P/C Insurers’ Faith in Predictive Modeling Continues to Grow

March 1, 2016

Clarification: This article first ran a year ago, in February, 2015, on Carrier Management.com.  It is not Willis Towers Watson’s most recent survey on insurers’ use of predictive modeling. For a report on Willis Towers Watson’s most recent survey, please see P/C Insurers Expanding Use of Predictive Models, Big Data Across Functions.

Property /casualty insurance executives believe that over the past six years predictive modeling has become an essential or very important driver of improved performance and they are broadening their use of the application into new areas of operation, according to a recent survey by Towers Watson.

Consulting firm Towers Watson, which offers predictive modeling services, first surveyed P/C insurers on their predictive modeling techniques in 2009. Since then, a growing proportion of insurers has reported positive impacts on rate accuracy (98 percent of insurers in 2014 versus 68 percent 2009), loss ratios (91 percent versus 57 percent) and profitability (87 percent versus 57 percent).

The survey also notes that predictive modeling applications are being deployed more broadly, beyond just risk selection and pricing. While less than 30 percent report they are currently using predictive models to evaluate fraud potential, claim triage, litigation potential or target marketing, an additional 36 percent anticipate doing so over the next two years across all those applications.

According to the firm, which provides modeling products and services, the increasing percentage of insurers now using predictive modeling for underwriting, risk selection, rating and pricing continues the long-term growth trend across every line of business compared to last year. For personal lines, auto experienced the most growth (97 percent in 2014 versus 80 percent 2013). Two commercial lines (property and auto) sustained year-to-year increases of 19 percentage points in the use of modeling. Specialty lines exhibited the largest increase (44 percent versus 13 percent).

“Insurers’ profitability in the competitive P/C market is hard earned,” said Brian Stoll, director, P/C practice, Towers Watson. “However, many are recognizing the value of predictive modeling to favorably impact loss costs, expenses and premium growth.”

In fact, he said, 92 percent of personal lines carriers say sophisticated risk selection and rating techniques are an essential driver of performance, while 86 percent of small to midsize commercial lines say it’s either an essential or a very important driver.

Insurers surveyed say price integration (overlay of customer behavior and loss cost models in setting prices) is one area where progress has been slow. Two-thirds aren’t using price integration for any products, while most have not yet moved on to price optimization for products.

“The disparity between insurers’ optimal use of price integration techniques and the actual level of implementation is surprising. Insurers may be missing the strong competitive advantage that model integration provides,” said Klayton Southwood, director of the P/C practice, Towers Watson.

In other findings:

  • Nearly two-thirds (65 percent) of respondents characterize their companies as data-driven organizations. For insurers that don’t, access to data and data warehouse constraints are the primary reasons, as opposed to philosophical considerations or disinterest.
  • Personal auto carriers are making progress with usage-based insurance (UBI) offerings, with nearly one-quarter (23 percent) having launched products or planning to in the next year.
  • A clearer road map is emerging for insurers to implement UBI in their auto operations — nearly three in five personal auto (59 percent) and commercial auto (58 percent) insurers say they’re in planning stages or considering adoption.

The survey results represent the views of 52 U.S. insurance executives in both personal lines carriers and commercial lines carriers.

Source: Towers Watson

Get Insurance Journal Every Day

Latest Comments

  • March 16, 2016 at 12:15 am
    UW says:
    This is just statistically incorrect. "Many," "some," "my clients," etc are irrelevant; the only thing that matters is the population on a whole which makes up the risk pool. ... read more
  • March 15, 2016 at 12:24 pm
    UW says:
    Damn, nice one Ron; although it's not hard, it's funny seeing this clown made to look foolish. He says he prefers the accounts that don't switch, because the younger people ca... read more
  • March 9, 2016 at 4:36 pm
    Agent says:
    Does this mean that you will actually pay your taxes this year or do I have to pay them for you like the past several years. We do well in spite of the Obama economy, not beca... read more
See all comments

Add a Comment

Your email address will not be published. Required fields are marked *


More News
More News Features