[antasvara]

There are a lot of interesting dynamics in this market.

For example, CAT bonds are generally tied to the specific natural hazard ("this bond triggers if a hurricane of Category 3 or higher land falls in this segment of Florida") or to industry losses, as estimated by an agreed upon source.

This means that a CAT bond is correlated with, but not directly informed by an insurer's actual loss experience. Traditional reinsurance (so an insurer themselves getting insurance) will usually be tied to specific policies, so their experienced loss is what determines payout.

However, depending on the insurer's policies, traditional reinsurance may be unavailable or much too expensive (either due to the large limit needed, the risk level of the policies, or any number of other reasons). Depending on the trigger, a CAT bond can also pay out faster because you don't have to wait to see the claims from 100k home insurance policies.

From the technical side, most large reinsurers license CAT modeling software from one or both of the same two vendors: Moody's RMS or Verisk. The biggest reinsurers will develop their own models, and there are other modeling vendors that they may license for particular perils (EQEcat for earthquake and KatRisk for flood come to mind), but the big two are pretty widely accepted in reinsurance markets.

That means if your policies are "odd" in some way (uncommon construction type, power facility, etc.), depending on how a reinsurers chooses to model them (or how the model specifically handles them) can have a big impact on your reinsurance pricing. If you know something about your policies that can't be incorporated into a vendor model very well, you may get better pricing on a CAT bond.

These are just some of the considerations! There are so many more things that go into it. But I think it's super interesting to think about.

Source: I work in this side of the industry, specifically in natural catastrophe modeling.

[rrjjww]

Correct on every point and great insight, hello fellow insurance person.

I will clarify that CAT bonds can have industry loss triggers OR actual indemnity triggers. If an indemnity trigger then the insurer has to prove the actual loss. But you’re right on ILWs (Industry Loss Warranty) in that there is additional model/basis risk considerations.

Insurance companies try to minimize this basis risk. Because while sure it’s great to be in the situation where your CAT bond recovers when you didn’t have large losses, it’s NOT good to be in the position where you had big losses and you don’t recover. Certainty of recover can affect things like how much regulatory credit you get for your reinsurance.

[rogerrogerr]

Curious what major trends you’re seeing in your line of work. Guessing global warming induced increase in weather related hazards in some places - anything that would be surprising to people outside the industry?

[Zenst]

Having worked on reinsurance software in the 90s, one question that springs to mind, which came to light from the asbestos claims era, was brokers commission. What did happen was brokers would package up risks and sell those off (taking commission) which would see other brokers bundle those up and again package them and others up into a bundle and sell those off. So when a claim came down the line, that was huge, like the asbestos claim in period https://en.wikipedia.org/wiki/Lloyd%27s_of_London which saw such a diluted risk and brokers commission leaching all profit, brought many down financially due to exposure.

So interested how things are today regarding brokers endlessly packaging up risks they sell on, rinse repeat. I'm aware of certain changes that came about to reinsurance brokers in both the Lloyds and London Markets on the back of the asbestos claim era, but not sure of the the CAT model risks/insurance regarding brokers endlessly packaging up to offset risk exposure vs regulations limiting how much they can do that - more so the USA market.

So curious - is there a risk from brokers diluting risks for commission profits in this market or is that saftly covered against and regulated?

[NoboruWataya]

> CAT bonds are generally tied to the specific natural hazard ("this bond triggers if a hurricane of Category 3 or higher land falls in this segment of Florida") or to industry losses, as estimated by an agreed upon source.

A big part of the reason for this is that if payout is tied to actual loss, it starts to look a lot like actual insurance, which is specifically not what you want. Because while anyone* can buy a bond, only insurers or reinsurers can write insurance. This is something that needs to be considered whenever an insurer (or anyone, really) tries to transfer risk to a non-insurer.