UNIVERSAL INSURANCE HOLDINGS, INC. (UVE) stock intrinsic value fair value estimate & owner earnings analysis

Universal Insurance Holdings sells homeowners insurance, mainly in Florida, through two insurers, Universal Property & Casualty Insurance Company and American Platinum Property and Casualty Insurance Company. It also runs its own claims adjusting and distribution businesses. At the end of 2025 it had 895,927 policies in force across 19 states and $2.15 billion of premium in force, 72.7% of it in Florida. North Carolina, Georgia, South Carolina, Massachusetts and Alabama are the next largest states. By mid-2026 policies had grown to 934,371. Almost all of the growth was outside Florida: in the second quarter, written premium rose 14.4% in other states and 0.8% in Florida. Three things drive the economics: how much premium the company can charge per home, how much of that premium it pays reinsurers each June 1 for hurricane protection, and how large and how heavily litigated its claims turn out to be. The company gives away a large share of its premium, about 31% in the first half of 2026. That makes the reinsurance renewal the most important event of each year. For 2026-2027 the company bought a $2.623 billion first-event tower, about $50 million larger than last year, with the same $45 million retention. $352 million of that coverage is already locked in through the 2027-2028 season. Excluding reinstatement premiums, reinsurance cost as a share of direct earned premium rose 10.9% in 2025. We measure earnings as the cash the business produces for its owners. We start with operating cash flow and take out the swings in reinsurance balances, loss reserves, unearned premium and receivables, because that cash belongs to future claims and is not earnings. We also deduct stock compensation and the small amount of capital spending needed to maintain the business. Gains and losses on selling investments are excluded. Every year reconciles exactly to the reported cash flow statement. On that basis the last ten years produced, per share: $2.89 (2016), $2.93 (2017, Hurricane Irma), $3.57 (2018), $1.01 (2019), negative $1.49 (2020), $0.82 (2021), negative $0.27 (2022, Hurricane Ian), $2.42 (2023), $1.85 (2024, Debby, Helene and Milton) and $6.46 (2025, a quiet season). The ten-year median is about $2.14. The years from 2019 to 2022 were shaped by Florida's litigation crisis. Assignment-of-benefits abuse and one-way attorney fees inflated claim costs during that time. Senate Bill 2-A (December 2022) and House Bill 837 (March 2023) ended both practices. Since then the company has earned $2.42, $1.85 and $6.46 per share, and the two middle years carried heavy hurricane losses. We therefore judge that today's business earns about $3.00 per share in a year with a normal share of storms. That is above the ten-year median, because the old legal regime is gone, and well below 2025, because 2025 had almost no storms. The first half of 2026 supports that level and does not raise it. Net income was $113 million against $77 million a year earlier, and the net loss ratio improved from 71.4% to 64.4%. Book value per share rose to $22.89. However, the first half of the year falls outside the main hurricane season. On or about October 10, 2026, Hurricane Isaias came ashore in Florida as a Category 2 storm and crossed the Panhandle toward Alabama and Georgia. The company has not yet disclosed its losses from that storm. On the current program, its own share of a first event is capped near the $45 million retention plus reinstatement costs, which is roughly $1.20 per share after tax. Because this is an insurer, borrowing that funds policyholder obligations is part of the business. We value earnings after interest and do not subtract debt separately. The holding company has one bond. In June 2026 it repaid its $100 million of 5.625% notes due November 2026 with $100 million of new 7.75% notes due June 2031. The refinancing went through and the debt is small relative to $637 million of equity, but it costs about $2 million a year more in interest. The regulated insurers could not pay ordinary dividends to the parent in 2026, so the dividend and buybacks are funded from the holding company's own cash and fee income. Growth credited: none from market expansion or price. Policies outside Florida have grown steadily for years, and we credit that only as a slow 2% a year rise in per-share earnings, helped by the buybacks the company has run every year. We give no credit to further reinsurance price declines, rate increases or a quieter climate. The base case starts at about $5 per share in 2026, reflecting a strong first half less a Category 2 landfall inside the retention. From 2027 it uses $3.00 in a normal year, growing 2% a year. That matches the trailing rate in direction: owner earnings grew sharply in 2025, and 2026 is assumed to fall from $6.46 rather than grow. The earnings continue for 30 years with no growth after year ten, then stop. That values the shares at about $48.80. The bear case is built from what Florida homeowners insurers have actually been through, and the industry's own record supplies it. It assumes a major landfall on a densely insured coast within two years, as Ian was in 2022 and Irma in 2017. That year's earnings are consumed through the retention and reinstatements, and the reinsurance renewal after it costs 30-50% more, as it did after Ian. It also assumes litigation partly returns, through bad-faith suits or a narrowing of the 2023 reforms, so earnings repeat the 2019-2022 pattern of three to four years near or below zero. After that, earnings recover only to about $1.50 per share. On that path the shares are worth about $13.30. If capital fell short and the company had to raise equity at trough prices, the value would be about $11.50. Even this case leaves value below the current $22.89 book value per share, which is what we would expect if a Florida insurer's capital were impaired by a cycle of this kind. The bull case assumes the reforms hold and the market stays calmer. It assumes reinsurance costs ease from the 2025 peak. The out-of-state book keeps growing at its recent pace and spreads the hurricane risk. On that path normal-year earnings run near $5 per share, close to the post-reform average excluding the worst storm year, and grow 3.5% a year. That is worth about $87. We are only moderately confident in the earnings level. Ten years of history run from negative $1.49 to positive $6.46 per share, and only two of those years were free of significant storms. The worst plausible outcome is a single very large Miami-Dade or Tampa Bay hurricane that exhausts the $2.6 billion tower, or a second large storm in the same season, alongside a legislative reversal. The capital structure would very likely survive that through dilution rather than failure, since debt is just $100 million against $637 million of equity. Shareholders would still lose much of their stake. The three things we are most likely to be wrong about are reinsurance pricing, litigation trends and storm frequency. If all three go against us, the shares are worth roughly the bear figure in the low teens. Over the holding period, the company pays $0.64 a year in dividends and keeps buying back shares. The price assumes a holding period of three to five years. We set the margin of safety price at $24. That price is about book value, $22.89 in June 2026, for a business that has earned more than $3 per share in a normal post-reform year. If earnings settle near $3.00 and the business is valued like the base case within about five years, plus dividends, the base case implies a return above 20% a year. If a major storm lands first, the price is still close to the company's tangible capital, so the bear case implies a limited loss unless it plays out in full. We set the price no higher because the first half of 2026 does not yet include the main storm season, and the cost of Isaias has not been disclosed. Compared with our previous review, the business has changed in three ways. The 2026 notes were refinanced at a higher coupon. The 2026-2027 reinsurance program was placed with a slightly larger tower and an unchanged retention. And the first Florida hurricane landfall of 2026 has now occurred. We also brought our normal-year earnings estimate closer to the post-reform record and away from the older ten-year median. As a result, the base value is slightly lower, the bear case is somewhat lower because it now includes several years of litigation-era losses, and the bull case is lower because we no longer extend 2025's quiet-year results.