Fridson: High-yield valuations are rich any way you cut it

This commentary is written by Martin Fridson, a high-yield market veteran who is chief investment officer of Lehmann Livian Fridson Advisors LLC as well as a contributing analyst to Leveraged Commentary & Data.

Our monthly Fair Value update once again shows high-yield bonds to be extremely overvalued. That is an unwelcome conclusion for asset gatherers focused on management of noninvestment-grade corporate debt. They might well ask, “How can an option-adjusted spread (OAS) of +467 bps (as of Dec. 15, 2022) possibly be too low? The default rate is well below its historical average.”

Actually, our empirical work finds the trailing 12-month default rate has only a minor impact on the spread. A more formidable challenge would be that our current valuation finding is based on historical analysis that encompasses both economic expansions and recessions. In recessions, many high-yield issues trade on dollar price, based on expected recoveries after default. Their spreads, which are therefore not truly the basis of their valuation at such times, become astronomical, skewing the index-wide spread upward. Therefore, one might hypothesize, inclusion of recession months in the calculation of the regression formula biases the estimated Fair Value spread upward.

This week’s piece addresses such concerns by analyzing the present high-yield spread solely in context of previous non-recession periods. We find that the evidence upholds our present verdict of high-yield overvaluation. That discussion follows an update based on our standard (all-economic-conditions) analysis. Let us begin by reviewing the methodology underlying that analysis.

Description of our Fair Value methodology
The valuation findings summarized in the chart below are drawn from the updated methodology presented in "Fair Value update and methodology review." In brief, we find that 80% of the historical variance in the ICE BofA US High Yield Index's OAS is explained by six variables:

• Credit availability, derived from the Federal Reserve's quarterly survey of senior loan officers.

• Capacity utilization.

• Industrial production.

• Current speculative-grade default rate.

• Five-year Treasury yield.

• A dummy variable for the period covered by quantitative easing.
  

Each month we calculate a fair value spread based on the levels of these six variables. The extent of high-yield overvaluation or undervaluation is determined by the difference between the actual OAS and the fair value number ("estimated"). We define material under- or overvaluation as a divergence of one standard deviation (124.5 bps) or more from fair value. The monthly difference between the actual and estimated OAS is tracked in the first chart below.

Current findings
Our Nov. 30, 2022 Fair Value estimate for the OAS on the ICE BofA US High Yield Index is +813 bps, up from +784 bps one month earlier. The key factor in that increase was a drop in the five-year Treasury yield, which is inversely correlated with the spread, from 4.24% to 3.79%. Also pushing modestly in the direction of a wider Fair Value spread was a dip in Capacity Utilization from 79.9% to 79.7%. The S&P Global default rate, a backward-looking factor with only a minor impact, was unchanged at 1.6%. Similarly, the change in Industrial Production from -0.1% to -0.2% had only a small influence. The dummy variable for quantitative easing remained at zero (QE not in force).

On Nov. 30, the index’s actual OAS was +455 bps, down from +463 bps at the end of October. That meant the spread was 358 bps tighter than our Fair Value estimate, a gaping divergence of 2.9 standard errors. In October, the corresponding numbers were 321 bps and 2.6 standard errors. These huge departures from Fair Value are plotted in the chart below. Zero on the vertical axis indicates perfect equivalence between Fair Value and the actual OAS. The green and red horizontal lines mark disparities of +1 and -1 standard error, respectively.

 
Note that according to the analysis described in "High-yield vs. investment-grade workout periods," high-yield bonds are not overvalued relative to investment-grade corporates. The ICE BofA US High Yield Index’s Nov. 30 OAS of +467 exceeds the investment-grade ICE BofA US Corporate Index’s +139 bps by 328 bps. That is comfortably above the 265 bps differential, below which we recommend underweighting high-yield in a portfolio that also includes investment-grade. The implication is that non-default-risk-free HY and IG corporates are both trading too rich, as benchmarked against default-risk-free Treasuries.

The explanatory variable that sets a higher bar for risk premiums than corporate bonds currently meet is our Credit Availability measure. It is the most powerful factor in our regression formula, based strictly on empirical analysis, not on our subjective view of what seems important. This factor is revised just once a quarter, when the Fed’s quarterly survey of senior loan officers is released. In the update for the third quarter, the percentage of banks tightening their lending standards for medium-sized and large borrowers minus the percentage easing their standards jumped to 39.2%, from 24.3%. That represents very tight credit conditions by historical standards. Of the 103 quarters prior to the present one covered in our analysis, the tighten-minus-ease differential was less than the present 39.2% in 90 quarters, or 87.4% of the time.

Still, skeptics can point to the above-described possibility of skewing of our regression model’s output as a result of its inclusion of recession months in the analysis. Let us now address that objection. We proceed by focusing on non-recession months in which credit was about as tight as at present and ascertain the typical level of the high-yield spread under those conditions.

Focusing historical analysis on non-recession periods
We began this analysis by identifying all months in which Credit Availability was within five percentage points of the latest third-quarter survey level of 39.2%. From that initial sample we eliminated all months that the National Bureau of Economic Research classified as recessionary. Note that in the second quarter of 2020, one month (April) was eliminated on that basis. Consequently, our final count of months, 11, is not divisible by three, as would be the case if the sample consisted entirely of complete quarters.

The table below ranks the qualifying months by their option-adjusted spread (final column). Highlighted in blue is the median month by OAS — June 2020, at +644 bps. For each month, we display values of the most important determinants of the spread, not according to our opinion but as determined by regression analysis.

One could argue that the extremes in OAS are explained by special circumstances. The fourth quarter of 2000 values, ranging from +779 to +916 bps, were attributable in part to a collapse of the telecommunications sector. Over the course of 2000, the ICE BofA US High Yield Telecommunications Index’s OAS soared from +570 to +1,288 bps. Over the same period, though, the ICE BofA US High Yield Excluding Telecommunications Index’s OAS widened from +455 to +852 bps. It is unlikely that the increase from the fourth quarter of 1999 to the fourth quarter of 2000 (indicating increasing risk) in the Credit Availability measure from 9.1 to 43.8 was sheer coincidence.

Incidentally, observe that during the fourth quarter of 2000, OAS increased as Capacity Utilization and the five-year Treasury yield declined, lending support to our model’s validity. Remember, the Treasury yield and the high-yield spread are inversely correlated, contrary to what some readers may have been taught.

The low-side OAS outliers of Nov.-Dec. 2020 coincided with exceptionally low Treasury yields. Observe, however, that those narrow spreads coexisted with Capacity Utilization levels lower than the Nov. 30, 2022 figure of 79.7%.

Given the multiple and period-specific factors influencing the spread, the chart’s mid-range OAS values offer the best guide to what the spread ought to be with Credit Availability at its present level. Granted, the June 2020 median figure (+644 bps) reflects an economy just emerging from recession with the help of strong suppression of interest rates via quantitative easing. Those effects are respectively reflected by an exceptionally low Capacity Utilization (68.9%) and five-year Treasury yield (0.29%). On the other hand, in the fourth quarter of 1998, Capacity Utilization was somewhat higher than the present 79.7%, at 81.9% to 82.4%. This means that all else being equal, we would expect the OAS to have been greater at that time than currently. Similarly, the fourth quarter of 1998 five-year Treasury yield, which ranged from 4.24% to 4.50%, was consistent with a somewhat narrower OAS than the Nov. 30 yield of 3.79%.

In actuality, the high-yield spread ranged from +544 bps to +652 bps in the fourth quarter of 1998, well above the Nov. 30, 2022 level of +455 bps. By the most generous interpretation then, the latest month-end high-yield spread was 544 – 455 = 89 bps (0.7 standard errors) tighter than historical experience implies that it ought to be. That gap increases to 132 bps (1.1 standard errors) if we base the analysis on the average spread for the three months of the fourth quarter of 1998. It seems to us an inescapable conclusion that with the 2-to-10-year Treasury curve inverted to the tune of 71 bps and with numerous other established precursors pointing to recession, the high-yield market is currently compensating investors inadequately for risk.

Research assistance by Ravi Shah.

ICE BofA Index System data is used by permission. Copyright © 2022 ICE Data Services. The use of the above in no way implies that ICE Data Services or any of its affiliates endorses the views or interpretation or the use of such information or acts as any endorsement of Lehmann Livian Fridson Advisors LLC's use of such information. The information is provided "as is" and none of ICE Data Services or any of its affiliates warrants the accuracy or completeness of the information.

Featured image by Things/Shutterstock

This article originally appeared on PitchBook News