The use of predictive modeling by property/casualty insurers is now pervasive a new study from
The survey, which queried 69 P&C insurers in the U.S. and Canada, found 85 percent of respondents saying they use or are planning to use predictive modeling. Brian Stoll, Towers Watson director and the survey’s coauthor says one of the reasons for the ascent of predictive modeling is that insurers have seen improvements both their top- and bottom-line results from the technology, encouraging broader usage.
Regarding bottom-line results, 83 percent of respondents indicated that predictive modeling had improved rate accuracy, and 76 percent said the technology had a positive impact on loss ratio. Results were less pronounced on the top line, with 49 percent of respondents saying the modeling had helped expand underwriting appetite and 46 percent indicating that it had helped renewal retention.
“The optimism expressed by the senior executives responding to the survey suggests the range of future uses for predictive modeling is broad, and will include not only pricing and product innovation but also new refinements in areas such as underwriting, risk selection, claim applications and target marketing,” Stoll says.
The survey did reveal some interesting discrepancies according to line of business. Predictive modeling is most ubiquitous in personal lines auto, where some 88 percent were using or planning to use the technology. On the commercial lines side, usage was more speculative. For example, in workers compensation 41 percent of respondents were currently using modeling while 31 percent were planning to use it. The technology is even more nascent in specialty lines—15 percent using the technology, 37 percent planning to use it and 48 percent indicating they do not use and have no plans to use.
“Commercial lines respondents expressed significant skepticism when asked about the appropriateness of standard industry class plans and exposure bases,” the report states. “Most of these carriers have limited confidence in the ability of standard industry class plans to segment business appropriately.”
