Some clarity on cat bonds

Wherein I chide Felix Salmon, the fine Reuters financial blogger, for his views on the cat market.

Mind you, I agree with his conclusion – that the catastrophe bond market is likely to remain small compared to the catastrophe reinsurance market. I just disagree with his reasoning.

Salmon is playing off of a Business Week article that describes nicely what a catastrophe bond is:

An insurance company issues bonds to financial investors, such as hedge and pension funds, that are willing to place a bet on the probability of a disaster occurring at a particular location and during a specific time frame. During the life of the bond, the insurer pays investors a coupon interest rate. If nothing happens, the insurer returns the money when the bond matures. If the fates are cruel, cat bond investors kiss off all or part of the principal.

And this is newsy because defaulting cat bonds will fund well less than 5% of the estimated $30B or so in Japan earthquake/tsunami losses.

Salmon maintains that “catastrophe bonds are the capital-markets security of the future, and they always will be,” maintaining that:

  1. “Insurers will always accept lower returns on their capital than the kind of ROI that hedge-fund cat-bond buyers are looking for.”
  2. Bondholders want the default to be triggered by the objective scale of an event (a Category 4 hurricane passing over latitude X and longitude Y with a windspeed of Z) while reinsurers pay losses based on the damage a storm does, which is harder to control, as it’s subject to vicissitudes of building codes, contractor costs and judicial fiat. The chance that the objective scale would be triggered when the damages don’t occur – or vice versa – is basis risk, and the reinsurer who wants to lay cat risk off into the capital markets doesn’t want to be stuck with it. Nor do bondholders.

I agree that cat bonds won’t dominate the markets, and basis risk is a big reason.

But I don’t think the former point holds up well. The catastrophe reinsurance business can have very attractive returns; however, you have to accept variability in those returns. The purest example of catastrophe reinsurance is probably Renaissance Re, and it has an average ROE of 23% since its formation in 1993, according to this investor presentation (pdf), found here. That’s a bit better than the typical cat bond, which has returns around eight percentage points above Libor, according to this Business Insurance article.

And traditional cat reinsurance should have better returns, because cat bonds bear less risk. Few Japan bonds defaulted because most had triggers tied to earthquakes closer to Tokyo. And in general, cat bonds are structured so the probability of default is remote. From the Business Week article:

“It’s almost like hole-in-one insurance,” says Nelson Seo, co-founder of Fermat Capital Management in Westport, Conn., which oversees about $2 billion, including cat bonds. “It’s been very good returns, and most of the investors in this space have been very happy with it.”

Bond investors don’t like defaults, and a cat bond structured to default when, say, a Category 2 hurricane hit Florida wouldn’t find much market support.

That fact by itself means most catastrophe losses will remain in the (re)insurance industry. Beyond that, I’m told, cat bonds are more difficult to structure than traditional reinsurance treaties, particularly if the bonds are subject to regulatory scrutiny.

A reinsurance treaty, by contrast, is a contract among a half-dozen parties, perhaps fewer. As the reinsurance world is small, the parties know each other well. The contracts are highly standardized, swapping in well-known clauses that vary in predictable ways. Processing the paperwork is relatively easy – so easy that it’s easy to forget to do it. The industry has needed policing to ensure that the contracts get signed promptly.

Besides, quite a few hedge-fund managers like the returns on cat reinsurance, even at lower layers. They embrace the variability since it is lightly correlated with overall market returns.

They can invest directly in a company or fund a sidecar – a short-term reinsurance company that takes a portion of the catastrophe business that a larger reinsurer writes. Alterra recently created one with Stone Point Capital, a private equity firm, in the wake of the Japan earthquake.

The bonds do have their place – they let reinsurers lay off their accumulated risk in extreme events, like a Tokyo earthquake. And like Salmon, I don’t think they will ever replace reinsurance. However I think the reasons have a lot to do with probability of default and less with the promised returns.

Incidentally, Salmon’s post gets bogged down in a couple of issues that, while not critical to his main point – or mine –  I’m going to address, mainly to clarify things.

First, net written premiums for worldwide property business didn’t total $425 billion last year. That number is actually a reasonable estimate of 2010 written premiums for all U.S. property-casualty business, based on June results.

So that number would include auto insurance, medical malpractice and more than 20 other lines of business. It would exclude (of course) business written outside the United States. (I don’t blame the blogger. The pdf he relied on was vague. But here’s p/c results as of second quarter, and I don’t think the numbers are similar by coincidence.)

I mention it because Salmon then compares it to Swiss Re’s estimate that catastrophes cost insurers $43B worldwide in 2010, leaving the impression that cat reinsurers ran 10% loss ratios. (And last year was a heavy one for cats.) Salmon recognizes some of the incongruity, noting that property insurers pay losses in addition to cats, but he really has an apples-to-oranges problem, with worldwide property cat losses being compared to total U.S. p/c premium.

But Salmon was in a tough spot. I honestly don’t know where you’d get worldwide cat premiums (Guy Carpenter?) Few companies that write the business report the premiums separately, and many treaties cover catastrophe risks but that coverage is rolled in with less headlining losses, such as fire.

But for an estimate of cat loss ratios, you can turn to Renaissance Re’s annual reports – a rough proxy to be sure, but a good sense-check. It reports results from its cat segment separately from its other business, and scrolling through its 10-K (p. 119 of the pdf found here), I calculate RenRe’s cat segment posted net accident year loss and LAE ratios of 70%, 12% and 43% over the past three years, which averages out to 42% – a number quite close to what some companies budget for cat losses.

Tagged ,

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: