A $523M bank halves its expense base by walking away from mortgage origination — at a trough it admits it cannot call.
NPB · Earnings Call · 2026-10-21
An inflection point, self-declared
Northpointe Bancshares is small — $7.7 billion of assets, a $523 million market cap — and this week it announced it is selling the thing that until Tuesday defined it. A non-binding letter of intent with CMG Financial and affiliates would transfer the distributed retail and consumer direct mortgage origination channels. No loans, no servicing rights, no material assets — a people-and-infrastructure transfer.
It is a genuine strategic pivot, and it is company-unique. It is also a theme management stopped talking about: distributed retail and consumer direct sat at the very top of this company's keyword stack as recently as January; refinance activity led the pack in Q2. Now the growth engine is being sold. “We have reached an inflection point that no longer made the business channel viable” — Chuck Williams, Chairman and CEO · 2026-10-21, Chuck Williams said — and when Crispin Love pressed on whether this is the trough, Williams did not pretend to know:
It is hard to predict whether we are in the trough or not. We have been at this business since the end of COVID... it is just an inflection point to where the volumes could no longer carry the infrastructure overhead.
That is an unusually honest framing, and it is the whole dossier in one sentence: this is not a sale into strength, it is a sale out of exhaustion. The replacement thesis is specialty finance — warehouse-style lending through the Mortgage Purchase Program, specialty servicing, and opportunistic asset purchases.
The arithmetic behind the pivot
Q3 net income was $13.9 million, or $0.39 a share, against $21.3 million and $0.60 in Q2. Two discrete items did $7.5 million of pre-tax damage: a $4.3 million combined fair-value markdown on three rate-sensitive assets, and roughly $3.1 million of operational loss from an isolated external deposit fraud incident. (Note a small internal inconsistency: Chuck says $3.1 million, Brad books $3.2 million in the expense line. Minor, but worth flagging when a fraud loss is still under FBI review and an insurance claim is pending.)
Strip those out and the operating story is decent. Net interest income of $42 million rose 16% year over year, and non-interest expense of $35 million was up 11%. Management guides full-year 2026 expense to $140–146 million and 2027 to $80–95 million — an annualized savings of $50–55 million, mostly commissions, salaries and benefits.
Here is the tension. Brad claims “the expense savings will be greater than the revenue losses. So that will result in a net overall P&L benefit” — Brad Howes, Chief Financial Officer · 2026-10-21. True in sign, thin in magnitude. Net gain on sale of loans, excluding fair-value noise, was $11.5 million in Q3 and falls to essentially nothing in 2027 — call it $46 million annualized. Against $50–55 million of savings, the net improvement is a rounding error before you assume the savings cost you any of the $8–10 million of MPP fees that survive. The non-interest income line being amputated is not a footnote: at roughly 52%, Northpointe has been running a bank with an unusually large fee-and-gain engine bolted on. Half of it is leaving.
The rate-hike pinch is not theirs alone
Net interest margin came in at 2.32%, down a basis point, and the mechanics are instructive. Brokered CD rates repriced in advance of the Fed move; the assets did not. Kevin Comps was explicit: “our yields and our MPP business reprice at a one-month lag, so the benefit from the increase in SOFR is not reflected in our third quarter results” — Kevin Comps, President · 2026-10-21. Higher average deposit costs — a keyword that has sat in this company's stack for several quarters — did the damage first.
What matters is that this is not a Northpointe story. The market's own top keywords for the current quarter read like a single rate-regime trade: hiking cycle, retail deposits, Fed rate hike, deposit spread. Recent reporters echoed it almost verbatim — HWC flagged deposit costs, RF and ORRF and SMBK all leaned on rate hikes, CNS explicitly discussed the pace of Fed hiking cycles. Northpointe is riding a sector wave, not creating one.
And the wave reversed on them. In January, Brad built the 2026 guide on “2 rate cuts that get embedded in the middle part of the year” — Bradley Howes, Chief Financial Officer · 2026-01-21. Today the company assumes one additional 25bp Fed funds rate increase in early 2027. That is a full sign-flip in the rate posture in nine months, and it explains why the 2027 margin guide of 2.4–2.6% is a bet on mix rather than on the curve.
What replaces the machine — and the tell
MPP balances are guided to $4.1–4.3 billion at year-end 2026 and $4.7–5.0 billion in 2027. The constraint that mattered a quarter ago has loosened: in July Brad warned “growth is going to slow as we bought up against our limitations on the capital side” — Bradley Howes, Executive Vice President and Chief Financial Officer · 2026-07-22. The held-for-sale book winding down frees exactly that capital for redeployment. What has not loosened is pricing. Average MPP yields of 6.27% (6.49% fee-adjusted) continue to compress, and MPP yields have been a downward-trending keyword for this company since at least April. Chuck, in July: “There's competitive pressures from a limited... not expanding volumes in the space, while we continue to grow pretty substantially” — Charles Williams, Chairman and Chief Executive Officer · 2026-07-22. Selling origination does not fix that; it concentrates the company in the more crowded half of the mortgage ecosystem.
The offset is opportunistic purchases — the $85.4 million seasoned AIO pool bought at a discount for $69.9 million of net balances, loans originated 2005–2008, ~5% coupon accreting to ~7.5%. Cheap assets, cheap overhead, no origination infrastructure. That is the model.
The setup
At $523 million of market cap against $612 million of common equity, Northpointe trades well below book — an analyst on the call put it below tangible book outright. The stock is down 15.2% in 90 days and 21.6% from its July high of $19.31. Against that, management raised the dividend 20% to $0.03 and said, in Brad's words, “we do not have a share repurchase plan today, but we are in the process of putting one together” — Brad Howes, Chief Financial Officer · 2026-10-21. An inflection point, a sub-book multiple, and a buyback in the works — the ingredients are there. The question is whether a company that just discovered its origination overhead could not be carried can execute a leaner, more competitive warehouse-and-servicing model fast enough to matter before the sector's pricing pressure eats the savings. Efficiency has never been the problem — the efficiency ratio of 27.3% already looks like a different industry. Revenue durability is the problem now.