What a bank examiner actually looks for

I spent twenty years as an FDIC examiner. People imagine the job as forensic: an auditor with a flashlight, looking for the hidden loss. It is not that. The hidden loss, when it exists, is usually in plain view — a concentration in one kind of loan, a deposit base that could leave in a week, a vendor nobody has read the contract with. The examiner's real job is to find out whether the people running the bank have seen it too, and what they did about it last quarter.

That is the whole test, and it fits in a sentence: does management know what could hurt the bank, and can they show you, on paper, what they did about it? I examined something like two hundred banks. The good ones could answer in under a minute, usually by handing me a document. The bad ones talked for an hour about their culture.

In 2008 I was examiner-in-charge on four bank closures. I remember each chief executive's face on the Friday evening when we walked in with the receivership papers. Not one of them was a crook. Every one of them had been surprised by something they could have imagined a year earlier — a construction-loan book that depended on prices never falling, a funding model that depended on brokered deposits always being available. The failure was not of honesty. It was of imagination, applied early enough.

When I came to Meridian Trust in 2017 as its first dedicated chief risk officer, I built the function around that one lesson. The result is the Red Folder: twelve pages, once a quarter, listing the five things most likely to hurt the bank in the next twelve months, in order, with a page on each saying what we did about it since the last Folder. The board chair reads every page. So do the branch managers. Page count is itself a control — thirteen pages is a number nobody finishes.

What goes on page one changes. In 2019 it was interest-rate risk on a fixed-rate mortgage book. In 2022 it was the core-system migration. In March 2023, when two large banks failed in a week, it was liquidity — and because the CFO and I had written that page eighteen months earlier, the response was already a plan rather than a meeting. This quarter page one is a vendor whose parent company has changed hands. Most banks would put that on page four. I have watched a vendor's change of ownership end a bank's online banking for nine days. Page one.

The second thing an examiner looks for is whether bad news travels upward fast. In a healthy bank a teller who notices something odd about a customer's deposits can get that observation to the CRO by the end of the day. In an unhealthy one, bad news is filtered at every level until it arrives at the top as good news. We test this at Meridian the way Dr. Osei tests it at his hospital: when something goes wrong, the review asks what should have caught it, never who. People report what they will not be blamed for.

The third is concentration, which is simply the question 'what is the one thing that, if it went wrong, goes wrong everywhere at once?' For Meridian the honest answer is agriculture: a fifth of our loan book is farms, and farms have one payday a year. Robert Tanaka made sure we understood that before we lent a dollar to a hazelnut orchard. The Folder tracks the hazelnut price, the Willamette Valley weather and the export market every quarter, not because I farm, but because a bad November would be the bank's bad November too.

I grew up on a dairy farm outside Tillamook. In 1986 my father lost the milk contract and the family kept the land only because a local bank chose to restructure the mortgage rather than foreclose. I met the examiner who reviewed that bank the next year, and decided at fourteen that I wanted his job. I got it, and then I crossed the table. The question I ask is still his: do you know what could hurt you, and can you show me what you did about it? If you work at a bank and cannot answer that in a minute, start writing. Twelve pages. No more.

One more thing an examiner always asks for, and that I always have ready: the list of loans to insiders. Any credit to a director, an officer, or a member of their immediate family is a related-party loan. Our policy is stricter than the regulation: the full board approves it, the interested director leaves the room and does not vote, the minutes record the recusal, and the loan is listed by name in the Red Folder until it is repaid. A board that cannot show an examiner that list, with the recusals beside it, has a governance problem whether or not it has a credit problem.