CNA FINANCIAL CORP (CNA) stock intrinsic value fair value estimate & owner earnings analysis

CNA Financial is a Chicago-based commercial insurer, about 92% owned by Loews Corporation. Its three underwriting segments are Specialty (management and professional liability, surety and related coverages sold through brokers and agents), Commercial (property and casualty cover for small business, construction and middle-market customers) and International (Canada, the U.K., Luxembourg and the Hardy Lloyd's syndicate). Two other pieces sit outside the core: a long-term care book that is in run-off, and legacy asbestos, environmental and mass-tort liabilities in Corporate & Other. Revenue grew from about $9.4 billion in 2016 to $15.0 billion in 2025, and net income was $1,278 million in 2025. How we measured earnings. An insurer's operating cash flow is flattered whenever its reserves for unpaid claims grow, because that money is owed to claimants and is not profit. We therefore removed the yearly changes in insurance reserves, receivables, accrued investment income, deferred acquisition costs and the filing's "other, net" line (the filing does not itemize that last line, so we treated all of it as working capital). What is left is net income plus genuine non-cash items, less the equipment and software spending needed to stay in business (about $81 million in 2025). Stock-based compensation (about $41 million in 2025) is not added back, so it remains a cost. Interest on CNA's roughly $3.0 billion of senior notes is an ordinary operating cost for an insurer and is already deducted in earnings, so no separate debt is subtracted; the values below are for the shares directly. The notes mature between 2027 and 2035 with no single large wall, and in 2026 the company replaced its 4.50% notes due March 2026 with a 5.20% issue due 2035. The record. On this basis earnings per share were $3.98 (2016), $3.73, $4.00, $3.22, $2.39 (2020, hit by the pandemic and long-term care reserve pressure), $3.50, $3.37, $4.04, $2.47 (2024) and $4.33 (2025). 2024 was depressed by a $293 million after-tax non-cash pension settlement loss, so we left that year out when setting the normal level. The median of the other nine years is $3.73 per share, roughly $1.0 billion a year, and that is the level used. Eight of the ten years fall between $3.22 and $4.34. Growth. Earnings per share rose about 76% in 2025, but that was a rebound from the pension-depressed 2024 (against 2023 the rise was about 7%). We assume 0% growth in year 1, below what was just reported. Revenue has grown roughly 5% a year, yet earnings per share have moved only from $3.98 to $4.33 over nine years and nearly all earnings are paid out as dividends (about $1.05 billion in 2025), so we credit no growth: rate increases, share gains and higher reinvestment yields are not locked in, and we did not count them. Base case: $62.17 per share. Earnings of $3.73 per share (about $1.0 billion) continue indefinitely with no growth. The named support is a franchise that has produced between $3.22 and $4.34 per share in eight of the last ten years, through hard and soft insurance markets. Business value equals equity value here (no separate debt deduction), or about 1.5 times book value of roughly $41 per share. Bear case: $40.00 per share. Earnings drop to the 2020 trough of $2.40 per share (about $650 million) and stay there. The named risks are all visible in the filing: unfavorable prior-year reserve development of $185 million in 2025 (versus $48 million in 2024), a rising loss ratio in lines affected by social inflation, catastrophe losses of $240 million in 2025 and $358 million in 2024, legacy asbestos and mass-tort strengthening in Corporate & Other, and the auditors' flag on the judgment behind the long-term care reserves. The result is about equal to book value. The debt is carried inside these earnings; nothing is left over after lenders in this case beyond that, so the difference between business value and debt is not separately positive or negative. Bull case: $72.33 per share. Earnings hold at the 2025 level of $4.34 per share (about $1.18 billion), which is the best year of the decade and was reached despite the 2025 reserve charges and catastrophe losses. Consistent with the rest, no further growth is added. What we tested. We challenged the base case with the 2025 report: unfavorable reserve development and a higher underlying loss ratio in some commercial lines did not stop 2025 earnings from coming in 16% above our normal level. First-half 2026 net income of $532 million (second quarter $321 million versus $299 million a year earlier) compares with $573 million in the first half of 2025, which is in line with our normal level. What changed since the prior analysis. We previously estimated $23.80 / $50.80 / $82.40 using higher discount rates and 2-3% growth assumptions. We now discount every case at 6%, use a stricter earnings measure that strips out reserve growth, and credit no growth that is not already secured. The bear and base cases rise, the bull case falls; this is a change in method as much as in the business. What would make this wrong. A large reserve charge in the long-term care or legacy liability books, a prolonged soft insurance market or repeated catastrophe years would push earnings below our bear level. A hard market that lifts underwriting margins durably would justify more than the bull case. Margin-of-safety price: $35.00. Insurance reserves are estimates that can move sharply, results swing with catastrophes and the pricing cycle, and the bear case sits near book value. The margin-of-safety price is about 44% below the base case and 12% below the bear case, which corresponds to an earnings yield of roughly 10.7% on the normal level.