Non-modeled perils
Catastrophe loss sources that no model in the portfolio covers, missing from the average annual loss and the PMLs, which the underwriter has to load separately in price and capital.
Non-modeled perils are the catastrophe loss sources that no model in the portfolio covers, and which are therefore missing from the average annual loss (AAL), the PMLs and the whole loss table. Thailand, fall 2011: the Chao Phraya flood swamped the industrial estates around Ayutthaya and Bangkok. The major vendors had no flood model for the country. The insured bill went past 10 billion dollars, and a large share landed on the Japanese insurers of manufacturers with plants there. The term covers three different things. Whole perils with no model in a region: flood in several Asian countries, wildfire outside the markets where it has been built, hail in many places. Sub-perils of a modeled peril, such as the Christchurch liquefaction in 2011, which earthquake models of the time handled poorly. And loss components: contingent business interruption, claims handling expenses, post-event cost inflation when that option is switched off. The market handles them with loads. A percentage goes on top of the modeled AAL, a deterministic scenario serves as a floor, or the cedent's claims experience serves as the benchmark. Lloyd's asks syndicates to document these loads in their capital model. They do not show up in the ELT handed to the reinsurer, which has to rebuild them to compare like with like. A practitioner's caveat: the load is set by expert judgment, then negotiated. In a soft market it shrinks from one renewal to the next with no change in the risk, and that is often where a program's underpricing hides.
Textbook case, made-up figures: a Southeast Asian property program shows a modeled wind and earthquake AAL of 4 million dollars. The underwriter adds 25% for flood, which has no model in the region, and 10% for contingent business interruption. Selected AAL: 5.4 million. At the next renewal, the cedent gets the flood load cut to 15%. The AAL drops to 5 million, roughly 7% less technical premium, without a single change in exposure.
non-modeled perils, non-modelled perils, unmodeled perils, non-modeled perils load
Follow the research
One letter a week. First, the week's underwriting note, when one is published: a dated, argued decision, signed by Paul Bernardski. Then what the observatory published on cyber risk, political risk and insurance. No account, no tracking.
No spam, one-click unsubscribe, no data ever sold.