Newsletter Sign-up

Have you tried to understand what catastrophe risk models actually do? It used to be the only people who did understand them were paid good money by insurers, reinsurers and their biggest brokers to use the technology to estimate CAT exposures and manage portfolios billions of dollars wide.

At least that used to be the case. Insurance brokers big and small--paid big bucks to do their own thing--are now wielding the models for risk managers and their other clients.

Tom Larsen, senior vice president at EQECAT Inc., or "E-Q-E," as the firm's called, says that brokers have "always" used models.

"Both insurance and reinsurance (brokers) are using CAT models all the time," Larsen says. "For one, it's to help a client understand what their risk is. But then, two, it's to come up with a risk-financing plan. What the brokers will work on is, 'Here's what your risk is. Here's how we can come up with a plan for you that's the most economically viable.'"

When we get the view of the third major modeler, Boston-based AIR Worldwide Inc., we see that the trend lies somewhere in the middle. Sure, big brokers have always used them, but AIR has noticed an upswing in not only the mammoth institutional brokers using its products, but the smaller players as well, says Tom O'Brien, senior risk modeling consultant at AIR.

O'Brien attributes the increase to the catastrophobia lingering from the carnage of the 2004 and 2005 hurricane seasons, along with the fallout from New York Attorney General Eliot Spitzer's investigations and the hardening property market.

He also points to brokering trends whereby the middle market is becoming more full-service, and brokers in general are investing more effort into clients' risk-transfer, risk-mitigation and administrative needs.

"Brokers are re-evaluating how they do their business," O'Brien says.

TRICKLE-DOWN CATASTROPHES

Anne Anderson, managing principal with startup broker Integro, has used models since the early 1990s, primarily as a broker on major property accounts working out of San Francisco. She is familiar with all of the major vendors and currently is a client of AIR. From her vantage point, the big brokers have always used models, usually offering them to their bigger clients with the biggest portfolios as an additional tool on an à-la-carte menu. Now, though, everybody's hip.

"Honestly," she says, "I think everybody's probably using them in some fashion right now."

"All large brokers do," says Bruce Norris, senior vice president at Hilb Rogal & Hobbs, "without question."

Add to that list now midsize and regional brokers, they say. Anderson also mentions upstart brokers focusing on the risk management side of the market (like Integro), wholesale brokers and brokers who specialize in getting coverage in tough areas, such as earthquake in California or windstorm in the Gulf.

The hullabaloo after Hurricane Katrina, when calls for modeler blood could be heard at industry events and in trade articles, seems to have been forgiven on all sides. It turns out that models are a trusted and essential tool for underwriting and risk management after all. And they're understood now.

"People are comfortable with models now," says EQE's Larsen. "Boxes are becoming less black."

For those brokers with the big clients with big portfolios, CAT modeling is simply a powerful tool that's now indispensable.

"It's incredibly valuable when you have larger portfolios," says Norris, "because it's hard to understand where the spread of risk is, concentration of values, and that's when (the modeling technology) really starts to help you understand the exposure."

For Anderson, one of the main reasons models are going big-time is their widespread use in underwriting circles.

"As the underwriting community has become much more careful about managing its aggregated liabilities, they've used and depended on the tool to a much greater extent," Anderson says, and that has pushed the need to be familiar with the product down to the broker level.

"You simply couldn't determine a good strategy for your client if you didn't know how the underwriter was going to view your client in the market," she says.

It's these same reasons that explain why midsize and regional insurance brokers are getting in the mix. Additionally, AIR's O'Brien suggests the Spitzer bomb has something to do with it.

Many brokers who were gonzo for modeling while at the big brokers left for personal or career reasons to smaller firms. They brought their enthusiasm for the technology with them. Anderson is a prime example; she joined Integro in July 2005 and promptly established that new firm's modeling capabilities. But these reasons alone cannot explain why midsize and smaller regional brokers have tapped into modeling's less-black box.

"It's not a cheap thing to do," says Dan Farster, vice president at Hays Cos., a second-tier U.S. insurance broker based in Minneapolis.

A DIFFERENTIATOR

Why then did this midsize brokerage, with 18 locations and about 375 brokers, decide to take the hit to license a CAT model?

Sure, they have sizeable clients with large schedules with thousands of locations, some of which are located in tremor- and wind-exposed areas, Farster says. But the major reason Hays Cos. started dabbling two years ago in the world of 475-year-returns and PMLs was because a client asked for it.

The client, a real estate company, had rubbed elbows with modelers at a RIMS event and come back to Hays with the idea. The broker then shopped around at the modeling firms, and found one willing to negotiate on price, AIR. Since then, says Farster, who is in charge of CAT modeling at Hays, his firm has broadened its license to allow use of the models with all of its clients. And clients want the models.

Brokers now use a little modeling magic as part of their sales pitch to attract new business. Farster, for instance, mentions a large grocery-store chain that hired Hays this year. The broker's models weren't the "big swinging thing" that bagged this new client, but they were a determining factor, according to Farster.

For large brokers, models aren't even a pitch--they're a prerequisite.

"One of the things about modeling," says Anderson, "is it may not be a differentiator to have it, but it would be a differentiator to not have it."

Again, it's all about risk managers and their underwriters. "Risk managers will be increasingly aware of the use of models by insurers and the roll they play in underwriting," says RMS' Vandermarck.

HRH's Norris estimates that four out of five large clients expect modeling capabilities from their broker, and ask for it. Regarding middle-market clients, he says, "They're getting excited about it." At HRH, there's a push to bring more and more of the technology to these smaller clients, no matter the revenue they bring in.

This push begins during courting. As Norris explains it, he can put together a slick "hazard report" for a prospective client. Into the model goes the would-be customer's facility addresses, say, and out comes info on the 250-year-returns on every peril that the properties face. The tactic works especially, he says, with a client that faces obvious catastrophe risk, like a company in Washington state that may need earthquake cover but is unsure.

Farster agrees. Take that recently acquired grocer client. "We'd run a couple examples for them ahead of time when we were in our negotiation phase and brought them examples of what we could do for them," he says, "which they seemed to like."

WHAT MODELS TELL US

When prospecting hits gold, brokers can tap into the full power of the models for their new clients. They can wield the modeling programs to create hazard reports, which raise clients' awareness of how elevation, soil type, flood zones and other local factors can affect their potential losses.

They can generate ring analyses that graphically demonstrate dense concentrations of property and personnel where their clients face potentially scary losses.

Brokers can run the programs--which conjure 10,000 or more extreme simulations of what could hit client portfolios next year--to grasp what the likelihoods are of certain levels of losses (100-year losses, 500-year losses) and to uncover losses by quantifying each and every property's annual average loss.

Not only can clients grasp their worst-case loss scenarios, but they can see how their retention is affected by limits, deductibles and even their properties' construction--and how altering them can extract more value from their premium.

All the broker has to do, roughly speaking, is compile the data in a client's portfolio--the building types, years of construction, occupancy types, number of stories--into a spreadsheet, feed the data into the model at a secure Web site and tell the model what scenarios to run. The model does all the heavy lifting. Out come the results.

Next comes the "sanity check," says Norris. The broker must look over the results and gauge their accuracy. Models, after all, are all about "garbage in, garbage out."

"All you need to do is flip a zero--$100 million versus $1 billion on one location--that's going to skew it," says Norris. "Or a 40-story wood frame versus a four-story."

The need for this review gets greater the larger a schedule is, he says, when fact-checking each line of the spreadsheet would be as tedious and cruel as proofreading the draft of the latest Harry Potter tome.

Once a broker is sure he and his numbers are sane, it's time to go to the risk manager, armed with the model's output, and say, here are the facts, buy the insurance that I say--right?

"No way," says Anderson. Modeling is just a tool, she says, "a good place to start to help them fine-tune their business decisions."

Adds Anderson, clients should also take into account their own risk appetites, and that thing called a budget. Then there's the chance that underwriters may not offer as much coverage as the model calls for, or that a client's loan covenant dictates another level of insurance buying.

So it's never, "Here's your number. Here's how much insurance you need to buy," says Norris. "It's not an end-all answer, but it gets their arms around it. It helps them make informed decisions when they're purchasing their limits--how much to buy. It helps them understand the pricing ... where the pricing's going and why it's being priced that way."

"It becomes a financial transaction to the company, not just a transfer of risk," adds Norris.

MATTHEW BRODSKY is associate editor of Risk & Insurance®.

May 1, 2006

Copyright 2006© LRP Publications