Comparing strategies for retirement wealth management: Mutual funds and annuities by Pang and Warshawsky in the Journal of Financial Planning

This was a wonderful, must read and long overdue paper in the latest issue of the Journal of Financial Planning which will help advisers (and sophisticated do-it-yourselfers) better understand available decumulation strategies and the fundamental flaws associated with GMWBs. (One of my earlier blogs on GMWBs (GMWB II) is the most read page at RetirementAction.comexcept for the Home, Nortel Pensions and AboutUs pages). Clearly it’s a topic of interest to many professionals. This paper coming from individuals of stature in the actuarial field will no doubt help persuade more professionals to clearly understand where there is value and where not.

The authors compare six strategies that individuals in retirement might use, such as:

1. Systematic withdrawal (a fixed percentage of the portfolio each year)

1. Immediate fixed annuity

1. Immediate variable annuity

1. Immediate variable annuity with GMWB

1. Mix of fixed immediate annuity and systematic withdrawal plan (partial annuitization)

1. Mix of systematic withdrawal plan with staged but ultimately complete annuitization

After clear explanations of the various approaches the authors proceed, using historically based distributions, to compare outcomes for the six strategies in terms of real wealth and real income associated with the strategies using specific fee assumptions.

Not surprisingly, the conclusion is that fees are a key determinant in the outcome when comparing approaches and that the heavily hyped GMWBs don’t “dominate systematic withdrawals in real terms” even at the high mutual fund fees (1.2%) assumed for the comparables (instead of low ETF/index-like fees). Options #1 (it wasn’t clear if withdrawal rate was set to 4.5 or 5%) and #5 are clearly the best in most typical cases where there is some level of bequest motive. The use of historical (1962-2008) data, while may not be predictive in nature, added credibility to the results.

If the work was expanded, additional strategies for considerations that would be very interesting:

1. Substitute low cost index funds or ETFs for mutual funds. Today with the shift to passive over active implementation, these funds can be obtained from reliable sources for fees of 0.1-0.2% (instead of the 1.2% used for the mutual funds). This is the cost structure available to retail investors today and is thus the appropriate baseline. The GMWB (which you’d be hard pressed to get for under 2.5% and more like 3.5% in Canada) inferiority moves from being subtly clear to overwhelmingly so.

1. Include consideration of a pure longevity insurance product (payout stream only if you live past 85) rather than a traditional annuity (with an investment component). (As readers of this website know, I have been looking at alternatives to annuities for Nortel’s Canadian pensioners which can be expected to deliver superior outcomes to annuitization. For those interested in these alternatives the simulated example in Doomed Nortel pensioners? Outside-the-box pension options and path to pension reform might be of interest, as it discusses the “longevity insurance” option coupled with a fixed percent systematic withdrawal plan. Here options are compared in terms of their real NPV to end-of-life, thus including both the income stream and the residual estate value.)

1. Simulation of a phased (like in #6) but applied to the partial (like #5) annuitization might have interesting characteristics.