NORTHPOINTE BANCSHARES INC (NPB) stock intrinsic value fair value estimate & owner earnings analysis

Northpointe Bancshares owns Northpointe Bank of Grand Rapids, Michigan, a mortgage-focused bank that operates from a single branch and lends and gathers deposits across the United States through digital channels and loan offices. It has two main businesses. The Mortgage Purchase Program funds independent mortgage lenders: the bank holds their newly closed home loans, typically for less than 30 days, until they are sold to Fannie Mae, Freddie Mac, Ginnie Mae or other investors. Those balances doubled in 2025 and grew another 36% in the year to June 30, 2026, and they are now more than half of the bank's loans. The bank also passed $489 million of these loans to other banks as participations at June 30, 2026, a sign that client demand exceeds what its own capital can carry. The bank reports no charge-offs in this program since it began in 2010. Retail Banking originates home loans nationwide, holds All-in-One loans (a first-mortgage credit line linked to a checking account; $732.6 million at year-end 2025, up 19.7%) and gathers deposits online. Its older residential mortgage portfolio, 28% of loans at year-end 2025, is being allowed to run off. Over 23 months starting in 2022 the company exited correspondent lending, moved standard loan servicing to subservicers and sold most of its servicing rights, and in February 2025 it listed its shares and raised new common equity. How we measured earnings. A bank's operating cash flow is swamped by mortgages it makes and resells within weeks, so we measured owner earnings from net income: we added back depreciation and intangible amortization, deducted the capital spending needed to maintain the business, and removed one-time items (gains from repurchasing debt below par, a 2024 loss on the servicing sale, and a $3.2 million non-cash write-off in 2025 when the Series A preferred was redeemed). Loan-loss provisions, servicing-value declines and stock compensation are treated as real costs. On that basis owner earnings to common shareholders were $0.99 a share in 2023, $1.97 in 2024 and $2.19 in 2025. Because the share count and capital base changed sharply with the 2025 listing and the preferred redemption, we value the bank by the return it earns on its equity. Measured on total shareholders' equity, that return was about 8% in 2023, 13.2% in 2024, about 14.5% in 2025, and 15.3% and 14.4% in the first two quarters of 2026. We treat 2023 as describing a different business: it was the last year of the repositioning, when the bank was still winding down a correspondent lending operation whose costs exceeded its revenue (noninterest expense fell from $153 million in 2023 to $115 million in 2024). It remains useful as evidence of how badly a severe mortgage trough can hurt, and it sets our bear case. A bank must keep adding capital as its loans grow, so in every case we subtract the equity each year's growth requires; only what is left counts as cash to owners. Debt and preferred stock. For a bank, deposits ($5.23 billion at June 30, 2026, of which $2.74 billion were brokered CDs), Federal Home Loan Bank advances ($1.43 billion) and the holding company's $112 million of subordinated notes are the raw material of the business. Their interest is already charged in the earnings we value, so we do not subtract them again. The $25 million Series B preferred stock ranks ahead of the common, so we value the earnings available to all shareholders and subtract the preferred at its $25 million liquidation value. Its 7.25% fixed dividend switches to a floating rate after December 30, 2026. Base case, $36.25 a share. The business is worth about $1,309 million to all shareholders; the preferred claims $25 million, leaving $1,284 million for common shareholders, or $36.25 on 35.4 million diluted shares. Owner earnings per share grew 11.2% in 2025. For the first year we assume a 14.3% return on the $612 million of equity at June 30, 2026, just below the second quarter's 14.4%; that is about $2.42 a share for common holders, growth of 10.4%, below the 2025 rate. For years 2 through 10 we assume 13.5%. That sits below the 14.4% to 15.3% the repositioned bank has earned recently and above the 13.2% it earned in 2024, a weak year for mortgage volumes, because ten years will very likely include another downturn and because the margin narrowed in the second quarter of 2026 to 2.33% from 2.42%. We credit loan and capital growth of 5% a year for five years and 3% for the next five, all funded from retained earnings. That is at or below what the bank has funded itself: equity grew 3.2% in the 2023 trough and 7.4% in 2024 with no new capital, and the bank keeps nearly all its earnings. We did not credit the pace of 2025, when warehouse balances doubled, because the company attributes part of it to capitalizing on disruption among competing lenders, a share gain that cannot be assumed to repeat. After year 10 we assume no growth and a 12% return for a further ten years, roughly as long as the warehouse program's 16-year record supports, after which the return above our 6% discount rate fades to nothing over the following decade, leaving the bank worth its book value. About 60% of the base value comes from years beyond the first decade. Bear case, $19.49 a share. Returns fall back to the roughly 8.3% of 2023 and stay there for ten years with no growth, after which the bank is worth only its book value. The business is then worth about $715 million to all shareholders, less the $25 million preferred, or $690 million for common holders. The risks behind this case are all in the filings: the warehouse business depends on independent mortgage lenders, who suffer in the same downturns as the bank's own lending; the company warns that Fannie Mae or Freddie Mac could develop real-time funding that would reduce lenders' need for warehouse facilities; brokered CDs are 52% of deposits and wholesale funding 63% of the total, and access to brokered money depends on the bank staying well capitalized (tier 1 leverage 8.30%, common equity tier 1 9.19% at June 30, 2026); and the allowance for credit losses is only $9.4 million, 0.15% of loans, so a housing downturn that hit the All-in-One and residential books would require a larger reserve. Nonperforming assets were 1.15% of assets (0.80% excluding government-guaranteed loans), down from 1.35% a year earlier. Even in this case earnings of about $51 million a year cover the roughly $8 million of annual interest on the subordinated notes many times over. Bull case, $44.40 a share. Capital and loans grow 10% a year for three years, 7% for three and 5% for four, which returns near 15% with nearly all earnings retained can fund and which the warehouse participations show demand for. Returns hold at the 15% of early 2026 for five years and 14% for five more, then 13% with no growth for another ten years before fading to a book-value return. The business is worth about $1,598 million to all shareholders, less $25 million of preferred, or $1,573 million for common holders. What changed. Our previous estimates were $19.49, $35.51 and $44.40. The base case rises to $36.25 because we now measure the bank's normal return from the business as it has existed since the repositioning finished, rather than counting the 2023 wind-down year in the middle of the range; that lifts the return for years 2 through 10 from 13% to 13.5%. The first-year return is set at 14.3%, just below the latest quarter. The bear and bull cases are unchanged. No new financial statements have been released since our last review. The shares appear to be worth about $36.25 in our base case, with a range of $19.49 to $44.40. The margin-of-safety price is $20.00, close to the bear case and about 1.2 times tangible book value of $16.94 a share. From that price, the base case implies about 23% a year if the shares reach the base value within four years, while the bear case implies only about 5% a year. The margin of safety is wide because this is a leveraged, mortgage-cycle bank that relies heavily on brokered funding, has only three full years of public financial statements, and draws most of its base value from years beyond the first decade. The main ways this could be wrong: a lasting drop in warehouse volumes or margins, a government-sponsored real-time funding product, credit losses on the All-in-One and residential books against a thin reserve, or a loss of access to brokered deposits if capital slipped. The company reports no losses on its warehouse program, but the bear case does not depend on that record continuing.