The breadth and depth of predictive modeling applications have grown, but of equal importance, the percentage of participants reporting a positive impact on profitability has dramatically increased, Towers Watson's most recent predictive modeling survey finds.
Our 2014 Predictive Modeling Benchmarking Survey indicates the use of predictive modeling in risk selection and rating has increased significantly for all lines of business over the last year, continuing a long-term trend. For instance, in the personal auto business, 97% of participants said that in 2014, they used predictive modeling in underwriting/risk selection or rating/pricing, compared with 80% in 2013, a 17-percentage-point increase. For standard commercial property/commercial multiperil (CMP)/business-owner peril (BOP), the number jumped 19 percentage points, to 51%, during the same time period (Figure 1). In fact, the percentage of participants that currently use predictive modeling increased for every line of business covered in the survey.
Figure 1. The use of predictive modeling in risk selection/rating has increased signiﬁcantly for all lines of business over the last year
Does your company group currently use or plan to use predictive modeling in underwriting/risk selection or rating/pricing for the following lines of business?
Sophisticated risk selection and rating techniques are particularly important in personal lines, where models have now penetrated most of the market. An overwhelming 92% of survey participants cited these techniques as essential drivers of performance or success. To a significant degree, this was also true for small to midsize commercial carriers, with 44% citing sophisticated risk selection and rating techniques as essential, and another 42% identifying them as very important.
Even as the use of predictive modeling extends to more lines of business, there is an increasing depth in its use. Predictive modeling applications are increasingly being deployed by insurance companies more broadly across their organizations as their confidence in modeling increases. For example, 57% of survey participants currently use predictive modeling techniques for underwriting and risk selection, and another 33% have plans to use them over the next two years. Although a more modest 28% currently use predictive modeling to evaluate fraud potential, a sizable additional 36% anticipate using it for this purpose over the next two years. Survey participants report plans to deploy predictive modeling applications in areas including claim triage, evaluation of litigation potential, target marketing and agency management. These applications will favorably impact loss costs, expenses and premium growth.
The Bottom Line
Eighty-seven percent of our survey participants report that predictive modeling favorably impacted profitability last year, an increase of eight percentage points over 2013 (Figure 2). The increase continues a pattern of growth over several years.
Figure 2. Companies implementing predictive models have increasingly seen favorable proﬁtability impacts over time
What impact has predictive modeling had in the following areas?
A positive impact on rate accuracy helps explain the improvement. In fact, the percentage of carriers citing a positive impact on rate accuracy has increased every year since 2010, when 70% cited a positive impact. In three of the past four years, the percentage-point increase in carriers citing a positive impact has hovered around 10%. In this year's survey, nearly all (98%) of the respondents reported that predictive modeling has improved their rate accuracy. Improved rate accuracy has both top- and bottom-line benefits: It boosts revenue because it enables insurers to price more effectively in very competitive markets, retaining existing customers and attracting potential customers with rates that accurately reflect their level of risk. At the same time, rate accuracy drives profit because it also helps carriers identify and write more profitable business, and not focus solely on market share and price.
More accurate rates also positively impact loss ratios, which have improved in parallel, according to our survey participants. In 2014, 91% of survey participants cited the favorable impact of predictive modeling on loss ratios, an increase of 14 percentage points over 2013. When premiums more accurately reflect risk, losses are more likely to be properly funded.
The bottom-line fundamentals — profitability, rate accuracy and loss ratio improvement — identified in our survey are complemented by top-line benefits. Positive impacts were registered on renewal retention (55%), underwriting appetite (46%) and market share (41%).
The Next Step
Sophisticated risk selection and rating are cited as essential by many of our participants, but our survey indicates that despite favorable trends, insurers are still far from leveraging sophisticated modeling techniques to their fullest, even in pricing. Two-thirds of participants aren't currently using price integration (the overlay of customer behavior and loss cost models to create metrics that measure different rate scenarios) for any products. A few are past price integration and are currently implementing price optimization (harnessing a mathematical search algorithm to a price integration framework to maximize profit, volume and other business metrics) for some products.
The disparity between what is viewed as the optimal use of modeling techniques and the current level of implementation needs to be bridged if insurers want to leverage predictive modeling as a competitive advantage to identify and capture profitable business. Increasingly, insurers are making greater use of analytics including by peril rating (which replaces rating at the broad, line-of-business level with specific rating by coverage), proprietary symbol (customizing vehicle classifications for personal automobile policies), and territorial and credit analysis.
Those insurance companies that can't employ sophisticated risk identification and management tools face the possibility of losing profitable business and adverse selection.
More Progress Is Still Possible
Profitability is hard earned in the current competitive property & casualty market, and predictive modeling is recognized by a steadily growing number of companies as an invaluable tool to improve both top- and bottom-line performance that ultimately reflects in earnings growth. Our survey suggests that insurers are increasingly comfortable with predictive modeling and are using it in a growing number of capacities. However, participant responses also indicate that there are still many benefits offered by predictive modeling and other more sophisticated analytical tools that have not been achieved, such as treating data as an asset and more effectively using predictive modeling applications to improve claim and other functional results. Improving performance on these issues alone could make a significant difference in the profitability of insurance companies and offers all the more reason to explore new ways to benefit from data-driven analytics and predictive modeling.
About the Survey
Towers Watson conducted a web-based survey of U.S. and Canadian property & casualty insurance executives from September 3 through October 22, 2014. The results discussed in this article represent the views of 52 U.S. insurance executives. Responding companies represent a significant share of the U.S. property & casualty insurance market for both personal lines carriers (17%) and commercial lines carriers (22%).