Earthquake cat models have a structural blind spot: they systematically exclude the class of rupture that accounts for 66% of all insured earthquake losses since 2016 — approximately 13.2 billion USD in total. According to a peer-reviewed paper by MS Amlin researchers Luke Wedmore and William Sturgeon, supershear earthquake ruptures — in which a fault propagates faster than the S-wave velocity, generating a seismic Mach cone — are absent from seismic hazard models, building design codes, and insurance catastrophe models globally. Every capital number, every cat bond pricing sheet, and every reinsurance treaty rate that flows from current earthquake models is calibrated against a hazard function that does not capture this phenomenon.
What Supershear Ruptures Do — and Why Models Miss Them
When fault rupture velocity exceeds the shear wave speed of the surrounding rock — typically around 3.5 km/s — the energy release pattern changes qualitatively. Rather than spreading evenly from the rupture front, shaking is concentrated into a Mach cone propagating ahead of the rupture tip, similar in physics to a supersonic shock wave. Structures in the Mach cone path experience shaking intensities that standard ground-motion prediction equations (GMPEs) — calibrated on sub-shear events — significantly underestimate. The result: higher-than-modeled damage ratios, wider geographic footprints, and rupture lengths that exceed magnitude-scaling expectations.
The phenomenon is not rare. Around 36% of major strike-slip earthquakes globally since 2010 have involved supershear rupture, according to MS Amlin’s analysis. The affected events include the 2016 Kaikoura (New Zealand), 2018 Palu (Indonesia), and 2025 Myanmar earthquakes — collectively responsible for the majority of the 13.2 billion USD in insured supershear losses recorded since 2016. Yet no commercially deployed cat model from RMS, AIR, or Verisk explicitly incorporates supershear hazard into its seismic event set or ground-motion modules.
Myanmar 2025: A Real-World Validation of the Gap
The March 2025 Mw7.7 Myanmar earthquake provided an almost forensic validation of MS Amlin’s analysis. According to USGS research, the event produced a 475km surface rupture along the Sagaing Fault — approximately 230km longer than standard magnitude-scaling relationships predict. The rupture propagated at supershear velocities greater than 5 km/s — well above the ~3.5 km/s S-wave speed that marks the supershear threshold. Standard PML models, which use magnitude-based rupture length scaling, would have clipped this event at a fraction of its actual footprint.
The broader historical record reinforces the concern. Swiss Re’s ongoing tracking of insured catastrophe losses shows earthquake events consistently outperforming model expectations. MS Amlin’s paper identifies the 1906 San Francisco earthquake (Mw7.9) as a retroactively classified supershear event — a modern analog for California tail risk that suggests atypically long, fast ruptures represent a recurring, not freak, hazard pattern on the San Andreas Fault. Luke Wedmore warns explicitly that “there is a significant chance that earthquake risk in California is markedly underestimated” given the supershear dynamics present on the San Andreas system.
The Capital Implication: Understated Losses at the 500-Year Return Period
The modelling gap translates directly into capital adequacy risk. MS Amlin’s portfolio analysis found that incorporating supershear effects increases losses at the 500-year return period by 30–60% — the precise level used for Solvency II SCR calibration and ICS stress testing. At the 200-year return period — the standard reinsurance attachment reference — losses increase by 5–10%. In extreme scenarios on specific fault segments, earthquake risk may be underestimated by up to 60%.
The ILS market faces a particularly direct exposure. According to MS Amlin’s analysis, over 6 billion USD in purely US earthquake-exposed catastrophe bonds are outstanding — pricing seismic risk via models that omit supershear. Over 17.6 billion USD in US multi-peril cat bonds carry material California earthquake exposure. Attachment probabilities and expected loss calculations for these instruments may be systematically understated at the 500-year return period. Palomar’s earthquake reinsurance tower at the June 2026 renewal and the broader California earthquake reinsurance market have been priced in this environment.
Who Needs to Act: Cat Modelers, Reinsurers, and Regulators
The MS Amlin white paper, published in the Journal of Catastrophe Risk and Resilience, is authored by Luke Wedmore and William Sturgeon and formally titled “Supershear Earthquakes, An insurance blind spot”. Its practical recommendations centre on three categories of action.
For cat model vendors, the paper calls for integrating supershear ground-motion amplification factors into GMPE modules for strike-slip fault scenarios, and expanding fault source characterisation databases to flag supershear-susceptible segments by slip rate and geometry. For ILS investors and reinsurers, it recommends demanding model documentation that explicitly addresses supershear treatment before committing capital to earthquake-exposed portfolios. For regulators, the implication is that ORSA and ICS stress tests for earthquake-exposed (re)insurers should include a supershear scenario overlay until commercial models demonstrably incorporate the hazard — a disclosure requirement consistent with the model risk management frameworks already operating under Solvency II internal model approval.
