Catastrophe reinsurance is concerned with large-loss events that an individual insurer might not be able to cover.
Having a (geographically) dispersed portfolio is helpful when covering natural perils, particularly as a reinsurer. Hurricanes are a particular threat to the US insurance market because of their potential to cause severe damage and disruption over a wide region: not least, the populated (and insurance-rich) eastern seaboard.
Reinsurance spreads the load to companies whose larger and wider portfolios should be better able to withstand big hits.
The reinsurance market may itself be vulnerable to an exceptionally widespread high-loss catastrophe, or a series of linked catastrophes.
Worries about vulnerability to super-catastrophe are traditionally checked by a transposition approach. This takes a known extreme event and hypothetically relocates it to have maximum impact, e.g. What if Hurricane Andrew had hit further north?
The chief factor overlooked is the scope for a spatially more extensive event than any in the recent record. A physically larger (or more sustained) extreme event may present a double-whammy. The event affects an unusually wide area and, in addition, is of unprecedented severity at its core.
Paradoxically, a further factor is that many structures or communities are so well designed/protected against extreme events that there is a loss of awareness of how to respond when an exceptionally severe event occurs: adding to losses.
Many catastrophes - natural as well as man-made - arise from a combination of factors. While the direct cause of loss (e.g. extreme storm) may not be unprecedented, secondary factors (e.g. antecedent wetness) can trigger spatially extensive or sustained impacts without precedent.
Reinsurers have expressed interest in studies of collective risk, but do not appear moved to develop mathods specifically addressed at understanding the tail properties of the expected distribution of annual aggregate losses. There is little competitive advantage to such understanding. Reinsurers set rates that will bring them the type and volume of business they require.
Nevertheless, research on the aggregate loss problem could be valuable. First, it could help reinsurers to judge the relative vulnerability of Portfolio A and Portfolio B to a very large aggregate loss with regard to a particular insured peril. You can view a PowerPoint presentation on this, or ask me to send you it.
Second, suitably developed, research on the aggregate loss problem could allow a regulator to assess whether catastrophe reinsurance rates are healthily competitive or riskily competitive, with regard to natural perils. Projected climate change might be the trigger to such a review.
Although not as eye-catchingly important, another reinsurance problem touches on temporal dependence issues. This is the definition or distinction of independent events needed to apply certain layered policies and one-reinstatement limits in catastrophe reinsurance.
If you are interested in any of the above, please do get in touch.