Natural Catastrophe Risk - Cats land on their feet
Catastrophe risk delivered positive returns in 2008 amid rising downside correlation but was not immune from credit exposure, finds Martin Steward
Despite journalists' clichés about "financial storms" and "seismic shocks to the economy", there is no scientific evidence that volatile financial markets cause hurricanes or earthquakes (or vice versa). That is why underwriting re-insurance or buying bonds that capitalise such re-insurance seems so robust as a diversifying investment. Sure enough, through 2008 insurers continued paying premiums and returns from cat risk finished the year positive.
But the asset class was not untouched by the financial crisis. Just look at the primary market in cat bonds: after Hurricane Katrina this $1bn (€0.7bn)-per-year niche market turned into a $7bn-per-year snowball, but once the credit crisis got into full-swing - zilch, nothing between summer 2008 and February 2009. There was cause as well as correlation: when an insurer goes from a $100 to a $3 stock in the space of a year (as XL Group did), or has its asset portfolio ravaged, its capital no longer supports the underwriting that feeds re-insurance. But now consider what kind of correlation that ought to create. On top of the wave of post-Katrina cat bonds all expiring, ceteris paribus, that kind of primary market supply squeeze would lead to rising secondary market prices. Cat bonds should be negatively correlated with financial assets in a crisis.
Unfortunately, ceteris was anything but paribus. A chunk of that post-Katrina issuance had been picked up by multi-strategy hedge funds. As their credit assets crashed, their cat bonds were prime candidates for liquidation to meet margin calls. More seriously, LIBOR returns on cat bond collateral were generally guaranteed with a total return swap, with monthly margining at best. It was bad enough that four bonds used Lehman Brothers as counterparty to this swap; also, the collateral contained asset-backed securities, creating a credit duration mismatch that came home to roost when Lehman went bust. Credit risk began to be priced-in across the market.
The majority of bonds are now backed by high-quality money market funds. That has depressed yields, but negligibly - and in any case that lost yield was clearly a credit risk premium. Some investors will miss that spread - especially hedge funds with LIBOR-plus fee hurdles - but the majority of ILS managers will not. As Rob Procter of London's Securis Investment Partners observes: "We are mandated to run insurance risk, not credit duration risk."
The incentives are similar for bond-sponsoring insurers: the better the collateralisation, the more likely claims will be paid out. Furthermore, if they ditch the swap-based structure they can rip 50 pages out of the bond prospectus and save a heap of fees. However, as Niklaus Hilti, head of insurance linked strategies at Credit Suisse observes, only a couple of money market funds meet the requirements to back-up cat bonds. "The industry will have to discuss whether it makes sense to put more and more collateral into these funds," he says. Credit Suisse wants more diversity of collateralisation solutions - if not the daily-margined third-party repo structures some have mooted, then perhaps more like recent issues that used KfW floating-rate notes.
Given the credit exposure, the fact that prices bottomed-out at about 85c on the dollar is impressive. Sandro Kriesch, co-founder of ILS specialist Twelve Capital, has regressed the dollar BB-rated cat bond market against high yield bonds back to 2002. Remove July 2008 through March 2009 and correlation comes in at 0.20. But even with the crisis included it only rises to 0.37. Clearly buyers were willing and able to brave the storm. "Although the primary market came to a halt, the end of 2008 was probably the most active period for the secondary market that Swiss Re has ever seen," confirms director of insurance-linked securities Jean-Louis Monnier. "In late 2008 and early 2009 the market had buyers who hadn't yet been hit by the liquidity crisis," says Michel Queruel, CEO of Bermuda-based ILS manager Pentelia Capital Management.
So who were those buyers, and will they always be around? If sponsors withdrawing from the primary market put downward pressure on secondary market yields, they compounded that by adding cat bonds to their own portfolios: if they can source higher yields from bonds than they can generate from underwriting re-insurance, this potentially provides a pricing floor. But the extent to which they increased their holdings through 2008 is uncertain.
Everyone agrees that the insurance-company buyers were less influential than the ILS specialists, for whom the crisis was "a time of great opportunity", in the words of Greg Hagood, co-founder of Nephila Capital, the Bermuda-based cat risk specialist 25%-owned by Man Group. How did those managers know they were not catching falling knives? The liquidity crunch in credit was largely about investors tearing up old pricing rules - often based on models that assumed financial markets obey the laws of nature. Pricing models in cat risk also assume the market obeys the laws of nature - but of course in this case that assumption is correct. Some argue that this transparency into expected losses makes it easier to agree prices. Others observe that by no means every market player employs prohibitively expensive cat models. More pertinently, the problem was not one of pricing insurance risk, but credit risk: those who bought at 85c must have been happy that these risks were already priced-in.
The more obvious explanation for liquidity holding up is that, while these managers saw redemptions, serious new money was rolling in from institutional investors. Alasdair MacDonald at Towers Watson tells of $2bn of client capital being directed into ILS managers over the past year. They had to put this to work in some kind of cat risk (rather than, say, higher-yielding distressed credit), and bonds were the obvious choice: their two to three-year term structure and functioning secondary market meant that yields drifted up relative to traditional re-insurance, presenting a clear arbitrage.
"When the sell-off from multi-strategy investors started, the higher implied yield in cat bonds made them more palatable," says Luca Albertini, CEO of London-based ILS manager Leadenhall Capital Partners. "In 2009, the lack of new issuance and new money coming in pushed implied yields lower and lower, putting investors at risk of mark to market gains reverting to par due to their relative short maturity. So as soon as the yields fell close to traditional re-insurance the industry showed a pricing floor as traditional re-insurance became more attractive again."
But if replacing hedge funds with pension funds helped keep a lid on yields, it raises the question of whether a bond market that still only amounts to $15bn (the total cat risk market is perhaps $400bn) can meet this extra demand.
"Solvency II will make insurance companies more and more likely to want to go to the capital markets," says Christophe Fritsch, head of ILS at AXA Investment Managers. On the other side, Hilti and Hagood point to the fact that the number of bonds sold is limited simply by the hassle of placements; and that, to get a rating, bonds can only be structured on very low-probability risks, for which limited capacity is required.
Even if they follow these managers' recommendations to allocate across the re-insurance spectrum, and not just in bonds, investors should not expect today's 10-12% yields to persist under the weight of their capital. But the real question is whether correlations will rise, too. We have seen how ILS managers make relative-value decisions between different forms of cat risk. What happens when most ILS capital comes from investors making relative-value decisions between cat risk and credit-related yields?
Most investors probably will not be allocating to cat risk tactically, or even strategically as part of a broader fixed income bucket, but rather strategically as a diversifying alternative asset class. This should reduce the risk that capital will flow based on opportunistic relative-value decisions and increase correlation to the credit cycle. The fact that cat risk managers generally do not offer sufficient liquidity to trade in that way should also help. But these investors may well trade to re-balance portfolios: a pension fund with a 2% strategic allocation to cat risk could find itself with a 5% allocation as the rest of its portfolio is ravaged by a crisis. If it decides to sell half its holding, the correlation of cat risk to financial assets will depend upon the ILS sector's ability to absorb that supply. In future there could be rebalancing-driven sell-offs similar to the hedge funds' deleveraging-driven sell-off in 2008, but on a much larger scale. The argument against this possibility - that a 2-5% allocation is too small for the investor to rebalance - is strong, but it remains a risk for long-term investors to bear in mind.