Black swans in plain sight

Predicting and assessing catastrophe risk is at the heart of reinsurance, but how well has that function been performed this millennium, and what lessons have been learned?

The terrorist attacks of 9/11 22 years ago stand out as terrible man-made events that were unexpected in nature and effect. Since then, we have seen a slew of natural catastrophes which were unmodelled either as to their frequency, or the potential for loss from secondary perils likely exacerbated by climate change.

Catastrophe models are based on mathematical assumptions driven by reliable historical data, but do not work as well for man-made catastrophes, especially for emerging risks such as cyber. Probable maximum loss modelling has always been asked to recognise war and disease as material risks, but the models failed to anticipate the systemic effects of political decisions such as national lockdowns caused by the Covid pandemic, or sanctions leading to over 400 leased aircraft not being returned from Russia to foreign leasing companies.

Are those political decisions and their effects so surprising in hindsight?

They should certainly be expected moving forwards, and can no longer be regarded as black swan events. Indeed we are told to now expect, amongst other things, heightened geopolitical risks, more severe cyber attacks potentially allied to kinetic or hybrid wars, more pandemics, more unpredictable natural catastrophes, and generally more uncertainty and risk.

After major unexpected losses the (quite natural) knee-jerk reaction of (re)insurers is to exclude the peril: the underwriting equivalent of shutting the stable door after the horse has bolted. Longer term, however, it is self-evident that the industry needs to offer solutions to risk events if it is to be relevant. Hence, the advent of the Terrorism Risk Insurance Act and a dynamic private terrorism market post 9/11.

Cover for systemic cyber threats in a now largely digitised world has become a major industry topic. However, blanket cyber exclusions cannot work and there remains concern about cover for state-sponsored cyber attacks and the related burden of proving attribution. Coverage for secondary or sub-perils will continue to emerge, however, the extent to which carriers appreciate the potential for aggregated risk from a systemic cause triggering those covers remains to be seen.

In all catastrophic and systemic loss scenarios, the scope of cover and any applicable exclusions will depend on the drafting of the grants of cover and breadth of the exclusionary terms regarding both sub-perils and indirect causation. Too often, there can arise avoidable dislocations of risk as between inwards and outwards covers, including for the major systemic perils of war, nuclear and cyber. The blackest of swans in the plainest sight include political risks which may not form part of war exclusion sub-perils, an issue highlighted in the ongoing Russian aircraft leasing coverage disputes. There remains an inconsistent approach towards radioactive contamination exclusions.

One would expect that after Fukushima and with the ongoing concerns about the safety of other nuclear power plants in Ukraine especially, there would be greater focus on the limitations of traditional nuclear exclusions, especially as regards consequential or secondary perils triggered by a nuclear incident.

The various natural and man-made catastrophes of this millennium have also confirmed that the reinsurance market’s approach to aggregation is cautious and unsuitable for systemic loss scenarios with multiple operating causes. In our ever more complex and connected world, traditional event-based UNL cover will continue to create an uncomfortable hard line for cedants who have had to ingest claims arising from black swan events.

ILS reinsurers have become an important part of the catastrophe market, but are understandably reluctant to provide products for risks which cannot be priced and modelled reliably. We have learned that even if natural cats can be predicted, many man-made ones cannot. The starkest truth is that it is usually what we have not expected that ends up hurting the most. Curiously, there is no definition of catastrophe under English law. This may be a good thing from a reinsurance perspective, as our concept of what it can be is likely to have to evolve as more black swans are discovered…

Simon Kilgour is a London-based partner in the insurance and reinsurance group of leading law firm CMS. He specialises in reinsurance and created the award-winning ReWord database which has been used by many of the leading global reinsurers to get comparative advice on over 1,500 clause variants.

ReWord was recently updated to include analysis of war, nuclear, terrorism, cyber and communicable disease exclusions.