When financial professionals talk about investing, the conversation invariably turns to the question of “risks.” This is a term of art only vaguely related to its everyday meaning. Ordinary people don’t talk about “upside risk,” for example. That’s the possibility that your investment will perform better than anticipated — not a problem for an ordinary person, but a source of deep anxiety for an investment professional.
When it comes to bond (or “fixed income”) investing, risks are meticulously separated into “credit” or “default” risk (the possibility that the issuer will default on its obligations), “interest rate” risk (the possibility that changes in interest rates will increase or decrease the price of the bond), “inflation” risk (the possibility that unexpectedly rising prices will erode the value of the interest and principle before maturity), and “reinvestment” risk. As I mentioned, each of these risks can have either sign: if you expect 10% of your bonds to default but only 5% ultimately do, you experienced upside default risk. More importantly, if everyone else expects 10% of issuers to default, but you are able to somehow predict in advance that only 5% will, you have an opportunity to underpay for unexpectedly (to everyone else) good bonds.
Lately I’ve been thinking about reinvestment risk, and wanted to figure out just how seriously it should be taken.
What is reinvestment risk?
Consider a 10-year, $10,000 bond that pays an annual coupon of $1,000: you deposit $10,000, and over the course of ten years you’ll get back $10,000 in interest plus your $10,000 deposit. This bond has an interest rate of 10%, but it doesn’t necessarily have a yield of 10%. To calculate the yield, you need to know what interest rate, if any, your annual coupon is reinvested at. If interest rates rise over the 10-year period, you’ll be able to take advantage of rising interest rate by spending each $1,000 coupon on bonds with higher and higher interest rates. If interest rates fall over the 10-year period, then each reinvested coupon will earn less and less interest than the one before it.
Thanks to the magic of financial engineering, there are ways to avoid taking this risk. So-called “zero coupon” bonds, for example, don’t pay interest, but are instead sold at a discount when issued, thereby “locking in” the yield (although still leaving you exposed to credit and inflation risks). Likewise, bank and credit union certificates of deposit are typically issued with a guaranteed yield over the life of the deposit, so your reinvestment risk is reduced to the single point of maturity (to manage this risk some people create CD “ladders” of varying maturities).
Of course, unless you’re a gazillionaire, or being ripped off, you probably don’t own individual bonds. Hopefully your “bond” portfolio is some combination of low-cost, well-diversified mutual funds or ETF’s. But even so, the concept of reinvestment risk still applies: instead of the interest rate of an individual bond issuer, you’re exposed to changes in the interest rate offered by an entire asset class, whether that’s investment grade, high-yield, international, or emerging market bonds. Every month or quarter when your dividend arrives, it will be reinvested at either a higher or lower interest rate than it was in the previous period, and those changes will have some effect on your overall return.
But how big an effect?
How worried should you be about reinvestment risk?
To find out, I ran a simple test: if reinvestment risk has an important impact on returns, then the total return of a bond mutual fund with dividends reinvested should differ from the yield at the beginning of a given period. For example, the SEC yield of Vanguard’s Total Bond Market Index Fund Admiral Shares (VBTLX) on December 1, 2010, was 2.6%. With dividends reinvested, the total annualized return through November 30, 2020, was 3.74%. Since these are nominal returns and investment-grade bonds let’s set aside the inflation and credit risks and focus on the two components of this return: the interest rate and reinvestment risks.
Interest rate risk applies price action: the price of the investment is higher at the end of the period than at the beginning of the period because interest rates fell between 2010 and 2020: a share of the mutual fund cost $10.72 in 2010 and $11.64 in 2020. That’s a 8.6% appreciation over the decade even if the fund didn’t pay any dividends at all.
But of course it did pay dividends, which by assumption we’ve reinvested in the same fund. The SEC yield on the fund varied over the period from a high of 3.38% on November 27, 2018, to a low of 1.11% on August 21, 2020.
Subtracting the price action of 0.86% per year (8.6% over the decade), we can arrive at a crude retrospective calculation of this fund’s reinvestment risk over the 10-year period: the upside reinvestment return was about 11% — a total interest return of 2.88% instead of the 2.6% SEC yield on the day the investment began.
With this methodology in hand, let’s do it with a bunch of funds over a bunch of different time periods!
Keep in mind that we’re trying to separate out reinvestment risk: over short periods, interest rate (price) action is going to swamp the effects of reinvestment: in the three years between December 1, 2017 and November 30, 2020, shares in VBTLX appreciated 8.2% — but the fund’s yield was just 2.52% at the beginning of the period! If you want to gamble on interest rate action there are lots of opportunities to do so. Instead, I want to focus on what kind of expectations are reasonable.
I looked at four Vanguard mutual funds: the Total Bond Market Index Fund, the High-Yield Corporate Fund (VWEHX), the Total International Bond Index Fund (VTABX), and the Emerging Markets Government Bond Index Fund (VGAVX). You can see the calculations and results here, and make a copy to run your own calculations (and fix any of my errors).
There are three interesting numbers for each fund and date range: the beginning SEC yield, the total return, and the amount of the total return that isn’t accounted for by price action, which I am calling “reinvestment risk.” When that value is positive, then reinvesting dividends accounted for more of the appreciation than the mechanical application of share price changes. When it’s negative, then reinvesting dividends would have made your investment underperform compared to the price action of the fund itself.
Let’s take two extreme examples. If you invested in the Total Bond Market Index Fund on November 12, 2005, reinvested dividends, and sold on November 11, 2015, you’d have realized an annualized total return of 4.6%. But that return is actually lower than the return on the underlying funds, which you bought when they yielded 4.87%. If you had access to a different investment vehicle yielding 4.87%, you would have been better off not reinvesting the dividends and instead moving the funds to that alternative investment.
In contrast, if you bought the Total International Bond Index Fund on December 1, 2017, reinvested dividends, and sold on November 30, 2020, your reinvested dividends actually improved your returns. This shouldn’t be surprising: shorter periods, and more volatile assets, should experience more interest rate/price volatility than more stable assets held over longer periods of time.
On the one hand, this is a totally superficial glance at the bond market. I picked four low-cost Vanguard mutual funds and 3 time periods for each fund, more from convenience than anything else: I had historical information available about them. Moreover, you’d be well within your rights to say that there was something “special” about the particular funds and particular periods I picked. War, financial crisis, plague, what relevance could these datapoints have to someone making investment decisions about the future, where war, financial crises, and plagues will all have been abolished?
I joke, of course. The only point of the exercise is to highlight the signs and the magnitudes: most of the time, over long enough time periods, you will get “roughly” the return on your bond investments as you’d predict from the starting SEC yield. Under some conditions you would be slightly better off not reinvesting dividends (as long as you have an alternate, higher-yielding investment vehicle in mind), and under other conditions you’ll get an unexpected windfall from rising interest rates, but it’s not unreasonable to anchor your return expectations, over long enough periods of time, to the starting SEC yield.
That yield might be too low! That’s why I save in high-interest rewards checking accounts instead of bonds. But the possibility of future interest rate fluctuations shouldn’t on its own discourage you from investing in bonds — assuming the available yields satisfy your return needs and expectations.