Help

BI’s Article search uses Boolean search capabilities. If you are not familiar with these principles, here are some quick tips.

To search specifically for more than one word, put the search term in quotation marks. For example, “workers compensation”. This will limit your search to that combination of words.

To search for a combination of terms, use quotations and the & symbol. For example, “hurricane” & “loss”.

Login Register Subscribe

Catastrophe modelers increase transparency as buyers push for data

Reprints
Catastrophe modelers increase transparency as buyers push for data

As hurricane season approaches, third-party catastrophe modelers are upgrading their offerings, affording risk managers and insurers a greater array of data and more control over the assumptions that underpin the models' predictions.

“The trend is that the models are becoming more open and more accessible,” said Bill Keogh, Stamford, Conn.-based reinsurance broker and risk management advisory firm TigerRisk Partners L.L.C.

“As time goes by, modeling companies seem to be more willing to share more details about how they came up with their assumptions behind some of the really important assumptions that drive results,” he said.

Alan Calder, London-based group head of catastrophe risk management at Aspen Insurance Holdings Ltd., credits the rise in model transparency to demands from users in the wake of large losses in recent years.

“From the customer point of view, there certainly is awareness that numbers coming out of the model are not necessarily all of the answer,” Mr. Calder said. “There have been too many surprises and different issues people have now become aware that they require some ability to get inside the black box and customize parts of it.”

Mr. Keogh agreed that users are taking a more active role concerning the risks they are modeling, but he said time and resources are needed to understand the models' nuances.

“Building models is a very complex pursuit,” Mr. Keogh said. “You want to take advantage of the asset, but not necessarily buy off on every assumption.”

Scott Clark, risk and benefits officer at Miami-Dade County Public Schools, said he has been using the blended approach for some time.

Mr. Clark said he employed a second outside modeling firm to validate the outcome of modeling runs conducted in 2011 that indicated a $1 billion increase in the probable maximum loss compared with the previous insurance renewal.

%%BREAK%%

Armed with new data from the second modeler that showed a PML close to the previous range, he returned to his primary modeling provider, Risk Management Solutions Inc., which confirmed that its new PML was erroneous. The error was a result of the new version interpreting Insurance Service Office Inc. building codes at the lowest settings for building construction.

When the model was rerun with RMS' own codes, the PML was at a similar prior-year level.

Bill Churney, Boston-based senior vice president at catastrophe modeling firm AIR Worldwide Corp., said he sees certain clients that want more information about the assumptions underlying model output, but users' ability to process that information varies.

“Some people are still relatively new in using model output in decision-making,” Mr. Churney said. “On the other end of the spectrum are companies that have been using models for some time and are eager to dig in on the details of the model.”

While the scientific and technical expertise required to open models and customize them according to unique specifications previously resided mainly with large insurers and reinsurers, recent modeling firm advances are making it easier for a broader base of users, Mr. Calder said.

“We are reaching a takeoff point where customization is becoming more feasible,” he said. “Up until now, it has been more difficult for the general property/casualty marketplace.”

Mr. Keogh said blending data from separate sources makes sense given the inherent complexity of catastrophe modeling.

Claire Souch, London-based senior vice president of model solutions at RMS, said the company recently released a new risk management environment, RMS(one), that enables clients to manage all of their risk models, exposures and analytics in a single place.

“In the past few years, we've seen our clients stretch the limits of the technology they have to run their catastrophe models on,” she said. “We have also seen the need for our clients to keep all of their exposures in one place and have much more transparency into our catastrophe models.”

Ms. Souch noted that the RMS(one) platform enables users to use catastrophe models from other outside providers.

%%BREAK%%

“Increasingly, users want to use views of risk from other perspectives,” she said. “There are definitely more people using a multiple-model approach today than in the past.”

Mr. Clark said he would like to see the modeling firms give more credence to risk managers' efforts to secure properties against loss. In the case of Miami-Dade County Schools, this entailed an extensive retrofitting of existing schools and replacing outdated schools with structures built to withstand hurricanes.

“One of our concerns years ago when modeling came into fashion was that we were concerned about companies that just relied upon the model and didn't reflect what we were doing from an owners' standpoint by investing in our infrastructure,” Mr. Clark said, adding that he is waiting to find out how RMS 13 may address those concerns.