As of today, exposure to loans from 'B-' rated obligors has reached a record level, constituting nearly 19% of U.S. broadly syndicated loan (BSL) CLO transaction collateral pools. This reflects the trends playing out in the overall U.S. corporate loan market (see chart 1).

The increase in loans from 'B-' rated obligors has drawn the market's attention, as ratings on these companies can be volatile. Even in relatively benign credit environments, about 10% of 'B-' rated issuers, on average, experience downgrades. During periods of economic stress, credit deterioration among 'B-' issuers can be particularly severe: In the last credit downturn, more than 40% were downgraded over the course of a given year, some of which defaulted (see "The Expansion of the 'B-' Segment Is Feeding Growing Vulnerabilities," Sept. 25, 2019).

In an effort to shed some light on the potential impact of downgrades and defaults within the cohort of 'B-' rated obligors on rated U.S. BSL CLOs, we created three scenarios of increasing severity, based on broad hypothetical outcomes. These scenarios are not meant to be predictive or part of any outlook statement, nor are they meant to reflect any of the stresses outlined in our rating definitions; they are specifically geared to address the questions and concerns voiced to us from CLO market participants.

(For additional color regarding our opinion of the potential impact to CLO ratings in a (global) macroeconomic downturn, see "When the Cycle Turns: How Would Global Structured Finance Fare In A Downturn," Sept. 4, 2019.)

Chart 1

So How Did We Get Here Anyway?

The increase in 'B-' obligors within U.S. BSL CLO collateral pools is simply a reflection of rating trends happening in the U.S. leveraged loan market. While the proportion of U.S. BSL CLO collateral consisting of loans from 'B-' obligors is at a record level of nearly 19%, that's actually lower than the proportion of 'B-' obligors among the larger speculative-grade corporate issuer universe, which now exceeds 20%. Just under half of these companies found their way to a 'B-' by being assigned this rating from the start (rather than by a downgrade). In prior years, an initial (corporate) rating of 'B-' was much less common. The count of companies initially rated 'B-' increased notably after the energy slowdown in 2015-2016, as highly leveraged issuers took advantage of the low rates and accommodating credit conditions, particularly in the tech and health care sectors, to issue debt at lower rating levels. As these 'B-' issuers became more prevalent within the leveraged loan market, their loans also found their way into reinvesting and new-issue BSL CLOs.

That said, a significant portion of the currently outstanding 'B-' ratings did get there via the downgrade route over the past year. Some such downgrades contributed to the growing CLO 'B-' exposure. From the start of Q4 2018 through the end of Q3 2019, just under 100 corporate issuers with loans held in CLOs had ratings lowered to 'B–'; these issuers make up about one-third of the current U.S. BSL CLO 'B-' exposure.

Our Analysis

CLO participants have tried to imagine the next economic downturn, assuming defaults spike, perhaps to levels similar to the credit crisis or the dotcom bubble. Some CLO participants are also interested in the effects of a milder scenario in which defaults don't necessarily spike, but downgrades do occur, inflating the 'CCC' buckets of CLOs. By the end of the third quarter of this year, the average 'CCC' category exposure in U.S. CLOs is around 4.7%, not too far off from historical levels during benign times (see chart 1, above). Of particular concern is the credit stability of the 357 'B-' rated obligors, which now account for 19% of U.S. BSL CLO exposure.

We created three scenarios that do not necessarily represent an economic downturn, but are meant to help explore the potential CLO impact given the concern over the 'B-' exposure. In and of itself, a downgrade to 'CCC' from 'B-' may represent a fairly mild event; however, in aggregate, such downgrades can have an amplified impact on the cash flow mechanics of the CLO, as well as the ratings assigned to the CLO tranches. Most U.S. BSL CLOs have an "allowable" bucket for 7.5% of the collateral coming from obligors with 'CCC' ratings. Beyond that, loans falling into the 'CCC' excess amount will see their par value haircut for purposes of calculating the overcollateralization (OC) ratio tests. This can divert interest payments away from the CLO equity or even the rated junior CLO tranches (see "S&P Global Ratings' CLO Primer," Sept. 21, 2018).

Given that the maturity wall of corporate debt coming due has been pushed out (CLO loan exposures currently have less than 5% maturing within the next two years), corporate interest coverage ratios are strong, and a large majority of loans are covenant-lite, we assumed in our first two scenarios there would be credit deterioration on the 'B-' exposure, but no defaults. In the third scenario, we overlay defaults on top of the 'B-' deterioration. Given the negative correlation in the supply and price of 'CCC' assets, we also assume incremental declines in the average price of assets rated 'CCC' for each of the scenarios in determining the impact to the OC ratio test cushions.

The scenarios

Our scenarios are outlined below in chart 2 (see the appendix for further details). We caution that the results are based on the application of the current models we use to rate CLOs as part of our criteria (see "Global Methodology And Assumptions For CLOs And Corporate CDOs" and "FAQ For CLO Managers, Trustees, And Arrangers: Implementation Of S&P Global Ratings' Updated Global CLO Criteria," June 21, 2019) on a sample of CLOs that were rated under the prior criteria. A rating committee applying the full breadth of S&P Global Ratings' criteria and weighing qualitative factors might, in certain instances, assign a different rating than the quantitative analysis would indicate.

Chart 2 (original)

Chart 2 (updated)

Scenario 1 

In this relatively benign scenario, we assume only the weaker 'B-' issuers get downgraded into the 'CCC' category. For purposes of this scenario, we classify a 'B-' obligor as weaker if it currently has:

• A rating on CreditWatch negative,
• A rating outlook of negative, or
• Loans that are trading below 90.

Altogether, this adds up to 19.7% of the 'B-' exposure, or 3.7% of overall CLO collateral. Again, we do not assume any further defaults within this scenario.

The credit deterioration to the weaker 'B-' issuers in this scenario leads to the average 'CCC' bucket of our CLOs nearly doubling, to 9.4% from 4.7%. More than 73% of the deals within our sample would have a 'CCC' bucket of over 7.5% and, hence, would be exceeding their allowable 'CCC' bucket. The excess 'CCC' loans' par value would be haircut to market value for purposes of calculating the CLO OC ratio tests. Given the increase in supply of 'CCC' assets in the loan market under this scenario, we assume a lower market price of 70. The average 'CCC' excess of 1.87% is exposed to market value haircuts, which will be 30 points. As a result of this scenario, the junior OC cushion decreases to 3.51% from 4.22%--down by about 71 basis points on average--and just 2.1% of the CLOs within our sample fail their junior OC test.

The impact of this scenario on the OC tests is muted because most CLOs currently have less than 7.5% exposure to 'CCC' assets; thus, they have some cushion before the OC tests are exposed to market value haircuts (one-fourth of the deals within our sample continue to have less than 7.5% exposure to 'CCC's even after this scenario is applied). (See table 1.)

Table 1

The impact on our CLO ratings is also relatively muted, as this is indeed a mild scenario. All the CLO 'AAA' and 'AA' rated notes within this sample have enough cushion to absorb the downgrades of the weaker 'B-' obligors. Further down the CLO capital stack, the ratings impact is greater, but not by much: Only a small portion (15%) of 'B' category CLO ratings are lowered by one notch, while a smaller portion of 'BB' category (9%), and an even smaller portion of 'BBB' and 'A' category (4% and 1%, respectively), notes are lowered by one notch.

The results of this scenario are, in some ways, similar to the impact of the energy slowdown on the pre-2015 vintage U.S. CLO 2.0s. A few handfuls of junior notes originally rated within the 'BB' and 'B' categories saw initial downgrades of one notch (or more in subsequent downgrades), while just one note originally rated within the 'BBB' category was lowered by one notch.

Scenario 2 

In this scenario, we take on the role of a wary CLO investor who views all of the 'B-' obligors within their CLO collateral pools as potential future 'CCC' obligors. Accordingly, we assume that all 357 'B-' obligors experience a downgrade into the 'CCC' category (but that none of them default).

Our hypothetical transition of all 'B-' CLO exposures (about 19% of total CLO collateral) to 'CCC' resulted in a significant deterioration to the credit quality of the portfolio, even though no defaults or par losses were applied. CLO 'CCC' buckets increased by 19% on average, resulting in more loans being haircut for purposes of the OC ratio tests and more CLO rating downgrades. All CLOs within our sample would now be exposed to market value haircuts to their OC ratios, as the entire portion of 'B-' exposures flood the 7.5% 'CCC' buckets, reaching an average 'CCC' bucket of 23.3%.

Since the supply of 'CCC' issuers has now increased more dramatically, we assume an average market value of 65 for these assets. The larger 'CCC' excess is now haircut by 35, causing junior OC cushions to fall to (1.76)% on average. A majority (86%) of the CLOs within our sample will fail their junior OC tests. (See table 2.)

Table 2

The majority of the notes from our sample originally rated at investment-grade levels (of 'BBB-' or higher) have their ratings affirmed, while a majority of the CLO notes originally rated with speculative-grade ratings are downgraded, some by multiple notches. More than half of the CLO 'BB' category notes and a large majority of the 'B' notes get downgraded by one or two notches in this scenario.

Some CLOs exhibited resilience across the capital structure, mostly because they had either less exposure to 'B-' obligors in their portfolios or a larger rating cushion to absorb the deterioration, or both. We also found that vintage can have an effect, with newer CLO collateral pools tending to have portfolios with less exposure to 'B-' obligors. We also noted a wide range of 'B-' exposures across the different CLOs and managers. Some CLOs are invested more heavily in the larger, more commonly held corporate names listed in our quarterly top 250 CLO obligors report (see "The Most Widely Referenced Corporate Obligors in Rated U.S. BSL CLOs: Third-Quarter 2019," Oct. 9, 2019). These obligors typically have higher ratings and potentially greater rating stability at a cost of lower spread.

Scenario 3 

Despite our sanguine speculative-grade default outlook in the real world (see "The U.S. Speculative-Grade Corporate Default Rate Is Expected To Reach 3.9% By September 2020," Nov. 19, 2019,"), in this scenario, we imagine the same credit deterioration in scenario 2 (all 'B-' obligors get downgraded into the 'CCC' range), but also assume that everything currently in the U.S. BSL CLO 'CCC' buckets (a total of 4.65% of CLO collateral) defaults. This is slightly above the base default forecast of 3.90% for U.S. speculative-grade corporate issuers. Additionally, for purposes of OC ratio haircuts, we assume the market price of the original 'B-' exposures that transitioned to 'CCC' under our hypothetical scenario will decline to 60, while the original 'CCC' exposures (all of which we assume will default) recover only 45%.

As a result, the average par loss on our sample from this scenario is 2.7% (recovery of 45% from the default of the 'CCC' bucket, 4.6% on average), while the 'CCC' bucket increases to 19.1% (this is lower than in scenario 2 because we assume the current 'CCC' exposure defaults). The 2.7% par loss will have a direct reduction on the OC ratio, while the excess 'CCC' bucket will result in further haircuts. Assuming an average market price of 60 for the 'CCC' assets, the average junior OC cushion drops to (3.5)%, while 97% of the sample fails one or more of their coverage tests. Only the deals that just recently closed, with very few 'B-' and 'CCC' exposures, are able to withstand this scenario within their junior coverage tests. (See table 3.)

Table 3

Still, even under this scenario, all but one of our CLO 'AAA' tranche ratings are affirmed. Relative to scenario 2, the added par loss from the 'CCC' defaults in scenario 3 result in significant rating transitions for the CLO notes within our sample, in particular the 'BBB' and lower categories. The 'AA' and 'A' category CLO notes remain fairly resilient, as a majority of them have adequate cushion to withstand the hypothetical downgrades. But a majority of the 'BBB' and 'BB' CLO notes now see downgrades of one or more notches, while all 'B' notes are downgraded.

Again, vintage and the current state of the CLO portfolio play a big factor in this scenario. Of our sample of CLOs, 5.5% are already experiencing market value haircuts (before the hypothetical downgrades), a handful have 'CCC' exposures currently greater than 10% (one deal within our sample currently has a 'CCC' bucket of 13% before the hypothetical downgrades). These deals would fare particularly poorly in this third scenario, as we assume the 'CCC' exposures will recover only 45%. Par loss, along with credit deterioration, will have a more dramatic impact on both the CLO OC ratios and the CLO ratings. Interestingly, only one tranche originally rated 'AAA' experienced a downgrade of one notch under this scenario. This deal was an older seasoned CLO that will exit its reinvestment period soon and has above average exposure to 'CCC' and 'B-' rated assets. An older CLO would have less time to benefit from interest diversion due to OC failures within our cash flows and, thus, was more vulnerable to a rating transition.

Vintage And Manager Matter

The rating distribution of a CLO portfolio can help in understanding how the CLO will perform through a downturn. Although the above three scenarios hardly classify as an economic downturn, the results may provide insight for CLO participants over the more imminent concern regarding the elevated level of 'B-' rated issuers held in CLO portfolios. Manager style has a significant impact on the rating distribution of a CLO collateral pool during benign times, with some CLOs having less than 10% exposure to 'B-', while others have over 30%. Not surprisingly, the CLO portfolios with weaker credit rating distributions fared worse under all three scenarios. Vintage also plays a key factor within our study. The older CLOs in our sample were more likely to have higher 'B-' and 'CCC' buckets, and also had less time to maturity, meaning these deals had less time to benefit from excess spread capture within our hypothetical scenario. However, expected senior note paydowns in the near future (not modeled) of these vintage CLOs will significantly alter the notes' resilience to further downgrades and or defaults.

Indeed, given the transparency of the CLO market, we can imagine a lot of hypothetical scenarios to test the resiliency of CLO notes. One key aspect that is difficult to predict, however, is the impact of the manager intervention during a downturn. We have seen that CLO managers have been able to preserve value during stressed periods (see "How Do CLO Managers Perform In Times Of Stress?" Sept. 6, 2016). Perhaps deals with more time left in their reinvestment periods may be able to benefit from this intervention, while deals that amortize in the middle of a downturn are exposed to more relative risk for the junior CLO notes. Indeed, we are starting to see some earlier-vintage amortizing CLO 2.0s begin to fail their junior OC tests for the first time in the third and fourth quarters of this year, most of which experienced deterioration during the peak of the energy slowdown a few years ago.

Appendix

Running the cash flows

For purposes of our scenario testing, we generated a quantitative analysis for our sample CLOs using the same tools we use when rating the CLO transactions under our current criteria. Our CDO Evaluator credit model assesses the overall credit quality of a portfolio of assets based on the rating and maturity of each asset, as well as the correlation between assets, and produces expected default rates at different rating levels--scenario default rates (SDRs). As part of our current criteria, the SDRs produced by the latest version of the CDO Evaluator model are approximately 3%-5% lower than the prior version given the same portfolio parameters (see "Credit FAQ: Understanding S&P Global Ratings' Updated CLO And Corporate CDO Criteria," June 26, 2019). Our S&P Cash Flow Evaluator model is used to assess the ability of the tranche in a given CLO to withstand default rates under various interest rate and default timing scenarios--break-even default rates (BDRs). While ratings are assigned by committee and our criteria encompass a variety of qualitative and quantitative components, looking at the output of these two models for a given CLO portfolio and structure can provide an indication of the ratings we would assign under a given scenario.

For the purposes of this study, we focused on reinvesting U.S. BSL CLO 2.0 transactions that have gone effective and started issuing trustee reports as of September 2019, all of which were rated under our criteria prior to the criteria update on June 21, 2019. We excluded amortizing transactions because senior note paydowns will significantly alter the capital structure of the transaction at closing. The results of this study are therefore based on portfolio information as of third-quarter 2019 and capital structures that were rated before June 21, 2019, which may not be indicative of the portfolios and capital structures of transactions that will go effective going forward.

Base case

We ran a base case using data as of third-quarter 2019 to determine the model-implied rating--the maximum rating level at which the CLO tranche can withstand the level of defaults as determined by CDO Evaluator given the parameters of the CLO portfolio (in other words, the highest rating where the BDR is greater than the SDR). For example, for a CLO tranche to attain a model implied rating of 'AAA (sf)', the tranche must be able to withstand (within our Cash Flow Evaluator model) the 'AAA' level of defaults as determined by CDO Evaluator, given the portfolio information as of third-quarter 2019 (i.e., the BDR of the CLO tranche [the Cash Flow Evaluator model output] must be equal to or greater than the SDR (the CDO Evaluator model output] at the 'AAA' level). The sample used for our study included 1,263 S&P Global Ratings-rated tranches from 301 CLOs where the model implied rating of the base case was within the same rating category as the current rating of the CLO tranche.

The scenarios

Given portfolio information as of third-quarter 2019, we made adjustments to the credit ratings of the portfolio assets (as described within the scenario descriptions above) to determine the new model implied ratings under each scenario. This means that the adjusted portfolio would be the starting point to which we apply our scenarios at each rating level. For example, lowering the ratings of the 'B-' obligors to 'CCC' in Scenario 2 would result in an increase to the SDR determined by CDO Evaluator, likely resulting in either a lower model implied rating for the CLO tranche or the same model implied rating at a lower rating cushion (a positive difference between BDR and SDR), all else being equal. (See table 4.)

Table 4

In this example, the 'AAA' SDR of the base case (with a 'CCC' bucket of 5.61%) was 60.12%, and the 'AAA' BDR of the senior note is 66.36%. Because the BDR is greater than the SDR, this note passes our model runs with a cushion of 6.24% within the base case. In scenario 3, we assume the 'B-' exposure gets downgraded to 'CCC' and the current 'CCC' bucket defaults with a recovery of 45%. The credit deterioration of the 'B-' downgrades result in an increase to the 'AAA' SDR, to 62.45%. The par loss from the default of the 'CCC' bucket results in a decline in the 'AAA' BDR to 64.32%. The BDR under scenario 3 is still greater than the SDR under the same scenario, so this note still passes our cash flow stress (though cushion is much lower, at 1.87%).

For each of the three scenarios above, the notch differential is the difference between the model implied rating under the base case and the model implied rating under each scenario.

Nonperforming model-implied ratings

For Scenario 3, some of the notch differentials for the junior CLO notes (originally rated within the 'BB' and 'B' categories) were five or more notches lower than the base scenario. In the scenario where the BDR of a tranche is lower than the 'CCC-' SDR, the model-implied rating is a nonperforming rating. Generally, a committee would not lower a rating to a nonperforming rating unless the tranche is at imminent risk of default, though we include it here for illustrative purposes.

Calculating the OC tests

In this study, we assumed a 7.5% 'CCC' trigger to determine the 'CCC' excess. In each scenario, we added the new 'CCC' assets that have transitioned to the 'CCC' category exposure from 'B-'. The excess 'CCC' bucket was haircut by 100 minus the average market price assumed for 'CCC' assets under each scenario. The 'CCC' haircuts are typically applied at the lowest price within the 'CCC' bucket; however, we applied the same haircut across all deals. In the third scenario, we applied a haircut of 100 minus the recovery value assumed for the new defaults. We calculate the impact of these haircuts on the junior-most OC test to determine if the OC ratio is still above its trigger (typically at the 'BB' or 'B' level). The junior OC test is typically the OC test with the least amount of cushion at closing, so it would be the first to fail if we were to enter a period of stress.