What was the key difference between these two periods? In the latter, rates generally fell, and in the former, they generally rose. 1953 ended with the 10-year at 2.48 percent, near an all-time low. By 1983, rates had moved all the way up to 11.67 percent (after peaking at over 14 percent during 1982). Then, rates fell again, finally closing 2013 at 2.67 percent, completing essentially a 60-year round trip.
Why not go back to 1926 like so many others do? We used the Ibbotson data (sourced from Morningstar Direct), but only back as far as we consider relevant to the task at hand. We were tempted to study that total period covered by Ibbotson’s data (back to 1926), as many do. But it occurred to us that what happened during the Great Depression and during WWII are a bit too far back to serve as models for what could happen from here.
We produced two sets of data to review:
1. The probability in each period of producing a return of 4 percent or greater over a five-year time period.
2. Isolating returns of high-quality bonds during each period to see the difference between their contribution to achieving the 4 percent rule during the rising-rate and falling-rate eras.
3. Projecting returns of the 10-year U.S. Treasury bond from current interest rate levels.
Why a five-year period? We think that’s about how much time the patient client will give an advisor to show they are worth paying for. We studied rolling five-year periods in which we shifted the date range by one month each time, then recorded the returns for the asset classes and portfolio mixes we studied. This gave us 360 different five-year rolling periods within each 30-year investment era, a large sample size.
Now, consider the plight of the retiree expecting to live for 30 years. Like Forrest Gump’s mother said about a box of chocolates, “You never know what you are going to get.” They may get an environment like Period 1, Period 2, anywhere in between or something completely different. The question is, what are they prepared for, and once they become conscious of what CAN happen, what WILL they do now to plan for the possibility that the last 30 years are NOT a good guide to the future?
Summary Of Results And Historical Charts
When comparing the two time periods (1953-1983 and 1983-2013), we present the more recent period first. We did this on the assumption that most investors and advisors are more familiar with the last 30 years than they are with the previous three decades of market history. Here is what we concluded:
1983 - 2013: Stocks were a good contributor, but bond returns were even better. This is a historical anomaly. The last 30 years were about as good as it gets for bonds. In fact, bonds were nearly “automatic” to generate 4 percent annualized over a five-year period. And the chart 3 shows that even deducting 4 percent at the end of each year for an assumed withdrawal, the bond allocation kept growing. Interestingly, in a period of general bullishness for stocks, the S&P 500 only met the 4 percent threshold less than three quarters of the time. This underscores the boom/bust character of stock index investing. In fact, a hypothetical portfolio of 50 percent in S&P 500 and 50 percent in a zero-return asset (shown on the chart as “50% S&P”) was successful in reaching the 4 percent goal nearly as often as the fully-invested S&P 500 Index.
1953 - 1983: Using bonds made sustaining the 4 percent annual withdrawal difficult. Stocks drove returns over this time, as rising interest rates were a frequent headwind for bonds. Success rates for long-term bonds during this period were less than half of what they were in the earlier era. And as shown in Chart 4, long-term bonds (Treasury and corporate, the types which are heavily weighted in many bond funds), when adjusted for a 4-percent end-of-year withdrawal, could not even get back to original $1 million … after 30 years! Even intermediate bonds suffered, and were less successful than the 50% S&P “portfolio.”