Seven Hills Realty Trust (SEVN) stock intrinsic value fair value estimate & owner earnings analysis

Seven Hills Realty Trust is a commercial mortgage REIT. It makes floating-rate first-mortgage bridge loans, mostly $20–50 million each, on middle-market properties across the United States. It is externally managed by Tremont Realty Capital, part of The RMR Group. At June 30, 2026 it held 27 loans with $765 million of commitments. These were funded by $470 million drawn on bank repurchase facilities with UBS, Citibank, Wells Fargo and BMO, plus $321 million of shareholders' equity, which is $14.15 of book value per share. About 19% of the book is office, and five loans (22% of the book) carry the company's higher-risk rating. No loan is in default, and the company has never realized a loan loss. How it earns. The loans yield about 7.7% (SOFR plus about 3.9%) and the borrowings cost about SOFR plus 2.15%. About two-thirds of the loans are funded with borrowed money, and that part earns a spread that does not depend on where rates go. The other third is funded with shareholders' money and earns SOFR plus a spread, so earnings rise and fall with short-term rates. Over time, each one-point move in SOFR moves earnings by about $0.14 per share a year. Minimum-rate clauses on the loans (a weighted average SOFR floor of 2.96%) soften the effect of a near-term drop, but those floors reset lower as loans are refinanced. In practice this is partly a bet on short-term interest rates, and a real part of why earnings fell from $1.21 to about $0.80 per share. We measure owner earnings as operating cash flow less working-capital swings and stock-based pay. We then make two adjustments that operating cash flow misses for a lender. We add the cash origination and exit fees the company collects, about $0.09 per share a year. We charge credit losses explicitly, because loan losses show up in investing cash flow, not operating cash flow. On that basis, owner earnings (before fees) were $0.98 per share in 2022, $1.23 in 2023, $1.21 in 2024 and $0.97 in 2025. They ran at about $0.75 annualized in the first half of 2026. What the rights offering did. In December 2025 the company sold 7.53 million new shares at $8.65 each through a rights offering (one new share for every two held), raising $65.2 million. That price was 49% of the $17.68 book value per share at September 30, 2025. Holders who exercised their rights bought about $14.50 of book value for $8.65, so they came out whole. Holders who didn't exercise saw book value per share fall from $17.68 to $14.55 by year-end. Tremont bought the 2.0 million unsubscribed shares (about $17.4 million) as backstop, so the manager captured part of that transfer. The offering also grew the manager's fees: the base management fee rose from $1.08 million in Q2 2025 to $1.30 million in Q2 2026, in step with the larger equity base. Base fees and reimbursed shared-services costs together come to about $7.5 million a year, roughly 30% of what the business earns before paying the manager. At the same time SOFR fell from about 5% in 2024 to about 3.75% now. The share count rose 51% while equity rose only about 20%, and rates fell. Together these cut per-share earning power from about $1.21 to about $0.75–0.90. That damage is real and permanent for shareholders who didn't take up their rights. The $0.28 quarterly dividend ($1.12 a year) is not covered by what the business earns: Q2 2026 distributable earnings were $0.23 a share, and owner earnings are about $0.80. Management says it expects to cover the dividend by the end of 2026 by growing the loan book to $950–960 million. Unless that happens, we expect the dividend to be cut to about what the business earns. Paying out more than it earns does not destroy value by itself, because the excess goes to shareholders. The harmful pattern is paying out capital at full book value and then refilling it by selling shares at half of book, which is what happened in 2025. Why the shares trade far below book. Mainly because the book earns too little. At today's rates the company earns roughly 5.5–6.5% on its book equity. An investor who wants 11–12% from a small mortgage REIT will pay only about 50–55 cents for a dollar of book earning 6%, which is about where the shares trade. That explains the discount without assuming any hidden loss. The shares were already at roughly 60% of book a year ago, before the offering, and their fall since then roughly matches the one-third drop in per-share earning power. Other things weigh on the price as well. The company's record of no realized losses is weaker evidence than it sounds. In 2023 it took over a property (Yardley) in place of a $14.8 million loan, recording it as owned real estate rather than a loss. In 2025 and the first half of 2026 it added $1.4 million of unpaid interest to loan balances instead of collecting it in cash. It also raised its credit reserve by $4.9 million in Q2 2026 on two office loans maturing this year. Roughly 30% of pre-fee earnings go to the manager. The company itself sold shares at half of book, showing it will do so again if it chooses. And at about $165 million in market value, the stock is too small for most institutions. Our value is higher than the market price mainly because we discount at a fixed 6% and do not borrow the market's much higher required return. Base case ($13.15). Earnings start at the first-half 2026 run rate of $0.75 per share. We add $0.09 of cash fees and $0.05 from lending out cash already raised, bringing cash back to the roughly $45 million the company held in mid-2025. The company's own current credit reserve of $14.4 million ($0.64 per share) is charged as losses over 2026–2028. After that we charge an ongoing loss rate of 0.25% of loans a year. No growth beyond today's book is credited: the $950 million target is a management goal, not a signed pipeline. Rates are held at today's level. The lender keeps recycling its capital as loans repay every two to three years, which is how it has operated since 2021. The result is about 6% on book equity, worth a little under book. Bear case ($6.40). About $35 million of losses ($1.54 per share) on the higher-risk loans over 2026–2028, which is roughly a 22% severity on those five loans. The ongoing loss rate is 0.5% a year. SOFR settles one point lower. Loan floors limit the hit to about $0.09 per share a year at first, rising to about $0.14 as the book is refinanced at lower floors. The case also includes a second below-book share issuance of similar size. The repurchase facilities can demand more collateral when loan values fall, which adds to the downside in a property downturn. Bull case ($15.85). The $44 million Dallas office loan and the other 2026 office maturities repay at par, with only about $3 million of losses. The company uses its $393 million of undrawn facility capacity to reach about $900 million of loans at the 1.86% net margins it earned on Q2 2026 originations. That lifts earnings to about $0.95–1.01 per share, which roughly covers the dividend. What would make this wrong. Office losses well beyond the current reserve, a drop in SOFR larger or longer-lasting than the bear case assumes, or repeated share sales at deep discounts to book. The margin-of-safety price is set well below the bear case. It does not depend on the market ever valuing the shares closer to book: from that price, earnings alone of about $0.80 a share imply roughly 14% a year.