ILS boom reshapes reinsurance amid surge in secondary perils

The gist

Insurance-linked securities are booming, transforming reinsurance from a disaster backstop into a data-driven shield against relentless, costly secondary perils.

What to know

2017: The Catalyst Year

The shock of 2017–2018 catastrophe losses forced a total overhaul of ILS and reinsurance structures, turning short-term fixes into long-term strategic capital integration.

The October–November 2026 push toward record ILS growth and tighter alignment with insurer capital models did not emerge suddenly; it was a reset that began when catastrophe losses in 2017–2018 exposed structural weaknesses in catastrophe sidecars and collateralized reinsurance. The 2017 hurricanes Harvey, Irma, and Maria, followed by major 2018 losses, revealed trapped collateral, reserve uncertainty, and investor losses in catastrophe sidecars and collateralized reinsurance, forcing market participants to revise contract terms and collateral provisions and move from simply adding capacity toward more strategic integration with insurer programs.

That redesign gained force in the 2020–2023 hard market, when traditional reinsurance capacity tightened and prices rose due to prior catastrophe losses, inflation, and frequent non-peak events, accelerating the use of catastrophe bonds, collateralized reinsurance, sidecars, and ILS funds as complementary capacity. By 2024, third-party reinsurance capital had reached about $107 billion, giving the market a large installed base before the October–November 2026 reports of record growth, while the same run of frequent non-peak events also underscored why reinsurance structures would need to adapt to changing loss patterns.

ILS Moves Center Stage

Insurers now treat ILS as an essential pillar of capital strategy, driving broader risk coverage and more sophisticated contract terms as alternative capital outpaces traditional reinsurance.

What changed in 2026 was not simply the size of insurance-linked securities, but their role inside insurers’ capital planning. Jason Bolding, CEO of Gallagher Securities, told Artemis.bm that “the question is no longer whether to engage” with ILS and non-traditional capital, but “how to bring it into your capital model effectively,” framing catastrophe bonds, sidecars and other structures as a “critical extension” of traditional reinsurance; Aon’s report that “ILS Capital Hits Record $144.5B” while catastrophe bond issuance rose 15% to $24.9 billion shows the scale now available to make that integration practical.

That capital is also reshaping what reinsurance can cover and how precisely it must be structured. Fitch said the “growth of the alternative capital and ILS sector is likely to continue next year” and expects “strong alternative reinsurance capital market growth and attractive risk-adjusted returns into 2027,” after “Capital levels in the insurance-linked securities market reached new highs in 2026 so far,” supported by catastrophe bond issuance and sidecar activity; as demand expands into non-peak risks, Fitch expects “higher limits, broader event definitions, longer hours clauses and increased aggregate covers,” while Artemis London 2026 stressed rigorous modelling for secondary perils.

Sources

Secondary Perils Break the Mold

Frequent, midsize disasters are overwhelming legacy reinsurance models built for rare events, exposing a critical need to redesign coverage for relentless, earnings-eroding losses.

Reinsurance was built to absorb exceptional shocks, and that legacy still shapes program design. Fortune describes it as the system’s “shock absorber,” the backstop that let insurers survive peak events from Hurricane Katrina to 9/11, when reinsurers covered 64% of the damage, and even the 1906 San Francisco earthquake, after which insurers paid more than 100,000 claims; but that peak-catastrophe architecture matters less if losses now arrive not as one historic blow, but as a steady stream of smaller disasters that erode earnings and capital year after year.

That is the mismatch Marsh Re says insurers now have to fix: reinsurance programs still prioritize the rare “big one” even though secondary perils are producing the bulk of catastrophe losses. According to Artemis.bm’s coverage of Marsh Re, “US-insured severe convective storm losses have exceeded $30 billion a year every year since 2020, and topped $40 billion in four of those years,” while “a season of 15 midsize hail events can do more damage to an insurer’s income statement than the one-in-100-year hurricane that still anchors many capital models”; “Importantly, this problem has heavily grown since the 2023 reinsurance market reset, which pushed retentions higher,” reducing reinsurers’ share and leaving more frequent losses on primary carriers.

Sources

AI Transforms Risk Decisions

Advanced analytics and AI-powered tools are turning vast, messy insurance data into actionable insights, enabling proactive resilience planning—but industry adaptation remains slow.

AI is making underwriting more forward-looking because insurers can now combine richer exposure data with tools that surface risk signals humans would miss. Adrian Hall told Risk & Insurance that insured natural catastrophe losses have exceeded USD 100 billion for six consecutive years, a “long-term trend rather than a temporary anomaly,” and Gasco similarly said “The reality is that the underlying risks businesses face are not softening,” noting that “The baseline for severe weather has shifted, with average annual losses over the past 10 years higher than in the” prior period. Hall argued that geospatial analytics, catastrophe models, and digital risk assessment let clients evaluate future climate scenarios, infrastructure vulnerabilities, and operational dependencies before losses occur, shifting the insurer’s role toward resilience planning and earlier intervention.

That shift depends on turning messy insurance information into usable underwriting inputs and then linking them to pricing and risk-engineering decisions. John Kellington said IDP is “the gift from the heavens” for the 80 to 90% of insurance data that is unstructured, while InsTech noted that “we use mathematical tools that are over 100 years old” but only now have the execution power to apply them at scale; he added that “nowadays we see only few startups trying to address that part of the value chain, AI first insurtech startups,” but after “10, 12 years of activity” there are “more startup on that space,” concluding the market is filling in. Yet Business Insurance reported Marsh’s survey found just 26% of organizations saw adaptation efforts recognized in negotiations, even as Prologis extended roof lives from 20 to 30 years and spent millions retrofitting earthquake-exposed properties.

Sources

Get the stories behind the trends

Deep-dive reporting and the weekly brief, in your inbox.