There is a relative advantage for discount (low-coupon) bonds over premium (high-coupon) bonds as investment vehicles. The proposed reason for the potential advantage is that in the case of default, two bonds that differ only by the coupon amount will receive the same amount of recovery (i.e., a percentage of the par value), despite the fact that the higher-coupon bond loses more valuable future cash flows.
The authors posit that standard duration and bond pricing techniques may not consider that in the case of a default, bonds that differ only in coupon rates receive the same amount of recovery despite different cash flow profiles. Both types of bonds will receive some recovery of the investment based on a percentage of par, but the bond with the higher coupon rate actually loses more future cash flow (which is priced into the bond) than the lower-coupon bond.
Interest rate spreads for identical corporate bonds that vary only by coupon rate are found to be higher as the coupon rate increases, which indicates that the market considers higher-coupon bonds riskier than similar bonds with lower coupons.
When computing the empirical duration of bonds using regression analysis (i.e., determining the price reaction to a shift in the US Treasury yield curve), the authors find the empirical duration to be less than the analytical duration, with the effect becoming stronger as the credit spread widens. When applying additional regression variables for discount and premium bonds with falling or rising Treasury yields, they find the rate sensitivity of premium bonds to be higher when yields rise and lower when yields fall. The rate sensitivity for discount bonds is lower when yields rise and higher when yields fall. The latter effect (a positive convexity effect) is very desirable because discount bonds react strongly to more favorable yield changes and do not react as much to unfavorable yield changes.
Given the positive convexity effect, a long–short portfolio is created (i.e., long discount bonds and short premium bonds) that is found to outperform by 3.5 bps per month, with a low volatility of 12 bps per month. Further testing demonstrates that discount bonds significantly outperform premium bonds in environments with increased default risk. The authors warn, however, that it is unclear whether any outperformance is in excess of transaction costs.
How Is This Research Useful to Practitioners?
The intuition of why low-coupon (discount) bonds are more desirable than comparable higher-coupon (premium) bonds is very interesting when considering possible default. Although the implemented long–short portfolio strategy may not generate excess return beyond transaction costs, long-only strategies may want to lean toward holding lower-coupon bonds.
How Did the Authors Conduct This Research?
The authors examine all noncallable bonds in the Barclays US Corporate Investment Grade Index from January 1992 through April 2014 on a monthly basis. The regression analysis also incorporates yields from US Treasury securities.
The portfolio analysis follows a bottom-up approach because it is more amenable to choosing companies with multiple issues of similar bonds than a top-down approach.
A preference for discount bonds over premium bonds because of what happens in the case of a default (the authors relate it to how a credit default swap is evaluated) is very interesting. I would like to see future research explore the positive convexity effect further.