PCS Shows How Cat Loss Estimates are Made, and Why They’re Important

By | September 20, 2013

The Property Claims Services (PCS) unit of Verisk Insurance Solutions could be described as the ultimate end product of risk modeling. AIR Worldwide is also part of Verisk, but “there’s a high wall between the two,” said Joe Louwagie, CPCU, SCLA, AIC, AIM – PCS’ assistant vice president.

He explained that PCS acts a sort of clearing house following a natural catastrophe. The company collects the data from primary carriers, reinsurers, brokers and third party data sources, such as investors, who may have been affected by the event.

PCS was originally set up 63 years ago, mainly to analyze losses from fire claims. It has since branched out to other types of nat cats, mainly weather events – hail, wind, floods and tornadoes, as well as fires. It included insurance linked securities (ILS) in its analysis portfolio in 1997.

“We can tell from companies and policyholders within two weeks [of a disaster] if the losses from the catastrophic event will exceed $25 million,” Louwagie said. In which case PCS takes on the task of collecting claims data, analyzing the losses the data shows, and then calculating the probable payments that will eventually be made.

The data log is kept open for 60 days, and if more data continues to come in, it is kept open, and reviewed every 60 days until Louwagie and his colleagues are satisfied that all of the information necessary to make an accurate final estimate has been obtained.

PCS issues regular bulletins on all of the claims it is analyzing, which in turn are used by the actuaries of the companies involved to estimate their monetary exposures. Eventually those figures are added to the data on all similar types of losses in a given area. They are also an intrinsic part of calculating the reserves that must be earmarked for claims payments.

After 63 years, however, the data PCS produces for any given event has taken on greater importance. Louwagie explained that natural catastrophes occur in greater numbers in some state, i.e. “high states,” and to a lesser degree in others, i.e. “low states.” A ‘high state’ has more data available to calculate their exposures and to put in place the necessary personnel and equipment to deal with catastrophes, while a ‘low state’ would have a less urgent need to do so.

Companies and state regulators also use the data to set premiums, and for re/insurers to make certain they have an adequate staff to deal with the disaster. “Claims departments use it [the loss data figures] to set their staffing levels,” said Louwagie. The re/insurers “use it to set staffing levels for underwriters and to calculate ‘risk segmentation;’ risk managers use it to calculate risk transfers.”

Tom Johansmeyer, PCS’ director of marketing, explained that the company has been successful, not only because it does a good job of collecting and disseminating accurate data, but also because it is see as a truly “objective source” for that data. As an example he cited PCS’ $18.75 billion estimate of overall insured losses from Superstorm Sandy, as having proven to be quite accurate. So accurate in fact that PCS felt confident in closing Sandy’s file.

Plans are in the works to expand beyond the U.S., specifically in advising the Korean Fire Protective Association, which serves the country’s insurance industry, to set up a loss index.

Louwagie stressed that in order for PCS to accomplish its mission two things are necessary: “First you need to have meetings with your customers, how can you serve them? You need to create an ‘interactive’ dialogue. Secondly, you need to keep high quality standards.”

Although the phrase “big data” is becoming somewhat of a cliché, it is the ability to collect data that has greatly improved the construction of more accurate – and hence more useful – catastrophe models. The data PCS collects goes back to the front end, and greatly helps to further refine those models.

As with other aspects of the re/insurance industry, everything seems to depend more and more on everything else.

Was this article valuable?

Here are more articles you may enjoy.