The Parametric New Kid on the Block

Meanwhile the Caribbean Catastrophe Risk Insurance Facility, which says it’s the largest parametric insurance program in the world, will pay $70.8 million to the island state, its biggest claim ever, potentially with more to follow.

The rather different characteristics of parametric coverage mean that Jamaica’s payments under these contracts were confirmed just days after the storm. The insured already has certainty, and soon will have the cash. In contrast, it’s likely to be many months before Melissa-related payments due under conventional catastrophe re/insurance policies are finalized.

Parametric reinsurance has finally come of age. After thorough testing in the international catastrophe bond market, where it provides sovereign protection (as in the Jamaica example), retrocession capacity and tranches of reinsurance for a handful of large insurers, it’s now a regular component of the commercial reinsurance toolkit. It has found places in programs all over the world, and is widely available to all cedants during the current renewal. Every cat-exposed cedant should at least consider it as part of their reinsurance backstop.

Parametric reinsurance is different to the core: it isn’t an indemnity product. Instead, parametric pays a fixed amount when triggered, like an industry loss warranty. But unlike an ILW, which pays when the aggregate insured loss from a catastrophe exceeds a specified total, parametric payments are triggered by a named-peril event of a stated intensity. Since the event, not the loss, triggers the cover, recoveries can be paid lightning fast, providing a swift injection of liquidity as the claims roll in, and eliminating the investors’ challenge of trapped capital. These advantages may soon make ILW contracts obsolete.

Here’s a simple example of how parametric reinsurance works. A treaty might pay a reinsured $10 million if a windstorm with central pressure below 910 millibars lands within a certain length of southeast Florida’s shoreline. This simple example usually leads to two very good questions:

1. How does that fit into my program?
2. What if the storm hits further up for coast, or has a pressure of 925 millibars, either of which would still inflict huge losses to my book?

A simple answer to the first question is ‘any way you like’. The beauty of parametric is that it has no attachment point. It responds to the event, not to the loss.

In our example, any insurer with exposure to Florida wind is likely to have coverage under a conventional reinsurance program. (We certainly wouldn’t recommend replacing conventional reinsurance entirely with parametric.) Perhaps the insurer has aggregate cover of $10 million above a retention of $5 million, since buying lower down the program remains very expensive, despite the softening market.

If the cedant is hit by a $10 million loss, the parametric would cover the retention and fund reinstatement premiums. If it was $20 million, the insurer would still have change. And if they were hit by a $25 million loss – the largest, perhaps, that their modeling predicted possible – they still would not suffer a capital event.

Put another way, parametric is like a sponge that sucks up losses, wherever they fall. It’s a leading example of the reinsurance sector’s growing willingness to adopt new practices and modernize.

Some CFOs have seen this as an outward buyer’s luxury, not a critical component of a prudent program. Others have turned to parametric because it was the only available substitute for a lost layer of conventional reinsurance, one that other markets seem unwilling to provide under a conventional contract at an economic price, and are eager to return to familiar territory.

To help get more parametric buyers over the line, several reinsurance markets have begun to offer parametric reinsurance that’s triggered not by the characteristics of the event alone, but by its occurrence plus its modeled impact on the reinsured portfolio.

For example, say an earthquake of a minimum magnitude occurs, or causes ground-shaking in excess of a set minimum, in a specified place. Typically the parameters of the event would be confirmed, in the first case, by an authority such as the USGS, or, in the second, by a dedicated in-situ sensor. The appointed model would be used to calculate the loss to the previously defined reinsured portfolio. If the modeled loss exceeds a specified amount – call it the attachment point – the agreed payout would fall due.

This structure makes parametric reinsurance look a lot more like a conventional layer. It has proved popular to cover hard-to-cede perils like SCS, but it has drawbacks. It is slower to settle than pure parametric, since modeling takes time. Plus, it reintroduces a black box into the equation. The transparent event-based trigger is muddied by an uncertain set of necessarily subjective algorithms.

The modeled trigger may well reduce the cost of the coverage, but it also decreases utility. In our opinion, it’s like having your gold medal bronzed.

Modeled triggers exist only because parametric reinsurance underwriters tend to price keenly to the true risk-based cost after extensive modeling, so suppliers can be comfortable with a price that also makes economic sense to the cedant. If they didn’t, conventional named-peril reinsurances would stand in their place.

For weather perils, underwriters tend to incorporate a highly informed understanding of the impact of climate change. For all, they are agnostic to common pricing factors such as loss record, market pricing and internal top-line budgets. It’s what the market called ‘technical pricing’ three cycles ago, and it’s a key characteristic of parametric reinsurance, one which proponents claim will dampen price cyclicality, although we see some evidence that prices may already be easing with the cycle.

Because of this pricing approach, we expect modeled triggers’ nascent popularity to wane as reinsurance markets soften, and undercut the technical price. The near-term exceptions may be those perils that continue to suffer constrained reinsurance supply, like wildfire, flood, and SCS. If and when conventional underwriters begin to cover those perils at prices lower than their parametric competitors, they are likely to unseat their new rivals.

The second question is about the near miss, often described as basis risk. It’s the genuine concern that losses happen, but do not trigger the parametric reinsurance policy.

In our simple Florida wind example, this would be a significant concern, because the contract wouldn’t pay out. But in practice, few new parametric treaties have such a simple trigger. Almost none are all-or-nothing. They almost always incorporate a sliding-scale payout which, for example, decreases with the landfall’s distance from the urban conurbation, or rises with the declining central pressure, or both.

Areas beyond the shoreline might be divided into zones, perhaps according to topography and distance from the shore. The presence of the cyclone at specified intensities could further increase the percentage of the limit paid. Melissa’s central pressure was so low at landfall that its presence in any zone of Jamaica was probably sufficient to trigger payment of the entire limit of the World Bank bond.

Other trigger-and-payment structures have been designed to match the ground truth of different perils. For example, a calculation agent determining payouts following a wildfire may examine satellite images to identify burnt cells within larger, portfolio-weighted reinsured geographical zones. A wider area may be subject to a lower limit to cover smoke damage.

All of the world’s major reinsurance markets now offer parametric reinsurance products, plus a couple of large, reliably backed MGAs. As the tool develops, the flexibility and creativity of those markets’ efforts to structure products which closely match cedant’s actual losses have been impressive and continuous. They are not perfect yet (nor is conventional reinsurance), but every iteration delivers dramatic improvements.

By manipulating the triggers, parametric reinsurance can be made more or less responsive to lower-frequency events. Like any insurance coverage, you get what you pay for: a year after the furor over the non-payment of Jamaica’s big bond after Beryl, the instrument’s response to Melissa shows that it has behaved exactly as its architects intended.

Little in the way of positives have come out of that terrible catastrophe, but it may prompt the leaders of more nations to acquire parametric risk transfer structures, which will help to close the global insurance gap. Meanwhile, parametric is a fantastic additional tool to help eliminate the peak-peril portfolio risks that can worry even the most prudent reinsurance buyers.