In this article, I look at why long-term investors would want to hold investment grade bonds in their portfolios. The main justification I see is uncertainty about equity returns, but it is unclear whether this is enough to justify a significant bond weighting in bonds in the policy portfolio. I will cover other less obvious reasons to hold bonds in a follow up article.
In this article, I am looking at how a long-term investor should approach the bond allocation within a policy portfolio (a target portfolio weighting; see this article for a further definition). I am writing this article for an intended audience of readers interested in personal finance, although the concepts are applicable for some institutional investors. However, most institutional investors do not have as long an investment horizon as I am considering here. As an example, I am thinking of a young employee who already has an adequate cash buffer (as discussed here), and does not expect to need to draw on the portfolio for 30 or more years. By contrast, someone who is near age 60 would not have such a long investment planning horizon. (These differences explain why people should not rely on generic investment advice, and should either get professional advice or do some planning exercises if they want to do it themselves.)
The Problem With Bonds - Equities Look Better
The problem with most developed market government bonds is simple - yields are low. At the time of writing, the 10-year U.S. Treasury yield is near 3%, and the 30-year is closer to 4%. The yields in Canada are similar, while Japanese yields remain at ridiculously low levels. But what I am writing about here is not applicable for all markets. Australian yields are higher, and so these bonds are less problematic. The Euro area governments are in the same boat as developing countries which do not control the currency they issue bonds in; default is a real possibility, and so bonds are just as speculative as equities. In this article, I am mainly considering government bonds, but what I am saying is also applicable to investment grade corporate bonds.
The low level of yields on bonds at shorter maturities (10 years or less) is understandable; bond yields are pinned down by the expectation that policy rates will remain at low levels for a considerable period. This is equivalent to saying that bonds (with maturities 10-years or less) are priced to give comparable returns to cash over the life of the bond. However, for a long-term investor, bonds should be compared to equities (stocks), not cash. (Note that I am simplifying things to just look at stocks and bonds; other investments tend to be a hybrid of those two asset classes.)
One basic way of looking at equities is using a dividend discount model. The simplest version of this model is based on assuming that dividends grow at a fixed rate forever. In this case,
Equity price = (current dividend)/((discount rate)-(growth rate of dividends)).
The implication of this relationship is that there is a 1:1 trade-off between increasing the discount rate used and the growth rate of dividends if the equity price is fixed. (The discount rate is the rate of return on holding the equity.) For example, if you increase the expected growth rate of dividends by 2%, the discount rate will rise by 2%.
At present, it seems that 3-4% represents a lower bound for average nominal GDP growth for the U.S. and Canada, and probably other countries as well (assuming they avoid replicating the policy errors of the Euro periphery). Even in this period of sluggish growth, average U.S. nominal GDP growth is running at 4%, although the last recession did manage to drop the 3-year average to 1% for a short period (chart above).
Also, developed equity markets have a high weighting in multinationals, which have considerable scope to participate in regions of the globe with stronger growth rates. We can then make a logical leap to assume that dividend growth will at least keep pace with nominal GDP growth, although there are risks associated with that assumption (I will have to discuss that elsewhere). The logic behind this is that aggregate corporate profits have to grow roughly at the same pace as GDP growth, as it seems unlikely that they will become an insignificantly small proportion of national income.
The implication is that the long-term return on equities will be the current dividend (or earnings yield) plus the growth rate, and the expected growth rate is already larger than the current long-term government bond yield. The implication is that equity returns should be well above bond returns on a long investment horizon. And that is not taking into account the possibility that future inflation could break above the 2% levels we have been seeing, which should be reflected in even higher dividend growth. (Although I discuss in another article the ambiguous relationship between inflation and equity returns.)
The question is this: why hold bonds when their returns will probably be trounced by equities on a long-term horizon? Reasons for long-term investors to hold bonds exist, but they are not too powerful. The rest of this article (and the follow up article) are attempts to find answers to this question
Uncertainty About Equity Returns- A Major Reason To Hold Bonds
Although equities have outperformed bonds on very long horizons, they have also underperformed for extended periods. If you bought the S&P 500 at the top of the market in March 2000, your returns to present have been pitiful (chart above; stock prices have risen at 1.3% annualised since the peak in 2000, but investors would have also received dividends to augment their returns).
Since you do not know whether you are buying equities at the beginning of a bear market, you can hedge your bets by holding some bonds in your portfolio. By holding non-equities, you have the capacity to buy equities if they fall in price. (Conversely, if you are holding 100% of your portfolio in equities, you do not have the capacity to buy more if they fall in price, you are already "all in".)
However, if you hold a bond, you have limited uncertainty about your returns if you plan on holding it to maturity. Unfortunately, those relatively certain bond returns stink at present. As a result, it is unclear that the advantage of reducing potential losses in the short to medium term is worth the drag on portfolio returns that may be created by holding long-term bonds.
In particular, if you are working and just beginning to build up a portfolio, the near-run equity losses should be a limited concern (once again, assuming you have a reasonable cash buffer). This is for two reasons:
- if stock prices drop now, your expected future savings will purchase more shares for amount of dollars saved, so you may actually end up better off;
- if the amount you save rises over time (as the result of salary increases), any losses you suffer now will be small relative to your future contributions over just a few years. (For example, if you save the same amount each year, a 30% drop in the stock market this year - which is a pretty vicious loss - will only amount to 10% of the total of savings over a 3-year period.
One way of looking at this (using some concepts from finance) is that people in this situation have an effective short position in future stock market prices - they know that they need to buy equities in the future, but at an unknown price. Having a high weighting in equities means that portfolio does a better job of cancelling out the uncertainty about those future purchase levels.
This logic is not really applicable to someone who has already built up a large portfolio relative to income, in which case the losses will be much larger than annual savings. In such a case, investment losses are much harder to recover from. But in most cases, people in this position will normally not have the extremely long investing horizons that I assume in this article.
I will discuss other (less obvious) reasons to hold bonds with a bit more detail in a follow up article.