LAS VEGAS–While insurers rely on risk models to predict their loss exposure, models can also be used to establish reserves after the claim and examine the variables that can reduce the cost, two actuarial experts explained at a conference here.
Gary C. Wang and Laura A. Maxwell, consulting actuaries with Bloomington, Ill.-based Pinnacle Actuarial Resources Inc., discussed the formulas and loss trends actuaries can use to help predict the cost of a claim event at the opening workshop of the National Underwriter Company's 11th annual ACE (America's Claims Event) conference.
"This is one of the coolest tools an actuary ever got his hands on," said Mr. Wang.
The session, titled "How to Use Predictive Modeling in Claim Organizations," discussed how using GLM (Generalized Linear Method) an actuary can estimate the cost of a claim and reserve for that risk, especially on long-tail events.
Mr. Wang said the method is most effective with frequent loss events simply because there is more data available to help predict what the loss trend will be over a period of years.
For insurers, it helps them determine if the event could penetrate layers of reserve and help them to make adjustments accordingly. However, it is less predictive with major loss events, he noted, because there is less data available to examine.
The method can be used to help disseminate whether the professional network an insurer uses is being effective throughout the claims process, it was explained.
It was also described as helpful to insurers for understanding the cost or benefit government legislation could have on a loss event.
He noted that the measure of severity can also be broken down into geographic location or other variables to assess the differences the claim presents.
He said the model allows the actuary to get a feel over time of what the loss trend will be, and where there are questions, it can give "hints where to go to find answers."
The models, said Ms. Maxwell, "are complex enough to produce an answer with accuracy, but simple enough to explain."
However, the consultants noted the model does not answer broad questions. To be effective the models are designed to answer a series of individual questions that can make a broader picture.
Mr. Wang described the process as putting ingredients into a pot. To arrive at an answer either data is put into the pot one ingredient at a time, or it begins as a jumble of data and pieces are removed individually until an answer is arrived at.
Working with the models, Mr. Wang noted, produces some basic observations of the claims process. He said the longer it takes to settle a claim, the more costly it is either from inflation or because the policyholder is seeking a better outcome. Involvement of attorneys on the side of the policyholder can also dramatically increase the cost of the claim.
He noted, too, that the cost of the claim can be affected by the insurer's provider network (doctors, attorneys, repair personal). However, he said, the cost of these professional services does not necessarily indicate quality. Instead, it may reflect the cost of providing these services in geographic locations or better quality of service. Using the models, claims executives can more closely examine the service provided and determine the quality received.
© Arc, All Rights Reserved. Request academic re-use from www.copyright.com. All other uses, submit a request to TMSalesOperations@arc-network.com. For more information visit Asset & Logo Licensing.