Dr Alan Punter reports on the development of insurance-linked securitisation to date.
Insurance-linked securitisation (ILS) can probably trace its origins back further than traditional reinsurance. The Code of Hammurabi, the Babylonian king who reigned about 3,700 years ago, covered the practice of “bottomry”, whereby a loan advanced on a ship was forfeit if the ship was lost at sea; this structure closely resembles the most common form of current ILS, often referred to as a catastrophe bond. The first traditional reinsurance contract was reputedly issued in 1370, to an insurer on a voyage from Genoa in Italy to Sluys in the Netherlands.
However ILS as we now recognise it really started in 1992. Following Hurricane Andrew in September 1998 the first insurance derivatives were issued on the Chicago Board of Trade (CBoT) in December that year. These were initially futures contracts linked to an index of US property losses produced by the Insurance Services Office (ISO). The contracts have subsequently been modified several times, with firstly option contract structures being introduced and secondly the index being changed to track just catastrophe losses as reported by Property Catastrophe Services (PCS). Various US regional and national indices, for quarterly and annual periods, are now available, but trading levels have been very low. In mid October 1999 there were just 6,737 open interest contracts, which represents insured limits of only around $13.5 million in total. A similar attempt to stimulate trading of reinsurance risk in Bermuda through the launch of the Bermuda Commodities Exchange in 1997 was suspended in September 1999 due to lack of trading volume.
Far more successful have been the issues of catastrophe bonds and contingent capital structures. Interest was originally stimulated by the California Earthquake Authority in 1995/6, which proposed to secure capacity of $1.5 billion by issuing earthquake risk bonds. Ultimately this layer of the scheme was written as a reinsurance contract, but the interest of the investment banks and reinsurance brokers, had been whetted. Some smaller catastrophe bonds were issued in 1995 and 1996, but the defining and first major deal was probably the USAA transaction in mid 1997.
The typical catastrophe bond structure is as follows, using the USAA 1997 deal as an illustration. A special purpose vehicle or reinsurer (SPV or SPR) is formed to act as a “transformer” between the capital market investors and the reinsured; in this case the SPR was Residential Re, incorporated in the Cayman Islands. The SPR issues bonds to investors, on which it pays interest periodically and subsequently, on maturity, repays the capital. The SPR also writes a reinsurance policy; in this case to indemnify USAA for 80% of any losses due to a single hurricane of Category 3,4 or 5 on the Saffir-Simpson index in the layer $500 million excess of $1 billion, i.e. a total policy limit of $400 million. The special feature of catastrophe bonds is that, in the event of the predefined catastrophic insurance event occurring, the issuer of the bonds may be allowed to defer or defease (i.e. be forgiven entirely) payment of interest and/or repayment of principal. Residential Re issued two types or tranches of bonds. The $164 million of Tranche A bonds were principal protected, and paid LIBOR+273 bps (i.e. 2.73% above “base” rate). The $313 million of Tranche B bonds were principal at risk, and hence paid a higher coupon of LIBOR+576 bps.
The need to include principal protected bonds in the overall structure has diminished as the investor market has widened, and almost all recent catastrophe bonds have been principal at risk. Pricing has also reduced. USAA repeated similar issues in each of 1998 and 1999, which were all principal at risk bonds, paying LIBOR+416 bps and LIBOR+366 bps respectively.
The development in ILS has been driven by demand and supply. Underwriters have been looking for alternative, stable capital. Investors have been attracted to bonds that have a higher coupon (i.e. interest rate) than other bonds with comparable risk of default, as well as the diversification provided by uncorrelated risk - the risk of an ILS bond defaulting is not correlated with the factors such as interest rates or exchange rates that determine the values of other bond investments. Capital markets are also used to pricing and investing on a long-term basis, whereas most insurance and reinsurance is conducted on a one-year basis. ILS development has also been supported by two other factors. Firstly the increasing availability of credible computer models for various natural perils, such as earthquake and windstorm, which can give investors some scientific assessment of the probability of catastrophes occurring, and hence the probability of default on the catastrophe bond. Secondly the development of methodologies by the credit rating agencies, enabling them to give catastrophe bonds investment grade ratings. The rating agencies have in some cases developed their own models, but have also performed extensive diligence on the various commercial packages now provided by the catastrophe modelling firms.
The advantages for insurance companies accessing risk capital by issuing catastrophe bonds include:
• Stability of price and cover, since bonds can be multi-year; some have been issued for five years, seven years and at least one has been for 10 years;
• Access to a new source of risk capital, in case adverse conditions in the traditional reinsurance ever cause cover to become either very expensive or unavailable;
• Clarity and certainty of response of contract - the triggering mechanism is generally simpler than under traditional reinsurance, and there is close to no credit risk where the SPV is fully collateralised.
The attractions for investors include:
• A high risk/reward equation, with coupons higher than comparably rated alternative investments;
• Underwriting risk that is uncorrelated with the other risks embedded in an investment portfolio of equities, bonds and other financial instruments;
• No moral hazard - earthquakes and windstorms are outside the control of the insurance company issuing the bonds;
• The ability to almost mark-to-market an ILS, because the events that could impair most ILS are so large that they would immediately be in the public domain;
• Liquidity - the issuers of the major securitisations such as Aon Capital Markets make a market in ILS, so that any investor can sell his bond holding at any time prior to maturity if he so desires.
Another development in the evolution of ILS has been the variety of trigger mechanisms. The earliest deals were mostly on an indemnity basis, i.e. the bond would exactly pay the ultimate insured losses suffered by the bond issuer, who was usually an insurance company. More recent deals have included index triggers and parametric triggers. Under an index trigger, the bond payment is determined with reference to an index of losses, rather than actual losses; the index may be industry losses, rather than the specific company losses. This principle is taken a step further under a parametric trigger, where the bond payment is determined by the “parameters” of the catastrophe. In the case of an earthquake, this would include the location of the epicentre and the strength of the shake; payment is then determined by reference to a formula or grid of values linking epicentre and strength. Several Japanese earthquake deals have now included this feature, as the names of the SPRs concerned betray, Parametric Re and Concentric Re.
A parametric trigger clearly gives rise to some basis risk, i.e. the actual payment from the bond may not exactly match the actual losses sustained by the bond issuer. However, we can expect to see further parametric trigger deals, since they contain features that attract both issuers and investors. Issuers of a bond do not need to disclose their actual portfolio of risks, either because they regard this as commercially sensitive information, or as reinsurers they do not have site specific information on every risk included in the treaties they underwrite. The risk to investors under a parametric trigger becomes more “pure”; it is solely the occurrence of the underlying natural peril (say earthquake or windstorm), and not the subsequent impact on an insured portfolio of properties.
Further developments have also included a number of swap transactions; these are cheaper to implement than public bond offerings, and so are more suitable for smaller deals, say in the range of limits up to $50 million. There has also more recently been growing interest in weather-related securitisations, where the triggers are determined by the number Heating degree days (HDD) or Cooling degree days (CDD).
One of the factors that has held back the number of insurance-linked securitisation deals has been the relative “softness” of the traditional (re)insurance market in recent years. It was always known how the more technical risk/reward approach to the pricing of risk in the capital markets would provide a “glass ceiling” to rates rising in the traditional markets, with the capital markets ready to come in as soon as traditional markets hardened sufficiently. However it is remarkable how much we have seen the interest in ILS increase over the last few weeks with just the early signs of some hardening in the traditional retro market.
Dr Alan Punter is an executive director of Aon Capital Markets based in London, a company dedicated to developing the securitisation of insurance risk.