Many fixed income investors are watching global economic developments (government bond yields at dramatic lows, the Federal Reserve poised to start hiking rates later this year) and evaluating ways to manage the risk that a rising rate environment could erode their bond allocations.
We would argue that this fear of duration is often misplaced, especially over the longer term, as investors tend to get rewarded for taking duration risk, even if rates go up (please see the February 2015 Viewpoint, “The Case for Duration in the Long Run”). But for clients who do want to take duration risk, we broadly speaking see three strategies, each with a different risk and return profile:
- Aim to structurally reduce interest rate risk by moving to a low duration bond strategy, which is anchored to a lower duration benchmark, swapping longer-dated for shorter-dated bonds (for instance, by moving from a benchmark that spans the entirety of bond markets to a version that includes bonds with maturities between one and five years; for this paper, we define a low duration strategy as having a duration band of one to five years).
- Aim to structurally limit interest rate risk via a duration-hedged approach, which keeps the bond portfolio itself intact but hedges interest rate risk with derivatives.
- Switch to a more dynamic, unconstrained bond strategy with an absolute return target and the flexibility to navigate changing investment landscapes with skill.
For investors who favor a more traditional approach in their bond portfolios – i.e., investors more likely to focus on managing duration risk via one of the first two strategies rather than take the unconstrained approach – comparing the potential benefits and risks of low duration versus duration-hedged strategies is a primary focus.
We find that for many investors concerned about rising interest rates, low duration strategies are a more attractive approach: They can be less complex, because they are less likely to require derivatives to hedge out the duration, and they tend to offer a higher yield along with lower volatility than their duration-hedged counterparts. This may seem counterintuitive, and it is the result of duration being only one of the risks in bond portfolios. Other risks, such as credit risk, remain, and historically when credit spreads widen, interest rates typically fall. This tends to mitigate the negative impact of spread widening on portfolio return, therefore lowering volatility. Also, because duration tends to be an important component of the return profile in a bond portfolio, adjusting exposure rather than hedging it away may make sense for many investors. Of note, these trends are generally consistent across bond strategies, including investment grade credit, high yield, emerging market hard currency and inflation-linked bonds.
Hedging duration risk can reduce yield and expected return
Bond portfolio duration can be hedged by paying a fixed rate on interest rate swaps or by taking short positions in bond futures. With yield curves upward-sloping in all major currencies (see Figure 1), duration hedging pays away a higher, longer-dated yield and receives a lower, shorter-dated yield. This reduces yield. For example, the yield on the Barclays Global Aggregate Credit index was 2.4% as of 27 February 2015, but yield on the duration-hedged version was 1.4%. We see a similar disparity in the Bank of America Merrill Lynch (BAML) U.S. High Yield BB-B Constrained index, which has a yield of 5.7%, versus 4.6% for the duration-hedged equivalent.
Yield curves are upward sloping for various reasons. Many market participants expect interest rates to rise at some point in the future, and this gets priced into the yield curve. In addition, investors generally demand higher yields for the risks involved in buying longer-dated bonds. When they hedge duration, investors forgo this term premium, and therefore may reduce the longer-term expected return on the duration-hedged portfolio.
This is important, because it means a duration-hedged strategy can outperform a non-hedged strategy only if interest rates rise more than what is embedded in the yield curve. Forgoing the term premium also means that duration hedging is likely to underperform a low (or standard) duration strategy over the long term. That said, for investors who understand the risk/reward profile of duration hedging, it may be viable as a shorter-term tactical strategy, requiring investors to time the switch into and out of duration hedging.
Performance in a rising rate environment: the mirror exercise
To illustrate the importance of this potential yield/carry return difference between a duration-hedged bond portfolio and its corresponding unhedged portfolio, we can simulate the performance of the Barclays U.S. Aggregate Credit Index (duration-hedged and unhedged) under a rising rates scenario. While we could use any hypothetical rising rate scenario, we opted for an illustration that may look familiar by assuming U.S. credit yields and spreads over Treasuries followed a path that was the mirror image of their path of the past 15 years. In other words, we evaluate how these indexes would perform if the forward path for rates and spreads retraced what we experienced historically, in reverse. For example, beginning at the 2.9% yield and 1.2% spread as of year-end 2014, in 2015 we would see a gradual rise in rates to close the year at 3.2% yield and 1.1% (where we were at year-end 2013), then a “reverse taper tantrum” in 2016 (reflecting May 2013), a major spike in spreads in 2020 (reflecting the Financial Crisis of 2008–2009), etc. until 2029 when yields stand at 1999 levels of 7.5% and spreads at 1.1%. (For another example of this mirror exercise that focuses on U.S. Treasuries, please see the February 2015 Viewpoint, “The Case for Duration in the Long Run.”)
Looking into the future along this mirror-image path, we can estimate the returns of the U.S. credit index – both duration-hedged and unhedged – based on the index yield, duration, spread duration, changes in yields and spreads, as well as historical losses due to defaults and downgrades. We know that U.S. credit has outperformed the equivalent duration-hedged strategy over the period 1999–2014. We estimate this outperformance was about 4% per year based on Barclays’ excess return data. But what about the estimated returns in the mirror scenario? Many would expect the opposite result, but Figure 2 shows that duration hedging again underperformed in this illustration, albeit by a smaller margin. The higher carry earned by the portfolio that held duration more than offset the capital losses resulting from the dramatic yield rise over the period.
Another counterintuitive finding is that duration hedging tends to increase overall portfolio volatility, even though it takes out one of the key portfolio risks, i.e., sensitivity to rising interest rates. The reason for this is diversification: Many factors, such as yields and spreads and – importantly – the correlations among them, drive returns and volatility in a bond portfolio. Portfolio return volatility is a function of volatility of individual risk factors such as credit spreads and interest rates, and negative correlation among them can often dampen overall portfolio volatility. Removing one risk factor can lead to higher overall volatility.
Low duration: a happy median?
Low duration strategies may provide a “happy median” – or a level of interest-rate duration that provides a better trade-off between a full market beta (with interest-rate duration) and a fully duration-hedged beta. Taking the U.S. high yield market as an example, Figure 3 shows the yield and estimated volatility of broad, duration-hedged and low duration indexes. In this case, duration hedging reduces the yield from 5.7% to 4.6% and increases the volatility from 11.9% to 13.1% relative to the broad high yield index. The low duration high yield index, in addition to reducing duration risk, also lowers volatility, and (in this example) actually provides a higher yield as a result of the high yield curve being inverted.
This analysis comparing unhedged, hedged and low duration strategies can be replicated for other fixed income sectors, such as investment grade credit, emerging market debt and inflation-linked bonds (see Figure 4). In all these cases, the yield-to-volatility ratio is most attractive for low duration and least attractive for the duration-hedged strategies.
The volatility numbers shown in Figure 4 are based on history, and one could argue that the initial conditions are materially different now than in years past and historical relationships between interest rates and spreads are less dependable going forward. This is an important point, but PIMCO’s analysis of the current environment suggests the negative correlation between spreads and yields will continue. This means a duration-hedged investment may very well be a costly way to defend a portfolio against rising rates.