An Ideal Retirement Spending Strategy?

Late last year, Steve Vernon, Joe Tomlinson, and Wade Pfau released a new piece of research (full version here, summary version here) that evaluated many different retirement income strategies according to several different criteria:

1. Average annual real retirement income expected during retirement,
2. Increase or decrease in real income expected during retirement (i.e., does the income go up for down over time),
3. Average accessible wealth expected throughout retirement (liquidity),
4. Rate at which wealth is spent down,
5. Average bequest expected upon death,
6. Downside volatility,
7. Probability of shortfall relative to a specified minimum level of income, and
8. Magnitude of such shortfall.

The report was pretty lengthy, so I put off reading it. But I recently had a long day of traveling (to the White Coat Investor conference), with plenty of time to read.

The report is primarily focused on discussing metrics for testing retirement income strategies, rather than recommending any particular strategy. And the authors repeatedly make the point that no strategy is perfect — it’s always a tradeoff between the different goals/metrics. That said, the authors do spend quite a bit of time discussing one particular strategy, essentially saying, “given our assumptions, this strategy is a pretty good one, when measured by the metrics discussed here.”

In short, the strategy works as follows:

• Delay Social Security until 70.
• For the part of the portfolio that is used to fund the delay, invest in something safe, such as a money market fund or short-term bond fund. (For example, if you are forgoing $150,000 of Social Security benefits by waiting from 62 until 70, set aside $150,000 in something safe in order to fund the extra spending necessary until age 70.)
• For the rest of the portfolio, use IRS required minimum distribution (RMD) tables to determine the amount of spending each year.

And with regard to asset allocation for the rest of the portfolio, the authors write:

“Our metrics support investing the RMD portion significantly in stocks – up to 100% if the retiree can tolerate the additional volatility (which is modest because of the dominance of Social Security benefits). However, the asset allocation to stocks for a typical target date fund for retirees (often around 50%) or balanced fund (often ranging from 40% to 60%) also produces reasonable results.”

Overall, this is basically a combination of several findings that we’ve seen repeatedly from other research over the last several years (including research by these same individuals). Specifically:

• Delaying Social Security is usually advantageous (especially for the higher earner in married couple);*
• It’s good to have safe money set aside in order to fund such a delay;*
• For the rest of the portfolio, a high stock allocation is reasonable (if you can tolerate the volatility) given that you have a significant “safe floor” of income from delaying Social Security;*
• It’s usually wise to adjust spending over time based on portfolio performance, rather than spending a fixed inflation-adjusted amount each year of retirement;
• It’s usually wise to adjust spending based on your remaining life expectancy (i.e., you can afford to spend a larger percentage of your portfolio per year when you are age 90 than when you’re age 60); and
• Using the RMD tables to calculate a spending amount each year does a reasonably good job of achieving the two prior points.

The authors also note that it’s wise to keep a separate emergency fund that would not be used to generate retirement income but which would be used as necessary to pay for unforeseen one-time expenses (e.g., home repairs). And they note that some people may want to make adjustments based on differing goals. For example, retirees who wish to spend at a higher rate during early retirement may wish to carve out a separate piece of the portfolio to fund such spending (and such piece of the portfolio should likely be invested conservatively given that it will be spent over a short period of time).

Again, no retirement spending strategy is perfect, because there’s always a tradeoff between competing goals (e.g., a higher level of spending now, as opposed to a higher expected bequest for your heirs). But a strategy roughly like the one discussed above does appear to be “pretty good” according to a whole list of different metrics.

In his summary write-up, Steve Vernon even concludes with the following:

“I’ve been studying retirement for my entire professional career, and at age 64, I’ve been thinking seriously about my own retirement. This actuary will be using a version of [the strategy discussed above], based on my 30+ years of study. My life-long quest may just be coming to an end!”

*With regard to these three bullet points, I find that it can be helpful to think of them in combination. That is, as you move from 62 to 70, you’re spending down your bonds to buy more Social Security. In other words, you’re shifting your portfolio from “stocks and bonds” to “stocks, a little bonds, and a lot of Social Security” — which is an improvement for most people given that delaying Social Security is, on average, a better deal than you can get from regular fixed-income investments and given that it helps reduce longevity risk.