California earthquake insurers may be materially understating their most severe loss scenarios because catastrophe models do not capture "supershear" ruptures, according to new research from MS Amlin.
The Lloyd's insurer found that adding supershear effects to representative insurance and reinsurance portfolios increased losses by 5% to 10% at a 200-year return period. At the more extreme 500-year return period, modeled losses rose by between 30% and 60%.
The findings are particularly significant for California, the world's largest earthquake insurance market, where the San Andreas Fault has the physical characteristics associated with supershear events and where researchers believe past ruptures have displayed the phenomenon.
"I do think California is susceptible to supershear earthquakes," said Luke Wedmore, senior research analyst at MS Amlin, who co-authored the study with research analyst William Sturgeon. "Not only do I think that, but we've observed them there in the past."
A supershear earthquake occurs when a rupture moves along a fault at or above the speed of the shear waves it produces. Wedmore compared the effect to the sonic boom created by a supersonic aircraft, with seismic energy arriving in a concentrated pulse.
The shaking can be more intense and cover a wider area. Buildings may also experience a double pulse from different directions, creating torsional forces that are difficult to engineer against.
"All that seismic energy from your P and S waves arrives at the same time," Wedmore said. "Because you've got much more energy, that is one of the reasons these earthquakes are much more destructive."
MS Amlin's paper, Supershear Earthquakes – An insurance blind spot, published in the Journal of Catastrophe Risk and Resilience, said the events accounted for up to 58% of insured earthquake losses since 2016, equivalent to $13.2 billion. Yet the effects are not represented in seismic hazard models, building codes or catastrophe models used for underwriting and capital decisions.
The researchers pointed to the magnitude 7.7 Myanmar earthquake in 2025, identified as a supershear event, which produced a surface rupture stretching 475km, around 230km longer than standard estimates would have predicted, materially increasing the area exposed to shaking.
California has several conditions believed to support supershear rupture. The San Andreas is a strike-slip fault, where tectonic plates move horizontally past each other, and much of it is notably straight. The geology also differs on either side, another characteristic associated with supershear behavior.
The 1906 San Francisco earthquake has subsequently been identified as likely supershear, Wedmore said, while the 1979 Imperial Valley earthquake may also have involved a supershear rupture.
The warning lands against an already substantial baseline of California earthquake exposure. The US Geological Survey's UCERF3 forecast estimates a 60% probability of a magnitude 6.7 earthquake in the Los Angeles region and a 72% probability in the San Francisco Bay Area over 30 years.
Wedmore said California may also be experiencing an unusually quiet period on its major faults. “Some of the science suggests that we are probably in the longest earthquake drought on California’s major faults in the last 1,000 years,” he said. “Regardless of that, we know earthquake risk is very substantial in California, and we know it is likely to experience supershear earthquakes in the future.”
That quiet period could create complacency after more than three decades without a loss on the scale of the 1994 Northridge earthquake. Wedmore noted that California experienced three major earthquakes within roughly 50 years: Fort Tejon in 1857, the 1868 Hayward event and San Francisco in 1906.
“While people are waiting for ‘the big one,’ we’ve been through such a period of quiet activity in California that, when major earthquakes do occur again, they could be much more frequent than we have experienced over the last 100 years or so,” he said.
One major challenge is that supershear behavior remains difficult to predict and computationally demanding to simulate. MS Amlin instead used a sensitivity approach, applying observations that around 36% of major strike-slip earthquakes globally since 2010 have involved supershear rupture.
Wedmore said the issue should be treated primarily as a tail-risk consideration rather than a direct pricing instruction. Insurers can begin by stress-testing portfolios, identifying long strike-slip faults capable of supershear rupture and testing alternative shaking patterns.
"The purpose here is not necessarily to influence pricing, because there are lots of other factors involved in pricing, but this is a tail risk," he said. "We suggest that earthquake insurance companies should consider it as a factor in their tail-risk assessments."
MS Amlin has already updated its own catastrophe models and view of risk to incorporate supershear effects, Wedmore said, and he urged the wider industry to follow before the next generation of US earthquake models is finalized.
With major model vendors preparing to update US earthquake models following revisions to the national seismic hazard framework, the paper argues a narrow window is open for the industry to close the gap, which will require collaboration between scientists, engineers, risk practitioners and the (re)insurance industry to produce practical solutions and guidance.
