Catastrophe risk management is not being given the attention it deserves by the industry – but it is “fundamental to the financial management of general insurers and the industry as a whole”, according to the boss of an Australian regulator.
Some insurers may be too heavily reliant on catastrophe models which can create a number of problems, APRA deputy chairman Ian Laughlin told delegates at the Aon Benfield Hazards Conference this week.
APRA expects insurers to demonstrate an understanding of the catastrophe model(s) being used and its limitations. It says that outputs from catastrophe models should be used as a base for further analysis and quantification; and that it is good practice for an insurer to assess model outputs against recent catastrophe events.
It requires insurers to guarantee that the data they use to estimate losses is sufficient, consistent, accurate and complete.
“It seems very clear that APRA expects explicit recognition of the limitations of catastrophe risk models and considerable further work to complement the output of models,” Laughlin told delegates. “It expects rigorous assessment by management and board. It seems clear that the insurer falls well short of APRA’s expectations.”
Laughlin stressed that “there is no such thing as perfect model” and that insurers should acknowledge and understand its uncertainties, or be at risk of misinterpreting results, over-interpreting results, and not preparing for reality.
Addressing how catastrophe modelling should be handled, he gave a fictional example of an insurer examining a number of models in the market, talking to brokers and model suppliers about their models and uncertainty in model outputs; and even talking to climate specialists.
Concluding with the improvements APRA wants from insurers, Laughlin said: “APRA wants each insurer to challenge itself about its governance and management of catastrophe reinsurance arrangements and to adopt very good practice.
“We particularly want the insurer to understand the strengths and weaknesses of any models it uses, and the degree of uncertainty in the results produced. And we want model outputs to be complemented by significant further analysis.
“Lastly, we want [the] insurer to satisfy itself that the residual risk is truly within its appetite. We are absolutely intent on seeing industry practice improve.”