California quake: not if, but when

While a late season hurricane is always a possibility it appears that the US has once again escaped the worst of the Atlantic storm season. Only Isaac—which cat modelling agencies estimate will cost between $1.5 billion and $2 billion—has made landfall. While severe thunderstorm and tornado losses in the first half of the year have again been expensive, and a summer-long drought has brought a spike in crop re/insurance claims, 2012 should not be a big year for catastrophes in the US.

Only an earthquake in California or a major terrorism event could change that outlook. With an earthquake in the state overdue, according to earthquake scientists, this remains a real possibility. But what size loss would it take to really impact the industry? With very high insured values in Los Angeles and San Francisco, the potential loss could be $80 billion plus, according to industry experts.

At the start of 2011, Guy Carpenter said it would take a $50 billion loss to make the industry pause for thought, that a $100 billion loss would trouble outliers—distressed reinsurers—while it would take a $150 billion loss to really turn the market. The reinsurance broker still stands by this assessment—which held up well last year, according to David Flandro, managing director and global head of business development at Guy Carpenter.

While last year’s ‘Year of the Cat’ cost the industry in excess of $100 billion, no reinsurers have entered run-off and the most any outlier lost was 35 to 40 percent of its capital. This contrasts markedly with 2005, where some companies lost 50 percent of their capital and three Bermuda reinsurers entered run-off following claims from hurricanesKatrina, Rita and Wilma—namely, Rosemont Re, Alea and Quanta, which were eventually joined by PXRE.

“The losses as a percentage of opening capital in 2011 were lower than in 2005 and part of that is because the sector was going through a liability crisis in 2005,” said Flandro. “In 2011 people were releasing reserves and experiencing investment gains and those two things were the primary drivers of capital growth.”

Position of strength

Since last year’s cat losses 2012 has, so far, been a benign year. With firmer rates on most property catastrophe accounts this has given re/insurers the opportunity to replenish lost capital. At the same time, the influx of capital market capacity continues in the shape of catastrophe bonds, collateralised reinsurance, sidecars and industry loss warranties. With such a strong capital position to start from, what would a Northridge-type event do to the industry today?

“It’s hard to put a fi nger on a precise fi gure that would turn the market in the wake of a California earthquake,” said Robert Hartwig, president and chief economist at the Insurance Information Institute. “The Northridge earthquake, which occurred in 1994, resulted in $19.1 billion in insured losses (in 2011 dollars). It is unlikely that an event of that magnitude would cause general market dislocation today (outside of California-specifi c property and reinsurance markets).

“It is also the case that earthquake risk in California is highly syndicated,” Hartwig continued. “By that I mean it is heavily reinsured and retroceded and there are quite a few cat bond issues on the market today with California seismic risk embedded in them. This suggests that the risk is very well spread on a global scale, reducing the likelihood of a major (dislocating) hit on the capital of US insurers.”

“The California Earthquake Authority [CEA] also has quite a bit of this risk on its book and it is worth noting that deductibles for earthquake coverage are high—about 15 percent—so risks will have significant retentions as well,” he concluded.

Remembering Northridge

California’s exposure to earthquakes is one of the most modelled and well understood perils in the world. But it was not always the case. Eighteen years ago when Northridge struck beneath the highly populated San Fernando Valley it occurred along an unknown blind thrust fault and caused massive upheaval in the industry, with the claims process dragging on for more than five years. With a much higher level of ground motion than expected, the earthquake also damaged many structures previously deemed earthquake-resilient.

Until this point the standard method of managing earthquake risk was to use a probable maximum loss (PML) approach that focused on the accumulation of insured exposure across the whole Los Angeles basin. Eighty percent of the risk within this accumulation zone was assumed to be commercial structures. However, 60 percent of the insured loss from Northridge was from residential lines of business.

Northridge changed how the peril was modelled and insured going forward, explained Dr Jayanta Guin, senior vice president of research and modelling at AIR Worldwide. “In the aftermath of the Northridge earthquake, the entire insurance market changed permanently; many insurers stopped offering earthquake insurance, or offered it only at a restricted level. In response, the CEA was created by the California legislature to make at least minimal earthquake insurance available on a broad scale.

“Also as a result of the earthquake, funding increased to research the threat of earthquakes in the region, and to understand them better from a scientific and engineering perspective. In fact, since Northridge, enormous strides have been made in our understanding of how buildings respond to earthquake ground motion and this has also made its way into the AIR model,” said Guin.

"The risk is very well spread on a global scale, reducing the likelihood of a major (dislocating) hit on the captial of US insurers."

More recently, scientists working in partnership with the US Geological Survey (USGS) revealed some new understanding on how earthquakes from faults such as San Andreas behave. They found that ground shaking from strike-slip earthquakes decreases more rapidly than had previously been thought with distance from the earthquake epicentre. Updates to vendor catastrophe models have incorporated the new data, showing reductions in insured loss estimates for California and much of the Western US.

The changes were less pronounced for Los Angeles, as the city is also subject to thrust earthquakes like Northridge, where ground motion behaviour is better understood. “Catastrophe models today are more sophisticated as a result of the exponential increases in computing power and the availability of ever more data at increasingly high resolution,” said Guin. “These, along with advances in our understanding of earthquakes by leveraging the latest science provided the opportunity to increase the realism of the models. As a result, earthquake models are becoming more detailed and are being used more broadly throughout the industry.”

Exposed residential properties

Guin pointed out that the take-up rate of residential earthquake insurance in California remains low despite the ever-present threat. He cautioned that the economic costs to society will “far outweigh” the insurance costs.

“My biggest concern has always been what it would do to the Californian economy, not just the insurance industry, and more so the knock-on effect of a major earthquake on the economy of the entire US,” agreed John Daum, executive director of Lockton Re.

In the current economic climate it is difficult to see how homeowners could be encouraged to buy more earthquake cover, he added. Fannie Mae and Freddie Mac—two large government-backed mortgage providers— still do not require homeowners to buy earthquake insurance. As a result, he thinks any major event will produce significantly higher commercial claims than on personal lines.

“The largest insurer in the State of California is not State Farm, it’s the banks—they have all that risk on their balance sheet,” said Daum. “One would think they’d start to look more aggressively at de-risking their portfolios. But can you imagine if the Federal government told Freddie and Fannie they would have to force their customers to buy earthquake insurance that is going to cost another $3,000 a year per consumer? Politically it wouldn’t happen.”

Guy Carpenter’s Flandro pointed out that a big earthquake event could impact both sides of re/insurers’ balance sheets. At a time when interest rates are low and investment returns already down, such an incident could tip the balance.

“A California earthquake is not going to be uncorrelated to the asset side of the balance sheet. If California experiences ‘The Big One’, that will affect the economy of California and thence global markets. A large event on the liability side of the balance sheet could also correspond to a big asset-side impairment. It’s clearly a significant risk that needs to be carefully modelled.”