Whether it's a money market fund or a long-term bond portfolio, one of the first things many investors — professionals and novices alike — want to know about a bond fund is its yield.
Many view a fund's yield as an approximation of the income a fund will produce over the next year, others use it as an indicator of a fund's total-return prospects. It isn't uncommon to hear an investor say, “This fund is yielding 9 percent; if I can just clip its coupons, the fund will return about 9 percent over the next year.”
At least one problem with this thinking, of course, is that calm, predictable periods in the bond market are relatively rare. Consider, for instance, that the average high-yield offering lost more than 2 percent of its value last year, while the typical long-term government portfolio posted a 13 percent gain. In addition, some funds are less predictable than others.
While it's clear that a yield is probably a poor performance predictor for portfolios with a fair amount of credit or interest-rate risk, it's less clear for the more conservative bond vehicles, such as short-term bond and short-term government funds.
So, do conservative funds' yields reveal anything about either the income or total return these funds will produce over the next year?
Morningstar's Approach
To answer this question, Morningstar researched separately its short-term bond and short-term government categories. Each fund's SEC yield was recorded at the beginning of the last six calendar years, as well as its expense ratio and income and total return for the next 12 months. Why short-term bonds? They are burgeoning with assets, as advisors with a yen for yield are increasingly using them to replace stingy yielding money markets. The scope was limited to funds that are still in existence; for funds that have more than one share class, Morningstar looked at only one of those share classes, typically the oldest. For each of the last six calendar years, we ran four linear regressions for each of the two categories of funds. (A linear regression attempts to determine a relationship between two or more variables and the extent to which one or more variables can predict another). In addition to seeing if a fund's yield held any predictive power for either a fund's income or total return for that year, we also ran regressions to see if a fund's yield and expense ratio together held any predictive value for the fund's income or total return over the next year.
The Results
The last six years did not contain many “coupon clipping” opportunities. Perhaps the calmest period was 1997, a year characterized by stable short-term interest rates and similar performances from the major sectors of the investment-grade bond universe. Not surprisingly, out of the six years that we studied, yield served as the best predictor of both a fund's income and total return in 1997. For short-term bond funds, yield was a statistically significant variable (the results should hold in at least 95 out of 100 cases) that explained 32 percent of the variance in the funds' income returns and 22 percent of the variance in their total returns. The results for short-term government funds were quite similar.
In more colorful years, the results were a good deal less compelling. Last year was vivid, as the corporate-bond market was stung by a series of headline-grabbing cases of fraud and as short-term interest rates fell sharply. Yields were not a statistically significant predictor of income returns in 2002 for either short-term bond or short-term government funds; in fact, yields were a statistically significant predictor only of short-term bond funds' total returns, but they explained just 6 percent of the variance of those funds' total returns — and they took the wrong sign (meaning that short-term bond funds that came into 2002 with higher yields actually had a lower total return). That result doesn't jibe with the assumption that those higher-yielding funds boosted their yields by holding a greater number of more interest-rate-sensitive issues, because interest rates plummeted last year. But it does make sense if those funds with higher yields held lower-quality bonds, which got pounded last year in the wake of corporate scandals and lackluster economic data.
Out of the six years Morningstar studied the two categories of funds, yields were a statistically significant predictor of income returns only 50 percent of the time, explaining in those instances as little as 12 percent of the variance in income returns and at most 32 percent. For total returns, yields were a statistically significant variable less than half the time, explaining anywhere from 10 percent to 22 percent of the variance of total returns.
Peer Pressure
Though yield is not entirely irrelevant, our study showed that it seldom explains much of the variance in short-term bond or short-term government funds' income or total returns. Yields provide a decent forecast of an individual bond's prospective income generation. Unfortunately, that logic doesn't translate very well to the world of bond funds.
Instead of using yields as a potential indicator of a fund's future performance, we think they are best employed vis-à-vis a fund's peer group. Managers often boost a portfolio's yield by buying longer-dated bonds (assuming greater interest-rate risk) or buying lower-rated bonds (assuming more credit risk). In that sense, a fund's yield often can be used, in addition to portfolio analytics, better comprehend around the level of risk a manager is currently courting.
Our study suggests that investors should not think of a portfolio of bonds in the same manner that they think of individual fixed-income securities. Also, investors shouldn't shop for funds in a category based on their yields, as yields are seldom a good indicator of either the amount of income or the total return a fund will produce over the next 12 months.
Writer's BIO: Brad Sweeney is a senior mutual fund analyst with Morningstar Inc., covering both taxable and tax-exempt bond funds.
Prakhar Bansal, a quantitative analyst with Morningstar Associates, contributed to this piece.
