Testing a Blended Approach to Retirement Income

My previous simulations of how a bucket strategy would have performed since 2000 used what I call "the strict constructionist total-return" approach. By that I mean that we reinvested all income and capital gains distributions as the portfolio kicked them off, and relied strictly on rebalancing proceeds to meet living expenses. 

But how would that strategy compare to a more income-centric one, where we spend income distributions from bonds and dividend-paying stocks as the portfolio produces them? I'll be testing versions of this approach this week and next. 

This week, let's take a look at how a hybrid of an income-oriented and a total-return approach would have performed since 2000, the start of the dotcom meltdown. For the purpose of the simulation, we assumed that the retiree would spend the income distributions from the portfolio. 

But this simulation differs from a hard-core income-oriented approach in a few key ways, and that's what makes it a hybrid. First, the portfolio wasn't assembled with a focus on income production. (We'll test such a portfolio next week.) Instead, it employs the same holdings in the same proportions that appear in my aggressive bucket portfolio, except that it doesn't include a cash bucket. After all, the idea of more income-centric strategies is that the portfolio will supply most of the necessary income. 

Second, as with our other portfolio simulations, the portfolio-maintenance system we used for the hybrid approach incorporates an annual rebalancing program. That means that we rigorously trimmed highly appreciated positions and added to depressed ones.

The Hybrid Ground Rules
To help facilitate apples-to-apples comparisons, we made many of the same assumptions for our hybrid retirement portfolio simulation that we did with the other bucket simulations. We assumed a 65-year-old couple with a $1.5 million portfolio and an anticipated 25- to 30-year time horizon. We also employed the same withdrawal-rate rules as in the other simulations, assuming a 4% withdrawal in year one of retirement ($60,000) and a 3% annual inflation adjustment to that initial $60,000. In an effort to improve the portfolio's longevity, we did without the inflation adjustment in years in which the portfolio lost value. 

In addition, as noted above, we used the same portfolio holdings. Because we took our simulations back to 2000, before some of our bucket portfolio's holdings were around, we swapped in some like-minded holdings in their place.  T. Rowe Price Equity-Income (PRFDX) stood in for  Vanguard Dividend Growth (VDIGX), we used  PIMCO Real Return (PRRIX) in place of  Harbor Real Return , and Oppenheimer Commodity Strategy Real Return took the place of  Harbor Commodity Real Return ST Instl .

The key differences in this simulation versus our pure total-return approach came in the realm of portfolio cash flow production, maintenance, and rebalancing. Specifically, we employed the following framework:

• Use income distributions from bond and dividend-paying stock holdings to cover living expenses in the year in which they are received.
• If income distributions exceed planned distribution rate, reinvest them into depressed long-term positions in the portfolio. If all of portfolio's long-term positions are at or over their original size, reinvest additional income distributions into short-term bond fund.
• If income distributions fall short of planned distribution rate, use rebalancing proceeds from appreciated positions to provide additional income. If rebalancing proceeds are insufficient, use withdrawals from short-term bond position.
• Rebalance annually. Trim positions back to their starting values when positions exceed 110% of their original size. Use proceeds to meet planned distribution rate, if income distributions are insufficient, and to top up positions that have fallen below their starting values. If both the planned distributions have been met and all positions are back to their starting values, move additional assets into short-term bond position.

The Results, Please 
The hybrid approach, combined with a version of our model bucket portfolio, performed well during the time period studied. Not only did it provide cash flow in line with our planned distribution rate, but it also ended comfortably above the portfolio's starting value--reassuring, in that the period studied, 2000-13, encompassed the bursting of the dotcom bubble as well as the financial crisis from 2007-09. In fact, the value of the ending portfolio in our hybrid simulation was higher than the ending value of the portfolio in the simulation where we employed a pure total-return approach. 

The declining interest-rate environment that prevailed over the period examined (2000-13) has proved a challenging one for pure income-centric investors, as they've been forced to choose between settling for a lower income stream or venturing into lower-quality credits. The hybrid approach addresses this issue by using rebalancing proceeds to make up the income shortfall, thereby, enabling a retiree to stick with a portfolio with the same risk/reward characteristics regardless of the direction of interest rates. Income distributions alone were insufficient to meet the planned distribution rate for most of the years studied--save for 2000, 2007, and 2008--and fell to almost half of the planned distribution rate by the end of 2013. But harvesting rebalancing proceeds during those years was more than sufficient to make up the shortfall. If interest rates trend up, the retiree may be able to meet a higher percentage of his or her income needs with income distributions alone. 

Additionally, annual rebalancing helped add a contrarian element to the portfolio's positioning. At the end of 2008, for example, we removed assets from our short-term bond fund and intermediate-term  Harbor Bond (HABDX) and plowed them into our hard-hit equity holdings,  Loomis Sayles Bond (LSBRX), and our commodity fund. All but the commodities fund went on to enjoy very strong gains in 2014. That type of rebalancing, arguably more than the rules we employed to generate cash flows, had a big impact on our portfolio's risk/reward profile.

Also, boosting this portfolio's performance relative to our bucket portfolio is the fact that the bucket portfolio has a cash stake, whereas this portfolio does not. Holding cash can provide valuable peace of mind in turbulent times (which can't be quantified in performance data). But it also has the potential to drag down returns, particularly given that cash yields shriveled to nothing over the time period examined. Meanwhile, Harbor Bond and  T. Rowe Price Short-Term Bond (PRWBX), which occupied larger positions in the hybrid portfolio simulation, were able to benefit from declining rates, which pushed up the prices of the bonds in their portfolios. 

For a detailed look at the year-by-year results of our bucket stress test, you can download this spreadsheet (Microsoft Excel required) or this PDF.

See More Articles by Christine Benz