Long-term bond returns under duration targeting is the name of a paper published earlier this year in the Financial Analysts Journal that recently caught the eye of The Value Perspective. Honestly. To its credit, it makes greater use of real-world data than many of the magazine’s offerings but it is still pretty heavy on formulae and Greek letters.
That being the case, what we plan to do here is focus on the conclusion of the authors Martin L Leibowitz, Anthony Bova and Stanley Kogelman that the total returns from investing in an investment-grade bond fund would – at the time of investing – appear to be predictable with a reasonable degree of accuracy.
That is the good news. The less good news is that, in order to see those returns, an investor may need to hold the fund for a sizable period of time. Furthermore, should interest rates rise at any point during that period of time, then the investor would experience losses before ultimately going on to achieve that return.
As a brief aside, ‘duration’ is the time it takes for a bond to repay all of the principal investment. Over time, as a bond gets closer to its maturity date, its duration falls – we are closer in time to getting our money back. Most people invest in bonds through collective funds, however, so each year these portfolios will contain some bonds that mature and repay their principal each year. As the fund manager will then buy new bonds, with longer durations, the overall duration of a fund will tend to stay constant over time.
Why, you may be wondering, is all this important? Well, some investors are buying into bond funds today and hoping for a rise in rates to improve their returns. Their logic runs that, although higher rates act to depress bond prices, as the fund’s older bond holdings mature, the fund will have the chance to invest in a new set of higher-yielding bonds.
According to Leibowitz and his co-authors, however, these investors may well be waiting a long time for their improved returns. Say you buy a 10-year duration bond index fund, yielding 2% today, and rates steadily rise by 0.5% a year, then – assuming the fund does not experience any defaults – the maths in the paper suggest your total return by year 10 will have been roughly 2% a year.
In fact, for any fund that has a constant duration, the implication of the paper is that the number of years from today when you can be most certain about the returns you are likely to achieve will be the same as the duration of the fund. The best guess of these total long-term annual returns would be your starting yield on day one.
One other point worth making here – especially in the current environment – is that anybody who chooses to ‘reach for yield’ today by buying longer-duration bond funds is going to have to be very wary of rate rises. Indeed, if rates do rise, the paper suggests investors would need to hold some funds for decades before the total return from their investment would be higher than the initial yield.
The paper’s authors do make some assumptions, such as interest rates moving in a steady direction – be that upwards, down or flat – and that investors specifically use a duration-targeted fund. Here on The Value Perspective we are inclined to be sceptical of all such assumptions although, as the following chart shows, some long-term data does appear to support their beliefs.
Barclays U.S. Government/Credit Index Yields and Shifted Returns over Six-Year Holding Periods
Source: Morgan Stanley Research; Datastream
The chart plots the yield of the Barclays Bond index as well as annualised compound returns for six-year holding periods for each month from 1985. “The six-year holding period returns have followed the same general declining pattern as the yields,” notes the paper. “However, far beyond simply sharing a general downward trend, these six-year returns evidently correlate very tightly with the starting yield.”
So while would-be bond investors may indeed be able to console themselves that the higher yields on offer from new additions to their fund will offset the negative impact of rising interest rates on the price of existing holdings, the question they need to ask is how long they might have to wait for those two things to balance out?
If Leibowitz, Bova and Kogelman are right in their calculations, you can only have any sort of certainty over a long – sometimes a very long – time horizon. So, should rates rise, you may have to suffer poorer returns than current yields for quite a while – albeit you will be reasonably secure in the knowledge that while the cavalry are coming, they are riding donkeys.