RMS Atlantic hurricane model update may adjust cat bond risk metrics

Catastrophe risk modelling specialist RMS has announced the launch of what it calls “significant updates” to its North Atlantic Hurricane (NAHU) Models, incorporating data from recent active years, including from 2017’s Harvey and Irma.

RMS said that version 21.0 of its North Atlantic Hurricane risk model now includes medium-term event rates, information gained thanks to the lessons learned from the 2017-2020 hurricane seasons, as well as a new alternative view of vulnerability for Florida Residential Lines.

We’re told by sources that the updated RMS hurricane risk model is likely to have ramifications for the risk metrics of catastrophe bond issues, as the learnings from active and impactful years such as 2017 may heighten the view of risk, in particular for Florida.

RMS has also built-in an alternative view of risk, in order to account for the Florida Building Code 25 percent Roof Replacement Rule, which it notes was expanded from the High Velocity Hurricane Zone (HVHZ) so as to cover the entire state of Florida in the 2017 Florida Building Code.

The new model will provide an updated view of hurricane risk for RMS clients, with ramifications for the risk metrics of insurance, reinsurance and insurance-linked securities (ILS) contracts and transactions.

RMS said that the new hurricane risk model update includes $6 billion of new claims data and insight, plus the latest statistics on recent hurricane activity in the basin.

RMS also becomes the first model vendor to update to a new view of Florida residential vulnerability, building in a range of impactful claims signals to help clients analyse their hurricane exposure in the state.

A separately licensed climate change model for North Atlantic hurricane risk in the version 21.0 models will also be released in June, enabling users to “quantify the near- and long-term financial impact of climate change on wind risk from the North Atlantic Hurricane Models across multiple Representative Concentration Pathway scenarios up to the year 2100,” RMS explained.

The updated medium-term rate (MTR) forecast, providing a rolling five-year forecast of future activity, includes data from recent impactful seasons, including hurricanes such as Harvey, Irma, Florence, and Michael from 2017-2018.

Changes in version 21.0 mean that the landfall rate increases, and RMS said, “The Version 21 MTR forecast increases in all North Atlantic regions, largely driven by the rate model updates informed by the incorporation of new data from 2019 to 2020.”

RMS believes that its medium-term rates provide a more accurate view of the current state of the Atlantic basin and its impacts on near-term event frequency, providing its clients with an enhanced understanding of their future risk.

These changes are said to affect the metrics related to outstanding catastrophe bonds, which when run through the new model can show elevations in certain risk metrics, such as attachment probabilities and expected losses, we understand.

As a result, the new model has ramifications for catastrophe bond and other ILS investors, as well as more broadly for insurance and reinsurance firms and could also have some implications for pricing once more widely used across the industry.

Improving views of risk through updates to risk modelling platforms can have ramifications for investor portfolios of cat bonds, which can model at higher expected loss levels under new modelling technology.

You’d expect a significant risk model update to have ramifications for ILS transactions and we’re told that while this update could heighten the risk profile for some deals, it may also lessen it for some others.

Of course, risk models are only directional tools and the data from them needs to be consumed alongside a users own view of risk and other research and modelling technology.

But models are important for the ILS market and reinsurance sector more broadly, meaning that as risk metrics could rise for deals run through this new model version, the implications may become noticed in cat bond issuance and also in how fund managers and investors manage their current portfolios, perhaps also having an influence on pricing, as market participants absorb the updated information from RMS’s version 21.0.

It will take some time for the market to adjust the impact of a model change like this, but ultimately modelling using the latest science and data available can only improve ILS fund managers and investors understanding of the risk they assume and gain a better view of the risk their portfolios already hold.

Mohsen Rahnama Ph.D., chief risk modeling officer and executive vice president, commented, “Risk is increasingly complex and connected. RMS is focused on providing the highest quality and most transparent, robust catastrophe models to the industry in this environment. With the new inland flood models and global flood hazard maps, we address an important set of regions where flood is the most important peril, and now cover 100 percent of flood premiums written worldwide. The significant update to the RMS North Atlantic Hurricane Models incorporates the latest science, and applies the learnings from 2017 onward.”

RMS also announced today that RiskLink version 21.0 will be available June 2021, while new inland flood risk models for China, New Zealand and Southeast Asia are also being released, plus global flood hazard maps for over 200 countries will become available in the second half of 2021.

RMS has also added its Canada Wildfire Model has been added to the RMS North America Wildfire Models Suite and said today that its RMS Cyber Solutions Version 5.1 is available now and includes core updates.