Water, wind, and wildfire. It’s been a devastating three months for the U.S.
Total insured losses from Hurricanes Florence and Michael, and the Camp and Woolsey wildfires are estimated by RMS in the range US$18.6 billion to US$28 billion (see table below):
|September 1||Hurricane Florence||$2.8 – $5.0 billion|
|October 8||Hurricane Michael||$6.8 – $10.0 billion|
|November 8||Camp Wildfire||$7.5 – $10.0 billion|
|November 8||Woolsey Wildfire||$1.5 – $3.0 billion|
|TOTAL INSURED LOSSES||$18.6 – $28 billion|
While California wildfires may seem far removed from Atlantic storms, for capital markets investors the fires may make the difference to how 2018 is remembered. Insurance Linked Securities (ILS) eyes are now trained on multi-peril aggregate catastrophe bonds.
Diversifying to Grow
Bundling multiple perils under a single coverage makes good sense for the capital markets.
ILS sponsors benefit from simplicity and efficiency of risk transfer. Multi-peril transactions mimic what traditional (re)insurance can provide, and the relative frictional costs of issuance are reduced. On the other side of the trade, ILS investors have historically been happy with the diversification benefits and returns of broader coverage.
Indeed, multi-peril ILS transactions are nothing new. U.S. hurricane and earthquake risks have been partnered for decades. These were quickly packaged with international peak-perils, like European windstorm and Japanese earthquake risk.
The mid-2000s then saw severe convective storm, winter storm, and wildfire introduced to the multi-peril mix. Before long, investor appetite for diversifying risk drove down spreads, encouraging the inclusion of volcanic eruption and even meteorite impact in cat bonds.
As comfort around multi-peril and non-modeled risk grew, so too did issuance. USAA, Chubb, and Nationwide Mutual have all contributed significantly to the multi-peril market, sponsoring the Residential Re, East Lane, and Caelus series respectively.
These transactions typically trigger on an annual aggregate indemnity basis. In other words, payouts are a function of reported claims across multiple events in a given year.
Named storms usually drive the risk in such deals. Thereafter, depending on region and attachment level, severe convective storms and earthquakes typically contribute most to expected loss, followed by winter storms and wildfires.
Annual aggregate structures can be particularly sensitive to attritional losses from frequent perils such as tornados. As a result, transactions tend to include a minimum event severity to mitigate the impact of small events. Nonetheless, annual aggregate structures can be subject to hidden risk if the perils are not well quantified.
Most multi-peril aggregate bonds in the market include wildfire risk as a modeled peril. Its contribution to risk, however, tends not to dominate the headline metric.
That may now need to change. While wildfire has historically been considered a source of attritional loss, the magnitude of individual losses in 2017 (Tubbs, Atlas, and Mendocino Complex) and 2018 (Camp and Woolsey) is forcing investors and issuers to rethink the potential for wildfires to contribute significantly on an occurrence basis.
Contributions from the 2017 wildfires to the 2014 Residential Re aggregate erosion, for example, were on par with the combined losses from Harvey and Irma. And the 2018 wildfires have highlighted the impact of the losses caused by Florence and Michael.
2017 revealed the potential for wildfire to cause extensive damage. Reconnaissance of the Wine Country wildfires supported an economic loss estimate of US$6 billion to US$8 billion.
As with all catastrophes, this and last years’ devastating wildfires have surfaced important insights. The events emphasized the effects of smoke in driving both damage and evacuation costs. The events also illustrated the significant role of embers and Diablo winds in fire propagation.
Stimulated by liability concerns, the summer of 2018 saw the first wildfire-only catastrophe bonds.
California utility providers are responsible for insurable wildfire losses that are caused by their infrastructure. This not only includes insured and uninsured property and casualty exposure, but also business interruption and fire suppression costs.
If PG&E is found to be responsible for the ignition of the Camp fire in Butte County, then the Cal Phoenix bond will provide a full US$200m payout to PG&E to help cover its liability. This coverage, though well received, would be dwarfed by the overall size of the loss, which RMS estimates is of the order of US$7.5 billion to $10 billion.
If the potential impact to wildfire-only transactions isn’t enough to make investors question their current understanding of the peril, then the distressed prices of some multi-peril annual aggregate bonds and collateralized reinsurance positions certainly will.
The scale of impact from the Camp and Woolsey wildfires has rightfully brought the peril into focus.
Wildfires have the potential to generate significant annual losses – not only in aggregate, but also as individual events.
The need for wildfire coverage is evident, as is the appetite for multi-peril aggregate risk. Now more than ever, high-resolution loss modeling, which comprehensively captures the frequency and severity of potential impacts, is required to deepen insight and rebuild confidence in our understanding of the peril.