The ILS industry was “surprised” by PCS coming out with a $5bn industry loss for Hurricane Milton, according to multiple sources, with managers head-scratching over how to bridge the gap between PCS and their in-house assumptions and modelled numbers.
The differential between PCS and other industry views raises questions about who is “off-the-mark" and how the gap can be fully explained to investors.
There are also potential ramifications for index-based ILS products, particularly if buyers’ confidence in the index is diminished.
Sources canvassed by this publication largely fall into one of two cohorts: either the one putting the Milton industry loss in the $20bn-$30bn range, or the $30bn-$40bn camp.
A smaller number of sources are tending toward Milton as a sub-$20bn event as more favourable signs on the loss development emerge.
Even so, other early signs still suggest a large gap to be bridged.
The Florida Office of Insurance Regulation (Floir) claims tally was already up to $3bn from 266,833 claims as of 31 October, three weeks after the event.
Other datapoints to emerge last week include the Florida Hurricane Catastrophe Fund’s projection of ~$4.5bn for its own share of the loss based on modelled outputs, pointing to a Florida homeowners’ industry loss of ~$15bn.
Meanwhile, analysis by Insurance Insider ILS indicates an approximate industry aggregate loss of $20bn-$25bn for Milton, when comparing the net loss estimates of nine insurance carriers for Ian and Milton. These nine were among the first to publicly release their net Milton losses during Q3 results season, and the combined tally is based on assuming a similar market share to that of their Ian losses.
PCS numbers do have a history of climbing.
Sources cite Hurricane Laura in 2020 as a storm where PCS significantly increased its number over time. In Laura’s case, it rose by 110% from the initial pick of $5.4bn to a final $11.3bn.
Hurricane Irma in 2017 – well-known for its lengthy and unexpected loss creep – rose by around 18% from the initial number to $27.4bn at the close.
However, the 4x gap between the initial PCS number and lower-end $20bn industry expectation is “not even in the same ballpark”, as one source commented.
Approached for comment, a PCS spokesperson said: “Our work is for our valued clients, and it speaks for itself.”
Investor communications
Because the gap between the PCS number and the modelled loss numbers is going to be hard to bridge, there is a thread of concern in the industry about investors who may raise the credibility issue over the estimates.
However, several sources canvassed indicated that, on the whole, investors would be inclined to take a lower loss scenario on face value simply as a positive outcome.
Sources suggest that within the context of generally positive expectations on returns, the PCS number may simply be “ignored”, with investors overall satisfied with the marks they are seeing.
It’s also true that even as the storm was threatening a $40bn-$50bn strike, ILS managers were generally signalling that they are well-positioned to absorb a cat loss of that size. That’s a strong stance to be starting from in any circumstances, including if a complex and nuanced narrative around revised loss estimations needs to be communicated later on.
But one cat bond manager cautioned there is “no upside” to communicating a loss number that may be underbaked, adding that PCS numbers have climbed in the past.
Logistical challenge
Even if the PCS figures do rise, this still presents a challenge for buyers of ILWs that have a chance of collecting if numbers go up.
Buffer loss factors are lower now than they were around the time of Irma, and combined with a low loss pick, could mean that capital is released sooner than buyers want.
In Milton’s case, the number would need to multiply by 4x or 5x for significant recoveries from ILWs to be triggered, sources said.
Typically, a follow-up estimate is not released until 60 days have elapsed, which would have run close to the 1 January renewal timeframe.
However, according to a PCS news bulletin to clients seen by this publication, the firm has said it will do an earlier resurvey than is typical and release a new number by 22 November, ahead of US Thanksgiving.
It said this would avoid a holiday season resurvey, allow insurers to have gathered more information and capture more third-party information that could provide insight into large commercial losses.
ILW market sources suggest a few hundred million dollars of limit is in scope to trigger at a PCS loss of sub-$25bn. Most Florida-specific ILWs are said to kick in at around $30bn, with more significant limit starting at around $60bn.
Meanwhile, Florida-exposed, industry loss-based cat bonds total $8.3bn of limit, according to data tracked by Insurance Insider ILS. Of this, around $6.1bn is retro ($3.9bn aggregate retro), while $2.2bn is reinsurance.
Where deductibles are eroded by less than the amount that model outputs suggest, there is a potential basis risk-related hit to sponsors.
The Milton experience could potentially re-open debate over buffer loss factors or lead ILW buyers to rework contracts to allow for leeway in the early development time.
Several ILS market participants noted that any loss of confidence in the index provider could smooth a path for more parametric ILS hedges to develop, although they add that basis risk remains a concern in this too.
Multiple mitigating factors falling together
Sources trying to reconcile the PCS number with manager and model outputs pointed to various factors to try to explain why the loss could tally lower than expected.
These include Milton having travelled parts of its track at lower windspeeds than are implied by its official cat 3 designation, as well as high-quality building standards and properties having already been damaged by Hurricane Helene.
One loss adjuster working the Gulf states said a high-pressure front that came in from the north of Florida and interacted with Milton had destroyed part of the eyewall of the storm, meaning that its impact on the ground was felt with less intensity than typical.
They added that flooding from Helene in North Carolina was much more severe than flooding from Milton in Florida, with washed out roads and bridges in NC preventing people from travelling around the region.
This was not the case with Milton where there was “no real severe damage”, they said.
Other sources said in intensity, Milton was close to a category 2 storm, and that modelling the storm track as a cat 2 produces a much lower loss number than when it’s modelled as a cat 3.
The National Hurricane Centre’s advisory on landfall at 20:30 EDT on 9 October recorded a cat 3-level maximum sustained windspeed of 120mph. However, the four datapoints given in that same advisory were in the cat 1 range of 74mph- 95mph or lower: Venice (78mph), Egmont Channel (77mph), Skyway Fishing Pier (67mph) and Sarasota-Bradenton International Airport (40mph).
High-quality construction work within the storm track may also be a factor, including where properties were rebuilt after 2022’s Hurricane Ian. Others say that Hurricane Helene had already damaged properties in Milton’s track and that these losses would be claimed under the first event.
The relatively low number of commercial properties in the Sarasota region was cited as another mitigating factor.
The question managers are grappling with is whether these factors would limit losses to the extent implied by the PCS initial loss number of $5bn.
Protection sellers will also be on the alert for shuffling claims between Helene on 26 September and Milton just 13 days later.
They will also be weighing up the possible significance of each factor and how they interact as they try to understand the true impact of this storm.