On the non-correlation, or otherwise, of catastrophe bonds. And drooling

The subject of correlation among asset classes is one which can get institutional investors excited. In financial circles, correlation is a measure of how two different investments move in relation to one another and is often used to describe how specific market or global events can affect two different kinds of asset class.

For an asset class to be considered truly non-correlated, it would mean that it has absolutely no reaction whatsoever to a certain market moving event which affects other asset classes in a portfolio. However as we all know, there are events in this world which are so severe, and have ramifications so far-reaching, that truly non-correlated asset classes are few and far between, perhaps even non-existent.

Still, the institutional investor search for a truly non-correlated asset class goes on, as they look for niche assets which can provide a guarantee of return when all else is moving south. But this search for non-correlation is a little like the hunt for the Holy Grail, you might find something you think has no correlation with the rest of your portfolio only for a world-changing event to come along and dash your beliefs.

The catastrophe bond market has often been held up to be an example of an asset class with no correlation to the wider economic environment and the majority of the rest of the investment universe. When stocks go up catastrophe bonds don’t follow them and when stocks go down it has no effect on catastrophe bond returns. This much is true in the majority of scenarios and any investment manager or advisor worth his salt is well aware of this fact.

So yesterday Bloomberg published an article featuring comments made at the Standard & Poor’s Insurance Conference held in New York. The article is titled ‘Drooling Cat-Bond Investors Overlook Risk, Montross Says‘, which is taken from comments made by Berkshire Hathaway executive, chairman and CEO of reinsurer Gen Re, Franklin “Tad” Montross.

Tad Montross said that pension funds that have increased allocations to catastrophe bonds in order to improve their returns may withdraw their capital from the market after a major disaster strikes. Certainly this is a possibility and this should be levelled at all third-party investors in catastrophe reinsurance, not just pension funds. As we’ve said before the market remains untested and there is no guarantees for how investors will react to significant losses.

However pension funds typically aren’t scrambling for returns when they allocate to catastrophe bonds, rather they are looking for a new diversified, alternative investment which promises a return comparable with some of the best asset classes at this time. The allocations pension funds make into cat bonds are typically very small percentages of their overall assets, we see no cases of rash deployment of capital to improve returns at this time.

Zeroing in on catastrophe bonds specifically, Montross said; “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.” A fair comment although it has to be admitted that of all the many investors we speak to at Artemis, we have yet to meet one that drools over a 3.5% return on a cat bond.

We actually know some investors who hold back from lower coupon cat bonds as they don’t meet their return ambitions. Similarly we know some ILS investment managers who feel some recent bonds may be a little underpriced and so did not invest in them. Of course these deals do meet some investors risk and return profiles, a sign of the broadening ILS investor base, and that is how they get to market successfully.

It’s worth noting that Bloomberg explains that Montross has previously criticised investors for pouring capital into the reinsurance sector. He’s called that capital fickle compared to the balance sheets that reinsurers have built. He’s quoted as saying that after a major catastrophe event investors could have an “emotional reaction” and exit the market, which is of course a possibility and even a certainty that at least a few investors will exit under extreme circumstances.

Montross then turns to correlation, saying; “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.”

Here we take issue. Catastrophe bonds have been discussed as non-correlated, uncorrelated, having low-correlation or a lack of correlation, but these are sophisticated assets which are sold to qualified investors and funds. Saying that a cat bond is completely uncorrelated or has no correlation at all to other asset classes would, we feel, seem a strange statement. Nothing is uncorrelated or safe when the largest, market moving and world-changing, catastrophic events occur, as we’re sure you’ll agree.

The example of a $150 billion earthquake loss is a good one. Imagine that San Francisco is levelled by a magnitude 9.0 earthquake, Los Angeles is severely damaged and much of the most expensive real-estate in California just doesn’t exist anymore. Catastrophe bonds would be hit hard by such an event, we know of a few that would certainly be triggered losing investors their full principal. Even if cat bonds weren’t triggered they would see mark-to-market losses on secondary prices and the wider reinsurance market would be in for a hefty bill.

The stock market would be hit hard too. Imagine losing the head offices of Apple, Google, Twitter, Facebook and most other Silicon Valley companies along with many of their most talented staff. That would be enough to hit the stock markets very hard indeed, let alone the devastation of one of the U.S.’s largest cities and thousands of people dead.

A $150 billion earthquake loss event, in an area where there are catastrophe bonds in force, would show a very strong correlation between the cat bond asset class and broader financial market movements. This is a completely true statement.

Our issue with the comment from Montross is that it is hard to believe that any pension fund investment manager has actually deployed capital into the catastrophe bond asset class believing that it is some magically uncorrelated asset which will never move in relation to other asset classes or financial markets. No responsible investment advisor, manager, hedge fund researcher or asset class analyst would ever claim that catastrophe bonds and stock markets will never move in tandem.

Of course there may well be a few investors who have managed to invest in catastrophe bonds who really do think they are totally uncorrelated. If there are, and we doubt that they would be many, this is not the asset class for them and we wouldn’t expect their capital to remain in the market after any sort of loss event occurs.

So what can be said about catastrophe bonds and correlation. You can say that the catastrophe bond asset class is largely non-correlated to all but the most extreme world events, that is a reasonably safe statement but there may be outliers in smaller events which could also impact the market (but we can’t think what they are right now). You can also say that as an asset class catastrophe bonds offer a return which is broadly uncorrelated with the wider financial market, because it is. You can even say that catastrophe bonds offer a low-correlated source of alpha in the alternative investment market, again a fact. If you compare cat bonds directly with other asset classes you can get even more certain of the lack of correlation, the fewer comparisons the lower the correlation is likely to be.

But, to suggest that catastrophe bonds are completely non-correlated, or indeed to suggest that pension fund investment managers believe it, is in Artemis’ opinion too simple a view point. Anybody, investment manager, advisor or otherwise, who promotes the asset class as truly non-correlated is not telling the full story.

The typical investor in a catastrophe bond transaction is sophisticated, armed with a wealth of experience deploying significant sums of capital into alternative asset classes, has information on the transaction, has access to risk models and is likely well-informed by investment counterparties. They are not average retail investors who may have read that cat bonds were a good bet in a reduced interest rate environment.

Montross also discusses risk models and states that these are giving “an aura of credibility” to pricing. Montross commented; “Anyone who’s in the industry knows that the models are always wrong. Directionally, they’re helpful, but we are trying to price a risk today that we do not know the cost of.”

This again is interesting as finding a catastrophe bond investor who would tell you that they 100% believed the output of a risk model would likely be like finding a needle in a haystack. Models are directional and by their very nature are wrong as they output probabilities and percentage confidence, not certainty. Within the range of probabilities output by a risk model is likely a form of the truth or something close to the truth, but it’s the ‘directional’ view of risk that is valuable to and expected by their users. Models provide an extremely valuable view of the risk which is typically just one part of an investment managers decision-making about whether to deploy capital into a cat bond or not.

There have been a number of recent cat bond deals which do seem to be more risky and offer less return than the market has historically been comfortable with, a function of the capital inflows and pressure put on pricing. However these are still meeting some investors mandates and hence get completed.

The market has matured, investor sophistication has increased and investors motivations or outlook on investing in catastrophe risk has changed or matured to some degree as well. There are now more investors in the market who look on allocating capital into catastrophe risk as a form of risk trading rather than underwriting, and as we said before these are sophisticated investors, this seems to be a natural progression for the market at this time of rapid change and growth.

Will mistakes be made? Yes, of course. Investors will over-expose themselves to certain perils, funds will buy into cat bonds that are perhaps more risky than their mandates demand and even pension fund investors will get burnt when Florida next has a major category hurricane landfall on a metro region.

These things will happen and people will leave the market because of it, but the general consensus is that while some money may exit more may enter. Until this happens we have no way to know how investors will react. It’s also worth pointing out that nobody has ever said that the cat bond market would last forever, new ways to deploy capital into catastrophe risk and reinsurance as an investment may well become more attractive to investors at some point in the future. The asset class is much broader than just cat bonds now, something that traditional reinsurers are now getting to grips with.

Finally, to suggest that pension fund investment managers are drooling over the catastrophe bond asset class is, it seems to us, a very strange statement to make. Their desks, if not their whole offices, must be flooded with drool, as the average time a pension fund spends on researching a new alternative asset class like cat bonds is typically measured in years.

This is not an asset class that a pension fund can jump into speculatively or on a whim. It takes time to research the asset, review the research, introduce it to the pension funds investment advisory board, find the right investment counterparties, gain approval to allocate to it, free up assets by offloading some of its portfolio and finally actually buy cat bond notes directly or allocate to a catastrophe bond fund or ILS fund manager. This is not something that a pension fund would undertake lightly.

The catastrophe bond asset class, and catastrophe or reinsurance risk-linked investments, provide minimal correlation to the rest of an investors portfolio, a great source of return and a valuable source of portfolio diversification. This is why they are popular and it is to be hoped that the vast majority of investors in cat bonds, pension funds or otherwise, are already fully aware of this.

Catastrophe bonds are not truly non-correlated, nothing really is. When talking about world-changing events, global crises and the largest of natural catastrophes, discussion of correlation becomes a moot point. As the famous scientific saying goes, ‘Correlation does not imply causation’, but that doesn’t mean correlation can’t happen in those one in a thousand edge case events.