At the start of 2020, S&P Global Ratings created its U.S. CLO Insights Index, an index of 410 reinvesting U.S. broadly syndicated loan collateralized loan obligations (BSL CLOs) that covered a wide range of CLO portfolios from different vintages and many different managers. The index is comprised of transactions that had already gone effective and started issuing trustee reports as of the start of 2020 and would reinvest through the entire year of 2020 (due to optional redemptions, upsize resets, special redemptions, and other data reasons, we removed six deals from the index during the year). We followed the remaining 404 transactions and tracked various performance metrics for all of 2020, a year that was dominated by the COVID-19 pandemic. Below, we look at the performance metrics of the CLOs in the index during 2020 and analyze whether certain portfolios were better positioned at the start of the year to withstand the credit impact of the pandemic.

COVID-19-Related Spike In Negative Corporate Rating Actions Significantly Disrupted 2020 CLO Index Performance

We analyzed the average change in U.S. BSL CLO portfolio performance in 2020 using 10 parameters (see table 1). Outside of turnover and exposure to widely held assets (the top 250 obligors), all of the other credit-related metrics of CLOs within the index listed in table 1 ended the year on a sour note due to the pandemic.

Table 1

We recap the 2020 performance metrics generally below.

Onset of COVID-19 has significant negative effect on performance of CLOs in index

The arrival of the pandemic and related shutdowns in early 2020 presented significant challenges to many of the speculative-grade companies with loans in U.S. CLO transactions. With some companies in the most-affected sectors facing the prospect of a second quarter with zero revenue, negative rating actions came swiftly. In late March and early April, corporate ratings experienced, over a period of a few weeks, a level of CreditWatch placements and downgrades that was comparable to the level seen over the entire course of the Global Financial Crisis in 2008-2009.

It wasn't long before the effects of these negative corporate rating actions were felt in CLO collateral pools within the index: 'CCC' buckets increased to a peak of more than 12% by the second quarter of 2020 from roughly 4% at the start of the year (before the COVID-19-related corporate downgrades), and the proportion of obligor ratings in CLO collateral pools on CreditWatch negative increased fivefold to more than 10% from less than 2%. By the middle of 2020, nearly a third of all corporate ratings within U.S. CLO collateral pools had experienced a downgrade or CreditWatch negative placement, and about a quarter of reinvesting U.S. BSL CLO transactions were failing one or more of their overcollateralization (O/C) ratio tests.

U.S. CLO performance begins to stabilize in the third quarter

The volume of negative corporate rating actions began to taper off in third-quarter 2020, and a handful of corporate issuers that had seen their ratings placed on CreditWatch negative in the first half saw their ratings removed from CreditWatch negative. By the end of 2020, average exposure to loans from 'CCC' rated obligors across the index declined to 8.7% from a peak of 12.3% in May 2020, while the average portfolio exposure to issuers with ratings on CreditWatch negative declined to less than 1.8% from a peak of 10.7% in April 2020.

As the corporate rating landscape and loan prices started to stabilize during the second half of the year, so too did CLO O/C ratio tests. The average junior O/C cushion of CLOs in the index improved to about 2.1% by the end of the year, up from a low of 1.1% in June 2020 due to improving credit ratings and loan prices of the underlying portfolio assets as well as senior note paydowns due to interest diversion for some deals. One metric that did not have an inflection point in 2020 was the average par loss, which has had a gradual decline for the entire year. After collateral turnover of about 37% during 2020, the average CLO ended the year with a decline in par of just under 1%, due to asset defaults and sale of assets at prices below par. The drop in par was more pronounced in the months preceding payment dates, as managers worked to rebalance their portfolios and improved O/C ratio tests ahead of CLO payment dates.

Some Types Of CLO Portfolios Were Able To Withstand The Effects Of COVID-19 Better Than Others

CLO managers each have their own unique styles to managing, meaning the managers of CLOs within the index will have assembled portfolios with different credit profiles. For example, some managers may run portfolios with lower S&P weighted average rating factors (SPWARFs) or higher weighted average recovery rates (WARRs), while other managers may have portfolios with a higher weighted average spread (WAS) or diversity, measured here through the obligor diversity measure (ODM). Exposure to widely held names (for instance, the top 250 obligors) can also differentiate portfolios as the larger widely held issuers tend to have higher ratings but lower spreads. If we break down our broad sample of 404 index deals into different cohorts, we can explore some of the nuance amongst the different types of portfolios and determine which ones ended up performing better during the 2020 pandemic-driven credit environment.

In table 2 below, we split our sample of 404 CLOs in half in multiple cohorts by color code, based on how the portfolios scored across various portfolio metrics at the start of 2020. We then looked at how each cohort of CLOs performed during 2020 by looking at the change in metrics during the year (across the same set of 10 parameters from table 1). For example, for SPWARF, did having a better credit distribution (lower SPWARF) presage stronger performance during the downturn? The average SPWARF of the portfolios in our sample of 404 CLOs was 2644 at the start of 2020. We divided the main sample of 404 deals into two roughly equally sized cohorts of about 200 deals each: one where the SPWARF was above 2644 (weaker credit) and the other where SPWARF was below 2644 (stronger credit) at the start of 2020. Then, for both cohorts (high and low SPWARF), we looked at the performance of credit metrics during 2020 to see if there was a material difference between the two cohorts. We then repeated this analysis across eight other portfolio characteristics. Within each column/metric, we've bolded in orange the cohorts with the greatest deviation between the two cohorts.

Table 2

Following the SPWARF cohorts (first pair of rows in table 2), we see the deals with weaker rating profiles (and therefore a higher SPWARF) at the start of 2020 experienced higher collateral turnover during 2020 relative to deals with stronger rating profiles (and a lower SPWARF). For CLOs in the higher WARF cohort, collateral managers turned over 39.09% of the assets during the year, versus 35.78% turnover for the lower SPWARF cohort. This may suggest that CLO collateral rating profiles heading into the pandemic had some relationship with the volume of turnover during the downturn. Looking at the last two rows within the turnover column (table 2), we see the average 2020 turnover of the two cohorts bifurcated by ODM (the measure to determine level of diversity) have the smallest deviation from the mean of the overall sample, suggesting the level of diversity of the portfolio at the start of 2020 had less to do with the turnover experienced in 2020. For each of the 10 performance metrics (the 10 columns), the cohort pairs that had the largest absolute deviations from the overall mean of the full sample (table 1) are highlighted in orange in table 2.

Some of our general observations of the beginning and end-of-year comparison in table 2 include:

• Portfolios with stronger credit-related metrics (lower SPWARF, lower 'CCC' exposure, lower nonperforming exposure, higher junior O/C cushion higher portfolio price, and higher WARR) at the start of 2020 generally had less turnover and less par loss over the course of the full year.
• Portfolios with weaker credit-related metrics (higher SPWARF, higher 'CCC' exposure, higher nonperforming exposure, lower junior O/C cushion, lower portfolio price, and lower WARR) at the start of the year generally experienced less credit deterioration, though they experienced more turnover and par loss (for example: portfolios that started 2020 with a higher 'CCC' bucket experienced a smaller 'CCC' increase in 2020, but experienced more par loss).
• Nonperforming exposure at the start of 2020 may have had a relationship across multiple credit-related metrics; for example, the level of par loss experienced in 2020 (four metrics highlighted in red).
• Weighted average price as of the start of 2020 (before COVID-19 risk was fully priced in) also had a significant impact across multiple 2020 performance metrics listed (for example, O/C change and change in weighted average spread, both highlighted in red; also see the Appendix for R-square values).
• Deals that started 2020 with greater exposure to thinly held issuers, less exposure to the top 250 (thinly held issuers generally have weaker ratings and higher spreads) experienced a flight to quality represented by the larger increase in top 250 issuers by end of 2020 (highlighted in orange).
• Obligor diversity at the start of the year did not significantly differentiate across the 10 2020 performance metrics listed.
• Portfolios that experienced greater turnover in 2020 tended to lose more par but experience less credit deterioration by the end of the year (see table 3 in the Appendix). This may be evidence of manager attempts to de-risk during the downturn.

CLOs Overall Got Through 2020 In Relatively Good Shape, Due In Part To Their Managers

Given the tsunami of negative corporate rating actions that characterized the first half of 2020, the market proved to be surprisingly resilient in the second half, as evidenced by the rebound of new U.S. CLO issuance later in the year. The worst of the 2020 credit impact from the pandemic was absorbed by pre-COVID CLOs with a relatively modest proportion of their notes being downgraded. This is partly due to CLO structural mechanics (which, in our view, worked as intended during the downturn), but we also believe it was due to the mitigating impact of active collateral management. The reinvesting aspect of the CLO asset class enables the portfolio to evolve during downturns, in some cases, helping to preserve par where needed for O/C tests, but also shedding par to preserve credit when it needs it most.

Given our data sample of 404 deals and their performance in 2020, we find that certain parameters like portfolio price and the exposure to nonperforming assets before the downturn may have had more of an impact on the performance metrics during the downturn. Interestingly, since COVID-19 was perhaps one of the most-correlated events within recent memory, affecting multiple sectors, differences in portfolio diversity may have had a limited impact on 2020 performance across this sample. Nevertheless, the data within our sample suggests many CLOs had enough cushion to absorb some par loss leaving, a large majority of the sample with passing O/C tests by the end of 2020. We will continue to use CLO indexes to track CLO performance throughout 2021.

As vaccine rollouts in several countries continue, S&P Global Ratings believes there remains a high degree of uncertainty about the evolution of the coronavirus pandemic and its economic effects. Widespread immunization, which certain countries might achieve by midyear, will help pave the way for a return to more normal levels of social and economic activity. We use this assumption about vaccine timing in assessing the economic and credit implications associated with the pandemic (see our research here: www.spglobal.com/ratings). As the situation evolves, we will update our assumptions and estimates accordingly.

Appendix

Table 3

Table 4

Table 5