Old Second Finally Turns the Credit Corner — Right as Its Margin Sermon Gets a Haircut

For four straight quarters, analysts have opened Old Second Bancorp's call by asking some version of the same question: when does the credit noise stop? Nathan Race asked it in July. He asked it again Thursday. The answer this time came with a number attached — $12.7 million of net charge-offs, $9.2 million of it a single warehousing-and-distribution credit the company has been working out "for the better part of a year" — and an oddly serene management team on the other side of it.

The Clean-Up Quarter That Wasn't Clean — And That's the Point

The isolated charge-off is the headline; the loan book is the story. Classified loans fell $57.8 million, nonperformers $16.8 million, and criticized loans dropped $61.3 million, or 38%, in a single quarter. Eccher's framing: “we think we have turned the corner an inflection point with a lot of this credit” — James L. Eccher, Chairman, President, and Chief Executive Officer · 2026-10-22 — and, more concretely, “we have got to go back to October 2019. Way before COVID to have classifieds at this level” — James L. Eccher, Chairman, President, and Chief Executive Officer · 2026-10-22.

That is a real break in pattern. In July, Eccher told the same analyst he thought the bank was “really close to having a very clean quarter on the credit front” — James L. Eccher, Chairman, President and CEO · 2026-07-23 — and then printed the elevated number. The sequence matters because the market has been trained, correctly, to discount that kind of optimism. What's different this time is that the forward-looking buckets moved: criticized and special-mention migration is the early-warning system, and it emptied out. The allowance followed, down to $64.2 million or 1.23% of loans from 1.34%, with Bradley Adams cautioning the reserve won't fall much further because the powersport vertical will always require replenishment.

The résidual stinger is that powersports — the acquired Evergreen specialty engine — keeps generating both the best and worst lines in the release. The loan book is, in Eccher's words, as good as it has been "in several years," but the consumer credit tail is structural, not cyclical.

Five Basis Points, and the End of a Four-Quarter Sermon

The more consequential shift on this call was in tone on the balance sheet. Adams has spent roughly a year telling investors the margin is flat, durable, and going nowhere but sideways — he once said he was “very bullish, a very elevated margin, for a very long time” — Bradley S. Adams, COO and CFO · 2026-07-23. In July he guided to 5.18% for Q3 and 5.15% for Q4 “that is my best guess” — Bradley S. Adams, COO and CFO · 2026-07-23. The bank delivered 5.27%, a 4 basis point increase and 22 basis points above a year ago, with cost of deposits at 102 basis points versus 133 a year earlier. On the printed scale, that is exceptional margin performance once again.

And then, for the first time, the hedge came with a direction and a cause:

the loan to deposit ratio is pretty much nearing the ceiling at this point. So for the first time in a while, we are participating at the marginal funding rate relative to our loan growth expectations. And I think that is probably the biggest driver in what I am talking about.

That is the whole story compressed. Loan-to-deposit went from 91.4% a year ago to 96.4% last quarter to 98.3% at September 30 — loan to deposit ratio is now the binding constraint, not rate sensitivity. Average deposits fell $78 million, or 1.4%, largely seasonal but with residual runoff of high-beta wholesale money left over from Evergreen. And $39.5 million of buybacks in the quarter literally retired free funding. Adams expects to give back about 5 basis points in Q4 and to grow deposits again, which is a subtly different message than "the margin can go up from here."

This is not a company-specific problem — it is the whole cohort's problem, and the market context says so. The last quarter's editor-curated global keywords read like a regional-bank anxiety index: Fed rate hike near the very top, hiking cycle back in the vocabulary, plus loan beta and deposit beta as standalone themes. Peer reporters this week tell the same story — AMAL discussing deposit beta, CIVB and SFBS flagging repricing fixed-rate assets, HWC talking rate hikes built in. Old Second's rate backdrop flipped with everyone else's: Adams noted expectations for cuts have "reversed course to under rate hikes" on Middle East instability and high energy prices. What's company-specific is that Old Second starts that fight from a 5.27% margin and an efficiency ratio that has compressed to 44.8%, so even a 5-handle margin is an extraordinary place to cede ground from.

The Things Management Quietly Demoted — and the One It Unveiled

Read the keyword list against prior quarters and the vocabulary has rotated. outsized payoffs is new and prominent: prepayments through three quarters have already matched all of 2025, running roughly double a normal quarter, which is why total loans fell $7.8 million despite what Eccher called one of the stronger production quarters in years. And the syndication portfolio — acquired with West Suburban, once $550 million — is being deliberately demoted: “We have reduced that by over 60% since the end of 21. Yeah. Roughly $285 million. Will we go to zero? No. it is just not going to be a material part of our balance sheet” — James L. Eccher, Chairman, President, and Chief Executive Officer · 2026-10-22. Translation: a multi-year drag on reported growth is nearly spent.

The genuinely new item is buried in the specialty lending discussion. Darin Campbell announced a signed multiyear endorsed program with Ducati USA, launched the prior week, with “Ducati is 1 of the, you know, the best motorcycle brands that is in the market” — Darin Campbell, Unknown · 2026-10-22 — a best motorcycle brands origination pipeline that management says gives a lift in Q4 and a "substantial boost" in 2027. For a $1.26 billion bank whose asset growth has been gated by payoffs, an OEM endorsement is a rare, hard-to-replicate growth lever.

Capital return remains the loudest theme. share repurchases ran at $24 million in Q1 and $16 million in Q2 per filings, then $39.5 million in Q3 — 1.5 million shares, 3.4 million year-to-date at an average $22.74, plus a 14% dividend increase. Adams says 5% of shares a year is repeatable "and still have the ability to execute" on the fill-in M&A he prefers. Note the tension: with the stock at $25.91 as recently as August 14 and +12% over 90 days, the buyback has been buying below the current quote, and tangible book value per share slipped to $14.75 from $14.77 — decline by design.

At 1.3x tangible book and 12.9x earnings, the setup is a bank whose earnings power (1.69% ROA, 15.77% return on tangible common) has outrun its multiple. The catch: a stock four and a half percent off its high with a fresh report that says the credit cycle's tail is over but the funding cycle's is just beginning is not a free option. It's a re-rating that now needs the deposit side to cooperate.