Monday, February 27, 2012

Joe Tomlinson’s “A Utility-Based Approach to Evaluating Investment Strategies”

Joe Tomlinson has written a very worthwhile article in the February 2012 Journal of Financial Planning called, “A Utility-Based Approach to Evaluating Investment Strategies.” He will visit GRIPS on March 15 to present this research and subsequent advances he has made to it.
An area of research that I am getting more and more interested in relates to developing tools to analyze how to find the best allocation for a retiree’s investment portfolio combined with single-premium immediate annuities (SPIAs), deferred annuities, and variable annuities with guaranteed benefit riders. This combines asset allocation with product allocation (as Moshe Milevsky calls the framework for choosing among different types of annuities).
How to analyze the appropriate asset allocation and product allocation? There are two basic ways. Are innovations for the framework to evaluate retirement income strategies best guided by using various criteria and risk metrics (such as failure probability or value of bequests) which allow direct retiree decisions for evaluating tradeoffs related to spending and bequests, or should the framework specifically incorporate the formal mathematical concepts of utility maximization to help guide retiree choices, perhaps under the recognition that the human mind is not equipped for making decisions among abstract and uncertain tradeoffs?
First let me give a brief answer to that. I think both approaches are useful and mutually reinforcing. Retirees should be free to choose as they wish to find the proper balance among competing tradeoffs. Most retirees will not be happy letting their decisions be guided by an indecipherable black box. They will want to feel they are empowered to make decisions. But with utility, their decisions can also be “double checked” to make sure they sufficiently understood the task of choosing among abstract and uncertain possibilities and found something matching a reasonable set of preferences.
In this article, Joe Tomlinson makes an important contribution. The article actually considers both approaches at the same time. Table 1 shows results for risk metrics (I’d like to come up with a better term for this – any ideas?), while Table 2 shows results for utility analysis.
He considers an inflation-adjusted SPIA with a payout rate of 5.05%, and he assumes that this matches the retiree’s desired spending. He also considers other strategies which combine stocks, bonds, and partial annuitization in varying ways. In later research, perhaps he can work more toward finding an optimal withdrawal rate as well.
In Table 1, he shows how the different strategies perform across 1,000 Monte Carlo simulations based on the average value of the bequest left at death, the probability that the strategy failed to provide wealth for one’s whole life, and the average amount of the losses in the cases of failure assuming that one could have kept spending into a hole.
In Table 2, he considers the matter of utility. He explained very well about the intuition for why he does this: “The problem is that three different variables need to be brought into the decision. This naturally leads to the question of whether a more informative decision tool could be developed by combining the measures. To do this we need to shift our focus from pure financial outcomes to the fuzzier concept of the satisfaction associated with those outcomes. We need to develop a utility measure.”
He uses a loss aversion utility function, in which running out of wealth has a stronger negative impact on utility than the positive utility experienced by leaving a bequest. This approach makes a lot of sense, and it can also be thought of more simply as a way to judge how important leaving a bequest is, compared to not running out of funds. Defining parameters for utility functions can be complicated, but Joe Tomlinson also describes a small survey he conducted to get a better idea about the degree in which losses hurt worse than parallel gains.
What he finds is a stark cut off: 100% stocks are preferred by someone who is not so hurt by losses and/or is more interested to leave a bequest. Meanwhile, 100% annuitization is preferred by someone who feels great pain by running out of wealth and/or is otherwise not interested to leave a bequest.
The article is quite accessible, and it includes lots of interesting gems, such as his discussion of the danger of bond ladders with uncertain lifespans, the role of the equity premium in driving the results, and his ideas for how to make guaranteed benefit riders that would be more worthwhile (“could be very useful if it became available in a low-cost, index-funds version, with an inflation guarantee, but without the commission loads and the charges for active management.”). I hear that, and hope we get there one of these days!