgment, if it ever comes, comes later.

The Flag Is Automated. The Consequences Are Not.

This is the engine under every mass-suspension headline, and California is just the biggest example of it yet. The agency runs the loan database against a set of fraud indicators, addresses that appear too many times, identities that do not match, patterns that look wrong, and the matches get flagged and suspended in bulk. That is a reasonable way to find fraud. It is a catastrophic way to punish people, because a flag is a suspicion and a suspension is a sentence, and the SBA collapses the two into a single automated step.

The agency's own showcase example proves both halves at once. Investigators pointed to a single San Diego address linked to 14 different small businesses that together pulled in more than 2 million dollars in fraudulent loans. That is real fraud, blatant and deserving of every consequence coming to it, and the pattern-matching found it exactly as designed. But here is the thing the podium will not say: the same system that catches the obvious 14-businesses-one-mailbox scam also flags the legitimate borrower whose shared office building, family address, or strip-mall suite happens to trip the same wire. The machine cannot tell the difference between a fraud ring and a coincidence. It only knows that the pattern matched, and it suspends on the match.

A flag is a suspicion. A suspension is a sentence. The SBA built a system that turns the first into the second automatically, for 111,620 people at once, and then calls the result enforcement.

What Suspension Actually Does To A Real Borrower

Strip away the press release and look at what lands on the person at the other end of the flag. Suspension from SBA programs is not a polite letter asking for clarification. It can freeze access to programs, mark the borrower as a suspected fraudster in federal systems, and route the case toward the Department of Justice, all before that borrower has spoken to a single human being who reviewed their specific situation. For the actual fraudster, fine, the system worked. For the legitimate small business owner caught in the net, the burden of proof has been quietly inverted. They are now suspected, flagged, and referred, and the job of proving the algorithm wrong falls entirely on them, with no hearing scheduled and no human assigned.

Out of 111,620 suspensions, the statistical certainty is that some meaningful number are false positives, ordinary borrowers whose only crime was matching a pattern. The SBA will never tell you how many, because counting them would require the individual review the agency skipped in the first place. The false positives are invisible by design. They are the cost the machine is allowed to impose because the headline number is too big and too satisfying for anyone to audit the denominator.

The Asymmetry, As Always

Now run the agency's two speeds side by side, because this is where every LOLSBA story lands. When the SBA wants to flag and suspend a broke California borrower, it does it instantly and in bulk, 111,620 at a time, no human review, no hearing, machine-flagged and done. When that same borrower needs the agency to actually look at their individual file and lift a wrongful flag, there is no timeline, no assigned reviewer, and no automated process working in their favor. The sweep is industrial. The appeal is artisanal. The agency built a factory for accusing people and left the exoneration desk unstaffed.

That asymmetry is the product, not a glitch. The SBA spent the pandemic shoveling money out the door with no verification, created the fraud, and is now performing the cleanup at the scale of a hundred thousand suspensions per announcement because the only way to look tough after failing that badly is to swing the net as wide as possible. Wide nets catch the guilty. They also catch the innocent. And the agency has decided that the legitimate borrower flagged by mistake is acceptable collateral for a number big enough to put on a podium.

The Number Is The Punchline

So here is where it lands. The SBA wants 8.6 billion dollars and 111,620 suspensions to read as the agency finally getting serious. Read it the other way. That number is only possible because no one reviewed those files individually, because the deciding was handed to a pattern-matcher, and because the agency is willing to treat a suspicion as a verdict for six figures of people at a time. The San Diego mailbox with 14 fake businesses deserved everything it got. The unknown thousands of real borrowers swept up alongside it deserved a human being, and got an algorithm instead.

The lesson will not be learned, because the lesson is inconvenient. An agency that can suspend 111,620 borrowers in a single sweep but cannot verify a single application before it sends the money is not bad at its job. It is extremely good at the job it actually does, which is generating big enforcement numbers on the back end to cover for a complete absence of verification on the front end. The machine that decided who got the money never checked. The machine that decided who gets accused never checks either. Same machine. Same indifference. Only the direction of the harm changes.

SHARE ON X SHARE ON FACEBOOK SHARE ON LINKEDIN