The Fed may have been able to momentarily tame the upward momentum of U.S. sovereign bond yields. Yet, with the forthcoming tapering, the relationship between U.S. high-yield bonds (HYG) and floating-rate bank loans (BKLN) will change. Given that bank loans are more immune to a rise in long-term interest rates than high-yield bonds (floating rates indexed to U.S. Libor against fixed coupons, that is no duration risk), there is a positive correlation between the outperformance of loans (BKLN, which tracks the S&P LSTRA index) to high-yield bonds and the change in the U.S. yield curve (10-year U.S. Treasury yield vs. three-month Libor), which is clearly visible in the chart below.
The sharp rise in long-term interest rates gave rise to a sharp underperformance of the U.S. high-yield market. Yet, over the last few days, there has been a noticeable disconnect between the change in the yield curve and the relative returns of both markets: HYG outperformed BKLN even though the yield curve continued to steepen.
The first explanation comes from the fact that the relative performance of senior bank loans to junk bonds is highly correlated with equity returns. As can be seen below, high-yield bonds returns are much more correlated to equity returns than those of senior bank loans.
Click to enlarge images.
Combining those three first charts, the outperformance of bank loans to junk does not depend solely on rising long-term interest rates but also on the stock vs. Treasury regime: It will be all the more pronounced that higher interest rates are deemed negative for stocks (and not the signal of stronger growth ahead). Said differently, overweight bank loans when you think that rates will rise, but also when you consider that it will affect negatively equity returns.
A second explanation comes from the fact that a buying opportunity was provided by the increase of junk bond yields in the wake of the Fed tapering frenzy. As can be seen below, yields increased much more than what would have been implied by the risk premium (proxied here by the Markit U.S. HY CDX). Now that the Fed has managed to stabilize (temporary) market expectations and succeeded in separating QE exit from rate hikes, investors are once again long high yield -- especially since, even if absolute effective yields are still below their 2005/07 average, the spread against Treasuries remains 50 basis points above that period's level.
This points is very important. The yield of a high-yield bond depends on the risk-free rate (UST Treasury) and a risk premium (spread). The return can therefore be impacted either by a rise in the risk-free rate or a higher spread (risk premium). As we will see below, in a low-yield environment most of the return of high-yield bonds were driven by spread variation. The disconnect between the risk premium (the CDS spread) and the yield that is visible in the chart above suggests that most of the recent underperformance of U.S. high-yield has been driven by the perception that the era of low U.S. treasury yields was over. Yet, once investors were convinced that the tapering would be smother than expected, there was clearly a buying opportunity.
Yet, the regime switch observed in the relationship between UST 10-year yield and the return of high-yield bonds has not reverted to the pre-taper level. It remains negative, as can be seen below: Higher (lower) UST yields leads to lower (higher) high-yield bond return.
This was not the case during the low-yield era: The main contributor to high-yield returns was the risk premium, not the risk-free interest rate (proxied by U.S. Treasury yield). In periods of higher risk aversion, the risk premium would increase (prices would fall) while U.S. Treasuries yield would decline, hence a positive correlation. This same correlation would also weaken in periods of weaker growth as the risk premium embedded in the spread would rise independently of where yields were going (especially since they were skewed by different phases of QE). This is clearly what the chart below shows. In any case, it would stay in positive territory.
As the news flow has recovered somewhat over the last few weeks, the negative correlation remains. It is clearly a sign that even if high-yield bonds have recovered in the wake of better equity markets, their return is seen to depend more on yield fear rather than recession fear for the first time since 2008.
If the Fed maintains a lid on long-term interest rates, the correlation would potentially go back to where it stood over the last few years. High-yield returns would continue to show some equity-like patterns. This is the reason why high-yield bonds have outperformed loans much more than what the implied volatility on U.S. Treasuries futures would suggests.
Recent moves are nonetheless a reminder that the era of low yield is over, and that the determinants of U.S. high-yield returns will be much more varied than the simple directionality of U.S. stocks. Therefore, as the leverage loan market remains far away from its excesses of 2006/07 and as U.S. long-term interest rates are poised to rise further in the second half of the year, there is a potential for bank loans to outperform over high-yield bonds. But the analysis above adds a very important point: The extent of the outperformance will not depend on the sole increase in Treasury yields, but above all on the perceived impact they might have on equity prices.