For time segmentation to work, there must be a clear procedure for how to extend the bond ladder. Unfortunately, with its varied implementation, that procedure is often overlooked. I will examine the potential for time segmentation by considering three different ways to implement it.

In Part 1, I provided the case for time segmentation strategies provided by their advocates, as well as how it fits into the spectrum of retirement-income approaches. I will consider three practical ways to extend bond ladders and put them to a quantitative test in Part 3.

Taxonomy of retirement income bond ladders

When it comes to retirement income bond ladders, Joe Tomlinson created an excellent taxonomy of different types in his November 2014 article in Advisor Perspectives. His list inspired me to create the more extended version shown in Table 1.

Bond Ladder

Bond Ladder

Retirement-income bond ladders are divided between one-time and rolling ladders. One-time ladders are spent down over time. The bond ladder is not extended as bonds mature and it gradually disappears. These one-time ladders are not used with time segmentation strategies.

Our focus here is on rolling bond ladders, which are extended over time to keep the length relatively constant as time passes. They are not meant to be fully wound down. As bonds mature with the proceeds (face value and any coupon payments) spent for retirement needs, new bonds are purchased with other financial assets to extend the ladder length. Rolling ladders provide the basis for time segmentation strategies.

Rolling ladders can be designed to be automatically extended by one additional year as each year passes, or extended only when certain conditions are met. Possible decision criteria for extending a rolling ladder could be based on stock market valuations, interest rates, market performance or whether the individual is adequately funded as determined by a capital-needs analysis.

If a time segmentation approach does not offer clear rules for extending the bond ladder over time, then it is not a true retirement income plan that can be tested and analyzed. I will investigate three different methods for choosing when to extend the bond ladder as retirement progresses: automatic, market-Based, and personalized.

Automatic rolling ladders

Automatic rolling ladders, which Stephen Huxley and J. Brent Burns called “rolling horizon” in their Asset Dedication book, keep the same time horizon perpetually by automatically extending the ladder length each year as a bond matures. This is done for as long as the growth portfolio has sufficient assets to extend the ladder length. An initial ladder length is chosen and the ladder is built. At the end of each year, the ladder is extended by one year so that the length remains fixed over time.

This strategy involves taking ongoing distributions from the growth portfolio, which may be 100% in stocks or including some combination of other assets. The distribution amount is not the full value of spending, but rather the cost of the bonds to be purchased based on the then-prevailing interest rates. If and when the growth portfolio depletes, the ladder will continue to provide income without being further extended until it is wound down completely and all assets are depleted.

Figure 1 provides the outcomes for a Monte Carlo simulation for stock returns and bond yields to show how an automatic rolling ladder might perform in practice. This retiree seeks to support an initial $40,000 annual spending goal with an annual 2% cost-of-living adjustment for as long as possible in retirement. The retiree has $1 million at the start of retirement to divide between a bond ladder and a stock portfolio. The retiree seeks to maintain a 10-year bond ladder providing the desired annual spending throughout retirement. With a 2.45% initial bond yield and assuming a flat yield curve, building the initial bond ladder requires 39.2% of the asset base. The other 60.8% of assets are left in stocks. While I allow interest rates to fluctuate randomly over time, I simplified by assuming a flat yield curve. Relatively to an upward sloping yield curve, a flat yield curve lowers the costs of short-term spending and raises the costs of long-term spending. But these factors will tend to offset one another and with low interest rates the differing assumptions will lead to relatively minor differences in ladder costs.

by Wade D. Pfau, read the full article here.