Tad Montross recently tried to throw some cold water on the cat bond party:
“With interest rates being where they are, I don’t think it’s a surprise that a cat bond with a yield of 350 or 500 basis points over Libor looks attractive. People are drooling for those.”
“What happens after the $150 billion earthquake, when Nevada is basically coastline to the Pacific? This whole issue that it’s a non-correlated asset class, which makes it so attractive as people look at their risk-return profiles, is one that really needs to be thought through very, very carefully.”
More plus commentary here.
So here’s a chance to think a bit about all kinds of issues.
Number 1, and most importantly, Montross is biased. Incredibly biased. He competes against cat bonds. If his comments move their marginal prices up he can make more money.
So beware, Dan Kahneman teaches us that we systematically underestimate the effects of bias on others. I could (should?) end the post here because all of his comments should be interpreted as naked self-interest even if it’s not consciously intended.
Number 2, correlation is a brutally simple statistic that glosses over a lot of complexity in risk management. At the link, Steve Evans makes the point that many investors are too sophisticated to actually take the “non-correlation” line at face value with cat bonds. Or any other asset class for that matter.
But I’ll try to make a stronger stand. Low correlation is real and here to stay. Forget sophistication, if you have a portfolio of cat bonds diversified by region and peril, even if a 150bn EQ hits, ripping California from the mainland or whatever ridiculous sci-fi fantasy you want, your cat bond portfolio will vastly outperform every other asset class. It’s diversified, people. There’s US wind, there European perils, Japan, etc, etc. The general economy, on the other hand, would be toast.
Which takes me to point 3: the ONLY source of meaningful correlation to other assets is through the holding portfolio of treasuries or Money Market Funds. If these tank, cat bonds follow. This got found out in the financial crisis where some bonds got into real trouble because Lehman stuck the assets into garbage “AAA” mortgage CDOs. Luckily it was a small subset and we got a chance to learn the lesson (allow me to say that all collateralized deals we did pre-crisis used super-safe assets and didn’t blink in the crash).
Anyone can imagine a sufficiently unlikely scenario where all assets correlate (asteroid destroys earth and it’s correlation city: all assets go to 0). The key is that you need to tie it into your own risk management system. On a 1-in-100 portfolio basis tresauries are still liquid and as risk-free as possible and so cat bonds are non-correlating. End of story.
Not Unreasonable 