A new study, co-authored by Dr. Benjamin Hohermuth of Schroders Capital, alongside researchers from ETH Zurich, MeteoSwiss, and Stanford University, suggests that standard vendor models used by the majority of insurance-linked securities (ILS) managers contain inherent limitations by overlooking changes in hurricane frequency.
“For decades, catastrophe models have largely relied on historical observations to estimate future hurricane losses. While these models remain essential tools for risk management and pricing, they were generally designed around a relatively stable climate baseline,” Hohermuth explains.
“As global temperatures rise, however, and to paraphrase the standard financial industry disclaimer, historical experience may not be a reliable guide to future risk. This in turn has important implications for how insurers, reinsurers and investors think about hurricane risk.”
The specialist observed that North Atlantic hurricanes rank among the most economically significant natural catastrophes globally and noted that losses have continued to increase substantially over recent decades, which has been highly driven by a combination of growing coastal exposure and changing climatic conditions.
“The challenge for risk managers is that climate change does not only alter the severity of storms. It can also influence how often storms form, where they travel and how frequently they make landfall. These shifts are difficult to capture using conventional modelling approaches that are calibrated primarily on historical data,” Hohermuth noted.
Hohermuth, alongside the paper’s other authors, looked to develop a simplified, but physically grounded, model capable of estimating how hurricane landfall rates may evolve under different levels of global warming.
At the core of the research is Schroders Capital’s in-house Schroders Capital ILS hazard model, which utilises historical hurricane data alongside climate data to create a framework for evaluating the potential changes in hurricane frequency and intensity as the climate continues to warm.
Additionally, by modifying event frequencies to reflect evolving climate conditions, the authors created an updated view of hurricane risk without requiring a complete rebuild of existing catastrophe modelling frameworks.
According to Hohermuth, one of the more stand-out conclusions from the paper is that changes in hurricane frequency may arguably be more important than many market participants currently appreciate.
Much of the climate discussion around tropical cyclones has generally focused on storm intensity, with scientific research increasingly suggesting that the strongest storms are likely to become more intense in a warmer world.
However, the specialist emphasises that research suggests that the frequency of hurricanes plays a larger role in driving losses, with more storms making landfall contributing more to loss growth than increases in intensity alone.
Of course, this is particularly noteworthy for the ILS sector, due to the fact that storm intensity risk is already widely recognised and often incorporated into modelling assumptions.
“Risk related to the frequency of storms over certain intensity thresholds may be less visible, but potentially just as, if not more, consequential,” Hohermuth added.
He continued: “The findings reinforce a broader trend that investors have increasingly recognised in recent years: climate change is becoming a material factor in catastrophe risk pricing.
“Understanding how climate trends translate into losses is becoming increasingly important for investors. If catastrophe models fail to capture shifts in the underlying risk characteristics, pricing and portfolio decisions could be based on an incomplete picture.”
The results of the study also portray how the relatively modest shifts in hurricane activity can translate into meaningful changes in expected losses, while also highlighting the importance of looking beyond headline loss metrics.
“The authors find that climate-related increases for more frequent loss events can be adjusted for using resampling approaches, while the impact on rare events is more difficult to capture and requires more detailed models. Schroders Capital continues to address this problem in an ongoing collaboration with ETH Zurich, using state-of-the-art weather prediction models,” Hohermuth outlined.
The specialist stressed that this distinction matters due to the fact that many different segments of the insurance value chain are exposed to different parts of the risk spectrum.
Primary insurers, reinsurers and capital market investors may experience climate effects in different ways, therefore it is important to understand the implications along the whole risk transfer chain.
“The study also reinforces our view of why an in-house ‘own’ view of risk matters; building on advanced, independent risk models that have been developed over long periods of time, but applying additional criteria and updated modelling assumptions. Recognising that vendor models used by the majority of managers can have inherent limitations, we believe this can enable us to make fundamentally better investment decisions on behalf of our clients,” Hohermuth concluded.
