Document Type


Publication Date



pension finance


In this paper we propose a new approach to sustainable public pension funding, as an alternative to: (i) traditional actuarial full-funding policies, on the one hand; and (ii) recent proposals aimed instead at stabilizing pension debt at current levels. Actuarial contribution policies aim to fund liabilities that are wrongly discounted at the expected rate of return on risky assets; and these policies promise to do so with amortization schedules that terminate in a precipitous future drop in contributions, which never materializes. Conversely, recent debt-stabilization proposals (Lenney, Lutz, and Sheiner, 2019a; 2019b) properly discount liabilities at a risk-free rate, but effectively untether contribution policy from those liabilities. Our analysis integrates properly discounted liabilities with investment strategies that may be risk-tolerant to some degree, in a policy framework that more transparently conveys the tradeoffs we face.

We begin with the fundamental equations of motion for assets and liabilities – how these two sides of the ledger evolve with contributions, asset returns, and newly accrued liabilities. From these equations we formally derive the characteristics of steady-state pension funding – which we take as the definition of sustainability. We also derive the set of contribution adjustment parameters that smoothly achieve steady-state – a non-trivial exercise. The resulting contribution schedules differ conceptually from the traditional setup of normal cost plus amortization. Building on previous work (Costrell, 2018, Costrell and McGee, 2020), we examine the steady-state implications of differentiating between the assumed return on assets (r) and the discount rate on liabilities (d). We integrate these insights into a semi-formal social optimization framework to sketch out a contribution policy approach that conveys the tradeoffs between intergenerational burden-sharing, the pursuit of returns, and the cost of risk-bearing.

Series Title

EDRE Working Paper

Series Number