OCC's Gould recruits Minnesota community banks against a nearly 300% rise in check-fraud filings

A rise of nearly 300% means check-fraud SAR filings in 2024 were about four times their 2020 level ^[15]. Non-mortgage fraud reports grew a little more than twofold over the same years ^[1]. By Gould's own figures, then, the check-fraud count grew by about twice the multiple of the broader measure ^[16]. Both figures count reports ^[1]. A SAR count goes up when fraud goes up, and it also goes up when banks get quicker to file, and these numbers cannot tell the two apart. The two accounts also differ on geography: the OCC release puts the doubling in Minnesota ^[1], while American Banker reported Gould describing a doubling in the Minneapolis area ^[2].

At the October 6 panel in Minneapolis ^[3], Gould was plain about where he expects fraud to be caught. "The community bankers gathered here today, some of whom we will hear from directly, are truly the front line of defense against fraud," Gould said ^[12]. He offered them two things. One was a new tool, and he urged them to use it ^[4][11]. The other was a reminder of the OCC's bulletin on FinCEN's 314(b) guidance ^[5]. American Banker reported that the OCC and the FDIC issued bulletins in July encouraging banks to opt into that sharing ^[9].

The push could go several ways from here. Small banks could make little use of the tool and the sharing, because both are opt-in ^[9]. The task force could build the reports it receives into a record of which banks caught what, and examiners would then have a way to measure detection. Or the 314(b) safe harbor could do most of the good. A bank that can warn a peer about an unproven suspicion without legal exposure ^[10] could flag a scheme before it reaches the next bank.

I think the safe harbor is the bigger change, and I expect banks to make little use of the new tool in the near term. Nothing Gould announced says examiners will now grade community banks on how well they detect fraud. The case against that view is the task-force path. An agency that collects fraud reports straight from banks is gathering the data a future exam would use. As American Banker reported it, Gould said the aim was to put the "force and gravitas of the entire U.S. government behind" the crackdown ^[14]. That sounds like more than a hotline. If the OCC issued an exam bulletin or took enforcement action over weak fraud detection at a community bank, that case would be right and mine wrong.

Gould also highlighted Treasury's Do Not Pay figures ^[6], which show what federal screening catches by itself. The $175 million returned is about 0.005% of the roughly $3.7 trillion screened ^[17], or about one payment in 81,000 ^[18]. The returned payments averaged about $13,000 each, close to four times the roughly $3,400 average of everything screened ^[19]. As Treasury reported it, the screen caught payments that would have gone to deceased individuals ^[7]. The Minnesota money Gould described went to living people. "Billions of dollars intended for hungry children, housing for disabled seniors, and services for children with special needs were diverted to people who cheated the system, some of whom are not even American citizens," he said ^[8].