US Severe Convective Storm Losses Top $35B as the August Derecho Outruns 2026 Cat Budgets

US severe convective storms have generated more than $35 billion in insured losses through August 14, 2026, after an early-August derecho tore across the Chicago metro and the Ohio Valley (Gallagher Re, August 2026). That already ranks as the sixth-most-expensive SCS year on record, built from at least six separate multi-billion-dollar outbreaks rather than one storm that trips a single reinsurance layer.

Six weeks earlier, the number was $22 billion. Gallagher Re's mid-year read, published June 18, put year-to-date US severe convective storm insured losses at $22 billion, tracking behind the elevated pace of 2023 through 2025 (Gallagher Re, June 2026). Aon's own H1 2026 tally, released around the same window, put US SCS insured losses at roughly $27 billion for the first half, notably below the $40 billion-plus recorded in the first halves of 2023, 2024, and 2025 (Aon, H1 2026 Global Catastrophe Recap). Both brokers described a season that looked, at the halfway mark, like a below-trend year for the peril. Then a late-June stretch and an August 9-12 outbreak, capped by a high-end derecho on August 11, added roughly $8 to $13 billion in five weeks and pushed the total past $35 billion. The lesson for anyone budgeting to an SCS peril in Q1 is that a quiet mid-year read is not a signal to release reserve capacity; it is a signal that the loss has not arrived yet.

A Quiet First Half, Then Six Weeks of Acceleration

The August derecho is the event that reset the year. Gallagher Re's report describes "a high-end derecho event" that produced "significant damage across the Chicago (IL) metro region and other notable population centers in Indiana, Ohio, and Kentucky" (Gallagher Re, August 2026). Winds approached 90 miles per hour through Chicago's south suburbs and northwest Indiana, and the outbreak spawned 11 tornadoes, including an EF-2 with peak winds of 115 miles per hour that tracked more than 25 miles across the Illinois-Indiana border (Insurance Business, August 2026). More than 1.1 million customers lost power. Illinois alone has recorded more than 200 confirmed tornadoes year to date, surpassing its full-year 2024 count of 142, a threshold the state had already blown past by mid-year according to Aon, which noted Illinois running at roughly four times its long-term tornado average and both Illinois and Indiana setting new annual records before the August system even formed (Aon, H1 2026 Global Catastrophe Recap).

Guy Carpenter's early read put the August event alone among the costliest single SCS outbreaks the market has priced in three decades: the outbreak "could rank among the ten costliest US severe convective storm events since 1995," a list whose tenth-ranked entry, a 2014 event, cost the industry $5.1 billion in insured losses (Guy Carpenter, via Insurance Business, August 2026). Gallagher Re's broader count for the year now stands at six separate multi-billion-dollar SCS outbreaks in 2026, a pace that trails only 2023 (11), 2024 (10), and 2025 (7) for the same metric (Gallagher Re, August 2026). Set against the full-year totals those three years eventually produced, $63 billion in 2023 and $64 billion in 2024 (Insurance Journal, July 2024), and $51 billion in 2025, the third-costliest year on record after those two (Swiss Re Institute, sigma 1/2026), the 2021-2025 five-year average for the peril lands at roughly $50 billion. Gallagher Re also estimates that once uninsured and underinsured property is counted, the August outbreak's total economic cost runs at least 25% above its insured figure, a multiplier consistent with how the broker has scaled economic-to-insured loss ratios for SCS events all year (Gallagher Re, August 2026).

Why an Accumulating Peril Strains a Cat Budget Differently Than a Named Storm

A hurricane makes landfall once, generates one occurrence date, and either pierces a carrier's per-occurrence catastrophe reinsurance attachment or it does not. Severe convective storm losses do not work that way. They arrive as a sequence of discrete, geographically scattered events, an April Plains outbreak, a June Midwest derecho, an August Ohio Valley system, each one individually well below the size that would trigger a per-occurrence treaty recovery, but collectively large enough to erode an aggregate deductible or an annual aggregate limit over the course of a year. That structural difference is why a peril that generated $35 billion in industry losses through mid-August can still leave most of that total sitting on primary balance sheets rather than flowing through to reinsurers.

Moody's Ratings has been explicit about where that leaves the retention math: "insurers will continue to retain a large proportion of losses from secondary perils," a function of reinsurers holding attachment points at levels calibrated to keep frequency-driven perils like SCS largely within the cedant's net retention (Moody's Ratings, 2026). Aggregate covers are the mechanism most directly exposed to this dynamic, and 2026 has already produced a live example of the interaction between one SCS season and the next: some aggregate catastrophe bond transactions covering the SCS peril saw their attachment levels eroded by winter storm losses earlier in the year, meaning the August derecho landed against aggregate layers that had less remaining capacity than they would have in a clean year (Artemis, August 2026). That is the actuarial argument for why frequency erosion, not any single event severity, is the variable primary cat budgets most need to track through the back half of a season: an aggregate layer with $200 million of remaining capacity in January can arrive at August with a fraction of that cushion left, purely from a string of sub-threshold events that never individually looked large enough to flag.

Reinsurers responded to that dynamic at the mid-year renewal by restructuring how aggregate capacity gets priced rather than simply raising rate. Gallagher Re's own July 2026 renewal report described 20% to 25% property catastrophe rate reductions alongside a more consequential structural shift: aggregate, multi-year, and multi-line covers are now pricing at attachment points cedants can actually use, a departure from the punitively high aggregate attachments reinsurers demanded after 2023's SCS-driven losses (see actuary.info's coverage of Gallagher Re's July 2026 First View). That repricing makes the 2026 season a live test of whether the new attachment structure holds up against a run rate that, by mid-August, is already tracking toward the top of the five-year range.

Net of Reinsurance: Where the $35 Billion Actually Lands

Individual SCS events in 2026 have mostly landed in the $1 billion to $5 billion range, well under the per-occurrence retentions many large national carriers carry on their primary catastrophe treaties. That means the bulk of the $35 billion is not flowing through occurrence-triggered reinsurance recoveries at all; it is accumulating claim by claim against net retention, quarter after quarter, until it shows up as a drag on underwriting income rather than a single reported cat-loss event. Allstate's own March 2026 disclosure illustrated the mechanism directly: $925 million in March catastrophe losses alone, part of $1.24 billion in Q1 2026 cat losses spread across 15 separate wind and hail events, none of which was large enough on its own to be the story, but which cumulatively drove the quarter (see actuary.info's analysis of Allstate's Q1 SCS signal). That is the retention profile a frequency-driven secondary peril produces across the industry: fifteen sub-events adding up to a number that matters, with no single occurrence large enough to have been individually reinsured.

Public entities with capital-markets-heavy cat towers show the other side of that structure. Florida Citizens renewed its $2.82 billion catastrophe tower in June 2026 at a 9.52% weighted-average rate on line, with 75.4% of the tower placed in catastrophe bonds and 24.6% in traditional reinsurance (see actuary.info's breakdown of Florida Citizens' 2026 capital stack), a structure built primarily around hurricane peak-zone risk rather than SCS frequency. That contrast is instructive: a program engineered around one dominant peak peril can transfer the bulk of its severity to capital markets at a defined attachment point, while a program exposed to a frequency peril like SCS has no equivalent single attachment to design around, because the loss accumulates across dozens of independent occurrences rather than concentrating in one. Cat bond investors have shown continued appetite for the asset class regardless: Q2 2026 issuance hit a record $11.3 billion and secondary spreads compressed to 5.71% by late June, breaking the usual seasonal widening pattern (see actuary.info's coverage of the Q2 2026 cat bond market). That demand is what let reinsurers lower aggregate attachment points at mid-year without needing to charge materially more for the capacity, a bet the August derecho is now testing in real time.

Hail and Wind Claims Develop Slower Than the Headline Suggests

The $35 billion figure itself is a preliminary industry estimate, and SCS claims, hail claims in particular, are among the slowest-developing property lines to move from initial reserve to final paid. A hailstorm does not produce visible roof damage the way a tornado produces visible structural damage; homeowners frequently do not notice or report hail damage for weeks or months, and once they do, claims queue behind a backlog of roofing contractors and independent adjusters working the same geography at the same time. That lag between occurrence date and reported date, and then between reported date and adjusted, paid amount, is exactly the mechanism that produces upward reserve development on hail-heavy accident quarters long after the initial cat-loss estimate has been published and moved on to the next headline.

The scale of exposure sitting behind that development risk has grown materially. Cotality's 2026 Severe Convective Storm Risk Report found more than 43.5 million US properties now carry moderate or greater risk from damaging hail, representing roughly $17.84 trillion in reconstruction cost value (Cotality, 2026 Severe Convective Storm Risk Report). The US recorded 142 days with damaging hail in 2025, seven more than 2024 and well above the 20-year average of 122 days, and more than 235,000 Texas homes experienced damaging hail that year alone, more than any other state (Cotality, 2026). Cotality's modeling puts the insured-loss potential of a single severe hail event, the kind that occurs roughly once every few decades, at nearly $30 billion, a figure the firm frames as comparable to a major hurricane landfall. That framing matters for reserving discipline specifically because it means a carrier's hail book can now produce a single-event tail loss of hurricane scale without a hurricane ever forming, and the claims from that event will develop on the slower, contractor-availability-driven timeline hail claims always follow rather than the faster timeline typical of wind or surge damage.

For reserving actuaries, the practical implication is that 2026's $35 billion headline should be treated as a floor, not a point estimate. Loss development factors selected against pre-2020 hail and wind accident years will understate the tail on 2026 accident-quarter claims if contractor capacity in the hardest-hit metros, greater Chicago and the Ohio Valley corridor most recently, remains constrained relative to the volume of claims filed. IBNR held against the second half of 2026 should reflect that a derecho landing in a dense metro area produces a claims count large enough to outstrip local adjusting and repair capacity for months, not weeks, extending the tail on what would otherwise be a routine wind-and-hail loss.

Rate Adequacy in the Most Exposed States

The states absorbing the bulk of 2026's SCS activity are also states where homeowners rate filings have been running hot for other reasons. Texas, the state Cotality flagged for the most hail-affected homes in 2025, has already seen carriers push through some of the largest homeowners rate increases in the country this year (see actuary.info's coverage of Texas's Q1 2026 homeowners rate hikes), and an SCS-heavy 2026 gives filed loss-cost trend even less room to be read as conservative. Illinois, which entered 2026 without the same reputation for hail and tornado severity that Texas, Oklahoma, and Nebraska carry, now has a state-record tornado count and a share of the derecho damage large enough to reset how carriers weight the state in a countrywide SCS loss-cost model. A single confirmed year of elevated frequency does not necessarily change a long-run trend selection on its own, but three or four consecutive years of Illinois running above its historical tornado average would, and 2026 is building that track record in real time.

The broader signal for primary carriers and reinsurers alike is that a $50 billion five-year average for US SCS losses is itself a moving target, not a stable anchor. Swiss Re's own sigma research found that the number of SCS events causing average annual losses of $1 billion or more was 59% higher in the five years to 2025 than in the five years to 2020, evidence that the frequency shift behind 2026's run rate is not a single anomalous season but a continuation of a trend the broker has now tracked across two consecutive five-year windows (Swiss Re Institute, sigma 1/2026). Cat budgets, aggregate reinsurance structures, and hail-heavy state rate filings that are still anchored to a $50 billion five-year average risk understating the peril precisely because that average is itself climbing, event by event, the same way 2026's own total climbed from $22 billion to $35 billion in six weeks.