Today, with high interest rates and slowing economic growth putting pressure on speculative-grade companies, the prospect of corporate rating downgrades into the 'CCC' range is a frequent topic of conversation with CLO investors, managers, and others. The overall average U.S. BSL CLO 'CCC' basket (including transactions that have closed since second-quarter 2020) has been gradually creeping upward since August 2022, when it stood at 4%, to just over 5% of total assets as of today. With the average BSL CLO having more than 30% of total assets coming from 'B-' obligors, it wouldn't take a lot of corporate rating downgrades to increase the average BSL CLO 'CCC' baskets significantly. However, with the current average junior O/C test cushion sitting at a healthy 4.5% though, it would take a lot of downgrades before many tests started to fail. In this article, S&P Global Ratings explores the hypothetical impact of different CLO 'CCC' exposures and market value declines on CLO junior O/C test cushions.

U.S. BSL CLO O/C ratios have had a good run to date. During the depths of the COVID-19 pandemic in second-quarter 2020, as exposure to assets from 'CCC' range companies peaked at 12.31%, the average junior O/C ratio test cushion dropped to 1.13%, and about a quarter of the transactions failed their junior O/C test and some diverted interest proceeds away from equity. Since then, things have gotten a lot better: by second-quarter 2022, the average reinvesting pre-pandemic BSL CLO exposure to 'CCC' obligors had fallen to 4.80%, and CLOs had steadily built-up par as managers took advantage of loan prices and traded collateral. The average junior O/C test cushion of reinvesting pre-pandemic CLOs had recovered back to 3.80%, and most BSL CLOs had built up a solid cushion ahead of whatever might come next.

A Refresher On Haircuts And CLO Overcollateralization Tests

Coverage tests are a defining feature of CLO transactions as they act as "temporary shock absorbers" when the underlying collateral experiences stress. As the name implies, CLO O/C (or par value) coverage tests measure how much coverage the collateral assets are providing various classes of notes issued by the CLO. The O/C ratio is calculated by taking the total par value of the assets, with some adjustments made, and dividing this by the balance of the tranche being tested and any tranches senior to it. The calculated ratio is then compared to a preset threshold for each tranche level O/C test to determine whether the test is passing or failing. If the test fails, interest proceeds are redirected away from the CLO equity (and any CLO tranches below the failing tranche level) and used to reduce the balance of the CLO 'AAA' notes outstanding. These paydowns to the CLO's senior tranche balance help move the failing test(s) back towards compliance.

During times of economic stress that have affected CLO collateral, like the Global Financial Crisis (GFC) in 2008-2009 or the pandemic-driven downturn in 2020, these tests have performed as they were designed to by restoring credit enhancement to the more senior CLO notes at the expense of the equity (or if the O/C test failures are significant enough, the CLO equity plus some classes of notes). To best capture any potential deterioration in the CLO's collateral, several adjustments may be made to the numerator of the O/C ratio, which measures the par value of assets available to support the CLO notes. Among others, these adjustments often include:

• Defaulted assets, which, for purposes of the O/C tests, generally get carried at the lower of market value or assumed recovery under rating agency methodology.
• Assets from 'CCC' range obligors above a 7.5% threshold with the lowest market prices (also known as "excess 'CCC' assets"), which typically get carried at market value. Note that most CLOs include obligors rated 'B-' on CreditWatch negative as part of the total 'CCC' exposure. It's also worth noting here that CLOs are not compelled to sell 'CCC' assets above the 7.5% threshold, although the manager may choose to; they just can't carry the excess 'CCC' assets at par for purposes of calculating the O/C tests.
• Assets purchased at a significant discount to par ("discount purchase assets") as defined by the CLO's documents, which get carried at purchase price until certain conditions are met.
• Long-dated assets may be carried at a lower value, as determined by the indenture.
• Interest deferral balances on payment-in-kind (PIK) assets generally do not receive par credit.

Other haircuts are also included in most CLO transaction documents but are in practice less likely to be applicable. The haircuts given above are more likely to be seen, although during periods with limited corporate rating downgrades and defaults they tend to have (at least on average) only a modest impact on O/C ratios. For example, in 2022, haircuts to the O/C test numerator as reported within CLO trustee reports measured less than 10 basis points as a percentage of total CLO portfolio par (see chart 1). The haircuts were primarily made up of recovery haircuts from defaulted asset exposures. A small number of transactions have already breached their 7.5% 'CCC' threshold and are making up a small proportion of the haircuts.

Chart 1

Stress Scenario Analysis Details

The vintage effect: pre-pandemic compared with post-pandemic CLOs

While averages are a useful barometer to indicate what's going on across the BSL CLO universe, it's worth remembering that metrics for different cohorts of transactions can differ significantly. There is a pronounced vintage effect when pre-pandemic CLOs (those originated in first-quarter 2020 and before) are compared with post-pandemic ones: pre-pandemic CLOs, having experienced par loss and credit deterioration, have higher 'CCC' asset exposure and lower junior O/C test cushions relative to the post-pandemic transactions. Within the pre-pandemic CLOs amortizing transactions have even higher 'CCC' asset exposure percentages and lower junior O/C test cushions, on average. These transactions can be vulnerable to concentration risk as better credits pay down sooner, leaving behind overall weaker credits and resulting in a greater proportion of 'CCC' in the portfolios. Several of the older amortizing pre-pandemic CLOs have already breached the 7.5% 'CCC' threshold and are seeing 'CCC' excess haircuts to the O/C test numerator on top of the par loss experienced during the pandemic.

Table 1

Within a given CLO, you can sometimes see divergent outcomes for senior and junior O/C tests. As principal cash from collateral amortization/prepayments are usually used to pay down the senior notes of the CLO, it raises the cushion of senior O/C tests and increases the credit enhancement available to support them. The positive effects of the paydowns mostly outweigh the negative effects of the haircuts for the senior tests of amortizing transactions. At the same time, the positive impact of senior note paydowns is less pronounced for the junior O/C tests; sometimes, the negative effects of the 'CCC' haircuts outweigh the senior paydowns, resulting in a decline in junior O/C cushion. In some cases, we have simultaneously downgraded our ratings on the junior tranche and upgraded our ratings on the non-'AAA' senior tranches of the same transaction (see "CLO Rating Movements: Same Credits, But Different Strokes For Different Notes," published Nov. 20, 2019).

Given this, we provide the results of the scenarios below through the lens of the three cohorts mentioned in table 1. Note that we excluded post-pandemic CLOs that are currently amortizing from the analysis because the sample size was too small.

Issuer ratings, loan prices, and CLO O/C tests

In addition to defaulted assets, CLO O/C ratio tests are sensitive to both the total par amount of the 'CCC' excess (above 7.5% of total par) and the loan prices of those 'CCC' exposures. Because most O/C tests are calculated with the lowest-priced 'CCC' assets being used as the "excess" (above 7.5%) amount, the O/C tests are particularly sensitive to the lowest-priced 'CCC' assets.

Loan prices can react to numerous things, including changes to the rating on the issuer. Table 2 below shows this for U.S. BSL CLO obligors that saw ratings lowered into the 'CCC' range since the start of 2022. Within loans from 'B-' rated companies, there has been a bifurcation in loan prices between issuers that are perceived as being stronger and those presumed to be at risk of being downgraded (by definition, into the 'CCC' rating category or worse). Anecdotally, some market participants have told us they think there is a correlation between the ratings on loans issuers and the prices on their loans that is stronger than it has been in the past, as CLO managers work to avoid companies that may end up in the 'CCC' range.

Table 2

In recent quarters, unsurprisingly, we find loans from 'B-' issuers with ratings lowered into the 'CCC' rating category experience loan price declines during the quarter of downgrade even where the broader loan market does not see declines. We also see price declines for these loans in the quarter before the downgrade occurs. Average loan prices of the issuers that would eventually get downgraded into the 'CCC' range during the quarter have been decreasing, to 78 at the start of 2023 from 95 at the start of 2022. The average price change seemed to be greatest across the 'B-' downgrades in fourth-quarter 2022, when there were higher overall levels of downgrades; these loans saw a 9.5-point price drop, or a 10.4% price drop, on average.

Exploring The Scenarios And Their Results

To explore the potential impact of various levels of 'CCC' downgrades on BSL CLO junior O/C ratio tests, we generated four hypothetical scenarios of different 'CCC' range and below exposures (10%, 15%, 20% and 25%). Because loan prices also impact how many tests end up failing and by how much, for each 'CCC' downgrade scenario, we've also made various assumptions on the prices of the affected loans to calculate the impact to CLO O/C ratios, covering a range of haircuts, from using current market prices (i.e., no haircut) up to a 50% market price haircut off current market prices, in increments of 10%.

First, a quick note on 'B' tranches and their associated O/C tests. Nearly all BSL CLOs today are issued with five credit classes ('AAA', 'AA', 'A', 'BBB-', and 'BB-') plus CLO equity. At various points in the past (and potentially again in the future), some BSL CLOs have also been issued with a 'B-' tranche in the capital stack. This is a function of transaction economics; when feasible, including a 'B-' tranche can increase leverage for the CLO equity, and (depending upon the spread the tranche goes out at) can be accretive to equity returns. In recent years, many U.S. BSL CLOs issued prior to the arrival of the pandemic in first-quarter 2020 included a 'B-' tranche at the time of issuance, while many of those issued since the pandemic have not.

To the extent a CLO includes a 'B-' tranche and that tranche has an O/C test associated with it, this will be the junior-most O/C test for the CLO (rather than the O/C test associated with the 'BB-' tranche above it); and given the place in the CLO capital structure and threshold associated with the O/C test, it is more likely to fail. In the results below, we include the proportion of 'B-' O/C tests that fail under each scenario, but it should be noted that 100% of 'B-' O/C tests failing does not mean that 100% of the CLOs are failing their junior-most O/C test, given that not all CLOs include a 'B-' O/C test. In addition, the sample size of post-pandemic reinvesting transactions with a 'B' rated tranche is small. We include them for the sake of making the data set more complete, but due to the sample bias, some of the average 'B' O/C test stats that may appear in the tables below may be out of line with the other O/C tests within the same cohort.

Scenario 1: BSL CLOs see 10% exposure to 'CCC+' and below

The average U.S. BSL CLO currently has just over 5.1% exposure to assets from 'CCC' range obligors and about 0.7% exposure to assets from obligors with a nonperforming rating; to achieve this scenario, another 4.2% of CLO assets--64 obligors--would need to see downgrades (if you're interested in how we selected which obligors in CLOs would see downgrades, see the Appendix section for details). The current CLO exposure to assets from companies rated 'B-' with a negative outlook happens to be about 4.1% of total CLO par, so this 10% scenario is like one where all CLO assets from 'B-' obligors with a negative outlook experience a downgrade.

With a 10% average BSL CLO exposure to 'CCC' and lower assets, we see:

• 20% of the CLOs within our sample do not exceed the 7.5% 'CCC' threshold under this scenario.
• Of the remaining CLOs that exceed the 7.5% threshold, the average 'CCC' excess amount is about 3.1%.
• There is a gradual decline in O/C test cushions as we apply successively higher market value haircuts on the downgraded assets.
• For CLOs that have 'B' tranches and have a 'B' class O/C test, many see failures and would presumably need to divert excess spread away from CLO equity on the next payment date. This was true for both the reinvesting and amortizing CLOs in our sample.
• 'BB' class O/C tests across reinvesting CLOs mostly still see a positive cushion, while amortizing CLOs see their cushions approach zero under the more punitive market value decline assumptions.
• See our interactive dashboard to view the results of this stress under the various haircut assumptions: https://www.spglobal.com/ratings/en/research-insights/sector-intelligence/interactives/u-s-bsl-scenario-analysis.

Chart 2

Chart 3

When we break down the reinvesting cohort into pre- and post-pandemic transactions, we find the post-pandemic transactions have less exposure to 'B-' issuers with very low loan prices, less current 'CCC' exposure, and higher O/C cushion in general, resulting in a significant difference in outcomes. Assuming the loans from the 64 'B-' issuers that get downgraded in this scenario experience a 10% market price haircut, none of the post-pandemic transactions within our sample fail any of their O/C tests, while almost 10% of the pre-pandemic reinvesting transactions are failing their 'BB' O/C tests (compared to 30% of the amortizing transactions).

Table 3

Scenario 2: 15% exposure to 'CCC+' and below

To achieve the 15% 'CCC' exposure required for scenario 2, another 9.2% of CLO assets would need to see downgrades (125 obligors; see the Appendix section for details). For context, during the peak of the pandemic downturn in 2020, the average BSL CLO 'CCC' basket reached 12%, while average exposures to defaulted assets reached nearly 2% across reinvesting U.S. BSL CLOs. As a result, about a quarter of these CLOs experienced one or more O/C test failures and some diverted proceeds away from equity, typically for one payment date.

In some ways, scenario 2 is similar to the stress seen at the peak of the 2020 pandemic. Some transactions saw their 'CCC' buckets swell to over 15% back then, resulting in one or more O/C tests failing. A small number of CLOs saw O/C test failures not just at the junior tranche level but also on investment-grade tranches, deferring payments to the non-investment-grade-rated debt tranches for a short period of time.

Chart 4

Chart 5

In scenario 2:

• The average 'CCC' excess amount is now 7.5%, three times larger than scenario one. This results in modest declines in O/C test cushions as we assume higher market value haircuts on the downgraded assets.
• The average 'BBB' O/C test cushion for amortizing CLOs approaches zero under the 50% haircut scenario, as does the average 'BB' O/C test cushion for reinvesting CLOs.

Given a market value decline assumption of 20% under scenario 2:

• Across amortizing transactions, about two thirds are now failing their 'BB' O/C tests, and about one third are failing their 'BBB' tranche O/C tests (meaning interest payments to 'BB' tranches will be deferred for these transactions).
• Across pre-pandemic reinvesting transactions, 45% are now failing their 'BB' O/C tests, and 24% are failing their 'BBB' O/C tests.
• Across post-pandemic reinvesting transactions, the picture is much brighter: just 5% are failing their 'BB' O/C tests and only 3% are failing their 'BBB' O/C test.
• See our interactive dashboard to view the results of this stress under the various haircut assumptions: https://www.spglobal.com/ratings/en/research-insights/sector-intelligence/interactives/u-s-bsl-scenario-analysis.

Table 4

All of the 'B' tranches from the post-pandemic reinvesting transactions within our sample failed under this scenario. The sample size of 'B' tranches in this cohort was small, as few post-pandemic transactions had a 'B' CLO tranche in their capital structure.

Scenarios 3 and 4: 20% and 25% exposure to 'CCC+' and below, respectively

We include scenarios where 'CCC+' and below exposures approach 20% and 25% in scenarios 3 and 4, respectively, where 206 and 302 of the currently rated 'B-' rated issuers with the lowest loan prices are downgraded (see the Appendix section for details). For scenario 4, about two thirds of the current 'B-' bucket would have to be downgraded, including issuers with loans currently trading as high as 95. There may not be much historical context for scenario 4, but perhaps there is a rough historical parallel for scenario 3. During the GFC, we saw 'CCC' and nonperforming buckets peak at around 11% and 7%, respectively, across the reinvesting CLO 1.0s during the time, where roughly half of these CLO 1.0s experienced one or more O/C test failures.

Chart 6

Chart 7

Table 5

Chart 8

Chart 9

Table 6

Under scenarios 3 and 4:

• The impact of various market value haircuts now have a much more notable impact on the O/C cushions as the excess 'CCC' exposures are now 12.5% and 17.5% under scenarios 3 and 4, respectively.
• Mezzanine and senior O/C tests begin to fail in larger numbers for both amortizing and reinvesting transactions under the harsher market value haircut assumptions.
• Senior note paydowns across amortizing transactions have improved the senior O/C cushions for several transactions, some much more than others depending on how much of the senior notes have paid down.
• See our interactive dashboard to view the results of this stress under the various haircut assumptions: https://www.spglobal.com/ratings/en/research-insights/sector-intelligence/interactives/u-s-bsl-scenario-analysis.

O/C Test Failures Can Cure

Almost all of the transactions that failed one or more O/C tests during the pandemic saw their tests come back into compliance by the end of 2020. Manager intervention, particularly for transactions still within their reinvestment period, has helped to reduce 'CCC' and default buckets, thus helping to improve or maintain O/C cushions. The diversion of cash to paydown senior notes also helps to restore O/C cushion and credit enhancement, all else equal. We find, within our study, that the vintage effect continues to be pronounced, as older transactions that have already gone through stressed periods now have less cushion to absorb additional credit deterioration before the junior O/C tests breach.

Across our sample, the junior O/C tests of the amortizing transactions have less O/C cushion and are more likely to experience failures in each of the various scenarios. Interestingly, despite having higher 'CCC' buckets, some of the senior and even mezzanine O/C tests of amortizing transactions have higher O/C cushion due to prior senior note paydowns, providing some stability in the interest payments for the mezz notes. Reinvesting CLOs generally have more junior O/C cushion than the older amortizing transactions, but again, the differences in 'CCC' exposures and O/C cushions between the pre-pandemic and post-pandemic reinvesting transactions result in very different failure rates under each scenario.

Appendix: How We Did the Analysis

In this study, we considered the impact to O/C ratios under different scenarios. We mainly focused on the impact from the market value haircuts on the excess 'CCC' exposures (the 'CCC' exposures with the lowest market values above the 7.5% threshold). We did not factor in haircuts from discount purchases, deferring exposures, etc., as these represented a smaller proportion of haircuts by the end of first-quarter 2023. We mainly focused on additional 'B-' downgrades into the 'CCC' category, along with various assumptions on drops in market value as a result of the stress scenario. We did not include within our scenarios downgrades to the nonperforming category (which would result in a different type of haircut--the lower of recovery value or market value).

Cleary, an economic environment that produces 'CCC' levels like the ones seen in the scenarios might also produce some defaults. While we didn't explicitly include defaults in the scenarios, the harsher market value price haircuts can implicitly include the assumption that some proportion of the downgraded collateral defaults, resulting in similar haircuts to the O/C numerator.

Table 5

To achieve the four target 'CCC' and below exposures (10%, 15%, 20%, and 25%) across the collective exposures of our full sample of U.S. BSL CLOs, we adjusted the 'B-' ratings on as many obligors as needed, starting with the lowest prices (based on end-of-March 2023 loan prices sourced from Markit). Note that this can produce CLOs with a range of exposures in each of the stresses (for example, within scenario 1, by lowering 64 issuers currently rated 'B-' to 'CCC', 10% of the collective exposures across all 778 CLOs within our sample are rated 'CCC' or below; though across the sample, the individual 'CCC' buckets ranged from 1% (from a recent issued transaction) to over 50% (from an older amortizing transaction)).

Related Research

• CLO Rating Movements: Same Credits, But Different Strokes For Different Notes, Nov. 20, 2019