Willis Re, the reinsurance division of Willis Group Holdings plc, is aiming to enhance its clients’ understanding of catastrophe risk with its own views of both modeled and unmodeled risks, the reinsurance adviser announced last week.
Unveiling the “Willis Re View of Catastrophe Risk” in a report published on Friday with the same name, Willis Re discussed the shortcomings of vendor models and described a two-part method to help insurers get a better handle of catastrophe risks.
- Where no vendor models exist, Willis Re builds new proprietary models.
- For perils and regions covered by vendor models, Willis Re enhances and validates these models.
Explaining the need for this dual approach in the opening of the report, Willis Re said that it licenses all of the main third-party commercial vendor catastrophe models.
“These models are complex risk quantification tools that are based on extensive scientific analysis. Despite this, there have been unexpected shocks when major loss events have occurred,” Willis Re said in the report, adding that “Willis Re seeks a full understanding of what is and isn’t captured in these models and how to consider the complexities and uncertainties in catastrophe risk modeling.”
In both aspects of this dual approach, the “Willis Re View of Catastrophe Risk” draws directly on the broader external academic resources of the Willis Research Network.
The report suggests that the Willis Research Network “is the world’s largest insurance-related network of academic institutions,” including universities such as the Oxford University, City University London and the University of Colorado at Boulder and entities including The Rand Corp., The National Atmospheric Corp., NOAA, NASA and many others, according to a page of the report.
The report includes worldwide maps with icons indicating exactly where Willis Re is supplementing, validating or adjusting existing vendor models and where it is offering independent proprietary models.
In the Americas, vendors have the landscape covered, the maps reveal. Still, Willis Re offers scenario-based tests for wind and tornado in the U.S., along with validation reports. And in South America, the graphics indicate that Willis makes model adjustments for earthquake in Peru and Costa Rica. The Americas map also reveals model adjustments for wind in the Caribbean.
Willis’s independent probabilistic models dot the maps of Europe, Asia and Africa. Notable areas where there are no icons indicating vendor model validation or adjustment—where there are only icons for Willis’s probabilistic or scenario-based models shown—are in the U.K. for hail risk; in India, Malaysia and Thailand for flood; and tsunami in Japan.
Re Brokers as Price Drivers, Not Middle Men
In an announcement about the overall “Willis Re View of Catastrophe Risk,” Willis Re said insurers are under increasing pressure from both regulators and rating agencies to demonstrate a full and independent understanding of their own catastrophe risks, along with reasoned quantification of their exposures. “This isn’t always straightforward given the multiple catastrophe models in existence, each with their own unique methodology and perspective,” Willis Re said.
The announcement comes at a time when reinsurance brokers seek to differentiate themselves by adding value to customers beyond just placing their reinsurance programs and acting as middle men to communication transaction terms.
Paddy Jago, global chairman of Willis Re and chairman of Willis Re North America, discussed the changes in an interview with Carrier Management during the Rendez-Vous de Septembre in Monte Carlo.
“The brokers have historically been ‘transactors’ of business. Historically, we’ve been an intermediary.
We’ve introduced risk…and stayed in the middle….
“Now, as a consequence of the environment and a need, we’ve become far more analytical. We’re able to better measure that risk far more deeply than ever before.
“I think that helps us be able to represent our clients and go to either the insurance market, and especially the reinsurance market, more ably equipped to have a debate about what the right price of the product should be.
“Previously, we simply transacted. We simply took one and presented it to another, and then we waited for the terms and conditions to come back.
“Now, we’re in a position whereby I think we start forcing and driving terms and conditions and, indeed, driving price. I think we better understand the product as well as our counterparts, the insurance and indeed the reinsurance companies,” he said.
During the interview, Jago also discussed natural perils like flood in the U.S., for which Willis Re, working together with the Willis Research Network, is gathering data to help not just insurers but government agencies to better understand and measure risk.
Jago also stressed the need for insurers to be as able to quantify their casualty exposures as they are their nat-cat exposures. Indeed, Willis sees this as a key issue for executives addressing investment analysts and other stakeholder representatives. He noted that when a CEO is questioned by analysts about what might happen in the event that a hurricane were to hit Florida or an earthquake to hit California, “it’s very easy for him to trip out a response based on a 1-in-100-year event.
“He can trip out a response that is a number, say 500 million, and then he can also put that number in the shape of what that will do to earnings, and what that can do to his capital, and why he buys reinsurance.
“If we turn that question and look at the casualty arena, if a CEO was asked within that same analyst call, ‘What would be the impact of a 1-in-200 casualty event”—asbestos, for example—”I don’t believe that the community at large can actually answer that question. I think that’s extraordinary. It’s a much more difficult question to answer.
“The property losses are limited in time and place. Casualty is no observer of time. It can run through many years. In terms of place, it can be a global impact.
“They need to get their arms around that,” he said, noting that Willis has turned its attention to developing models that will help these insurers better understand and measure their casualty downsides.
Earlier this year, Willis Re unveiled a new cyber risk model known as PRISM, which is described by Willis Re Senior Vice President Jess Fung and Alice Underwood, head of Analytics for Willis Re North America, in an article they authored for Carrier Management, “Measuring Risk in the Cyber World.”
During the video interview, Jago also referenced a $400 million global casualty reinsurance facility that Willis Re launched in May to provide reinsurance coverage for catastrophic casualty loss. Named PRIMO, the syndicated reinsurance facility—initially supported by 20 reinsurers—responds to increasing concern around accumulation and systemic risk with pre-agreed contract wording.
Unbridled Enthusiasm
Jago also offered his own personal insights on the state of the reinsurance market, based on his 40-plus years of experience. He predicted that some of the capital that has entered the industry will drift away when returns fail to live up to their expectations.
“Capital has entered the reinsurance industry at an extraordinary pace—some people will say with unbridled enthusiasm. I rode a horse once with unbridled enthusiasm. I’ve still got the bruises,” he said.
As for return prospects that will inevitably rein in the enthusiasm, Jago noted that Willis Re’s calculations put the return-on-equity for reinsurers captured within a Willis Re Index at roughly 7.8 percent for the first half of 2014 and 5.0 percent for first-half 2015. That’s the reality in a sector that talks about double-digit returns, he said, noting that the Willis Re figures reflect typical rather than actual (benign) cat activity but still include the impact of significant prior reserve releases.
What does he predict going forward?
“When that stops, which I promise you it will, and…when the cat activity starts to increase and go back again, which I promise you it will, [then] I promise you that those returns will fall away much further,” he said.