Karen Clark was sitting in front of a computer in themid-1980s, mulling over problems the actuaries weren’t workingon.

Commercial Union Assurance, her employer at the time, wanted tofind a way to manage hurricane risk. Clark was told to figure out away to determine the insurer’s probable maximum losses (PMLs), andshe immersed herself in research.

“I got hooked on it,” Clark says. “I knew I wanted to build acomputer model. I fell in love with the whole concept.”

Her research led to “A Formal Approach to Catastrophe RiskAssessment and Management,” a paper Clark wrote in 1986 for theCasualty Actuarial Society, in which she argued for a move towardprobabilistic models rather than the subjective rules of thumb thenused in underwriting.

Michael Wacek, at the time an actuary with E.W. Blanch, says he“immediately had that ‘Ah-ha! feeling’” upon reading the paper andtold his senior executives to hire Clark.

“I knew this was exactly the way to go forward,” says Wacek, nowan executive vice president at Odyssey Re.

“I told them no,” Clark says. “I wanted to start my owncompany.”

After you've read this, click here to return to the list and readmore

And Clark, having conceived the insurance industry’s firstprobabilistic catastrophe model, did just that, founding the firstmodeling firm, Applied Insurance Research, later AIR Worldwide.E.W. Blanch, the reinsurance predecessor to Aon Benfield, becameher first client.

“She helped change the landscape of the P&C business,” saysJohn Tierney, who worked with Clark at Commercial Union and is nowchief actuary at Quincy Mutual. “The entire industry uses modelsfor pricing and risk management.”

Clark says AIR had 30 clients, all reinsurers, when a bit ofironic serendipity proved to be a big boon to AIR—and eventuallyspawned an industry of competitors.

Hurricane Andrew slammed Florida in 1992—and it revealed thatproperty insurers had “stopped tracking their exposures,” Clarksays. “They didn’t know their accumulations.”

Regulators and ratings agencies realized it, too. Andreinsurers, already accustomed to model usage, began requiring morespecific exposure data.

“It took off—it snowballed,” Clark says of AIR’s model use. “Itwas something insurers could use to better estimate their losspotential and provide information to regulators and ratingagencies.”

“The model gave reinsurers a guide for pricing their product andallowed property insurers to determine how much premiums to collecton the front end, in a much more sophisticated way,” addsWacek.

Clark declined many offers in the years following Andrew butdecided in 2002 to sell AIR to Insurance Services Office (ISO).

But within a few years, Clark says she came to realize that theuse of the models was straying from the original purpose she hadfor them.

That epiphany came with the model-bashing following HurricaneKatrina in 2005.

“I was shocked,” Clark says of the aftermath. “I began meetingwith [insurers and reinsurers], and I was horrified to learn thatat the board level there was not a complete understanding of thewide uncertainty inherent in the models.”

Enhanced computing power was giving model users a false sense oftheir precision, perpetrating an “illusion of accuracy,” sheadds.

In other words, the insurance industry was leaning on modeloutputs much more than Clark had ever intended.

“They are meant to give an estimate—a ballpark,” Clark says.“Companies were getting value from the tools but were relying onthem too heavily, in particular on point-estimates such asPMLs.”

In response, she launched Karen Clark & Co. in 2007.

“I was energized,” Clark says. “Again I was driven by the factthat I knew there must be a better way. We could provide additionalinformation and tools to help companies manage risk.”

Clark now strives to be an “independent, unbiased” provider ofconsulting services and tools that complement the models and allowcompanies to more efficiently develop proprietary views of theirrisk.

“It has been very exciting and gratifying,” Clark says. “It iscontinuously motivating to work with insurers and reinsurers todevelop new ways to better understand and manage catastrophe risk.”