16 October 2026 / 8 min read
Treaty underwriting team - Global RE Underwriting
Rates are falling. Nobody is arguing about that.
Almost every outlook published in the run-up to January points the same way. S&P expects property catastrophe pricing to keep softening through 2027, Autonomous Research has talked about reductions of 10% to 15% at 1/1, and Moody’s survey of primary insurers shows more of them expecting cheaper property cover than a year ago. Aon put global reinsurance capital at a record level in its mid-year report.
None of that is surprising. Capital is plentiful, the cat bond market is busy, and 2026 has been kind to reinsurers so far. Pricing is also still well above where it sat at the bottom of the last soft market, so there is room to give. The useful question is how much room, and what gets given away along with the rate.
The structure is what fixed the business
Looking back at 2023 and 2024, higher retentions did more for reinsurer results than price did. Tighter event definitions, clearer named perils and cleaner aggregate wordings stopped a lot of frequency loss from reaching treaty layers at all.
Those gains are easy to lose and slow to win back. We would rather give a fair rate reduction on a well-built layer than hold the rate on a layer that has quietly slipped down into working territory. S&P makes a similar point: the pressure in 2027 is increasingly on terms, not just on price.
Casualty is having a different year
Casualty is not following property down. Social inflation, larger jury awards and reserve strengthening on older accident years are still shaping how much capacity reinsurers want to offer, and ceding commissions are being questioned rather than pushed up.
Cedants who can show claims trend by line, limits profiles and reserve triangles tend to have a much easier conversation than those leading with a headline loss ratio.
What helps a submission at 1/1
Good data is still the biggest thing a cedant controls. Geocoded schedules, secondary modifiers, up-to-date replacement values and an honest explanation of any model adjustments all shorten the discussion and usually improve the outcome.
Our advice is simple: send it early. Programmes that arrive in November with complete information get more attention, and better terms, than those that arrive in the last week of December.