March 16, 2017 by Canadian Underwriter
Lloyd’s reports that a partnership with modelling company Arium has produced a data-driven methodology that will allow (re)insurers to model liability risk exposure probabilistically across entire portfolios.
The approach – the result of a three-year project between Lloyd’s and Arium to improve insurer understanding of liability risk exposure – seeks to help create “a more robust insurance industry, strengthening its ability to pay claims to its customers,” notes a joint statement this week from the companies.
“Casualty risks accumulate in a variety of different ways and affect many lines of business. This makes it challenging for insurers to approach casualty risk accumulation systematically,” notes the report, Stochastic modelling of liability accumulation risk.
The methodology “categorizes casualty events based on a company’s business activities – its products and services, operations and infrastructure – and maps the economic relationships that reflect the journey of products and services through the economy,” the joint statement explains.
By creating liability “storm tracks” – described as providing a new, structured way of analyzing casualty events, regardless of risk classification – insurers are then able to “model liability risk in more detail than they have previously, in a way similar to how they model catastrophe exposure,” the companies maintain.
“Breaking liability risk modelling down into blocks, which mirrors the way in which natural catastrophe risks are divided into region-perils, makes it easier to carry out the modelling,” the report states. “This has not been possible previously and represents a big step forward in better understanding liability risk exposure.”
One of the key benefits of this new approach, the report notes, “is that it can identify hidden risks that have not yet materialized, but could do so in the future.”
Like when modelling property catastrophe, generating thousands of potential casualty events allows insurers to assess liability risk through use of factors such as annual average losses, exceedance probability (EP) curves and heat maps that allow a visual identification of risk clusters.
“The approach allows a fully probabilistic view and returns all the outputs familiar from property catastrophe modelling,” the report points out.
“It can produce an EP curve that can deliver annual average losses and arbitrary loss quantiles, as well as heat maps/hot-spot analyses that make potential accumulations instantly recognizable,” it adds.
Jon Hancock, performance management director for Lloyd’s, suggests in the statement that “it is in everybody’s interest to classify liability risks as accurately as possible, and this methodology represents a real step forward for the industry.”
The effectiveness of the methodology, however, “is dependent on high-quality industry classification data, and I would encourage all brokers and other stakeholders to help with the collection of such data,” Hancock cautions.
Lloyd’s helped incubate Arium from concept stage, notes the joint statement, adding that it has since been purchased by AIR Worldwide to further commercialize the tool.
Now having a methodology for a stochastic model, “this framework can facilitate the further understanding of liability accumulations through combining expert knowledge with data-driven analytics,” suggests Robin Wilkinson, CEO of Arium and managing director of AIR Casualty Analytics.