Bonds benefited from a bull market that lasted more than 32 years. Interest rates peaked back on December 11, 1980, when the 13-week U.S. Treasury bill paid 17.14% per annum.
Today, this same instrument pays only 0.07%. The longevity, consistency and magnitude of the collapse in interest rates have been unprecedented in the 200-plus year history of our nation.
But what comes down must eventually go back up. What will this reversal in the direction of interest rates mean for bond investors over the next 20 years?
To provide one possible answer to this question, I examined the last 100 years of bond market history (U.S. Treasurys and corporates) trying to identify that single 20-year window that might closely resemble our next 20 years. Specifically, I sought to pick out a period using three indicators:
• Real interest rates. These are expected to increase at an accelerating pace.
• Inflation. Inflation and inflationary expectations are expected to increase. They follow an accelerating path, but will lag behind the increase in real interest rates.
• The term structure for U.S. Treasury bills, notes and bonds. This interest rate curve is expected to shift up and steepen considerably.
Using these measures, I selected the 20-year period from April 20, 1954, to April 30, 1974, as the one that best characterized the bond market environment we are likely to experience over the next 20 years. What did bond investors endure during that period?
Long-Run Bond Performance
To answer this question, let’s first review the long-run history of bond investing, looking not only at how they did by themselves but comparing them to stocks. To provide a more balanced view of their investment performance, I constructed a well-diversified portfolio spanning both the U.S. Treasury and investment-grade corporate markets. The portfolio consists of 50% U.S. Treasurys and 50% U.S. corporates. This mix broke down into:
• 25% high-quality domestic corporate bonds (represented by the Dow Jones investment-grade corporate bond total return index);
• 25% ‘AAA’-rated long-term domestic corporates (represented by Moody’s ‘AAA’-rated long-term corporate bond total return index);
• 11.08% 30-year U.S. Treasury bonds;
• 11.08% 10-year U.S. Treasury bonds;
• 11.08% five-year U.S. Treasury bonds; and
• 16.76% 30-day Treasury bills.
The U.S. Treasurys were mixed so I could maintain a fixed and constant 10-year average Treasury maturity. The portfolio was rebalanced back to specified weights at the end of each month. I chose these particular indices because they have dependable, quality monthly return series extending back to January 31, 1919. As a result, we have monthly return data for this series extending back for 93.93 years (1919 through December 31, 2012). When we compare the total period results for such a portfolio to the S&P 500, after we adjust for inflation using the Consumer Price Index, we get the following results.
For the 93.93 years ending December 31, 2012, (after adjustment for inflation):
Stocks (S&P 500)
Return
Average annual real return = 6.97%
Risk
Over any randomly selected 18-month time period, there was:
• A 29% probability of losing at least 0%
• A 27% probability of losing at least 1% (unannualized)
• A 26% probability of losing at least 2% (unannualized)
Over any randomly selected 24-month time period, there was:
• A 27% probability of losing at least 0%
• A 26% probability of losing at least 1% (unannualized)
• A 25% probability of losing at least 2% (unannualized)
• A 24% probability of losing at least 3% (unannualized)
Bonds (50% U.S. Treasurys, 50% U.S. corporates)
Return
Average annual real return = 2.86%
Risk
Over any randomly selected 18-month time period, there was:
• A 27% probability of losing at least 0%
• A 24% probability of losing at least 1% (unannualized)
• A 21% probability of losing at least 2% (unannualized)
Over any randomly selected 24-month time period, there was:
• A 26% probability of losing at least 0%
• A 23% probability of losing at least 1% (unannualized)
• A 20% probability of losing at least 2% (unannualized)
• A 18% probability of losing at least 3% (unannualized)
These summary statistics go a long way toward demonstrating stocks’ dominant role in providing the growth component (stocks’ higher average real return was 6.97%) in any well-diversified portfolio. At the same time, the statistics show bonds’ lower-risk, stabilizing attributes (bonds had a lower, 18% probability of delivering a loss greater than 3% during any 24-month long period), which are suitable for portfolios defined by short investment time horizons.
Bonds’ Very Difficult 20 Years, 1954-1974
Now let’s examine the 20 years ended April 30, 1974, presuming that this period could be remarkably informative for the next 20 years. As you will observe from the summary statistics below, during these 20 years, the return to bonds virtually disappeared, while their risk increased (on a proportionate basis) by 38%.
For the 20 years ending April 30, 1974, (after an adjustment for inflation):
Stocks (S&P 500)
Return
Average annual real return = 6.47%
Risk
Over any randomly selected 18-month time period, there was:
• A 26% probability of losing at least 0%
• A 24% probability of losing at least 1% (unannualized)
• A 22% probability of losing at least 2% (unannualized)
Over any randomly selected 24-month time period, there was:
• A 24% probability of losing at least 0%
• A 22% probability of losing at least 1% (unannualized)
• A 19% probability of losing at least 2% (unannualized)
• An 18% probability of losing at least 3% (unannualized)
Bonds (50% U.S. Treasurys, 50% U.S. Corporates)
Return
Average annual real return = 0.51%
Risk
Over any randomly selected 18-month time period, there was:
• A 39% probability of losing at least 0%
• A 33% probability of losing at least 1% (unannualized)
• A 29% probability of losing at least 2% (unannualized)
Over any randomly selected 24-month time period, there was:
• A 44% probability of losing at least 0%
• A 35% probability of losing at least 1% (unannualized)
• A 27% probability of losing at least 2% (unannualized)
• A 21% probability of losing at least 3% (unannualized)
These data are highly instructive and demonstrate quite clearly just how much this 20-year period differed from long-run averages for both stocks and bonds! Certainly, this was not an attractive period for either. Both asset categories suffered relative to their long-run average returns (the last 93.92 years). However, bonds suffered far more than stocks. And in fact, the risk profile of stocks was actually somewhat milder.
When measured proportionately (against the long-run characteristics) stocks fell by 7%, but the risk of stocks also fell, by 15%. In contrast, the return on bonds fell by 82% (it was almost eliminated), while the risk of bonds increased by 38%. During this 20-year period, by some measures, bonds ended up being measurably riskier than stocks. Notice how the probability of the well-diversified bond portfolio delivering an absolute loss (over any randomly selected 24-month period) was a full 44%. Bonds were a quite risky bet during these two decades.
Certainly, if we endure in the next 20 years an investment environment analogous to that of the mid-’50s to the mid-’70s, then now is the time to be reallocating a sizable portion of your portfolio away from bonds and into stocks.
Rob Brown, PhD, CFA, if the Chief Investment Strategist at United Capital Financial Advisers, in Newport Beach, Calif.
References And Notes
[1] All data was provided by Global Financial Data LLC of San Juan Capistrano, Calif., and is current as of January 21, 2013. The firm can be reached at www.globalfinancialdata.com.
[2] All returns are shown after adjustment for consumer price inflation as measured by the CPI.