Unmodelled events – or even part-modelled events – have the potential to affect both reinsurance and retrocession markets in ways that are hard to predict, writes Eric Sugier, Head of Property, Liberty Mutual Re

The death of the great stage escapologist Harry Houdini is a prime example of the impact of the unmodelled event. Part of Houdini’s act was to ask a member of the audience to punch him as hard as possible in the stomach. Houdini would tense his abdominal muscles in advance to resist the blows. But on one particular evening in 1926, the 52-year-old was unprepared when an audience member came to his dressing room and unleashed a volley of blows on his stomach. In considerable pain, Houdini went on stage but died a few days later from abdominal injuries.


Although Houdini knew how to prepare for a punch to the stomach, the speed with which the fatal blow was delivered and the fact that he was reclining on a couch in his dressing room at the time made it impossible for him to react as planned. This element of surprise rendered the blow much more damaging than it would otherwise have been. You can see the same effect when a Paris Metro train moves away from the station. If passengers are warned the train is departing, nothing happens; if passengers are taken by surprise – even though they know the train will move at some time – some may fall over. Surprise multiplies the power of an event.


Unmodelled events – or even part-modelled events – have the potential to affect both reinsurance and retrocession markets in ways that are hard to predict. Certainly, their ability to move the market is much greater – and ours is a market that requires considerable force to move it these days. Japanese earthquakes, windstorms in Florida or California – these are events that the market’s predictive models have chewed over time and time again. We’re prepared for them, like Harry Houdini tensing protectively before the punch.


With events like 9/11 and Hurricane Katrina, which did move the market, while some elements had been modelled, others were not. The scenario of a plane crashing into a building was a familiar one, but not a plane being deliberately flown into a building. Having most of the lines of business hit at the same time was not part of any scenarios. Similarly, hurricanes are part of reinsurer’s staple diet, but not a hurricane so large in the Gulf of Mexico driving into New Orleans or the resultant dam failure. The impact of these two events was so dramatic they were sufficient to drive some investors and their capital from the market, thus fuelling rate rises.


The models we use to stress test our portfolios and reserves have improved significantly in recent years. They are good tools with which to assess the potential loss contribution of one account to a portfolio, an event’s impact on an account and to examine relative valuations from year to year, but reinsurers must not rely on them too much on an absolute value. If one examines a similar scenario with different model providers, the variations in outcomes expose the differing assumptions the modellers have made in their creations. Even with the same model, there are significant differences from one version to the other. I come from a mathematical background, which should make me a strong advocate for models – but even I take them with a small pinch of salt.


One of the most fascinating unmodelled events that has major potential to turn the market is a sudden rise in interest rates around the world. Faced with that scenario, alternative capacity, which is always hunting for the very best returns, would in all likelihood move to some other form of investment. A sharp spike in interest rates would also impact positively the Catastrophe reinsurance market as almost all reinsurers have a significant life component in their portfolios backed by bonds’ assets. This would force them to more orthodoxy on the Cat underwriting. Interestingly, rating agency Moody’s noted this month that rising interest rates globally are set to be ‘credit positive’ for reinsurers.


September’s simultaneous windstorms in the Atlantic and Pacific were unusual but even this two-pronged assault was not severe enough to shift the market. If there was a warning note to be sounded, it was the sheer volume of rainfall created by Hurricane Florence and its storm surge. RMS expects the cost of Florence’s wind component to be in the region of $1.3bn–$2.6bn while flood will be $1.5bn–$2.4bn but other firms see losses from water at much higher levels.


Florence’s path through the Carolinas is prone to water surges. When a hurricane weakens, its windspeed drops but the pressure it places on the sea remains, resulting in sea level rising on the coast. This also leads to huge volumes of water being dumped as the hurricane can no longer keep them airborne, resulting in a flood event rather than a wind event. Another example would be 2001’s tropical storm Alison, which was pretty much a non-event from a wind perspective but generated serious rainfall and flash flooding in Texas. 


Sea levels may be rising slowly and are currently not affecting storm surges, but by the end of this century it could be a very different picture. The consequences of global warming may be unpredictable but hardly a complete surprise. More unstable atmospheric conditions will lead to stronger hurricanes and thus larger water surges.


Florence and Typhoon Mangkhut may not have been severe enough to disturb the post Monte Carlo mood of business-as-usual, but they hint at the surprises in store in the not-too-distant future. And even if the models capture some elements of these scenarios, we are learning that those elements which have not been assimilated can make all the difference to the eventual outcome. One of the reasons why we recently launched Liberty Mutual Re is to take a global approach that can create the type of holistic solutions the changing world will increasingly demand. 


Harry Houdini thought he was prepared for every eventuality, but he overlooked the unexpected. For reinsurers and their clients alike, the key is to recognise that no one can be prepared for every eventuality – life will always surprise us – but by acknowledging that our models are but a single weapon in our armoury, we take an important step in terms of true preparedness.