A version of paper was presented at the 8th Municipal Finance Conference at The Brookings Institution on July 15-16, 2019.
Risk sharing is an important component of today’s public pension system, as the state and local governments strive to balance growing pension costs and risks as well as the competitiveness of compensation to public employees. In traditional public sector defined benefit (DB) plans, the employer bears nearly all investment risk, longevity risk, and inflation risk during both working and retirement years. On the other hand, the employee tends to be the one absorbing these risks in traditional defined contribution (DC) plans. Under this dilemma, risk-sharing mechanisms such as contingent cost-of-living adjustments (COLAs), contingent employee contributions, and hybrid DB-DC plans, were created.
However, risk sharing has not been widely used in the U.S. public pension plans. Current examples include COLAs in South Dakota Retirement System that depend partly on plan funded status, COLAs in Wisconsin Retirement System that depend on investment performance, and employee contributions in Pennsylvania State Employees’ Retirement System that depend partly on investment performance. Nevertheless, many variants and alternative approaches to risk sharing are possible. It brings a sense of urgency to understand how risk-sharing mechanisms affect costs and risks to pension plans, governmental employers, workers and retirees.
In this paper, Don Boyd, Gang Chen and Yimeng Yin (Center for Policy Research, Rockefeller College, University at Albany) examine the impacts of selected risk-sharing policies on employers and plan members, using a model that simulates a pension fund’s year-by-year finances taking investment return volatility into account (i.e., a stochastic simulation model). The pension plan they model has demographic characteristics of a stylized typical U.S. public pension plan. They assume that it has reached a steady state, with new members each year replacing leaving members in a way that keeps the plan’s overall demographic structure stable; this assumption greatly simplifies their calculations while still allowing valuable insights.
The authors’ simulation results are preliminary but informative. Their main conclusions are:
- The contingent COLA policies examined in the paper, reduce the volatility of employer contributions only marginally. The impact of these policies is more significant during dramatic market downturns than during more normal market conditions.
- The examined contingent COLAs could create a significant benefit risk for retirees. During downturns, retirees could experience low benefits during retirement. The acceptance of contingent COLA policies depends on the risk tolerance and risk preference of plan members and policymakers.
- The examined contingent employee contributions policy, styled after policies in Pennsylvania state retirement systems, also has relatively little impact on employer contribution volatility and total employer cost.
- In some instances, introduction of a risk-sharing policy when a plan is deeply underfunded may be less about reducing risk and more about reducing cost. Employers may utilize the interaction between risk-sharing mechanisms and other plan policies to further reduce cost. For example, the funded-ratio-triggered COLA policies can create incentive for employers to seek a lower discount rate: the lower discount rate would result in higher actuarial liability and a lower funded ratio, making COLAs less likely to be triggered and therefore reducing future benefit payouts. It also could make it easier for a plan to take less investment risk.
The authors did not receive financial support from any firm or person with a financial or political interest in this article. None is currently an officer, director, or board member of any organization with an interest in this article.