Carriers need to raise their California earthquake rates to account for the many risks that their risk models do not yet account for, argue specialists.
John Charman, the outgoing president and chief executive of Bermudian insurance group Axis Capital, said in a conference call earlier in February that insurers are only able to get about 70% of the premium they need to write earthquake risk in the region comfortably.
“For primary business, California earthquake exposed accounts have not moved sufficiently for us yet,” he said. “It’s around two thirds of where it needs to be in our view.”
Siamak Daneshvaran, director of engineering in the impact forecasting division of broker Aon Re, agrees. He predicts insurers will be forced to increase their earthquake prices by about 50% over the next four to five years, as models become more accurate at helping clients estimate their exposures. The increase may happen sooner, he adds, if a big event triggers capacity shortages.
Part of the reason for the deficiencies in the models is the limited data scientists have to make assumptions with. Earthquakes, particularly ones of great magnitude, are rare. Catastrophe analysts need to look back thousands, not just hundreds, of years to build an accurate picture of the seismicity, or probability of occurrence, in a given location.
Sarah Jenny, a geophysicist in the catastrophe research team of Bermudian reinsurer PartnerRe, says recent events have helped scientists improve their understanding of how different soil conditions and ground motion can amplify seismic waves. “Peak ground acceleration in soft soil reaches up to 20 times the value measured on firm rock,” she says.
But there is still a lot scientists have not yet been able to model, which could leave insurers vulnerable to a big loss. According to Peter Nakada, managing director of the consulting arm of catastrophe modeller Risk Management Solutions, these unmodelled risks include: flood following earthquake and related water containment issues; subsequent toxicity and environmental implications; and the absence of earthquake defences at local police stations, potentially increasing the possibility of riots or looting. Jayanta Guin, v.p. of research and modelling at rival model vendor AIR Worldwide, would add tsunamis to that list because research suggests they are often correlated with earthquakes.
Underestimation of the loss potential is not the only reason earthquake rates are insufficient. Consumer demand for earthquake cover is also low for such a catastrophe prone region. Daneshvaran estimates that the take up rate among Californian homeowners is between 20% and 30%.
“Lots of people are not buying ‘quake insurance because they think the price is already too high,” he explains. “They don’t like the minimum 5% deductible on their house, because they think any loss will be below that deductible. There hasn’t been an earthquake since 1994. And so far there is no regulation requiring residential building owners to buy ‘quake cover for their mortgage, in the same way they might have to buy wind cover.”
California is due a large earthquake, somewhere in the region of Los Angeles or San Francisco, according to Jenny at PartnerRe. The New Madrid fault line is also due a loss, though it is unclear from historical records whether it would be the sort of loss that insurers could pay out of their earnings or whether it would eat into their capital.