At 2.9%, the 10-Year Treasury yield is near its highest level in the past 4 years. At the current rate of inflation (2.2% CPI in the past year), this translates into a real yield of 0.7% (real yield = nominal yield minus inflation).
Is such a real yield good or bad value for Treasury bond investors? And what does it suggest, if anything, for real returns going forward? Let’s take a look at the historical evidence…
We have monthly inflation data (CPI) in the U.S. going back to 1948. Since then, the median real yield on 10-year Treasury bonds has been 2.2%.
Source Data for all charts herein: FRED
Does that mean we always see real yields around 2.2%? No, far from it.
There have been many periods in which real yields have been significantly higher or lower than the historical median. This includes times during which real yields have been negative (14% of the time) and other times when real yield have been above 4% (17% of the time). Today’s real yield of 0.7% is below average, but far from unprecedented.
Following the peak in yields/inflation in the early 1980s, bond yields remained stubbornly high even as inflation cooled. This resulted in a bonanza for bond investors as they earned yields substantially above the inflation rate (5.7% median real yield from 1981-1989) and benefited from a decline in interest rates (15% in 1981 to 7% in 1989). In the 1990s, the good times continued. As inflation cooled even further, bond yields were again slow to react, leading to a 3.6% median real yield from 1991-1999.
Over the past 7 years, we’ve seen much more challenging times for bond investors, with real yields falling to a median level of 0.6%. This is lowest for any 7-year period since the late 1970s/early 1980s when nominal rates had difficulty keeping pace with spiking inflation. We have a very different environment today, with inflation holding steady around 2% while nominal yields have been suppressed by a lack of nominal growth and easy global central bank policies.
Do low real yields imply low real returns?
Not necessarily. There’s an difference between a current real yield (= yield – inflation rate) and a future real return (function of yield, future inflation, and changes in yield).
Real yields were negative in October 1980 and over the next decade bonds would produce their best real returns in history. This was a function of the high starting yield (12.5%), a persistent decline in interest rates (7.5% by October 1990), and a sharp decline in inflation (12.6% in October 1980 to 6.4% in October 1990).
Unfortunately for bond investors, the low real yields of today are not in any way comparable to October 1980. 10-Year yields hit all-time lows less than 2 years ago (July 2016) and since then both yields and inflation have been rising, not falling.
The 1950s may prove to be a more instructive comparison in retrospect. Inflation averaged only 2.2% from 1950-1959 but bonds would produce a negative real return of 1.5% per year as yields rose from 2.3% to 4.7%. As you can see in the table below, a decade of negative real returns is not unprecedented; we just haven’t seen it in some time.
Data Source: BLS, Stern.NYU.edu/~adamodar
But the times they are a changin’. Over the past 6 years (2012-2017), 10-year Treasury bonds have produced a negative real return. The last time that happened was from 1978-1983. At the end of 2021, we could very well be looking back a negative real return for bonds over the prior 10 years.
That’s not to suggest bonds have no place in a portfolio today, but that bond investors need to keep their expectations in check. The 35-year bull market in long-term bonds (1981-2016) generated abnormal real returns, particularly during the 1980s (6.6%) and 1990s (4.4%). But the factors that led to those strong returns (high starting yields, potential for falling yields/inflation) are no longer in place today. In fact, the complete opposite is true (low starting yields, potential for rising yields/inflation).
Given this backdrop, what real returns should bond investors expect? Below the historical average (1.7% from 1928-2017) would seem to be reasonable starting point.