Quake Conference Fails to Produce Consensus on Risk

Earthquake scientists at a conference in Pasadena last week were in wide disagreement over the near-future likelihood of major quakes in California.

They differed on whether earthquakes along the San Gabriel and Santa Susana mountains since 1970, including San Fernando in 1971 and Northridge in 1994, presage a much larger temblor for the Southland.

The lack of consensus at the conference, organized as an exchange of information with the insurance industry, disconcerted some insurers attending the meeting. Major changes in the system of homeowners earthquake insurance are pending in the Legislature, and the industry is worried about its exposure to big losses.

Dennis Doyle of the American Reinsurance Co., a moderator at the conference, said the insurance industry “either has to bear all catastrophes or decide not to write them.”

Both in California and the nation, Doyle said, covering earthquakes “conceivably could be a luxury of the past,” unless much more concrete information is obtainable on what losses can be expected.

Computerized models of future quake frequency and damage in various urban areas, which have been used to justify earthquake insurance rate increases, came in for considerable skeptical analysis at the two-day conference.

The scientific director of the Southern California Earthquake Center, UCLA seismologist David D. Jackson, said the models developed by private quake mitigation firms give too much credence to the possibility that quake rates may double compared to the past 200 years.

Jackson, who testified for consumer organizations in a recent state Department of Insurance rate case that examined a quake frequency model, said the rates such models seek to justify would be too high.

In that case, however, Insurance Commissioner Chuck Quackenbush upheld the model by Risk Management Solutions of Menlo Park. He accordingly approved a State Farm commercial earthquake insurance rate increase of 100%.

Caltech’s Kerry Sieh indicated that even such increases might be far less than what actually is justified.

Sieh said quake rates change sharply over the centuries, and that allowing for 10 times the rate of the past 200 years may be warranted.

This would mean that even the present round of premium increases, often near 100%, and reductions in the amount of coverage may not be sufficient to cover insurers’ potential losses.

Most insurers attending the conference said the quake models--prepared by about a dozen firms--represent an advance over the situation just 10 years ago.

A decade ago there were no such models and no guide for calculating rates for quake insurance.

But Fouad Bendimerad, a senior member of the Risk Management Solutions firm whose model was just approved, acknowledged to the meeting: “There is no exact model. There is no correct answer. Some models are better than others, but there is no exact one.”

*

Meanwhile, State Farm’s director of underwriting, Thomas Powell, said in a presentation that first estimates by some of the modelers of Northridge earthquake damage proved to be far too low.

Powell said damage from the Northridge quake exceeded even the most recent official estimates by the State Office of Emergency Services. He said the insurance industry payout has reached $14.5 billion, and the overall damage was probably $40 billion.

The state estimates have been $12.5 billion and $25 billion, respectively. Powell said the figures do not take into account repairs that have not yet been made and private, un-reimbursed losses.

The conference was sponsored by the Southern California Earthquake Center, which is headquartered at USC, and was paid for by grants from the National Science Foundation, the Federal Emergency Management Agency and the U.S. Geological Survey.

One purpose of the conference, said center director Tom Henyey, was to establish a dialogue with insurers on matters of common interest.

Much of the first day was given to examination by the scientists of whether there was a “quake deficit” in California.

The term refers to indications that the accumulated tectonic strain over many centuries--as calculated in scientific studies--exceeds that released by earthquakes in the last 200 years.

The strain that has built up might indicate that major quakes may occur soon to make up the deficit. But some scientists told the conference there might be other means of making up the difference, such as fault creep not caused by quakes.

The scientist who seemed most concerned that big quakes may soon be in store was Lynn Sykes of Columbia University’s Lamont-Doherty Earth Observatory.

Sykes said the San Francisco earthquake of 1906 was preceded over several decades by a number of sizable quakes on other faults in the Bay Area. He said there is a possible similar pattern of precursory quakes in Southern California.

But a visiting Russian quake scientist, Volodya Keilis-Borok, vehemently disagreed. He said he saw no such precursory pattern and remarked that if he did, he would not be delivering the warning in a public forum.

*

Also at the conference, Tony Crone of the U.S. Geological Survey’s Denver office, a specialist in mid-continental earthquakes, disagreed with recurrent interval studies often favored by the insurance industry that purport to show how often a quake may be expected on a particular fault.

Crone told of cases where no quake would occur for 100,000 years, only to have several big ones in a comparatively short time.

“Earthquakes are episodic,” Crone said. Particularly in mid-continent, between tectonic plate boundaries, “they do not reflect constant, long term behavior,” he said.

Although this finding bears more on quake prospects in the rest of the U.S. than in California, it is not reassuring to insurers trying to calculate rates. It indicates that devastating quakes, causing vast insurance losses, can occur unexpectedly in areas where no quakes have occurred in U.S. history.