With positive changes under way to improve both public and private carrier participation across the U.S. flood market, many are looking to seize the opportunity that the U.S. flood market presents. Insurers, reinsurers, and the capital markets are exploring this opportunity which, in turn, has created a thirst for knowledge. I had the opportunity to see this first-hand when I was invited by Trading Risk magazine to take part in a panel discussion at the Trading Risk ILS: Reloaded and Resurgent event in New York last month. Sofia Geraghty from Trading Risk served as our moderator, and Joanna Syroka, Director of New Markets at Fermat Capital Management, and Ian Hanson, Vice President of Willis Re, were also on the panel.
One point that the audience wanted to understand was the level of demand to take on flood risk from an investor’s viewpoint, and also whether U.S. flood risk can be a portfolio diversifier. From the insurance-linked securities (ILS) side, Joanna confirmed the demand is there, but as with any peril, the ILS market needs to be able to clearly understand and define the risk to get comfortable enough to invest.
For a successful private flood insurance market to develop and ultimately provide the opportunity to the capital markets, there needs to be sufficient demand from consumers. As Ian noted, education and effective communication about the risk are key to driving flood insurance take-up and, picking up this point, I noted that there is a real challenge in how the scientific community, government, and the insurance market talk to the end consumer about flood risk.
We need to articulate flood risk to consumers in a way that is easy to understand, but also does justice to its gradient nature in a way that resonates with a consumer. Initiatives like FEMA’s Community Rating System, which contains a strong education component, is certainly making steps in the right direction, but overall there is much more research and planning to be done to create an effective language for flood risk.
Flooding in New York after Superstorm Sandy in 2012
The core issue underlying all of the current obstacles for growth of the private flood market is the ability of the insurance market to feel confident in its understanding of flood risk, and cat models play a vital role here. A central point of the panel discussion was therefore focused on whether flood cat models are strong enough for the market to confidently quantify that risk. It may not come as a surprise that I think they are indeed sufficiently robust to support the growth and business objectives of the private market, but they do need to tackle a few challenges along the way.
Flood models are not as established as those for other perils, such as hurricane and earthquake, which is partially driven by the immaturity of the flood market itself. The emerging nature of the private market in particular, means that the availability of data about exposure, mitigation, and claims is not ideal. For example, some of the key drivers of loss, such as the presence of a basement and the height of the first-floor is not consistently available. In terms of flood mitigation, 85 percent of the 100,000+ miles U.S. river network does not have information about defenses in the publicly available databases. Both of these weaknesses can have significant impact on modeled flood risk as well as the pricing, risk selection, and portfolio management decisions that are based on that assessment, and RMS has developed proprietary databases and innovative modeling techniques in its U.S. Inland Flood HD Model to address them.
Modeling a complex and granular peril like flood is another challenge and computational advancements have been instrumental in this area. RMS models all sources of precipitation simulated over 50,000 years through the entire hydrological cycle for the 48 contiguous states, a process that requires ~10 billion grid cells. Cost effective technology that allows for this type of computationally-intensive modeling on a timescale sufficient to support dynamic business decisions simply wasn’t broadly available 5+ years ago. We combined those technological advancements with more than 20 years of experience in modeling flood to develop the U.S. Inland Flood HD Model, which is the largest model that RMS has developed to date.
Lastly, the close relationship that inland flood has with tropical cyclone activity adds to the modeling challenge and can drive important issues from the business side of the equation. As Joanna noted during our panel discussion, U.S. flood risk has the potential to be a portfolio diversifier, but the market must then be able to fully understand how inland flood, coastal flood due to storm surge, and hurricane winds are correlated. Inland flooding from tropical cyclones contributes 40 percent of the economic average annual loss in hurricane-exposed states.
It is therefore critical for flood models to account for all sources of inland flood risk from both tropical cyclone and non-tropical cyclone sources, but to also accurately account for the relationship between wind and water from tropical cyclones so the industry can appropriately assess the correlation between these risks. The RMS® U.S. Inland Flood HD Model accounts for that relationship by explicitly coupling with the RMS® North Atlantic Hurricane Models. This allows the industry to properly assess the correlation between the two risks, in terms of geography, tail risk and other factors, with the specific purpose of allowing the industry to manage losses from the two perils and using flood to diversify.
This model development work has not been easy and has taken sixty person-years to achieve, but we are excited to see the model support the growth of the U.S. flood market and generate opportunities for businesses today. It is part of the larger trend where the advancements in flood modeling over recent years are significant and the confidence and adoption of flood models is growing.