Jun 1, 2020
• As is the case in equity markets, credit is showing strong evidence of short-term overvaluation
• Further downgrades are inevitable
• The impact of economic pain is far from clear
• A properly articulated bond strategy could avoid the pitfalls and take advantage of high spreads
Credit markets have recovered well from the point where the Fed announced its intention to intervene directly in the US credit market by investing in investment-grade and high-yield bonds mutual funds and ETFs. The recovery has been broad-based to include not only the areas directly targeted by the Fed, but also emerging market debt. Has the extra liquidity injection caused a fundamental shift in the market that will eliminate all of the remaining risks that the ‘Pancession’ (technical term: an economic recession, of as yet indeterminate severity and length, caused by a global pandemic) will throw at it? Or is there more to it than meets the eye?
We start by reminding ourselves that there are two main ways in which credit investors can come a-cropper in the market: spread-widening and outright default. Spread widening is the more benign of the two because a patient investor can recoup losses suffered in this way by waiting matters out: if the bond does not end up in default, spread widening is a temporary problem. Default is dangerous because it brings an irreversible loss of capital. It is thus evident that a sudden spike in the expected rate of defaults should lead to a similar spike in credit spreads. Investors require higher compensation for the predicted strain following a default or exit positions quickly to avoid the stress in the first place or a combination of both.
We saw just such a widening in this March and April in every credit market, except the highest quality sectors with fortress balance sheets. Some numbers will frame our analysis. In the US high yield market, the spread level reached up to 800 basis points over Treasuries. The spread has since compressed to 650 points. In historical context, the market last reached these levels in 2015/16. At the time, there was a surge in defaults in the high yield market driven by energy sector distress. The default rate then reached an annualised rate of 5.5%. In the Global Financial Crisis (GFC), spreads blew out to 1800 basis points. That was in anticipation of a default wave which did materialise; defaults peaked at an annualised rate of 14% in 2009/10. In the post-tech bubble recession of 2001, default rates touched 10% and spreads reached to 1000 basis points.
By those rough historical guidelines, the market’s current expected rate of default for this downturn is modest – somewhere higher than the energy bust of 2015, but not as severe as 2001, and certainly nowhere near as severe as the GFC. The current environment is just as difficult to assess as any previous market dislocation, if not more so. The collapse and subsequent recovery in credit markets have been much faster than the precedents mentioned above. The injection of support is an additional factor that makes direct comparisons potentially hazardous.
Using a simple observation of market spreads to infer the likely path of defaults is not necessarily the only correct way to assess immediate default risk in the market. Moody’s and S&P (and many other researchers) run economic models to compare default cycles and tease out relevant conclusions. On that basis, things look less rosy. For example, the ratio of credit downgrades to upgrades is useful in judging the rate at which credit quality of bonds in the market is deteriorating. The historical correlation of this measure is quite high. For the second quarter of 2020, the net downgrade ratio for US high yield calculated by Moody’s is close to 30%. The number for S&P should be similar, adjusting for some subtle timing differences and slight differences in the universe of bonds covered. The highest number on record was in Q1 2009 when the ratio hit 32.8%. By Q4 2009, the default rate was running at 14.5%. Note that market spreads had moved in anticipation of this default spike, not at the same time.
A further sense-check on spread levels is simply the state of the underlying economy. That calls for a view on how deep the so-called ‘pancession’ will be. In the past, a good indicator of the likely depth of a US recession was the rate of unemployment. We can safely say this time will be different because the jump in new jobless claims is so spectacularly bad. Next week will see the release of May payrolls numbers, which will likely set more records for severity. Trying to map this number to a realistic default expectation is a fruitless exercise. It simply does not compute.
Is the market headed for another setback, given strong circumstantial evidence that it is underestimating the potential for defaults? One counterargument is that the reaction time of rating agencies in this crisis has been up-to-the-job, compared to maybe being too slow in the GFC. Therefore, the worst of the downgrades are behind us, and things are about to brighten up. That would have been valid, except for the explicit signalling by the major agencies that many more downgrades are possible in the next two quarters. The agencies have a system that puts a bond on watch, after which a formal review is undertaken and published. The process can take up to 90 days. There is a large majority of bonds on negative watch (not to be confused with negative outlook) now. Roughly half of the negative watch bonds end up being downgraded by at least one notch.
So, if there are more downgrades and defaults in the pipeline, how is one to navigate credit markets for the rest of the year? As it turns out, not all high yield bonds are created equal. A screen of all the bonds currently showing distress – defined as trading at spread levels above 1000 basis points – reveals that two sectors dominate: basic industry and energy. The first is almost certainly COVID-driven, the second is facing the dual impact of an oil price war and a slump in global oil demand. On the flip side, some sectors seem to be almost unaffected using this measure. These would include banking, insurance and utilities. These observations match the performance experience in the high-grade market.
A final consideration to consider is the Fed’s intervention in the market. The Fed rules state that not all high yield bonds are eligible. Bonds rated low enough before March (in the single B range) stand to be offered no assistance. Our view is that investors should avoid such bonds. In the area of supported bonds, much remains to be clarified. For example, some recently defaulted bonds would have been held by ETF’s that the Fed now potentially owns. How does that work out? We are in territory as yet unexplored. It is also unclear how it will work if a bond suffers multiple downgrades in short succession.The market is pricing in a best-case scenario for all of these and other issues. Our preference would be for a conservative approach in this sector, with two central tenets: don’t go too low in rating, and steer clear of sectors showing pricing stress, however attractive the yields may appear in the short term.