Travelers posted $2.2 billion of second-quarter net income, a 46% jump that mostly reflects what didn’t happen: catastrophe losses of $518 million, roughly half the $927 million recorded a year earlier. The combined ratio improved 6.7 points to 83.6, the kind of underwriting margin that makes a property-casualty portfolio look like a bond fund with better tax treatment. For the first half of 2026, net income has doubled to about $3.9 billion and the combined ratio sits at 86.1, a 10.2-point swing from the same stretch of 2025.
The headline driver is reserve development, not pricing heroics. Favorable prior-year development of $578 million more than covered the quarter’s storm losses, and the business insurance segment alone contributed $319 million of that release, up from $79 million a year ago, thanks to better-than-expected workers’ compensation experience over multiple years and recent commercial property years. That segment’s underwriting gain more than doubled to $728 million on a combined ratio of 86.8, while net premiums written grew 3% to $6 billion. The math is clean: underwriting income rose $716 million to $1.7 billion.
Personal lines tell a sharper story. Six months ago the segment ran a 101.7 combined ratio; now it’s 81.2 at the halfway mark and 79.5 for the quarter. CEO Alan Schnitzer cited solid retention in auto and homeowners plus higher new business in homeowners. Net premiums written of $4.3 billion fell 8% quarter-over-quarter, though last year’s figure included $178 million from Canadian operations divested in the first quarter to Definity Financial for $2.4 billion. Strip that out and the organic trend is flatter than the combined ratio suggests.
The reserve releases are real money, but they are also backward-looking. Workers’ compensation tails are long, and “better than expected” is a phrase that ages poorly when inflation re-accelerates or social inflation reasserts itself. The catastrophe number looks benign only because the comparator was a noisy 2025; a normalized cat load would leave less room for the reserve cushion to flatter the combined ratio. Investors should also note the quiet build-out of an insurance-specific large language model, a signal that expense ratio pressure is next on the agenda.
What matters next is whether the 83-84 combined ratio range holds when the reserve tailwind fades and cat season reverts to mean. Retention and new business in homeowners are encouraging, but the premium base in personal lines is shrinking post-divestiture. If the LLM delivers underwriting expense savings faster than competitors, the combined ratio could structurally reset lower. If not, the first half of 2026 will look like a very good year that borrowed from the past.