Climate change has long been a topic of considerable debate. Its profile has been raised further recently by two key developments: first, the occurrence of several high profile losses where climate change may have played a contributory role (for example, the California wildfires) and second, regulatory initiatives, particularly in Europe, to stimulate the industry’s focus on this area, according to Matthew Eagle, Head of Global Model Solutions & Advisory, Guy Carpenter and Dr. Jessica Turner, Head of EMEA Cat Advisory, Guy Carpenter.
Although the climate-change debate is often polarized, it is both clear and indisputable that greenhouse gas emissions have increased and the planet as a whole has warmed considerably. There are some areas where theory, observations and climate model projections all align, resulting in stronger evidence of the consequences, although the impact on specific regions and perils is more mixed and, in some cases, less conclusive.
Expectations of more heatwaves and rising sea levels under most climate-change scenarios clearly bring an enhanced threat for properties exposed to wildfires and coastal flooding, for example. There is also a growing consensus that intense precipitation events will increase. Interestingly, global average rainfall amounts are not forecast to grow substantially, with wetter areas by and large getting wetter and drier areas getting drier. But crucially, the daily (and sub-daily) rainfall amounts are forecast to increase, in part down to the pure physics of the atmosphere holding more moisture in warmer climates and conditions becoming more optimal for storm development and increased intensity.
The Market’s Response
Commercial modeling companies are starting to respond by developing new models that better reflect certain climate scenario projections. JBA and Ambiental, for example, have developed climate-scenario models for UK flood. We expect others to do the same. End-users of current models need to examine what they use and find ways to make adjustments to better reflect future scenarios.
Regulatory scrutiny is also increasing in certain countries. The Prudential Regulatory Authority (PRA) in the United Kingdom has introduced new climate-change scenarios in their general insurance stress tests (GIST2019) for a range of perils in the United States and United Kingdom. Commercial modeling companies have been quick to respond to this particular initiative by carrying out analysis with their own models and releasing guidance to their clients on the potential impacts. But this on its own is not sufficient: after all, if all users rely solely on these estimates, the PRA might have just as well asked the vendors to do this work in the first place.
Analytics and Insights
Although users have fewer levers to adjust the models, Guy Carpenter has developed methods for translating scenarios such as a 10 percent increase in rainfall run-off into the impact on modeled Average Annual Loss and tail probable maximum losses by either adjusting event rates or the losses per event. This provides clients with mechanisms to test the sensitivity under different methods.
Exercises such as these are crucial to developing a better understanding of the risks and impacts associated with climate change. And by doing so, we can start to challenge the familiar refrain that climate-change scenarios for the 2030s, 2050s and 2100s are not pressing issues for insurers more focused on the short-term renewal of (mostly one-year) policies.
The risk of prevarication is simply too high. Climate risk covers not only physical risk but also transition risk and liability risk. Board level scrutiny is intensifying to ensure that companies are on the right side of this issue and that strategies are in place to safeguard future success in the changing climate environment.
Ultimately, the climate has already changed and carriers need to review their views of risk for all perils to ensure that the models they use are actually fit for purpose. Most models use historic data to calibrate their event frequencies and intensities. By their very nature, model outputs are backward looking and reflect the average of the previous period used. Consider this in the simple context of coastal flood and storm surge: if models rely on an average sea-level from 1970 to 2008 but expectations are for sea levels to rise by between 0.1 and 0.2 centimeters per annum, then clearly the model no longer reflects the reality of the risk today.
Guy Carpenter is well-placed to advise clients on responding to climate-change risk. We have multiple experts in atmospheric and climate science across the globe that have been actively researching and consolidating the scientific literature. In addition, we have participated in an important PRA working group on incorporating climate change into modeling tools and have created a response to the PRA stress test. Our solutions can help clients make informed decisions about pricing, capital and risk management strategies.