UNITED BANKSHARES INC/WV (UBSI) stock intrinsic value fair value estimate & owner earnings analysis

BUSINESS. United Bankshares is a $33.8 billion bank holding company based in Charleston, West Virginia, with branches across West Virginia, Virginia, Maryland, Washington D.C., North and South Carolina, Georgia and border counties of Ohio and Pennsylvania. It earns its money on the spread between what it pays for deposits (27.2 billion) and what it earns on 24.7 billion of loans, plus about 135 million a year of fee income. It has grown by buying banks with its own stock (Cardinal 2017, Carolina Financial 2020, Community Bankers Trust 2022, Piedmont Bancorp January 2025), which is why the share count roughly doubled since 2016 and per-share earnings power grew about 4.7% a year (owner earnings per share of $2.29 in 2016 to $3.47 in 2025) while total profit grew faster. It has raised its dividend for 52 consecutive years, through 2008-2010 and 2020. HOW WE MEASURED EARNINGS. We start from operating cash flow, remove swings in working capital (including the large mortgage-loans-held-for-sale flows of 2019-2022, when mortgage originations ran 2 to 6 billion a year and have since shrunk to under 0.4 billion), and subtract stock compensation and capital spending. Ten-year series per diluted share: 2016 $2.29, 2017 $1.98, 2018 $2.44, 2019 $2.69, 2020 $2.85, 2021 $2.62, 2022 $3.04, 2023 $3.06, 2024 $2.93, 2025 $3.47. For a bank this cash-flow figure adds back the provision for loan losses, but actual loan losses are a real cost that never passes through operating cash flow, so we subtract a normal level of net charge-offs: 0.18% of roughly 24 billion of loans, about $0.30 a share (actual: 0.06% in 2024, 0.19% in 2025, 0.25% to 0.28% in 2014-2016). We also subtract $0.11 a share for the amortization of tax-credit investments, which is added back in cash flow but was paid for in cash. Starting point: $3.47 - $0.30 - $0.11 = $3.06 of distributable owner earnings per share. We use 2025 (the first full year after Piedmont) rather than an average of the 2021-2024 years, because the five quarters since (net income of $121M, $131M, $129M, $124M and $131M) confirm the higher level is real; we did not lift it to the first-half-2026 run rate. For reference the plain median of 2021-2025 is $3.04, nearly identical. Growth check: owner earnings per share grew 18.5% in 2025 (Piedmont, a wider margin of 3.78% versus 3.49%, and a fuller year of earnings). The base case assumes only 2% for the first year, below that actual rate, so no reversal needs to be argued. A bank must keep part of each year's profit to hold its capital ratios as it grows (tangible book value is about $24.60 a share); every case counts only what is left after that retention. Banks are valued on earnings after interest, so no debt is subtracted: deposits are the raw material of the business. The holding company's own borrowings are small (about $0.5 billion of long-term debt, $0.2 billion short-term). BASE CASE: about $55.70 per share. Distributable earnings of $3.06 grow 2% a year for ten years (to about $3.73 of earnings, $3.30 distributable in year 10), then hold level with no further growth credited. The 2% is funded from the bank's own retained profit (roughly 16% of earnings retained at a 12% return on tangible equity) and sits well below the 4.7% a year it has actually delivered; we left out any credit for further acquisitions, share gains, buybacks funded from today's excess capital (CET1 of 13.4%, 16.3% equity-to-assets) and margin expansion because none is contracted or small and repeatable. We assume the franchise keeps earning at the year-10 level afterward because the evidence is a deposit base and dividend record that have continued through every cycle in the filings; no growth is credited beyond that. BEAR CASE: about $40.40 per share. Two flat years as in 2022-2024, when deposit costs held earnings per share flat for three years. Then a full credit cycle modeled on 2008-2012, when the industry's commercial real-estate losses peaked: United's nonowner-occupied commercial real estate loans are $8.3 billion (34% of loans, against $5.5 billion of equity), and the filing already flags office loans (two Northern Virginia office loans were charged off for $21.8 million in 2025; nonaccrual loans rose from 0.26% to 0.39% of loans). We assume cumulative losses of 4.5% of those loans ($375M) and 1.5% of all other loans ($246M), $621M before tax and $478M after tax, or $3.45 per share spread over four years ($1.30, $1.00, $0.70, $0.45), plus a permanent 0.35-point squeeze in the margin from rate cuts (about $0.58 per share a year). Capital would stay above requirements (CET1 would fall by roughly 1.7 points from 13.4%), so no new shares are issued. Earnings then recover at 1.5% a year and hold level after year 10. The bear value of $40.40 is well above tangible book value because the bank keeps earning about 10% on its tangible equity and keeps paying its dividend through the downturn, so the franchise is worth more than its capital. BULL CASE: about $68.00 per share. The current run-rate holds (net income of about $515M a year, $3.50 per share distributable) and grows 4% tapering to 2% over ten years, in line with the 4.7% already delivered, with the Piedmont integration and the repricing of fixed-rate loans (the margin rose from 3.49% to 3.78% in 2025) carrying through. No windfall or market recovery is assumed. CHANGE FROM OUR LAST ESTIMATE. Our earlier figures (bear $28.90, base $42.35, bull $53.45) were built on a different basis and are not comparable. The business itself has not changed materially since: second-quarter 2026 net income of $131M was in line with the preceding four quarters, and the company repurchased $136M of stock in the first half of 2026. WHAT WOULD MAKE THIS WRONG. A commercial real-estate downturn deeper than 4.5% of the nonowner-occupied book, a sustained fall in the margin, or another stock-funded acquisition at a poor price. MARGIN-OF-SAFETY PRICE: $27.00. We are fairly sure of the earnings level because five straight quarters support it and the dividend has been raised for 52 years, but less sure of loan losses: a third of the loan book is commercial property, and the plausible worst outcome (office losses well beyond the 4.5% we model together with margin compression) would still leave the bank profitable and well capitalised with 13.4% CET1, so the structure survives, but earnings could sit near $2.50 a share for years. If we are wrong on credit losses and margin together, our downturn case puts the shares at about $40 while the dividend keeps being paid. The price assumes a holding period of about three years, with a $1.52 dividend (about 5.6% at $27) paid while tangible book value builds. At $27 the shares are priced at about 1.1 times tangible book for a bank earning about 12% on that equity; if over three years the shares reflect about $42 (a little above the downturn-case value and far below the base case) with roughly $4.50 of dividends, the implied return is about 20% a year. Above that price the margin for the credit risk described is too thin for a 20% target.