Predictive Modeling Boosts Insurers’ Bottom Lines

North American property & casualty insurance carriers currently utilizing predictive modeling continue to see positive ongoing improvements in bottom-line results, according to findings from Chicago-based global professional services company Towers Watson's second annual predictive modeling survey. The conclusions are underscored by the fact that predictive modeling continues to gain momentum among insurers, with most carriers either expanding current implementations or planning new or additional predictive modeling applications, notes the firm.

The web-based survey, conducted in the fall of 2010, targeted 109 executives from U.S. and Canadian property & casualty carriers. Respondents were asked how predictive modeling is being used to support rating/pricing and underwriting/risk selection plans at their companies.

Of the 43 U.S. companies queried by Towers Watson, 88% said the use of predictive modeling enhanced rate accuracy (up from 77% when the survey was first conducted in 2009); 76% said they realized an improvement in loss ratio (versus 57%), and 68% said that it improved profitability (versus 55% a year ago).

Additionally, 42% said the use of predictive modeling — a tool that uses advanced statistical modeling techniques, along with critical company and external data related to individual policyholders, competitors, marketplace conditions and customer behavior — has furthered the expansion of their company's underwriting appetite (up from 40% in 2009), while 39% indicated it helped increase market share (versus 33%).

"Effective implementation of predictive modeling enhances risk selection and pricing — leading to greater insurer profitability and the potential for growth in market share," said Brian Stoll, Towers Watson senior consultant and the survey's co-author. "As additional companies utilize predictive modeling applications, they will be able to leverage new rating variables and sources of data, and apply the results in new and innovative ways. Companies need to remember that the competition is constantly improving in this area, so a company needs to keep moving just to maintain its relative sophistication and move aggressively to improve its competitive position."

The current use of predictive modeling in the U.S. is up by roughly 10% across all lines of business except commercial property/business owners package (BOP), which remained relatively flat:

Among personal lines, 83% of auto carriers said they use predictive modeling (versus 76% in 2009), while 61% of homeowners carriers have implemented predictive modeling (versus 44%).

Turning to standard commercial lines, predictive modeling usage in workers compensation has increased to 32%, up from 18% last year. Other increases include commercial auto (32% versus 21%). Commercial property/BOP was essentially flat at 25% (versus 23% in 2009).

Predictive modeling implementation in general liability lines has increased to 22%, up from 14% in the previous survey, and to 17% in specialty lines, up from 5%.

Personal lines carriers have been much more active in core predictive modeling activities over the past two years. Turning to the future, personal lines carriers are significantly broadening their focus, while standard and specialty commercial carriers are pursuing aggressive plans across all aspects of predictive modeling over the next two years.

U.S. personal lines carriers plan to move from focusing on first-generation predictive modeling analyses to a much broader array of analyses. For instance, 65% said they would enhance modeling approaches; 57% noted they would take action to improve data quality, and 54% indicated they would work to increase their internal data capture.

U.S. commercial large account/specialty carriers have had a clear focus on product pricing and plan to expand their focus to multiple dimensions going forward. Sixty-three percent revealed they would extend predictive modeling to additional product lines, while 74% said they were going to enhance modeling approaches.

"Personal lines and commercial lines carriers are seeing that predictive modeling benefits greatly exceed the costs and, overall, their plans clearly indicate a desire to move forward," said Klayton Southwood, a Towers Watson senior consultant and the survey's other co-author. "Throughout the industry, insurers are planning to be much more active in enhancing their underwriting acumen. Although personal lines carriers are currently more aggressive than their commercial lines counterparts across many dimensions, most commercial lines companies are planning to broaden their focus on a wider range of predictive modeling issues in the near- to mid-term and strive to narrow the gap."

Among other key findings from the survey:

While most carriers invest in some level of competitive market analysis, only a minority of carriers are pursuing more complex and informative qualitative and quantitative competitive market analyses, and working to fully understand the competition.

For the 18% of personal and/or commercial automobile carriers in the U.S. currently using or planning to implement telematics in the next two years, 83% said that they are using — or would use — telematics to measure annual mileage, 72% are using — or would use — the technology to provide information to insureds to improve driving behavior (e.g., loss control), and 72% indicated they would use telematics to track how the vehicle is being driven (e.g., speed, acceleration).

U.S. carriers have generally been successful in securing regulatory approval for their pricing predictive models — keeping them proprietary and confidential has been a bigger challenge. When asked how often they have encountered difficulty securing regulatory approval for new iterations of pricing predictive models, nearly half (49%) said less than 10% of the time, while an additional 31% said it occurs only 10% to 25% of the time. However, when filing for regulatory approval, 48% said they face challenges in keeping the details of their predictive models proprietary and confidential at least 25% of the time.

 

For reprint and licensing requests for this article, click here.
Core systems Data and information management Analytics
MORE FROM DIGITAL INSURANCE