Tuesday, February 28, 2012

"Pensionize Your Nest Egg" Part 2

Today I’m finally getting to Part 2 of my review for Moshe A. Milevsky and Alexandra C. Macqueen’s 2010 book, Pensionize Your Nest Egg: How to Use Product Allocation to Create a Guaranteed Income for Life. Part 1 of my review can be found here.
There is a famous episode of South Park in which Butters tries to bring disarray to the town of South Park. But for every evil scheme he considers, it turns out that The Simpsons already had a plot point about the same thing.
Moshe Milevsky is The Simpsons of the retirement planning research world. I really need to sit down and read everything he has written about retirement planning, because the second half of this book is almost surely based on one of his research articles, and it already effectively accomplishes what I had been thinking about trying to find a way to do myself.
In the second half of the book, Milevsky and Macqueen describe a very elegant set of risk metric tools (again, does anyone have a better name for “risk metric tools”?) to compare retirement income strategies related to both asset allocation and product allocation.
The first risk metric is “The Retirement Sustainability Quotient” (RSQ). This is basically the strategy’s probability of success (related inversely to the probability of failure). They define RSQ as:
RSQ = (Fraction of Income that is Pensionized) +
       + (Fraction of Income that is NOT Pensionized) x
        (1 – Portfolio’s Probability of Ruin)

You can increase your RSQ by:
-annuitizing more of your assets
-lowering your withdrawal rate
-adjusting your portfolio asset allocation to minimize the failure probability for a given withdrawal rate

The second risk metric is “Financial Legacy Value” (FLV). It’s not defined precisely, but I imagine it must be the average bequest value at death for the strategy across the Monte Carlo simulations, or something very close to that.
You can increase your FLV by:
-annuitizing LESS of your assets
-lowering your withdrawal rate
-increasing your stock allocation
-buying life insurance

Spending less will raise the RSQ and the FLV. With a low enough withdrawal rate, annuitization isn’t really necessary because it will lower the FLV, but the increase to the RSQ will be miniscule. With a high withdrawal rate, it does mean that one’s retirement plans are not realistic or sustainable, and the FLV will basically be zero either way, but annuitization will at least help increase the RSQ somewhat.
For a given withdrawal rate, there tends to be a tradeoff between the RSQ and the FLV. Annuitizing more assets means increasing the RSQ but lowering the FLV. Increasing the stock allocation may lower the RSQ (if the withdrawal rate is not too high), but it will increase the FLV.
(Those above sentences may read better if you replace RSQ with "success rate" and FLV with "expected bequest".)
Related to what I was discussing yesterday in my review of Joe Tomlinson’s article, retirees need to decide how they evaluate the tradeoffs between the RSQ and the FLV, and they also need to jointly decide on a withdrawal rate, a product allocation (between systematic withdrawals from their savings and various types of annuities), and an asset allocation that will get them to a level for the RSQ and the FLV that they feel comfortable with. That, simply, is how to go about determining one’s retirement income strategy.
Also, related to Joe's article, the RSQ and FLV are two risk metrics. Perhaps as a way to do more research about this (though Milevsky may have already done it - I wouldn't be surprised) is to add a third risk metric: the amount of spending enjoyed in retirement. Spending more means a more enjoyable retirement (a good thing) but also a lower RSQ and FLV (two bad things). Utility analysis could be used to find the most satisfactory combination.
I think this is a very good book that very clearly explains how to go about finding a sensible retirement income strategy. As I mentioned in Part 1, the book was written specifically for Canadian audiences, but that will help you empathize with how they feel always reading about the US (and jealously eyeing our inflation-adjusted single-premium indexed annuities). You should be aware of that, but it really doesn’t distract too much, and I give this book a big thumbs up!


  1. With extensions of work I've done and some different extensions to work that Wade has done with Michael Finke and Duncan Williams (see March JFP), I think we are going to be able to use utility measures to do all of the following: (1) factor in different levels of consumption,(2) recognize not only the probability of plan failure but also the magnitude e.g., 1 year of poverty versus 10, and (3) recognize the utility impact of variability of bequests (a sure bequest of $500,000 is worth more in a utility sense than a 50% chance of zero and a 50% chance of $1 million). (If I interpret Milevsky's measure correctly, he uses Expected Bequest, which would be $500,000 for both these cases.)

  2. Thanks Joe!

    About the variability of bequests: isn't the overwhelming main factor on the variability of bequests going to be the uncertain age of death (at least from the perspective of portfolio leftover funds)? I'm curious how you will capture this factor (3), as it isn't something I've thought about.