Reply to “Survivor Bonds: A Comment on Blake and Burrows”

Survivor bonds are bonds whose future coupon payments depend on the percentage of the whole population of retirement age (say 65) on the issue date still alive on the future coupon payment dates. Such bonds, if they were to be issued, would make ideal assets for matching the liabilities of life annuity providers in the presence of mortality risk. Kevin Dowd makes the following comments concerning the proposal made in the above article in the JRI for the state to issue survivor bonds: * He believes the proposal is feasible in principle but recommends that the bonds should be issued by the private sector rather than by the state on the grounds that the mortality risk contained in them can be hedged using, for example, reinsurance, dynamic hedging, or even securitization; * He believes that the relevant population for determining the coupon payment calculation should be the population of surviving annuitants, rather than the whole population; and * He proposes extending the concept of a cash survivor bond to cover survivor derivatives. We did consider whether the private sector would be in a position to issue survivor bonds, but we did not explicitly consider all the key hedging techniques mentioned by Dowd. Although these techniques are the standard ones for laying off risks, they work best when the risks being laid off are specific or idiosyncratic in nature, even if they are also potentially large risks such as the weather. The risks with survivor bonds are aggregate risks of potentially very long duration, and the issuer cannot revise the terms of the extant bond in the light of experience. The issuer can only change the terms of any subsequent bond issued, so any costs of underestimating mortality improvements are shared between the shareholders in the company issuing the survivor bond and the buyers of future survivor bonds. The greater the share of this cost passed on to future survivor bondholders, the less attractive these bonds will be to investors. While future survivor bond purchasers might not themselves experience mortality im provements that are unanticipated, they will nevertheless experience a mispricing of these privately issued bonds, and this provides another justification for the government to issue the bonds if it does not attempt to pass on the costs of underestimating mortality improvements to future purchasers. We did consider one hedging technique proposed by Dowd, namely life assurance (or endowment) contracts. These would work reasonably well only if mortality improvements were evenly spread across all ages. However, the mortality improvements over the last 20 years experienced in developed countries have been concentrated in the older age groups, and this uneven improvement in mortality has lowered the effectiveness of this particular hedge. An analogy would be the different effectivenesses of using short-term bonds to hedge long-term bonds when there are parallel and nonparallel shifts in the yield curve. For these reasons, we felt that it would be more efficient for the state to issue them. Furthermore, nothing can prevent private sector companies from issuing survivor bonds and using the hedging techniques suggested by Dowd, yet no such bonds have been issued in the modern age by private sector issuers anywhere in the world. Dowd argues that pharmaceutical companies and other private sector firms geared toward the elderly market would be natural issuers of survivor bonds. This may be so, but I envisage two problems with this suggestion. The first deals with the depth of these companies' pockets. …