Why Not Trade Pension Claims?

Abstract

Trading pension claims would serve many purposes. Beneficiaries would be able to diversify the idiosyncratic credit risk of their plan sponsors. And systematic risk could be reallocated to comply with individual risk–return preferences. The result would be an alignment of companies’ and pension fund managers’ incentives to keep fund plans fully funded—in line with beneficiary interests—which would lower agency costs and costs of government bailouts of defined-benefit plans and would improve the general welfare. As an accurate valuation of pension liabilities, trading would provide a measurable yardstick for plan managers.

Trading defined-benefit (DB) pension claims would serve many purposes. Beneficiaries would be able to diversify the idiosyncratic credit risk of their plan sponsors. And systematic risk could be reallocated to comply with individual risk–return preferences. The result would be a realignment of companies’ and pension fund managers’ incentives with the interests of beneficiaries—namely, to keep fund plans fully funded. Full funding would lower agency costs and the costs of government bailouts of DB plans and would improve the general welfare. As an accurate valuation of pension liabilities, trading would provide a measurable yardstick for plan managers.

We propose a flexible, low-cost mechanism for transferring and transforming credit risks. It is based on a financial structure that we call a “collateralized pension claim obligation” (CPCO). The CPCOs (and possibly corporate entities) would trade pension claims to optimize diversification and balance their assets according to a number of characteristics (age, industry, geography, etc.). Just as loans or mortgages are pooled and tranched in traditional collateralized debt obligations, pension claims would be pooled and tranched in a CPCO.

For example, suppose that, for any of several reasons, a beneficiary became reconciled to the idea that he would never collect the full face value of his pension claim and was uncomfortable with the specific risk of the plan sponsor. In this case, the pension beneficiary could trade in his claim to the CPCO.

In this way, the CPCO would accumulate claims on many companies and would become naturally diversified. In exchange, the beneficiary would receive a claim on the CPCO. These claims would be organized in tranches. The senior tranche would be almost riskless because it would be protected by the other tranches. Mezzanine tranches of intermediate risk would offer lower, weaker guaranteed levels of income against a higher expected income, with the realized future income to be tied to the future value of the CPCO’s assets. The equity tranche of the CPCO would be provided by speculative funds.

A participant would receive a claim on the CPCO tranche (or a mix of CPCO tranches) that she could select according to her affinity for risk and within the context of her overall wealth. Should the participant’s risk preference change, she could exchange her claim against that of another tranche that better reflected her new risk–return profile.

Once a market for pension pools has been developed, designing a benchmark to be used for asset manager evaluation and compensation would be straightforward.

We are grateful to Lucie Tepla for comments on an earlier draft of this article and are also grateful to Ian Edwards for first raising the topic, to Andrew Smithers for previous collaboration on the issue, and to members of the IXIS Bank staff in New York City for discussions of the feasibility of the scheme proposed here.

Notes

¹ Bader (2004) made the point that underfunding as an extra source of financing for corporations is an illusion.

² Moral hazard is the increased risk of detrimental behavior (and thus a negative outcome) because the person who causes the problem does not suffer the full (or any) consequences of the behavior and may actually benefit.

³ The claims traded in by beneficiaries would be limited to those already accrued to them, excluding those related to future employment or future, individual, salary raises. The purpose of this limitation would be to remove beneficiaries’ incentives to reduce their work effort after trading in claims.

⁴ See Vasicek (1987); Lucas, Klaassen, Spreij, and Straetmans (2001); O’Kane and Schloegl (2001).

⁵ In this section, the terms “expected value” and “probability” refer to risk-adjusted or risk-neutral probability distributions.

⁶ Mortality tables for the United States can be found in the IRS Cumulative Bulletin or in the National Vital Statistics Reports. Similar mortality tables are available from governments or private insurers in other countries.

⁷ The number of tranches can be chosen at will.

⁸ Expected losses per tranche depend on the correlation of the claims in the pool and on their recovery levels (i.e., the funding levels in case of bankruptcy). The Pension Insurance Data Book 2003 (PBGC 2003) shows an approximate historical average funding level of 40 percent in case of bankruptcy.

⁹ The factor ξ corresponds to 1 minus the expected loss on the claim handed in, divided by 1 minus the expected loss of the chosen tranche [i.e., (1 – 0.07)/(1 – 0.0025) = 0.932].

¹⁰ To improve the engineering of standardized pools, credit default swaps could also be used.

¹¹ EIB is the European Investment Bank.

¹² For a detailed approach to valuing mortality risk for life insurance contracts and pensions that involves stochastic interest and mortality rates, see Cairns, Blake, and Dowd (2004).

¹³ This position should be compared with the proposals of Geanakoplos, Mitchell, and Zeldes (1999).

¹⁴ According to the official PBGC website (http://www.pbgc.gov/about/about.html).