Repository | Book | Chapter
(2015) Artificial life and computational intelligence, Dordrecht, Springer.
An evolutionary algorithm for deriving withdrawal rates in defined contribution schemes
Kerem Senel , Jason West
pp. 286-296
Risk-averse investors typically adopt a fixed spending strategy during retirement to prevent against the premature depletion of their retirement portfolio. But a constant withdrawal rate means that retirees accumulate unspent surpluses when markets outperform and face spending shortfalls when markets underperform. The opportunity cost of unspent surpluses associated with this strategy can be extreme. We employ a genetic algorithm to find optimal asset allocation and withdrawal levels for a retirement portfolio. Using US and international data we compare this approach to existing strategies that use basic investment decision rules. Our results show that allocations to riskier assets early in retirement generates rising incomes later in retirement, without increasing the probability of ruin. A rising income profile remains optimal under different levels of risk aversion. This finding disputes the safe withdrawal rate conventions used in contemporary financial advice models.
Publication details
DOI: 10.1007/978-3-319-14803-8_22
Full citation:
Senel, K. , West, J. (2015)., An evolutionary algorithm for deriving withdrawal rates in defined contribution schemes, in M. Randall (ed.), Artificial life and computational intelligence, Dordrecht, Springer, pp. 286-296.
This document is unfortunately not available for download at the moment.