Secondary perils such as severe convective storms, wildfire, winter storms and flood now cause most insured catastrophe losses in North America.
But insurers’ reinsurance programs, models and pricing are still designed mainly for rare, extreme events, according to a new report from Marsh Re, the reinsurance and capital business of Marsh.
Marsh Re said the gap between how insurers buy reinsurance and how losses actually happen has become a fundamental strategic challenge, and that closing it is fast becoming a way for insurers to stand out from competitors.
US insured losses from severe convective storms have exceeded $30 billion every year since 2020, and topped $40 billion in four of those years. The January 2025 Los Angeles wildfires produced the largest insured wildfire loss on record. Marsh Re said a season of 15 midsize hail events can do more damage to an insurer’s results than the one-in-100-year hurricane that still anchors many capital models.
The problem has grown since the 2023 reinsurance market reset, which pushed retentions higher. Reinsurers’ share of catastrophe losses has fallen as a result, leaving more of the cost of frequent, smaller events on insurers’ own balance sheets.
That trend has been visible in third-quarter results. Several US carriers, including Hanover and CNA, reported higher catastrophe losses driven largely by convective storms and other weather events, rather than major hurricanes.
Marsh Re said catastrophe models for secondary perils are improving, but validation remains uneven. Its testing found that loss estimates can change significantly between model releases, and that results for the same portfolio can vary widely, particularly in less mature areas such as urban conflagration.
“The insurers that outperform over the next cycle will be the ones that bring the same discipline to hail, wildfire, and winter storm that they apply to capital planning and reinsurance strategy,” said Kim Roberts, North American peril advisory leader at Marsh Re and the report’s author.
The report recommended that insurers combine vendor model output with their own claims experience and validate results before using them to make decisions. It also calls for better mitigation data at the level of individual properties, from roof age and construction to community resilience measures.
Finally, it suggests redesigning reinsurance around the accumulation of losses, through annual aggregate covers, second-event layers and lower-attaching structures. Marsh Re acknowledged these structures remain hard to place profitably for many programs, and said capacity will go to portfolios that are modeled accurately and completely.
For retail brokers, the shift has direct consequences for clients. Insurers carrying more of the cost of hail and wind losses themselves are likely to push those costs through to policyholders, through higher deductibles, wind and hail deductibles as a percentage of value, roof age surcharges and tighter underwriting in hail-prone regions.
The report’s focus on mitigation data also matters. Insurers that use property-level information on roof condition and construction to price risk will reward clients who can show they’ve invested in resilience. Brokers who help clients document roof replacements, impact-resistant materials and other improvements can help them secure better terms.
For reinsurance brokers and cedents heading into January 1 renewals, aggregate and lower-attaching covers will be in demand, but capacity will be selective. Insurers with well-validated models and strong data are likely to secure better structures and pricing.
