In a previous post, we observed that while seemingly everyone talks about “asset classes,” there is little detail on what an asset class actually is.
We offered the following definition:
An asset class is a set of assets that have similar expected returns, risks, and that share a high degree of covariance.
The next step is to see if we can apply this definition to find sets of assets that have the same or similar such characteristics.
Bonds as an Example of an Asset Class
As frequently is the case in finance, bonds provide a (relatively) simple example to consider when looking at the concept of asset class.
Recall that an asset class is a set of assets that have similar expected returns, similar risks, and the returns of which have high correlation with each other.
It should be easy to see that with bonds, the expected return will be primarily a function of the yield-to-maturity,[1] and risk will be highly sensitive to duration (i.e., sensitivity to interest rates). Additionally, for bonds deemed to have credit risk, to expected return will be sensitive to that perceived credit risk.
Most bonds can be defined by just three parameters:
- Yield
- Duration, and
- Credit risk
Empirical evidence suggests that a high percentage of bond performance can be explained[2] (in a statistical sense) by these three parameters.
Yield
For ordinary bonds (i.e. excluding bonds that are convertible into equity, or that have some equity-like feature), the only sources of return are payments of the coupon, and the gain (if any) between the purchase price and the redemption or payoff amount.
Thus, yield (usually calculated as yield-to-maturity) is theoretically the best predictor of the return to owning the bond. Here is a graph of the expected return, based on yield at the beginning of the period, and the actual return, based on what happened, for US Treasury Ten Year Bonds during the 80 years from 1928 to 2018.

The “r=.953” is the correlation coefficient. It represents an “R-squared” of over .9, which means that the beginning yield explains over 90% of the actual annualized yield which was realized over the next ten years.
That’s pretty incredible for investments. A single number, the beginning yield, predicted the actual return over ten years with 90% accuracy. I think many people under-appreciate this simple fact about bond returns.
Risk
Everything else equal, the price of a bond will move in the opposite direction to a change in interest rates. For example, if interest rates go up, bonds will go down.
This change in bond price is a mathematical necessity. And the sensitivity of a bond’s price to changes in interest rates can be calculated with a high degree of accuracy by a single number, called the bond’s duration.
Duration
There are two versions of duration in common use.[3] All of them are expressed in years. The simplest example of duration is the duration of a zero-coupon bond. The duration of a zero-coupon bond is equal to its maturity. For example, the duration of a ten-year zero-coupon bond is ten.
Duration can be used to estimate the sensitivity of a bond’s price to a small change in interest rates. To find the likely response of a bond’s price to a small change in interest rates, multiply the percentage change in interest rates by the bond’s duration. For example, a bond with a duration of 10 can be expected to increase in price by about 10% if interest rates fall by 1%.
Similarly, if interest rates were to rise by 1%, the price of a bond with a duration of 10 would be expected to fall by about 10%.
Credit Risk
Credit risk is the risk that a bond will not pay all its scheduled payments on time, in full and in cash.
Bonds issued in a fiat currency (i.e. a currency that can be printed by a government at will, such as the US Dollar) by the currency’s issuer, such as US Treasury bonds, are generally considered to carry zero credit risk.
All other bonds are considered to have some degree of credit risk.
While bond credit risk can be reasonably accurately estimated most of the time, the bad surprises mostly occur during difficult times, when investors probably want and need the supposed safety of bonds the most.
In the US, the credit rating agencies have a pretty good historical track record. Highly rated bonds have rarely defaulted. As the rating of the bond decreases, the historical defaults did increase.
However, this is not a strong a safety net as might be hoped, because historically the rating agencies have usually downgraded the bonds prior to default.
While you might avoid a technical default loss, if you own highly rated bonds that are downgraded, the market price of those bonds is likely to fall. You have not avoided the credit risk, even though the bond did not default.
Other Bond Risks
Some bonds are subject to other risks, perhaps the most salient of which has been liquidity risk.
Liquidity risk is the risk that there might not be a liquid market for a bond. For example, during the height of Global Financial Crisis, the bid-off spreads for even some highly rated corporate bonds became very wide. I recall one instance in which the market for commercial paper (a very short term debt instrument) of Toyota Motor essentially dried up. This was despite the fact that seemingly no one doubted the credit quality of Toyota.
Overall
Historical evidence suggests that for corporate bonds, about 85% of the changes in price can be explained by interest rate risk (duration) and credit risk.
In other words, because the same, known, factors drive bond returns, and do so in a predictable way, it makes sense to think of “bonds” as an asset class.
Next steps
To speak with us further, call us at 703 437 9720, email us back to set up a time to talk, or click here to request an advisor guide.
[1] Yield-to-maturity is a calculated number, based on the coupon yield of the bond and the current market price. Yield-to-maturity makes it much easier to compare bonds across coupon yields.
[2] A statistical explanation typically consists of a linear regression model that provides a good fit. In the case of bond returns, the observed returns in the market (looking backward) would be the “dependent” variable, and the three characteristics: duration, coupon and credit risk would be the “independent” variables. If you want to read 75 pages about this, see Common Risk Factors in the Cross-Section of Corporate Bond Returns, by Bali, Bali and Wen, 2018 in the Journal of Financial Economics, and available at https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2819910.
[3] Macaulay Duration is measured in years and is the weighted average of the present values of the bond’s scheduled cash flows. One consequence of this definition is that, everything else equal, duration of a coupon bond is lower at higher interest rates. Modified Duration is defined as Macaulay Duration divided by 1 + the bond’s yield to maturity.

Leave a Reply