Temperature level trends serve as better indicator of probabilities of loss

By examining trends in temperature levels, rather than by examining extreme events or regional trends, insurers can better identify changing probabilities of loss, according to Andrew Dlugolecki, a research fellow at the U.K.’s University of East Anglia Climatic Research Unit.
In Dlugolecki’s paper, The Climate Change Challenge, published by the Geneva Association, he notes the most direct risk to insurers from climate change is that the probabilities of loss are changing so quickly.
Extreme events are infrequent, regional loss trends are unclear and while a global trend has been identified, “it is weak,” he wrote.
“An alternative approach is examining trends in temperature level, since it is fundamental to many extreme events, such as windstorms, convective storms, droughts and rainfall,” Dlugolecki wrote. “Higher temperatures can be expected to produce more extreme conditions.”
He illustrated his point by noting the frequency of hot months in Central England has been increasing over the past century, and comparing that with the return periods of 10-year, 20-year and century-level events.
“In the present decade, 10-year events are happening every three years, 20-year events are coming every four years, while 100-year [events] happen every 12.5 years — i.e., their probability has risen from 1% to 8%,” he continued.
“This phenomenon is typical of a shifting distribution; the more extreme the event, the faster the relative change in its behaviour.”
The 1,000-year event now has a return period of just 83 years, and should therefore be a consideration when doing contingency planning, he added.