Columbia Finishes the Buyback It Promised — Then Triples Down
NIM finally crossed 4%, deposits grew without price concessions, and a completed $700M repurchase became $1.2B with no expiry — while the shares sit 11% below their July peak
COLB · Earnings Call · 2026-10-22
A promise kept, then escalated
For four quarters Ivan Seda had been telling investors the same thing about net interest margin. On the April call he was specific: “we will be roughly at that 4% marker in Q2 of this year, and then continue to step up from there” — Ivan Seda, Chief Financial Officer · 2026-04-23. Q2 printed 3.93% — mangled by one-time credit interest reversals and a securities-accounting drag he spent an unusual amount of airtime explaining. Q3 arrived at 4.11%, and on an adjusted basis expanded 12 basis points sequentially. The promise, finally, is a number. Capital is the louder story. Columbia completed its prior authorization during the quarter — 6.5 million shares for roughly $200 million — ending an 8% reduction in the common share count over the life of the program and more than $1 billion returned in twelve months. The board then authorized $1.2 billion more, with no twelve-month clock attached. Clint Stein did not hedge: “We did not throw a number out there that we do not intend to utilize” — Clint E. Stein · 2026-10-22. Read that against two prior data points and the arc is clear. A year ago, Ivan said of the old $700 million program that “we could go through the entire authorization over the course of the next 12 months” — Ivan Seda, Chief Financial Officer · 2025-10-31 — and they did, ahead of schedule. And in July, Stein ruled out whole-bank M&A entirely and called buybacks “the single best investment that we can make” — Clint Stein, Chief Executive Officer · 2026-07-23. The new authorization is the same sentence, made bigger. Repurchase of common stock ran at $215 million in Q1 2026 and $203 million in Q2 — a cadence management guided to hold at $150–200 million per quarter, with flexibility to accelerate on weakness.Crossing 4% and immediately guiding down
Here is the tension. Margin hit 4.11% and guidance for Q4 is lower: about 4.05%. The cause is self-inflicted and deliberate — a hole in the capital stack being repaired.So the same quarter that raised total risk-based capital to 13.8% via $250 million of new sub debt also imports a few basis points of negative carry into Q4, plus the usual seasonal deposit outflow. Investors should treat a two-to-three basis point give-back as the price of a cheaper capital structure, not a deterioration in the franchise — the low-4% range is the guided runway into 2027. The accounting subplot is the top-ranked keyword in Columbia's own lexicon this quarter: discount accretion. Last quarter it was a 4-basis-point headwind as higher rates slowed CPR speeds on discounted MBS. This quarter it simply did not repeat — Seda's s-curve argument that prepayment speeds had already reset to minimal levels. The bond book's front-book purchases still carry coupons 75–80 basis points above the back book, so investment portfolio repricing remains a slow, mechanical tailwind rather than a swing factor.We completed that 10-non-call-5 transaction in mid September... There is about $240 million worth of Tier 2 capital in the form of trust preferred securities that will be redeemed. About 80% of that pool... will be redeemed between now and the end of this year... we have got somewhere in the ballpark $200 million of excess tier 2 capital. That is our most expensive form of capital.
Winning the funding war without paying for it
Against a global tape where Fed rate hike, beta assumption and brokered deposits decline crowd the market-wide keyword list — and where regional-bank peers spent the week discussing deposit betas and wholesale runoff — Columbia's third quarter is a genuine outlier. Total deposits rose to $52.7 billion from $52.1 billion; customer deposits grew nearly $1.2 billion; and management cut wholesale funding by $1.5 billion between June 30 and September 30. The average cost of interest-bearing deposits fell a basis point to 1.95% in a quarter with a Fed increase in it. Christopher Merrywell was almost dismissive about how: “there is no special products. there is no special pricings” — Christopher Merrywell · 2026-10-22. The levers were treasury management penetration around 58% of commercial customers, working-capital reviews, and a next-best-offer engine pushed through Salesforce. The fee evidence is concrete: treasury management +7% year over year, international banking +19%, commercial card +23%, merchant card +11%. Non-interest income as a share of effective revenue reached the high end of guidance at $87 million operating, with mid-$80s guided for Q4.What Columbia stopped saying
Two quarters ago the company's own keyword set was dominated by integration vocabulary: custodial trust, system conversion, transactional portfolio, loan sale, CDI amortization. Almost all of it is gone. What replaced it is geography and plumbing — physical presence, metro areas, underserved markets, fine tuning, relationship reviews. The Pacific Premier lexicon has retired; the asset now shows up as Southern California, which had been an outright decliner in the Q1 keyword set and returns this quarter because nearly half of commercial deposit growth came out of the legacy PPBI footprint. Commercial real estate's fall is equally telling. It was the company's number one keyword last quarter, framed as a competitive market driving elevated payoffs. This quarter loan decline slowed to $46.9 billion from $47.2 billion, and Tory Nixon's residual complaint is about other people's underwriting: “We have seen a bunch at a SOFR plus 100. I have yet to see 1 dip below 100, but I know it is coming” — Torran Nixon · 2026-10-22. One caution flag: tangible book value fell to $18.10 as AOCI moved against them with rates. Stein's framing — targeting roughly 8% TCE ex-AOCI against a 12% total risk-based capital binding constraint — is consistent with prior calls, and Basel III relief is explicitly not in the $1.2 billion plan. That is appropriately conservative.The setup
The stock is down 3.2% over 90 days and 11.1% from its July 16 peak of $33.61, against a full-history drawdown of 40.8% from the 2021 high. At 1.2x price to book, a company generating an ROTCE above 17% and a 1.4% return on average assets is repurchasing its own equity at a discount to what it earns on it — an efficiency ratio of roughly 42% says the machine is lean. The philosophical bet is stated plainly, and it is the one thing here that is genuinely company-unique in a quarter saturated with AI talk:Not one AI keyword appears in Columbia's top thirty this quarter. Whether that is discipline or a missed cycle is the question the next four quarters will answer — and the $1.2 billion says management is betting on itself either way.Technology enhances our relationship model, It does not replace it. We remain confident in our strategy because our customers are attached to relationships not a robot.