Post-Retirement Financial Planning with Discrete Dynamic Programming: A Practical Approach

Abstract

This paper finds the optimal consumption and asset allocation strategy for an Australian retiree who aims to maintain a level of minimum consumption. I use a discrete dynamic programming algorithm, with historical stock return distribution and a regime switching investment model, while taking into account mortality and realistic age pension means testing. The financial plan produced with this approach is compared to the 4\% consumption rule and financial plans produced with alternative assumptions. In numerical examples by looking at the outcome assuming the retiree retired in one of past 86 years and followed these financial plans. Dynamic programming method produces flexible financial plans which are robust to the form of investment return distribution assumed, and provide superior outcome compared to constant consumption and investment rules. Taking into account of the required consumption floor or age pension means testing in dynamic programming process helps to smooth consumption, but provides little improvement in term of average consumption level or financial safety. And the optimal investment strategy is found to depends essentially on current market states only. Base investment strategies on retiree's age, wealth, future market states and transitions provide little improvement in the outcome of the retiree.