By MATTHEW BRODSKY, senior editor/Web editor
The estimates range from 10 percent of U.S. risk managers and 20 percent of the Global 5000 to half of the Fortune 1000--companies that know what they're doing with their property-catastrophe exposures.
"A good percentage of them are pretty smart," says Tom Chan, CEO of Global Risk Miyamoto, the hazard risk control consultants. Yet other risk managers, he believes, are less educated or less sophisticated. They might let brokers lead them down a wrong path, only caring how much property insurance is needed to "cover themselves."
From what Doug Frazier has heard, most risk managers just go to their brokers, their brokers model their property data and then the risk managers buy insurance based on what the brokers come back with.
"They are outsourcing the process of managing catastrophic risk either to their broker or a modeling company," says Frazier, a senior adviser to Arup, the global engineering and consulting firm, and former chairman, CEO and co-founder of EQE. "The industry has grabbed onto a more convenient, easier solution."
San Francisco-based Andy Thompson, leader of Arup's risk consulting practice, echoes the sentiment, saying, "These models are valuable tools, but they were not designed for small portfolios or facultative risks. It is like trying to guess how a specific neighborhood or, worse, an individual will vote based on countywide voting trends."
So risk managers shouldn't really lean that much on models to understand their property-catastrophe exposure. Their primary tools should be their own judgment and experience, followed by others' engineering and operational expertise.
"I think the models are good, but I think they're being misused. And I think the modeling companies are encouraging the models to be misused," says Frazier. "They are trying to manage a long-term risk with a short-term solution."
(For technical reasons why models might not work for risk managers, see riskandinsurance.com.)
Part of this "misuse" comes from the brokers who run the models for the risk managers.
"It's kind of the blind leading the blind," says Chan.
Frazier recalls a story about a client who wouldn't pay to harden its new headquarters because its insurance broker said the added protection wouldn't factor into the model output.
WHAT RISK MANAGERS OUGHT TO DO
Frazier offers up the moral of his story: Risk managers need to drive the process, not modelers or brokers. They need to spend more time collecting and validating data, understanding their risk and developing a long-term corporate sustainability plan for catastrophic risk. Insurance and modeling have their place alongside cost-effective mitigation solutions and facility and operational planning. Also factor in corporate objectives in terms of supply chain and reputational risk, Frazier adds.
Then rinse and repeat every year to make sure your plan still meets the needs.
The first step should be for risk managers "to determine what is driving their risk," says Thompson. "We usually find that 20 percent of a corporation's facilities drive 80 percent of their risk--critical manufacturing and distribution centers, for example. These are the types of facilities for which risk mitigation and facility planning can greatly drive down an organization's cost of risk."
"As a risk manager, you have a sense of what's driving your risk. If you don't, you're not a good risk manager," says Frazier.
The next step is to take that sense and double-check it with modeling. That is what models are best at, according to Chan; they help define your risks, such as the return period of a hazard in a particular geographic area. They help run scenarios: OK, which facilities would get hit with a California quake? With a Category-4 hurricane in Houston?
Models can help risk managers confirm their suspicions about what their "20" are--as in the so-called "80-20" rule that Thompson referred to.
For the 20, Chan, Frazier and Thompson are adamant--risk managers cannot just get by with what the models say. They must get engineers on-site to inspect these properties.
We're talking structural engineers whose day job is to design similar facilities, who speak building codes, who can analyze a structure to see how it really was completed versus what the contractors say they did. They are steeped in local building practice, which can differ from region to region. And more importantly, they know how all this adds up to the way your building performs in a disaster.
The engineers can bring to bear sophisticated analysis tools that capture detailed characteristics of the insured properties, according to Karen Clark, president and CEO of the eponymous risk consulting firm. "It is these details that are the contributors to damage, and these things really aren't picked up in the models," she says.
The benefits of such inspections are manifold, according to Arup's Thompson. For starters, the detailed reports and data produced by visiting engineers can be uploaded into the overall property portfolio to give risk managers, and their underwriters, a truer sense of their exposure. The engineers can also pinpoint weaknesses in structures, recommend mitigation measures to rectify them, then implement said measures.
The end result is a long-term strategy to be more resilient in the face of catastrophe, along with potential short-term benefits of reduced loss estimates and lower insurance costs.
"There are monetary benefits now," says Thompson.
NOT JUST A SALES PITCH
Of course, you could read this and say, "Nice free commercial for the engineers, Risk & Insurance®!"
Sure, our sources in this story are engineers suggesting that risk managers consume more engineering services. But as Frazier points out, engineering is one component in this long-term plan, a part that shouldn't be too much for risk managers.
"It is not an overburdening," Frazier says. "They need to know the 10 to 20 properties that are driving the risk."
If it's any indication, the modelers and brokers whom the engineers seem to bash also agree on most of it.
Says Tom Larsen, senior vice president at Eqecat Inc., a model can come up with a good number, but a great number can be drawn up by having someone on the ground to see how a facility really fits together.
"Modeling and structural engineers do work well together," he says.
Darrell Barker, Larsen's colleague, an engineer, and vice president of extreme load and structural risks for ABSG Consulting Inc., Eqecat's parent company, mentions one of the things that sets EQE apart is that "we have the engineering." Risk managers can use models to determine where their hotspots are, he advises, then bring in engineers, who can devise mitigation measures if necessary as well as "PML Plus" figures for each facility, which can then be manually added into the portfolio.
Ravi Singhvi, assistant vice president of risk modeling with wholesale broker Napco LLC, also talks about the importance of risk drivers and site vists. "This is where we would team up with an inspection company and search for more refined data," he says. "This ranges from calling the local municipalities and speaking to the fire chiefs and the county surveyors while doing up-to-date valuation studies."
Of course, Singhvi disagrees brokers don't know how to use models (at least at his firm) and that risk managers rely on them too much:
"I wouldn't say the risk managers rely too heavily on it, more so the fact that the risk managers don't know enough about the model."
Matthew Grant, group executive of global client development at modeler Risk Management Solutions Inc., denies that modelers exert undue influence on corporate risk managers. How could they when risk managers don't really understand how models are used, he says? Risk managers rely on modelers and brokers for help in many cases because of scarce resources.
Still, Grant suggests boots on the ground are critical for key or unusual and highly engineered facilities. "A well-put-together engineering report on a building will definitely help an underwriter out," Grant admits, adding that RMS works closely with global engineering firm Thornton Tomasetti.
He stresses, though, that any information from engineers must be collected with models in mind, as modeling is the language underwriters use to compare one risk with another.
"It's actually a big disconnect," says Clark. "The models can be off, and if you know that, how can you get better information back into the insurers' PML?"
One way, she says, could simply be an executive summary for those engineering reports, highlighting what makes these properties a better risk.
"Hopefully, the insurance company will recognize that the engineering analysis is a lot more credible," she says.
December 1, 2008
Copyright 2008© LRP Publications