Continue active refreshing of this index's data?

Continue active refreshing of this index's data?

Blog Posts — March 19, 2020

Treasury yields are up and stocks are down… And that spells double trouble for corporate bonds

by Christoph Schon, CFA, CIPM

In times when “all correlations go to one”, some asset classes are hit harder than others. In a “normal” flight-to-quality environment, corporate bonds are likely to benefit from lower risk-free rates, which offset at least part of the higher risk premia. For securities with better credit quality, this may even result in a positive overall return. However, given the recent breakdown of the traditionally inverse relationship of stock and sovereign-bond markets, this is no longer the case—and credit instruments are taking hits across the entire rating spectrum.

Since the beginning of the equity sell-off in the last week of February, credit spreads have widened rapidly and significantly. The graph below shows the 5-year spread levels over USD swaps of Axioma Fixed Income Spread Curves for North American corporate issuers. While we have yet to reach the heights of the global financial crisis in 2008/2009, corporate risk premia have already surpassed the last two peaks of the Eurozone debt crisis in 2011 and the Chinese stock-market turbulence in 2015/16. The latest surge also represents the fastest rise in spreads since autumn of 2008, with, for example, the yield pickup of double-B issuers soaring 300 basis points in four weeks.

Average 5-year spreads over USD swaps for North American corporates

Source: Axioma Fixed Income Spread Curves

We analysed a typical USD-denominated corporate bond portfolio in Axioma Risk to quantify the impact of the recent surge in credit premia across the rating spectrum. The table below shows the predicted volatility of each rating category at two points in time: at the peak of US stock market on February 19 and, more recently, on March 16. We used the same calibration periods as in our weekly Multi-Asset Class Risk Monitor[1] to demonstrate the impact on both short-term and long-term risk. Numbers are in percent per annum, normalised to an average duration of seven years.

Predicted volatility of USD corporate bonds with a 7-year duration

Source: Axioma Risk

Not too surprisingly, increases in short-term risk were much stronger (4-10 times) than those for the longer calibration period (1.5-4 times). However, the duration-times-spread approach used in the Axioma Granular Fixed Income Risk Model assures that volatility forecasts adapt more quickly in periods of extreme variations in credit-risk premia than in more traditional models, based on absolute spread changes. In the latter case, predicted risk would have been, on average, one-third lower.

In times when volatilities and correlations deviate significantly from their regular patterns, scenario analysis can be a useful supplement to traditional risk modelling. We therefore used the stress-testing capabilities in Axioma Risk to analyse the impact of a further 10% drop in US shares prices on the corporate bond portfolio. To estimate the likely movements of risk-free yields and credit spreads, we applied three recent calibration periods that reflect how risk-factor correlations have changed over the past few weeks. The first period (January 2-24) represents a time of relative optimism, when there was still a lot of excitement about the China-US trade deal. In the second period (January 24 to February 19), things were starting to get bad, but the coronavirus was still seen as a primarily Chinese problem. Finally, as the full extent of the crisis began to sink in after February 19, share prices started to fall rapidly, while credit-risk premia exploded.

Simulated USD corporate bond returns for a 10% drop in US share prices

Source: Axioma Risk

The results from the stress test display a similar pattern as the risk analysis. In the first (pre-crisis) scenario, we can still observe the presumed homogeneity of investment-grade bonds versus their lower-quality counterparts. The reason for the positive performance of the former was that their risk and returns were mostly driven by changes in risk-free rates, while the latter were hit by higher spreads[2]. However, this distinction seems to have gradually diminished over the past eight weeks.

BBB-rated bonds were the first investment-grade category for which simulated returns turned negative, even in an environment where Treasury yields would have still fallen in response to lower share prices. By the time risk-free interest rates bottomed out and then rebounded in early March, despite continued selloffs in the stock market, even the highest-rated issues started to display negative returns.

In our whitepaper on The Stock-Bond Correlation, we noted that credit spreads are usually negatively correlated with both government bond yields and share prices. For most of the time, this has not been a dilemma, as the latter two would generally move in the same direction—upward as risk appetites increase, and downward when investors become more risk averse. However, in the current environment where sovereign yields seemed to have finally bottomed out, the relationship with the stock market seems to have prevailed. This is most notable in the sub-investment grade universe, where risk premia have traditionally been more closely linked to share-price performance.

The fact that the value of investment-grade debt now also seems to be increasingly driven by stock-market performance is very worrying, in particular at the lower bound of the spectrum. As we noted in our recent blog post Equity Markets Fell, Are Angels Next?, about 50% of all investment-grade corporate debt now carries a BBB rating. In case of a rating downgrade, issuers in this category would not only be adversely affected by a re-evaluation of their creditworthiness, but would face additional selling pressure from funds, which are barred from holding sub-investment grade securities by their mandates.

[1] Daily returns over 60 business days (~3 months) for short-term risk and five years of weekly returns with a half-life of one year for longer-term risk.

[2] For a more detailed analysis, see our whitepaper High-Yield Bonds: Analyzing the Risk and Return Tradeoff When Rates are Negative