Personal auto insurers make the most use of predictive modelingtechniques, while those at commercial carriers employ them less,according to a poll by Towers Perrin.

That finding was made in an online survey of 90 executives at 81property and casualty carriers conducted by the consultingfirm.

Seventy-six percent of the personal lines carriers said theyview the tool as essential, Towers Perrin said. The firm noted thatwhile only 56 percent of standard commercial carriers said modelingwas important, that group of insurers is taking a closer look atpredictive modeling.

The survey found current use of predictive modeling variessignificantly by business segment 3/4 led by personal auto at 68percent and homeowners at 42 percent.

Many commercial lines carriers responding said they areinvesting in predictive modeling across all segments, includingcommercial auto (44 percent), workers' compensation (43 percent)and commercial property (36 percent).

According to Towers Perrin, the poll findings "indicate adramatic evolution in insurers' pricing strategies, ashistorically, many firms relied solely on traditional cost-basedpricing approaches to set prices."

The firm added that "while cost is still a criticalconsideration, robust modeling techniques are helping carriersreflect an ever-changing competitive landscape and customerbehavior in their pricing structures."

Klayton Southwood, survey co-author and Towers Perrin seniorconsultant, said that "as personal lines have paved the way, webelieve commercial lines will move faster and with moreeffectiveness, given the lessons learned and the currentenvironment."

"The predictive modeling trends we've seen in property andcasualty are consistent with those beyond insurance, wherepredictive modeling is growing in such diverse areas as governmentoversight, health care, education and myriad others," he added.

Fifty-one percent of respondents in the survey, when asked whatprimarily drove their firm's use of predictive modeling, citedprofitability considerations.

The second reason cited most often was competitive pressure at22 percent. Other reasons mentioned included growth considerations,good business practices and regulatory changes.

Executives in the survey said they are using, or plan to extendtheir use of predictive modeling beyond rating/pricing andunderwriting/risk selection--most notably in the areas of claims(65 percent), catastrophe exposure (62 percent) and marketing (51percent).

"The vast majority of personal lines carriers see rating andunderwriting sophistication as critical 3/4 or at least veryimportant 3/4 as our research shows that there exists a strongrelationship among rating, pricing, underwriting, risk selectionsophistication and financial results," said Brian Stoll, TowersPerrin senior consultant and co-developer of the survey.

Mr. Stoll added that "the synergies that exist between personaland commercial lines suggest that predictive modeling will quicklybecome more pervasive among all types of carriers."

Among other key findings from the survey:

o Leading the list of hurdles for increasing rating orunderwriting sophistication are challenges associated with data (29percent), followed by production systems (20 percent) and culturalissues (15 percent).

o Most medium and larger carriers have full-time predictivemodeling units devoted to rating/underwriting modeling support.

o Small carriers tend to devote five or fewer full-timeemployees, while larger carriers dedicate at least six.