Learn Predictive Tech Or Die, Insurers Told

Insurers who don't master both the art and science of predictive technology are at risk of falling behind their competitors, the head of Insurance Services Office cautioned at an industry conference.

"We are at the beginning of an intellectual and technological arms race that will remake the insurance business," warned Frank J. Coyne, chairman, president and chief executive officer of Jersey City, N.J.-based ISO, according to a copy of his remarks supplied by his firm.

Mr. Coyne spoke at the ISOTECH technology conference in New Orleans.

He told his audience: "Insurers that succeed in developing and deploying advanced analytics all along the value creation chain can look forward to a long and prosperous future if they continue to invest in new and better analytics and new and better data. Other insurers face a far different future, with even the best of the best today being at risk of falling behind as new technologies come on line."

Mr. Coyne noted that "escalating competitive pressures cut the number of insurers serving the United States by a third in a decade and a half," and carriers face a prospect of ever-increasing catastrophe losses.

Rapid technology advances, he advised, mean "huge changes are just over the horizon for lines ranging from personal auto to workers' comp."

Next-generation personal auto loss models are adding variables "based on the characteristics of individual vehicles--such as horsepower--with the necessary data obtained by decoding vehicle identification numbers," said Mr. Coyne.

He noted the use by comp insurers of predictive modeling to identify the claims with the greatest potential for becoming ultra expensive and to target them for intensive case management by health care professionals.

Leading-edge insurers, he said, are using predictive analytics, such as scoring and data-visualization tools that perform link analysis, to settle meritorious claims more quickly and fight fraud that cost insurers approximately $28 billion last year.

According to Mr. Coyne, "New technologies, such as text mining, are poised to unlock the almost limitless and highly informative but unstructured data in claim adjusters' notes--information previously beyond the reach of those using scoring to detect potential fraud."

New access to that unstructured data, he said, will trigger further refinements to the advanced predictive models already beginning to transform marketing, underwriting, and pricing.

Discussing the growing damage from natural catastrophes, Mr. Coyne noted the $62.3 billion insured catastrophe losses in 2005 and said catastrophe modeling shows that "the effects of exposure growth have far outweighed any effects of global warming and cycles in sea surface temperatures."

Census data, he said, "shows that 63 of the 100 fastest growing counties are in hurricane-exposed coastal states, with 50 of them being in five hurricane-prone states. In fact, 12 of the fastest growing counties--including the very fastest--are in Florida, the state hit by more hurricanes than any other."

New homes, he said are becoming more expensive and elaborate. Mr. Coyne cited analyses by AIR Worldwide that indicate catastrophe losses will double about every ten years just because of exposure growth. "Events causing $100 billion or more in insured losses are already easy to imagine," he said.

"If a Category Four hurricane made landfall in New Jersey and swept its way through the metro New York area and points north, insured losses could total about $150 billion. A magnitude 8.3 earthquake with an epicenter near San Francisco could cause insured losses totaling about $180 billion."