Christopher McGhee looks at recent developments in the catastrophe bond market

Catastrophe bonds were developed in the mid-1990s to facilitate the direct transfer of catastrophe insurance risk from corporations, reinsurers and insurers to investors. They are designed to protect insurance companies from severe, catastrophic events such as Hurricane Andrew and the Northridge Earthquake, by providing an alternative to traditional reinsurance that is targeted at layers of risk with very low annual loss probabilities, typically less than 1% per annum.

The catastrophe bond market is dominated by East Coast hurricane, California earthquake, European winter storm and Japanese earthquake in terms of the perils and geographies most securitised, though there is increasing examination of securitisation of other geographic locations and perils by potential transaction sponsors. The leading risks, measured by capital placed, are East Coast/Gulf Coast hurricane followed closely by US (including California) earthquake. This is to be expected because these perils are among the largest natural catastrophe exposures for the insurance industry worldwide. Not surprisingly, these risks have also been the most expensive to transfer into the reinsurance markets because they are also the most concentrated risks in reinsurers' portfolios.

Catastrophe bonds have evolved significantly since the early days of the market. In the beginning, issuers lacked the knowledge of what would work in the capital markets as well as how different this market is from traditional reinsurance. In addition, investors were scarce and required substantial education before making a commitment. Today, the catastrophe bond marketplace features a solid core of experienced investors who are becoming increasingly confident in the industry's ability to better model expected losses for a given peril. Furthermore, lower coupons and transaction expenses have driven down the cost of issuance, making the bonds more competitive with the reinsurance market. From the issuer's perspective, the fully collateralised nature of catastrophe bonds provides a measure of comfort at a time when many are focused on reinsurer market security.

Why securitise?

For the issuer community, catastrophe bonds work best as a source of capacity for companies with large risk transfer needs - usually the largest insurance and reinsurance companies. Most, if not all, of these large companies have invested the time to become well acquainted with catastrophe bonds and the marketplace for them, whether they have ultimately elected to issue them or not. Catastrophe bonds pay interest and return principal the way other debt securities do, as long as the issuer doesn't get hit by a major catastrophe that causes losses above an agreed-upon limit.

Using the mortgage markets as an analogy may be the easiest way to understand why insurers and reinsurers, rather than insureds, have been securitisers of catastrophe risk. Single consumers or commercial entities generally find that selling their mortgages directly to the capital markets is too expensive. Instead, they obtain financing from financial intermediaries that package bundles of mortgages together and then sell these to the capital markets. Bundling the mortgages minimises transaction costs and helps structure the risk into separate transactions that meet the various investment preferences of different markets. This process is the most efficient method of accessing the markets that can bear a particular kind of risk.

The insurance risk securitisation market uses a similar packaging method.

Insurers and reinsurers can efficiently provide the necessary skills (product distribution, single risk underwriting and claims handling), which are expensive for investors to acquire and maintain. Insurers can also accumulate large pools of risk through the issuance of insurance policies, warehousing the policies until they are large enough to be packaged into transactions that can achieve economies of scale in transaction costs.

With their large balance sheets, insurers and reinsurers can also retain risks that are not attractive to capital market investors, because they are perceived as either too risky or requiring significant insurance underwriting expertise. Insurance securitisation follows a classic financial intermediation model where financial intermediaries assume risks, retain a portion of them and pass along only those risks that are attractive to investors.

Catastrophe bonds provide full collateral for the risk limits offered through the transaction, and investors may obtain a potentially higher yield with catastrophe bonds than they might have with alternative investments that hold the same level of risk. Another benefit is that catastrophe bonds provide good diversification and show no statistical correlation with other investments, for example, equities or corporate bonds.

These days, catastrophe bonds are currently targeted to provide risk transfer capacity for layers of risk that attaches at the 1-in-100-year (1% per annum) and exhausts at the 1-in-250-year return periods (0.4% per annum). This segment of risk is attractive to both investors and ceding entities for a couple of reasons. First, at this high-severity, low-frequency level, buyers of protection through uncollateralised transactions become increasingly concerned about the counterparty credit risk of their reinsurers.

Second, reinsurance pricing at this level has often been driven by minimum rate-on-line charges, which have made purchases less than economical.

Most sponsors of catastrophe bonds elect to have them rated by one or more rating agency, which aids in the marketing of the bonds, as investors know that the agencies assign their ratings after assessing each bond and its risks based on each rating agency's own proprietary model. The number of potential investors for all types of bonds is typically directly related to the rating of a bond - the higher the bond rating, the larger the overall investor base. There is a sharp reduction of the investor pool for bonds rated below investment grade (below Best's Ratings BBB-/Baa3), although there is still a substantial pool for high non-investment grade bonds (those bonds in the Best's Ratings BB/Ba range). The majority of catastrophe bonds have been rated in the non-investment grade, Best's Ratings BB/Ba range, typically having an annual expected loss in the range of 0.45% to 1.33%. Best's BBB/Baa rated catastrophe bonds generally have a substantially lower probability of loss, in the range of 0.20% to 0.45% per annum.

For a catastrophe bond to achieve a higher rating than Best's Ratings BB/Ba and attain the lowest investment grade rating in the Best's Ratings BBB/Baa range, the ceding entity most likely needs to hold a large retention above its reinsurance program before the catastrophe bond would be triggered.

The BB/Ba range thus represents a balance between what ceding entities want and the ratings that will attract a large, active investor base.

Broader acceptance

In the early days of the market, most insurance and reinsurance companies that considered using catastrophe bonds strongly preferred indemnity-based triggers, where payouts are based on the size of the ceding companies' actual losses. Although this approach gave ceding companies the lowest possible basis risk and most closely replicated traditional reinsurance protection, it also had certain complexities. For example, not only is the ceding company required to provide detailed disclosure on the protected portfolio - sometimes including information that it prefers to keep confidential for competitive reasons - but providing this information in a form that is suitable for a catastrophe bond offering circular can be labour intensive.

From investors' and rating agencies' perspectives, the indemnity-trigger approach requires that they understand a ceding company's portfolio of risk assumed through the writing of insurance and reinsurance. This can be very difficult, especially for complex commercial insurance and reinsurance portfolios. With payouts based on the companies' own losses, bond investors also need to be comfortable that the ceding entity will settle catastrophe claims in a way that would not be a disadvantage to investors.

Consequently, the clear trend with catastrophe bond transactions has been toward the greater use of index triggers. From the investors' perspective, a payout linked to a well-constructed non-manipulatable index eliminates concerns about the ceding entities' claims-handling practices or about general information disadvantages of the investor relative to the ceding entity. From the ceding companies' vantage point, use of index triggers may eliminate the need for burdensome and undesirable disclosure of proprietary underwriting information.

Increased understanding

When the catastrophe bond market began, the primary investors were insurance and reinsurance companies. Because these bond instruments were considered to be exotic, few capital market participants understood them or knew how to evaluate their associated risks.

As non-insurance investors became increasingly educated about catastrophe bonds, the investor base widened. Now, many investors view them as an attractive alternative to other fixed-income investments. Investors are particularly attracted to catastrophe bonds as a source of returns that are uncorrelated with the debt and equity markets. Today, the catastrophe bond market includes a wide variety of investors, including commercial banks, mutual funds, institutional money managers, dedicated catastrophe bond funds and insurance and reinsurance companies.

As investors have become more familiar with catastrophe bonds and their potential benefits, demand has been rising, as evidenced by the increase in capital invested in dedicated catastrophe bond funds. Total funds under management dedicated to investing in catastrophe bonds are expected to exceed $3bn in the near future. In addition, coupon-to-investor and transaction expenses are trending downward so that in some instances, catastrophe bonds are becoming competitive in cost with traditional reinsurance. As investors continue to better understand the value proposition inherent within this class of bonds, the capital base is likely to continue growing at a healthy rate.

State of the market

The way in which catastrophe bonds are structured has evolved significantly since the 1990s, and there is now a well-defined set of attributes that are designed to satisfy the competing demands of investors, rating agencies, regulatory agencies and issuers. In fact, 54 catastrophe bonds issues have been completed since 1997, with total risk limits of almost $8bn.

In 2003 alone, there were a total of eight transactions, with three originating from first-time issuers. Moreover, it was a record year for the catastrophe bond market, with total issuance of $1.73bn, an impressive 42% year-on-year increase from 2002's issuance of $1.22bn.

2003 continued to support the trend toward larger transactions, with the average issue size hitting a new high of $217m. This is a logical occurrence in a maturing marketplace and indicates fewer small, experimental transactions; a deeper investor base and issuers' greater comfort with relying on the catastrophe bond market as a source of substantial capacity.

Furthermore, the ability to spread fixed issuance costs over a larger base and investor preference for larger bond issuance amounts (providing more opportunities for liquidity) have tended to encourage larger transactions.

The three-tranche Zenkoryen Phoenix issue of $470m is the largest transaction to date, far surpassing the average issuance of $217m.

Another particularly notable transaction in 2003 was the Formosa Re issuance, the first catastrophe bond issued for Taiwanese earthquake risk. Central Re sponsored the Formosa Re issuance, making it the first non-Japanese Asian peril securitised to date. The three-year, $100m indemnity transaction covers potential losses to the Taiwan Residential Earthquake Insurance Pool portfolio of residential earthquake insurance policies. In addition, this transaction is the first to use a variable coupon feature dependent on changes to the underlying risk portfolio.

In 2004, USAA sponsored the issuance of a $227.5m catastrophe bond through Residential Reinsurance 2004 Ltd. The bond provides USAA with catastrophe reinsurance protection from hurricane and earthquake events in the US.

Initially, USAA intended to raise $150m, but the offering size was increased over 50% due to strong investor demand. This is the eighth year in a row that USAA has sponsored a catastrophe bond, making them the largest sponsor to date at $1.91bn.

Right for your risk?

Securitisation offers a large source of diversified capacity, often with superior credit quality. Such products usually work best where there are large capacity requirements and the risks are fairly straightforward - a combination of attributes that can lower the cost of issuance while being attractive to investors. As a general rule, however, traditional reinsurance is better for complicated risks and those in which capacity needs are smaller. Nonetheless, the capital markets are sure to continue supplementing the capacity that is currently available to ceding companies and will play an important role in covering potentially devastating risks like earthquakes and hurricanes.