COLB — Q2 2026 Financial Report Analysis

Bigger bank, same earnings per share

Columbia Banking System, the Tacoma-based parent of Columbia Bank, earned $208 million in the second quarter of 2026, up 37% from $152 million a year earlier — but diluted earnings per share were $0.73 in both quarters. The reason is the August 31, 2025 acquisition of Pacific Premier Bancorp: it added loans, deposits and fee businesses (net interest income rose 32%), but Columbia paid for it largely in new shares, so the average diluted share count rose 36% (from 210.0 million to 286.5 million). Profit grew about as fast as the number of shares splitting it.

Against the previous quarter the picture is better: EPS rose from $0.66 to $0.73 as merger costs fell away and Columbia bought back 2.3% of its shares in three months.

At a glance

• Net interest margin 3.93%, up from 3.75% a year ago. The bank now earns noticeably more on each dollar it lends and invests, mainly because it is paying less for its funding — the cost of interest-bearing liabilities fell from 2.79% to 2.23% over the first half.
• Diluted EPS $0.73, flat year on year, while net income grew 37%. Shareholders own a bigger bank, not yet a more profitable share.
• Problem loans rose to 0.42% of assets from 0.30% at year-end, which management attributes to a single agricultural borrower. Charge-offs (loans actually written off) stayed low at 0.25% of loans.

The quarter in numbers

Sources: Form 10-Q for the quarter ended June 30, 2026 (income statement, MD&A), and the July 23, 2026 earnings release (efficiency ratio, Q2 2025 balance-sheet and charge-off comparisons, operating EPS). CET1 is the 10-Q's final figure as of June 30, 2026 (the earnings release had estimated 11.6%); the 10-Q compares it with year-end 2025 (11.8%), not with June 2025.

How the bank makes money, and what changed

A bank like Columbia mostly earns the gap between what it charges borrowers and what it pays depositors and lenders. That gap, divided by the assets earning interest, is the net interest margin (NIM). Columbia's rose to 3.93% from 3.75% a year earlier. The 10-Q gives the cause directly: "lower funding costs and a favorable balance sheet mix shift toward lower-cost customer deposits and away from higher-cost wholesale funding sources." For the first half, the cost of interest-bearing liabilities fell 56 basis points (0.56 percentage points) to 2.23%, while the yield on earning assets fell only 14 basis points to 5.42%. The Federal Reserve's 0.75 points of rate cuts in 2025 helped the funding side more than they hurt the lending side.

Quarter on quarter, NIM slipped 3 basis points to 3.93% from 3.96%. Management says that entire drop came from $4 million of interest income reversals — interest the bank had booked on loans and then took back, which typically happens when a loan stops being expected to pay. Excluding that, the margin was flat: a lower yield on taxable securities (4.22% to 3.97%, from changed prepayment assumptions) was offset by cheaper deposits (cost of interest-bearing deposits down 8 basis points to 1.96%).

The efficiency ratio — non-interest expense as a share of revenue, so lower is better; 55% means 55 cents of every revenue dollar goes to running the bank — was 55.15%, slightly worse than 54.29% a year earlier but better than 58.03% in Q1. Non-interest expense fell $19 million from Q1 to $375 million: $15 million of that was lower merger and restructuring expense after the systems conversion finished in Q1, and the rest was acquisition cost savings. The earnings release says all cost synergies from the Pacific Premier deal were "essentially complete" by June 30.

Fee income rose to $88 million from $65 million a year ago. The biggest single driver was financial services and trust revenue, which went from $6 million to $15 million — the 10-Q credits Pacific Premier's custodial trust business. Bank-owned life insurance income rose to $9 million from $5 million, including $3 million of death-benefit proceeds on a single policy, which won't recur on a schedule.

The balance sheet is shrinking on purpose

Loans fell to $47.2 billion from $47.7 billion at the end of March, and deposits fell to $52.1 billion from $53.5 billion. Both are deliberate, by management's account:

• Loans: Columbia is letting "below-market-rate transactional loans" (loans made without a broader banking relationship) run off, and non-owner-occupied commercial real estate payoffs were elevated. Commercial loans including owner-occupied real estate grew at a 5% annualized rate, partly offsetting that.
• Deposits: brokered deposits (money placed through intermediaries, usually expensive) fell to $978 million from $1.6 billion, and wholesale public deposits to $928 million from $1.2 billion. April tax payments also pulled customer deposits down.
• But borrowings went up: borrowings rose to $4.3 billion from $3.4 billion in the quarter. The 10-Q explains that Federal Home Loan Bank advances were cheaper than brokered deposits, so part of the "brokered deposit reduction" was a swap of one wholesale funding source for another rather than a pure win for customer deposits. The two retail deposit campaigns are the offset to watch: the first brought in over $600 million by April 30, and the second about $650 million through mid-July.

What the headline numbers hide

• EPS growth came from fewer shares, not more profit per share — but only quarter on quarter. Year on year, EPS was flat because the share count jumped 36% with the acquisition. Since then Columbia has been buying back stock: 13.1 million shares for $398 million in the first half, cutting average diluted shares from 296.8 million in Q4 2025 to 286.5 million in Q2 2026. About $202 million of the $700 million authorization remained at June 30, and it expires November 30, 2026.
• GAAP vs "operating" earnings. Columbia's non-GAAP operating net income was $217 million ($0.76 per share) versus GAAP $208 million ($0.73). The $9 million gap excludes $9 million of merger, restructuring and exit costs and a $3 million fair-value/hedging loss, less $3 million of tax. That gap has narrowed sharply from Q1 ($192 million GAAP vs $209 million operating), which is what you'd expect as integration ends.
• Amortization of acquired intangibles is large. Intangible amortization was $38 million in the quarter (up from $26 million a year ago) — a non-cash charge that gradually expenses intangible assets bought in acquisitions, such as the value of acquired deposit relationships. It is why return on tangible common equity (15.29%) runs well above plain return on equity (10.99%). ROTCE is still below the 16.03% of a year ago because tangible equity grew 43% with the deal.
• One-offs in both directions. Q2 2026 includes the $3 million insurance death benefit (helps) and $4 million of interest income reversals (hurts). The first half of 2025 included a $55 million legal settlement accrual, which flatters the six-month comparison (net income $400 million vs $239 million) — that's why this report leads with the quarter, where neither year has the settlement.
• A cost moved from interest expense to non-interest expense. A "deposit costs" expense line rose to $14 million from $2 million, which the 10-Q ties to HOA-related fees from Pacific Premier's homeowners'-association banking business. These are costs of gathering deposits that don't show up in the deposit rate, so the reported deposit cost slightly flatters the margin.
• Credit: low losses, rising problem loans. Non-performing assets rose to $273 million (0.42% of assets) from $200 million (0.30%) at year-end; management attributes the increase to "adverse performance in a single agricultural industry relationship," and notes $78 million of non-performing loans are government guaranteed. Separately, half of the quarter's $30 million of net charge-offs ($15 million) came from FinPac, the equipment-leasing unit, while the leases and equipment finance portfolio is only about 3% of total loans. The allowance for credit losses was $475 million, 1.01% of loans.
• Cash conversion is not a concern. Operating cash flow for the first half was $789 million against $400 million of net income. For a bank this measure is noisy (it moves with loans held for sale), but nothing here suggests earnings that aren't turning into cash.
• Tax rate barely moved: about 24.4% in Q2 2026 vs 25.1% a year earlier, so taxes did not drive the profit growth.

Takeaway: The Pacific Premier deal has done what it was supposed to on the income statement — a 3.93% margin, cost savings fully delivered, and profit up 37% — but per-share earnings are where they were a year ago, because the deal's share issuance absorbed all of that growth. From here, EPS growth depends on buybacks and on whether Columbia can grow customer deposits fast enough to stop leaning on FHLB borrowings.

What to watch next

Columbia will report third-quarter results on October 22, 2026, after the market close. Neither the 10-Q nor the earnings release gives numeric guidance; the 10-Q says the balance sheet is "modestly liability sensitive" (it benefits slightly if rates fall, because funding costs reprice faster) and names customer deposit trends and reducing brokered deposits and FHLB advances as the key drivers of the margin.

Three things will decide whether this quarter's improvement holds:

1. Whether the agricultural relationship turns into a loss. Non-performing assets climbed for two straight quarters. If that borrower is charged off, the 0.25% net charge-off ratio will rise; if it is resolved, the $4 million of reversed interest could partly come back.
2. Loan growth. Total loans have shrunk for two quarters as transactional loans run off. Revenue can't grow much on a shrinking balance sheet, so the 5% annualized commercial growth needs to start outpacing the runoff.
3. Buybacks once the program expires. With $202 million left on an authorization that ends November 30 and a CET1 ratio of 11.7%, the size of any new program will matter a lot for per-share earnings, which is exactly where the merger hasn't paid off yet.

My read: the cost side of the merger is finished and the margin is solid, so the next leg of earnings growth has to come from the balance sheet — loans and core deposits — which has been going the other way. Until that turns, expect little per-share earnings growth beyond what buybacks contribute.