Towers Perrin: Commercial Lines Will Increasingly Use Predictive Modeling

A survey by consulting firm Towers Perrin found commercial lines carriers are increasingly discovering the benefits of using predictive modeling in establishing price and other areas.

According to the survey of 90 executives from 81 property/casualty companies, 76% of personal lines carriers view predictive modeling as essential. Although 56% of standard commercial lines carriers say such modeling is important, Towers Perrin says more are taking notice of the pricing tool.

"Personal lines are generally heavily regulated whereas in commercial lines, you can move to change price and select risk without divulging everything you're doing," Brian Stoll, senior consultant at Towers Perrin and co-developer of the survey, told BestWire.

"The distinction in commercial lines is there is less heavy-handed regulatory activity," added Klayton Southwood, also a senior consultant and co-developer of the survey.

Stoll said many companies using predictive modeling in commercial lines also write personal lines, such as Travelers, Hartford and American International Group Inc.

"They know how it works," Stoll said. "They know the opportunity and efficiency."

Commercial insurers are already using predictive modeling. While announcing quarterly earnings, Gregory E. Murphy, chairman, president and chief executive officer of Selective Insurance Group, said the company was "concentrating on the factors we can control such as using predictive modeling on all lines of business" (BestWire, Aug. 4, 2009). In a column in July's Best's Review, Frank J. Coyne, chairman, president and CEO of ISO, said it could help anticipate, detect and prosecute fraud.

The predictive model for pricing uses data related to individual policyholders, competitors, marketplace environment and customer behavior instead of the more traditional cost-based pricing approach.

Southwood and Stoll believe commercial lines carriers are developing the data for models in order to accomplish strong financial results, as personal lines carriers have. The use of the model can be expanded from pricing and risk selection to catastrophe management, reinsurance, marketing and claims. Many carriers surveyed said they are already using predictive modeling in this way.

"You can look at data in a multi-variable manner," said Southwood. "You can find interactions and identify risk potential -- get at those relationships and better understand them."

The result is better rate accuracy, loss ratio and profitability, said Stoll. "You can expand your underwriting appetite and grow market share," he said.

Large carriers have a "big advantage" over small carriers because they have a lot more data and observations, and they can devote more employees to the task. Small and regional carriers "know they can't out-model" the larger carriers but the key for them is to keep an eye on competitors and "fast-follow."

Predictive modeling is most used in the personal automobile segment, with 68% of carriers saying they use it, followed by homeowners, with 42%. According to the survey, commercial lines carriers say they are investigating the use, including 44% in commercial auto, 43% in workers' compensation and 36% in commercial property.

"The synergies between personal and commercial lines suggest predictive modeling will quickly become more pervasive among all kinds of carriers," Stoll said.