Sustainable retirement income for the socialite, the gardener, and the uninsured

Abstract

This paper advances the literature on the sustainability of retirement income by making consump­tion a stochastic variable instead of a constant real value, as previous papers have done. The paper continues to make the rate of return and date of death stochastic variables, as Milevsky and Robinson (2000, 2005) do. The sustainability of retirement income depends on the nature of the lifestyle that the retiree chooses. The difference in shortfall probabilities or risk of ruin between the variable cases and the fixed consumption case is significant, and so the adviser needs to take this into account. The difference in shortfall probabilities between making consumption a nonconstant but deterministic amount, and making it also stochastic, is not as important, because it does not reduce risk enough to make more aggressive consumption rates secure. Finally, making consumption correlated with the rate of return, which implies the family adjusts consumption as its wealth changes, reduces shortfall probabilities to a moderate extent. In general, an initial consumption of more than 4% of initial wealth is not sustainable for any likely set of conditions. In the very best case, an initial consumption rate of 6% is sustainable, but we think that case will fit very few people.