Managers work to add value for investors by reinvesting and repositioning their portfolios to increase returns in benign economic environments, as well as protecting against downside risk during more turbulent times. As a result of the COVID-19 outbreak, and the social distancing measures put in place to contain it, the loan market has undergone turmoil during the past several months that is unlike any seen previously. During the global financial crisis in 2008-2009, negative rating actions on speculative-grade corporate loans (rated 'BB+' or lower) played out over a span of six quarters or more, whereas the current downturn has turned quarters into weeks, with large volumes of negative rating actions occurring in a compressed time frame. The resulting change in collateralized loan obligation (CLO) collateral pool credit profiles over the period has been dramatic, but the change has also provided managers an opportunity to prove themselves and add value.

Since early March, the average U.S. CLO has seen nearly 9% collateral turnover, as collateral managers attempt to reduce exposure to 'CCC' rated loans, build par, or both. Given the negative rating actions on speculative-grade loan issuers that have seemed to plateau in recent weeks (see "As The 'CCC' Curve Flattens And Loan Prices Stabilize, O/C Tests May Stabilize As Well," published May 21, 2020), now seems like a good time to take an early look at how CLO managers have navigated the turmoil and what the results have been thus far. The current downturn resulting from the COVID-19 outbreak is the first broad-based economic stress faced by CLO 2.0 transactions, which were originated in the years after the global financial crisis and should benefit from additional credit enhancement and other changes put in place after the 2008-2009 downturn.

Swift Changes In Collateral Pools

The credit profile of U.S. CLO collateral pools has shifted more rapidly over the past several months than at any other point in the history of the CLO market. Looking at the portfolio exposures across our CLO Insights 2020 Index of 410 U.S. S&P Global Ratings-rated reinvesting broadly syndicated loan (BSL) CLO transactions (see "U.S. CLO Exposure To Negative Corporate Rating Actions (As Of May 17, 2020)," published May 19, 2020, we see that since March 1, 2020:

• Ratings on around one-third of the 1,500 obligors in the collateral pools were either lowered or placed on CreditWatch with negative implications;
• Ratings on more than 170 of these obligors were lowered into the 'CCC' range (or to 'B-' and listed on CreditWatch negative, which we treat as 'CCC' for purposes of our CLO analysis); and
• Ratings on more than 30 obligors were lowered into the nonperforming range (from a performing rating of 'CCC-' and above).

All the credit metrics relating to CLO portfolio exposures within this study are notched for ratings listed on CreditWatch (if a rating is on CreditWatch negative, the rating is notched down by one subcategory, and if a rating is on CreditWatch positive, the rating is notched up by one subcategory).

Table 1

Due to these actions, the 'CCC' rated buckets in our CLOs increased to 12.25% as of May 25, 2020, from 4.13% as of March 1, 2020, while exposure to nonperforming assets increased to 1.35% from 0.63%. Together, these exposures increased by 8.84%. Based on trustee reports we received during this period, portfolio turnover across individual CLOs ranged from less than 3% to 34%, with an average of about 8.75%. As noted in the table above, there was an average decline in par during this time of about 21 basis points (bps) across the index, ranging from a decline of 2.4% to an increase of 0.99%. Due to the downgrades, market value haircuts, and the par loss, the average junior overcollateralization (O/C) cushion has declined from 3.73% at the start of March to 1.63% (a decline of 2.10%), with about 17% of CLO transactions in the index failing as of the publication date.

Table 2

If The Collateral Managers Had Done Nothing

Borrowing from the framework we utilized in our 2016 study, we can imagine what could have happened if managers left the portfolios static over the past several months during the time of the COVID-19 outbreak. By taking the CLO portfolios we had for the CLOs as of March 1, 2020, and then comparing the rating transitions as of May 25, 2020, in these static CLO portfolios to the actual current portfolios.

If U.S. CLO managers had done nothing to the portfolios over the past several months, by May 25, 2020, we would have seen:

• Exposure to companies in the 'CCC' rated category, on average, would have risen to 13.2% (0.93% higher than what actually happened);
• Exposure to companies with ratings below 'CCC-' (i.e., nonperforming loans) would have risen 1.7% (0.30% higher than what actually happened); and,
• Together, the 'CCC' category and nonperforming bucket would have been 1.23% higher than it actually was; and because there was no trading, par would not have declined.

Interestingly, had the managers done nothing, exposure to 'B-' rated CLOs would have been 0.6% lower, suggesting part of the proceeds from reductions in the 'CCC' and nonperforming buckets were used to move slightly up the credit scale and purchase 'B-' rated loans.

Table 3

Table 4

Weighing the benefits

By turning over 8.75% of the portfolio, managers were able to reduce their 'CCC' and nonperforming exposures by about 1.23% at the cost of 21 bps in par. These trades likely saved a few deals from failing their junior O/C tests.

By May 25, 2020, almost all of the deals within the index exceeded the 7.5% 'CCC' threshold. However, if the 'CCC' buckets were 1.23% higher, it is likely a few more deals would have failed their O/C tests. As of the CLO trustee reports available to us as of May 25, about 17% of the CLO transactions within our index are failing one or more of the tranche O/C tests, while another 7% were passing but currently have less than 50 bps of cushion for their junior O/C test.

On average, about 30% of U.S. CLO portfolio assets as of March 1, 2020, experienced negative rating actions by May 25, ranging from 22% to 46% for individual CLOs across the index. Because several of the corporate sectors that were more directly affected by the COVID-19 containment measures (e.g., airlines, hotels, restaurants, and leisure and entertainment) had higher ratings at the start of the year, portfolios with stronger credit quality (i.e., a lower S&P Global Ratings' Weighted Average Rating Factor, or SPWARF) before COVID-19 were not any less likely to experience negative rating actions than CLO portfolios that had weaker rating profiles ahead of the downturn.

As a result of manager trading, a majority of the CLO transactions reduced their exposure to the 200-plus issuers that have seen ratings lowered into the 'CCC' category and below (i.e., nonperforming assets) since March 1. In Chart 1 below, we plot the following across the index with ratings data as of May 25:

• On the vertical axis, the hypothetical increase in exposure to 'CCC' category rated assets plus nonperforming assets had the CLO managers not made any trades, and
• On the horizontal axis, the actual increase in exposure to 'CCC' category rated assets plus nonperforming assets plus par loss for the actual portfolios, which include trades the managers made.

Chart 1

We find most deals fall above the diagonal line, indicating most deals within the index experienced a smaller increase in 'CCC' and below exposure, adjusted for par loss, than they would have experienced if the manager stopped trading after COVID-19. A smaller proportion of the index fell below the line, which means the exposure to 'CCC' and nonperforming categories were higher than if the portfolio were static. This suggests that the manager may have purchased issues with weaker ratings, although as a mitigating factor these trades may have resulted in par gain.

Same COVID-19 Reality, Different Outcomes

Using the change in SPWARF to represent change in credit quality and par loss (see chart 2), there are various outcomes from a range of CLO manager strategies. Across the index, the average increase in SPWARF was 332, while the average change in par was (21 bps).

Chart 2

We can divide the outcomes into four quadrants:

• Lower left quadrant (risk off): These transactions experienced below-average credit deterioration but experienced par loss, suggesting the manager sold off lower-rated credits at a loss. Some transactions sold significant portions of their 'CCC' bucket, which may have helped O/C tests come back into compliance ahead of the April payment date. Because the O/C numerator is typically haircut by the worst market prices of the 'CCC' excess, by selling the high-recovery 'CCCs', the haircut to the O/C numerator due to the par loss from selling high-recovery 'CCCs' is likely less than the haircut applied to the 'CCC' excess with the lowest market value if the manager held on to the high-recovery 'CCCs'.
• Upper right quadrant (risk on): These transactions experienced above-average credit deterioration, but as a mitigating factor picked up par along the way. Many of the transactions had lower 'CCC' exposure and a higher junior O/C cushion as of March 1, meaning they could absorb more credit deterioration before becoming at risk of failing an O/C test. Given the higher flexibility, some of these transactions were less likely to sell their weaker assets and, instead, were more likely to purchase additional lower-rated assets, building par along the way (many of these transactions are those that are below the diagonal line in chart 1).
• Upper left quadrant (win-win): These transactions experienced lower-than-average credit deterioration and par gain. They likely benefited from asset selection that avoided many of the sectors that were directly affected early on. Some also benefited from having a significant amount of cash going into March 1, allowing the managers to build par in quality assets.
• Lower right quadrant (not where you want to be): These transactions experienced both above-average credit deterioration and par loss. They likely had significant exposure to directly affected sectors at the start of the COVID-19 outbreak, and also tried to limit the deterioration by selling at a loss. Many issuers within the affected sectors (i.e., hotels, restaurants, and leisure; entertainment; and airlines) had above-average ratings with tight spreads. As these sectors were directly affected by COVID-19, the values for these low-yielding loans dropped notably.

Still In Early Innings

Given the swift impact that COVID-19 had across multiple sectors, having a diverse portfolio was not enough to shield CLO portfolios from credit deterioration, even if the portfolio before the downturn was higher than average credit quality. Given the reduction in corporate rating actions since the peak in late March and early April, it seems like a good time to see how managers have responded and aligned their portfolios to the new corporate rating landscape of the two to three months since the start of the pandemic. Some managers have been willing to give up some par in order to reduce their 'CCC' exposures--with some transactions experiencing significant par loss due to 'CCC' reduction--only to pass an O/C test with a 0% cushion ahead of a payment date. On the flip side, we have seen some managers build par by purchasing more assets. As time goes on, we will learn more about asset managers' actions as they juggle between managing credit, par, and their CLO O/C tests.

Appendix

Table 5

Table 6

Table 7

Table 8