$31.94
$52.50
bear $31.94
base $46.40
bull $52.50
margin of safety $24.00
market price* $28.44
Columbia Banking System owns Columbia Bank, a commercial bank of about $65 billion in assets and $52 billion in deposits. Most of its lending and deposit business is in Oregon, Washington, California, Idaho, Nevada, Arizona, Utah and Colorado, with a small presence in Texas. Today's bank is the product of two deals. In the February 2023 merger with Umpqua Holdings, Umpqua was the accounting acquirer, so the history before 2023 is Umpqua's. Pacific Premier was bought for $2.36 billion in stock on 31 August 2025, which added about $16.6 billion of assets, mostly in Southern California. The loan book is about $47 billion. Multifamily is 22%, non-owner-occupied commercial real estate 17%, owner-occupied commercial real estate 15%, commercial term loans 14%, residential mortgages 12%, and FinPac equipment leases 3%. The bank earns the spread between loan yields (5.95% in 2025) and its cost of funds (2.61%), which gave a net interest margin of 3.83%, plus about $300 million a year of fees. How we measure earnings. We start from operating cash flow. We remove the cash swings from mortgages held for sale and from other assets and liabilities, stock compensation, and branch and equipment spending. Because operating cash flow adds back the whole loan-loss provision, we charge actual net loan losses instead. This counts losses when they happen and does not deduct the reserve build that comes with buying another bank's loans. On that basis the combined bank's owner earnings per share were $2.84 in 2023, $3.02 in 2024 and $2.96 in 2025, against $2.68 and $2.63 for Umpqua alone in 2021 and 2022. Each year ties exactly to net income adjusted for the same items. These figures sit above reported earnings for two reasons. Amortisation of the core-deposit intangibles from the mergers ($127 million in 2025) is a non-cash charge and is added back. The day-one reserve set up for acquired Pacific Premier loans ($70 million) is a bookkeeping entry, not a loss. Because the deposit intangibles are added back, we credit no growth from the deposit franchise. The 2025 figure still carries $148 million of merger and restructuring cost and a $55 million legal settlement accrual, neither of which should recur. We anchor on the 2023 to 2025 median of $2.96 a share. These are the years of the combined franchise. Earlier years describe a smaller, different bank. At about 285 million diluted shares that is roughly $845 million a year. The first half of 2026 supports this level: operating cash flow of $789 million, net income of $400 million, and $418 million spent buying back stock. Per-share owner earnings fell 2% in 2025. We hold year one flat rather than at that decline, because the dated merger costs and legal accrual that depressed 2025 are behind it. This is a bank, so its deposits and its $4.25 billion of Federal Home Loan Bank and other borrowings are raw material, not debt to subtract. The values below are owner earnings after interest, valued directly. Base case, $46.40 a share. Owner earnings hold near $845 million for three years. Over years four and five, loan losses rise from today's 0.27% of loans to a more ordinary 0.6% to 0.7%, which takes earnings to about $700 million. After that, earnings settle around $790 million to $800 million and stay there. What keeps that going is a structural franchise: a $52 billion, largely relationship-based deposit base gathered through more than 300 branches. That base has produced margins in a 3.2% to 3.8% band through the last rate cycle. We credit no growth. Its acquisitions have been large, stock-funded, step-change deals, not small repeatable tuck-ins, so acquisition growth is not in the bag. Loan growth would need share gains, which are not in the bag either. Bear case, $31.94 a share. This case runs the bank through one credit cycle of the depth western banks suffered in 2008 to 2010. Umpqua and its peers were heavily exposed to commercial real estate and construction in that cycle. We assume net charge-offs of 1.5%, 2.5% and 1.5% of loans over three years, about $2.6 billion in total, or 5.5% of the book. We add a $450 million reserve build in the first stress year and lower pre-provision earnings as rates fall and loan demand weakens. Owner earnings turn to losses of about $120 million and $170 million in years two and three, then recover over two years to about $670 million, below the base case because some of the franchise is lost. Capital starts with a common equity tier 1 ratio of 11.8% against an effective minimum of 7%. That cushion absorbs these losses, but we assume about 6% dilution at trough prices to rebuild it and a reduced dividend. The bear value sits well above tangible book value of about $19.22 a share. The reason is that the deposit franchise earns about $1.1 billion a year before tax and credit costs, enough to absorb more than 2% of loans in losses every year without touching capital. A bank with that pre-provision earning power is worth more than its equity even in a bad cycle. The bear case would be worse if losses landed at the 2009 to 2010 peak rates for construction lending. Construction and development loans are only 4% of the book today, against far more for Umpqua going into 2008. Bull case, $52.50 a share. The Pacific Premier cost savings land in full, which the first-half 2026 expense run rate already points to. The FinPac leasing losses that drove $61 million of 2025 charge-offs keep shrinking. Owner earnings run at about $900 million, with only a mild credit year along the way. We include no margin expansion and no market recovery. What would make this wrong. First, commercial real estate. Multifamily and non-owner-occupied property loans are 39% of loans, concentrated in California and the Pacific Northwest, and a deeper property downturn than 2009 would push value below the bear case. Second, deposit costs. If customers move money into higher-yielding alternatives, the spread narrows. Third, the earnings level itself has only three years of history as a combined bank, two of them carrying heavy merger costs. That is short, though the reconciliations tie out cleanly. Margin of safety price, $24. We are reasonably confident about the earnings level: three years of combined-bank cash earnings within 6% of each other, and a first half of 2026 that confirms it. The worst plausible outcome is a severe western property bust together with deposit flight, like the regional bank failures of spring 2023. The structure survives that: capital is 4.8 points above the minimum, borrowings are modest, and no single borrower or regulator decision dominates. Shareholders would still be diluted. If we are wrong on credit losses and on the margin together, the shares are worth roughly the bear case, low $30s, and a deeper property bust would take them toward tangible book in the high teens. The bank pays a dividend of $1.48 a year and is buying back stock; the dividend is about 6% a year of the $24 margin-of-safety price. At $24, the price is about 25% above tangible book and about 8 times owner earnings. A 20% annual return from there requires either the base case to show through within about four years, or simply no credit cycle in the first two years. If the bear cycle arrives early, the bear case implies a value of roughly $32 plus dividends, roughly equal to that price. The margin-of-safety price is set no higher than that, because the bear case is a real possibility for a bank with this much property lending. The previous estimates were built on a different earnings basis and are not comparable to these figures.