Barron’s ran an article by Aberdeen on 3-21-2014 about the risks to bonds if rates rise. A study was commissioned and done by Strategas which showed that in 35 years the Barclays U.S. Aggregate Bond Index only had three negative years. The study said if history was run backwards for 30 years to the time in 1984 when rates were very high (the 6 year Treasury was yielding 12.3% to 14.3% in 1984) that the index would have produced a 4.8% annualized return.
What was not mentioned in the Barron’s article is that the Barclays Aggregate Index (formerly Lehman Index) has a duration of 5.19 years which implies about a 6 or 7 year maturity. What is much riskier is a 30 year zero coupon fixed bond which would have a duration of 30 years. Another concern I have is that during the 1979-1987 era when Volcker ran the Federal Reserve interest rates were artificially high to the extent that the “real” rate for long term Treasuries was about 4% instead of the typical 2.1%. Real rates remained moderately high during early part of Greenspan’s rule from 1987-1991 and occasionally in 1994-95.
If an investor owns a six year maturity bond and replaces it upon maturity every six years then he will basically get the bond’s yield as his rate of return, assuming he purchased the bond at par when newly issued. However, in the actual world of buying an actual bond, instead of studying an index, the investor might pay an excessive Broker-Dealer bid-ask spread of roughly 1% to buy a bond, thus lowering the rate of return by 1% every six year period that he purchased a bond. This would reduce the total return by 0.17% a year before compounding over 30 years, which would be a significant difference from the index.
Since real yields were artificially high at double the historical real rate (4% instead of 2.1%) during the 1980’s then that helped this hypothetical study to produce good results. However, if rates ever do go back up to roughly the same pattern but without the Federal Reserve raising the “real” rate of interest then the total compensation to new bond investors who bought at par would not be as good as compared to the 1980’s. If the Fed has started a tradition since 1996 of giving the store away with artificially low rates for most of the past 18 years then it may be possible that a repeat (in reverse) of the past 30 years rate history would have significantly different outcome for bond investors. The high real rates of the 1980’s suppressed inflation and thus set the stage for bond market rally. Except for the Great Depression “real” rates were never as high as in the 1979-1987 era, so the 1980’s were a fluke experience.
Some statistical anomalies in bond investing that may not be the same as a hypothetical index is that when buying mortgage backed bonds they have a variable duration which varies according to the whims of consumers (borrowers) who vary duration according to the opportunity refinance at better rates. If rates rise then duration rises for mortgages. Thus during rate increases the severity of portfolio loss increases faster than in a simple linear way. In previous decades there were lesser percentage of mortgage backed bonds in bond indexes so an exact replication of the Barclays Index may not be practical since in modern times the index might have more variable duration mortgages in it which could more significantly damage the portfolio during a time of rising rates. In 30 years a lot more mortgages have changed from portfolio holdings by the originator bank (where they weren't in an index) to secondary market holdings by non-originator investors and thus they have pushed up the percentage of mortgages inside of bond indexes.
The best way to invest in bonds would be to keep a tight rein on duration, keeping portfolio duration to a maximum of 5, and avoiding mutual funds that have duration significantly over 5. Also avoid owning individual bonds so as to avoid the high spread charges that retail bond investors incur and instead invest in institutional class bond funds.
Investors should seek independent financial advice about the risks of bond investing. I wrote an article “Grantham wrong about bond bubble”.