By Bonnie-Jeanne MacDonald (Independent), Richard J. Morrison (Independent), Marvin Avery (Human Resources and Skills Development), Lars Osberg (Dalhousie University)
This paper examines the importance of pensions (employment and social security), taxes and government transfers for alternative retirement savings drawdown strategies, based on Canadian evidence. Using as examples single elderly Canadians at the 10th, median and 90th percentiles of the income distribution, we use a lifetime utility framework to evaluate an illustrative set of six popular drawdown strategies. Our longitudinal dynamic micro-simulation model for Canada incorporates risk aversion, stochastic markets, stochastic mortality and the interactions among sources of retirement income within the complex Canadian tax and social benefit system, enabling us to rank six commonly advocated drawdown strategies and to ask whether incorporating pensions, taxes and transfers alters those rankings.
Our primary finding is that consideration of pensions, taxes and transfers does often alter the rankings of drawdown strategies for retirement savings. Notably, and contrary to nearly all the research on this topic, annuitization is not always the best strategy once pensions, taxes and government transfers are modeled. Second, including consideration of pensions, taxes and transfers differentially alters the ranking of drawdown strategies – the effects are not uniform between males and females, across people at different points in the income distribution, and at different levels of risk aversion. Last, the drawdown strategy choice can be very important to the retirement financial welfare of some seniors, and nearly inconsequential for others – an issue ignored when only the drawdown of savings alone is considered.
Our findings show the importance of treating the evaluation of alternative drawdown strategies as a comprehensive and integrated problem by including all sources of income – including pensions, taxes and government transfers. Using restricted measures risks leading to simplistic, and possibly misleading, conclusions.

Source: SSRN