Defined Benefi tt o Defined Contribution and Back: Valuation of the Florida Pension Election

During the year 2002, The State of Florida’s 600,000 public employees were given the choice of converting their traditional Defined Benefit (DB) pension plan into an individual-account Defined Contribution (DC) plan with full control over asset allocation and investment decisions. To mitigate some of the risk and uncertainty in the decision, the State granted each employee electing the DC plan an additional option to switch back into the DB plan at any point prior to retirement. This option has been labeled the 2nd election by the State and the cost of re-entry is fixed at the accumulated benefit obligation (ABO) of their pension entitlement. Our paper presents some original analytic insights relating to the optimal time and financial value of this unique 2nd election. We start with a simple deterministic model to provide intuition and conclude with a stochastic model that derives a formal upper bound for the economic value. The conclusions from our analysis differ from the results of Lachance, Mitchell and Smetters (JRI, 2003). We argue that the 2nd election behaves less like a traditional downside-protected put option and more like a linear forward contract. We estimate that the value of this 2nd election is at most 30% of the DC contribution rate and only when exercised at the optimal time. Furthermore, for most State employees above the age of 45, the 2nd election has little economic value since the DB plan dominates the DC plan from day one. Of course, it remains to be seen what percent of Florida’s 600,000 employees will elect to behave rationally with their newfound pension autonomy. 1 The Florida Pension Election: During the year 2002, The State of Florida’s 600,000 public employees were given the choice of converting their traditional Defined Benefit (DB) pension plan into an individual-account Defined Contribution (DC) plan with full control over asset allocation and investment decisions. This new Public Employee Optional Retirement Program (PEORP) has been the focus of intense scrutiny by local and national media because it is the largest such pension conversion in the history of the U.S. and is being viewed by some observers as a potential laboratory for Social Security reform. Interestingly, to mitigate some of the risk associated with this decision, the State granted each employee electing the DC plan an option to switch back into the DB plan at any point prior to retirement. This option has been called the 2nd election by the State authorities and we will adopt this name. The cost of getting back into the DB plan is the accumulated benefit obligation (ABO) of their pension entitlement. The ABO is effectively the present value of that portion of the life annuity (pension) to be received at retirement, based on the number of years of service and salary at the time of computation. For future employees — i.e. those not in the plan at the time the PEORP was initiated — the buy back price will be the accumulated actuarial liability (AAL). Our paper presents some original analytic insights relating to the optimal time and financial value of this unique 2nd election. We start with a simple deterministic model to provide intuition and conclude with a stochastic model that derives a formal upper bound for the economic value. We are careful to distinguish between the financial economic value of the 2nd election — which is the focus of this paper — versus the more vague and controversial pension ‘funding cost’ of providing the 2nd election to the employee. While the former is related to portfolio replication and dynamic hedging of guarantees, the latter depends on various actuarial standards of practice, assumptions and cost methods that are beyond the scope (and interest) of this analysis. We refer the interested reader to the work by Haberman and Sung (1994) as well as O’Brien (1986) for stochastic models of pension plans that are focused on actuarial funding methods. Most importantly, the conclusions from our analysis differ from the results of Lachance,

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.