EQECAT’s President Paul Little explains how cat models have evolved from the general to the specific and what’s now required to create good ones. He also foresees cellular technology – iPhones, Androids, iPads – enabling insurance buyers to access cat models for their house or their cars, and place their coverage directly on line.
Little has headed EQECAT for about six months, but prior to that for 2 and a half years he was the senior VP for sales. In that role he “participated in the transformation plan for EQECAT.” The plan was funded with an investment from EQECAT’s parent company, the American Bureau of Shipping, who “saw an opportunity for our firm to come out with the next generation of models.”
He said that they were built on a platform “that we’ve had for the last 20 years, in fact next year will be our 20th anniversary as a company.”
Talking about the new model EQECAT introduced in August, Little said: “The main thing we focused on when we created RQE (Risk Quantification & Engineering) is maintaining the science and engineering that underpins the model. What we did as part of the investment was to look at the front end of the model – the user interface – that’s the part of the model that the client sees when they’re using it. And in that functionality we made it a lot more flexible and a lot more easy to use.”
He added that with the templates created for RQE, data, “regardless of the type of data a client has in terms of its format, we can bring that data into our model very easily, and it really reflects the nature of the risk that’s being taken on.” On the other, or back end, “which is the output, we actually changed and improved our financial model.”
Work on the model included extensive consultation with clients. Little explained that EQECAT formed a “client advisory board, made up of insurance brokers, insurers, reinsurers, and even one or two players in the ILS space, which has become so prominent today. Their input informed all of the functionality and the output that we’ve included in RQE.”
EQECAT was one of the first to issue a cat bond based on a parametric trigger. Little pointed out that a “normal insurance contract is based on indemnity, the actual loss to the company.” However, a “parametric bond is actually based on the characteristics of the event – the wind speed of a hurricane, or the magnitude of an earthquake.”
Using a hurricane as an example he explained that “you can design a contract so that what triggers the payment under the contract is the wind speed of the actual event. So, if you have a force 5 hurricane, that’s one with typically 140 – 150 miles per hour wind, you can actually have an insurance contract that pays in the event that the wind gets into that level.” In the same way you can design a contract based on the magnitude of an earthquake.
For mainly agricultural coverage – hail and rain – Little said those types of contracts are “mainly characterized in ‘degree days,’ depending on heating and cooling degree days; that would be the parametric trigger.”
These types of triggers are used in cases where the underlying data “is not sufficient enough to really characterize the peril in the manner where an investor wants to be comfortable that they understand the risk,” Little said. Measurements are made independently by wind stations, or in the case of earthquake, by the USGS. “They can’t be influenced that much.”
An indemnity contract is based on the actual loss, and this can be “determined a lot by claims handling practices.” They [the insurers] may be liberal, which benefits policy holders, or they may settle claims in a different manner with a lesser payment. Using an independent trigger avoids this situation.
At present EQECAT is focusing on deploying the new model to its existing client base. “We’re about 80 percent deployed,” Little said. The model was introduced in January 2013. “Over the last several months we’ve been installing the software,” Little said. “There’s been a lot of change management, comparing the old model, the former model, to RQE.” 180 countries are involved in the model changeover “all across the globe – from Japan earthquake, to U.S. earthquake; U.S. wind, and also into Latin America, South America and Africa.”
Little explained that outside of the property field, there have been a number of models on the life side, such as mortality tables. He also said “some bonds may be issued in the future that could be based on terrorism risk, or manmade disasters.” He also indicated that there has been some exploration into the necessary data that would be needed to build a model for a casualty bond.
“Even though it’s a long tail line, you can in fact develop distributions,” he said, “and that’s really the combination of looking at the historical losses and running simulations to see what are the likely outcomes and the things that are within control; the things that you would use for the basis of designing a bond, or any sort of financial instrument, similar to a reinsurance contract for casualty business.”
Modeling casualty exposures is interesting, because “there are a lot of companies that are looking for diversification of risk,” Little said. “They have maybe in some instances enough hurricane risk,” or other types of property risk, and they “are looking for other ways to deploy their capital; so therefore they’re looking for something – another type of peril – that’s not correlated with what they already have in their portfolio.”
He is also thinking ahead. Earlier this year in a blog [ www.eqecat.com ] he described the present “digital age” with the proliferation of iPads, iPhones, Androids and all of the other portable and held devices. “As technology advances, and individuals become more comfortable with accessing all sorts of analytics on their devices, that’s where they go for information,” he said.
He sees a “time in the not too distant future,” when an individual is looking at buying homeowners insurance or automobile insurance, which they increasingly do via the Internet. “I think they will start doing that with homeowners insurance, and in fact a cat model can provide data and information that an individual can use – with some education – to determine whether or not the quote they’re receiving for that insurance policy is competitive enough, given the likelihood of loss to their particular home.”
He explained that it wouldn’t be an actual individual cat model, but rather “access to the output. There are service providers, and there could be an app – let’s say an RQE app, as an example – they would have the output from that model to make an assessment, – make a judgment – as to what their risk is, and in fact you can probably run simulations in this environment.”
The homeowner and potential insurance buyer could then explore different scenarios that might affect the cost of insurance. As an example, Little pointed out that elevating a home or putting up shutters might mean it would be more protected from damages. Then the question is: “How much credit should I be given on my premium, by taking these mitigation measures?”
Cat models have reached levels of development that has made them an integral part of the re/insurance industry, especially for property. They are adaptable, however, for other functions. Little foresees greater use by the U.S. government,” which will be “using models more directly. One of the ways they’ll be doing that is to look at the federal flood [program].” It’s “in a significant deficit,” he said; because of the “rates that are being charged to individuals who are heavily exposed to flood.”
In the private sector major losses trigger increases in the cost of coverage, as well as the terms and conditions of the policies, “to reflect the risk. That’s not happening today in the federal flood program,” Little said. He sees a time in the not too distant future, when “the federal government might employ catastrophe models to measure flood risk and therefore estimate the likelihood of loss, and include that in the pricing of the federal flood product. In fact it may lead to the privatization of flood risk.”
EQECAT already has models for similar perils, notably California wildfires, which are also “used for forestry.” Risks such as hail for automobiles, or typhoons, can be modeled “to the extent that we have the data that can be measured and is good quality,” Little said. “We can develop a model that reflects the nature of that peril, looking at the actual assets that are exposed to it, develop a likelihood of loss, and the estimate how much that loss might be for an individual structure, and also for the entire industry.”
There are now a large number of cat models – re/insurers can license or construct their own and large brokers construct them too. Little explained that “we work with an insurance company directly, or we work with the reinsurance intermediary, who would run the models and provide the output to their clients, and likewise with a reinsurer for reinsurer licensed cat models.” They typically license three models, sometimes only two.
“They like to have as complete a picture as possible, because each of the modeling companies has a different methodology that may be used, and to the extent that they have as many answers as possible, they’ll have the best opportunity to reflect on the nature of the peril, and also to price the products that they sell.”
In answer to the last question about the Rendez-vous, Little said he has been in the business for 28 years. During that time he has developed numerous relationships within the re/insurance industry. “Most of my career was as a reinsurance intermediary,” he said. “This is an opportunity to really reconnect; see a lot of old faces, and actually spend time with individuals,” and to catch up on their families, any job changes, as well as keeping contact with the people in the industry, the rating agencies, and hearing everyone’s views on topics, such as alternative capital, which concern the industry.
By getting “all of those viewpoints first hand, you have, first of all for EQECAT, the ability to inform our product road map as to what we see as the priorities going forward, and where the need is, so that we can respond to that need. We have the opportunity to meet with our clients, and we understand how are model is being used, and then what we can do – the things we can do differently, or to improve what we’re doing today, so that they can get better value out of the model.”