On Friday, March 27, 2020, S&P Global Ratings held its "U.S. CLOs In The Time Of Coronavirus" webinar to provide investors and others with our views on U.S. leveraged finance and collateralized loan obligations (CLOs) in the current market environment. We held two live sessions covering the same topics: one during local business hours in Asia and the other during U.S. business hours. Each webinar began with an overview of the U.S. macroeconomic outlook before moving on to our forecast for speculative-grade non-financial corporate defaults, an overview of trends in the leveraged loan market, and the potential implications for U.S. CLO transactions. (To listen to a replay of the webinar and download the slides, see the following link: https://event.on24.com/wcc/r/2236630/758403004151FDCC6CA659BC1B0FA316.)

S&P Global Ratings acknowledges a high degree of uncertainty about the rate of spread and peak of the coronavirus outbreak. Some government authorities estimate the pandemic will peak around midyear, and we are using this assumption in assessing the economic and credit implications. In our view, the measures adopted to contain COVID-19 have pushed the global economy into recession (see our macroeconomic and credit updates here: www.spglobal.com/ratings). As the situation evolves, we will update our assumptions and estimates accordingly.

During both sessions, we received numerous questions from participants. We worked to incorporate responses to many of these questions into our presentation of the prepared content, or answer them during the "Q&A" session at the end of the call. However, given the breadth and quantity of questions received, we were unable to answer all of them due to time constraints during the sessions. We are using this credit FAQ to highlight key topics of discussion from the webinar and to address some of the remaining items. Questions have been edited for clarity and to combine similar items.

Corporate Ratings-Related Questions

The double-digit default forecast was a big jump from the previous forecast. Why is it so elevated when debt maturities won't begin in earnest until 2024? What factors do you expect will trigger defaults?

We expect severe strains on liquidity for various companies will lead to more defaults as firms try to preserve cash by entering into distressed restructurings or as they deplete the liquidity necessary to remain solvent. Our near-term concerns are focused on firms' ability to generate revenues amid the economic shutdown and the dramatic strain this will cause on cash flows and liquidity. We expect that if the recession's duration is indeed two quarters, many firms may opt to engage in distressed exchanges to get them through the difficult period, which would boost our default rate. Missed interest payments (rather than just the inability to pay upcoming debt maturities) should also increase.

There have been a variety of rating actions taken on corporate entities as a result of COVID-19 and the oil shock (including CreditWatch negative placements, outlook revisions, downgrades, or downgrades with a CreditWatch negative). What determines the type of rating action taken, and how can investors better understand the differences?

Our objective is always to ensure that our ratings and our outlook or CreditWatch statements appropriately reflect the current situation and our expectations for potential future actions. The type of rating action provides important signals about our expectations for the rating in the near-to-intermediate term, with additional information about key parameters provided in our Outlook or CreditWatch statements. For example, where we have recently revised the outlook to negative, we are signaling that we believe that the current issuer rating remains appropriate, but that there is a roughly one-in-three chance of a downgrade in the near term (typically within 12 months for speculative-grade issuers and around two years for investment-grade issuers). The language in our outlook statement will outline the relevant factors that could lead to a downgrade or a revision in the outlook to stable.

In contrast, a CreditWatch negative placement indicates a higher likelihood of a downgrade--roughly one-in-two (although this threshold doesn't apply strictly during periods of significant uncertainty)--over a shorter period of time (typically less than 90 days). As with outlooks, a CreditWatch negative statement will indicate the additional information or analysis needed before taking a rating action, the factors that might lead to a rating change, and often how many notches down the rating could go.

Of course, in cases of severe stress or outsized events, there are situations where we may downgrade our rating on a company and leave it on CreditWatch negative. This would indicate that we have already determined that the prior rating was no longer appropriate, and that the revised rating is also be at risk of a downgrade subject to our evaluation of the factors highlighted in the CreditWatch statement.

The speed of rating downgrades has been dramatic. What are your thoughts on being as quick to upgrade (or remove from CreditWatch negative) recently downgraded issuers if the environment improves or if the macroeconomic outlook becomes less dire?

Since Feb. 3 (the date we started tracking COVID-19 and oil price-related rating actions), in North America, there have been 639 rating actions (21% investment grade and 79% speculative grade through April 10). The pace of downgrades (shown below for all S&P Global Rating actions) on corporations and sovereigns reflects our view of the almost immediate impact social distancing to slow COVID-19's spread has had on credit quality. Our actions incorporate liquidity considerations, levers companies can control such as cost measures, as well as considerations on how credit quality would look during an eventual recovery (should the issuer have the wherewithal to survive the coming quarters).

We do not expect to raise most ratings as rapidly as we lowered them, since it's hard to see how a rebound in credit quality (and economic activity) will be as sudden as their demise. Many companies will have permanently lost revenue and cash flow that will never be recouped. Consumer behavior will likely be different, at least temporarily; accordingly, the pace of many end markets' recoveries is likely to be slower than the pace at which they fell apart. The prospective financial policies of companies will be key: as we saw after the Great Recession, companies became more financially conservative for a time, but this did not last.

It is important to remember that every company-specific rating action has forward-looking information attached to it, in the form of a written outlook or CreditWatch reasoning. These indicate the possible future direction of rating actions.

For a regularly updated list of rating actions we have taken globally on corporations and sovereigns, as well as summary table and support charts, see the "COVID-19- And Oil Price-Related Public Rating Actions To Date" link on our Coronavirus microsite: https://www.spglobal.com/ratings/en/research-insights/topics/coronavirus-special-report.

There have been a significant number of downgrades into the 'CCC' category. Has your view of what constitutes a 'CCC' category rating changed? How are issuers that will have zero revenue for some period, and ongoing costs, considered in your analysis?

Our criteria for the 'CCC' category has not changed. To summarize, in the 'CCC' category, we view the capital structure of a company as unsustainable: the lower a company is in the 'CCC' category, the more near term a default scenario could be, in our view. For example, if we believe a partial restructuring is likely before maturity, but beyond the next 12 months, a rating of 'CCC+' might be appropriate. If we believe a full or partial restructuring is likely over the next six months, a 'CCC-' rating might be appropriate.

For issuers affected most sharply by social distancing (resulting in effectively no revenues for some period), we focus heavily on liquidity and the issuer's ability to absorb negative cash flow for the duration of the crisis. This analysis is important across the ratings spectrum, but especially so in the lower rating categories. Much of the recent borrowing activity, including draws under existing bank facilities, has been aimed at financing activity liquidity to the balance sheet.

We encourage investors to look at our written outlooks or CreditWatch publications on individual companies to see company-specific liquidity comments and concerns.

Will your views on EBITDA addbacks be any different if, for example, lenders add back some COVID-19-related costs or even revenue losses?

We will continue to rate to our view of EBITDA rather than the adjusted EBITDA defined by loan documents that typically include a plethora of generous and subjective addbacks. In this crisis, we may even see attempted amendments to addback lost coronavirus revenues or costs.

Nonetheless, this does not mean that the addbacks that are allowed by loan documents won't affect our analysis at all. To start, these addbacks will improve the financial ratio calculation under existing financial maintenance covenants, which should reduce borrower-lender covenant pressure; we will factor this into our liquidity analysis. More importantly, our analysis will look beyond the expected period of disruption and incorporate our views of expected future financial performance and credit metrics. We are likely to weigh these forward-looking metrics heavily in our analysis for credits where we are not concerned that intense liquidity pressure may trigger an abrupt payment default or distressed exchange.

For more information on EBITDA addbacks, see "When the Cycle Turns: The Continued Attack of the EBITDA Add-Back," published on Sept. 19, 2019.

How much impact could government intervention (including Federal Reserve and CARES Act) have on credit quality?

This is an important evolving topic. We anticipate that any liquidity infusions into the financial system or specific sectors or companies would ease liquidity pressures. However, we do not automatically assume that these favorable liquidity actions would bolster credit quality or avert an eventual default on a rated instrument. We would need to understand the terms and conditions of company-specific infusions to evaluate the impact on issuer credit quality and on the existing debt. In the airline sector, for example, we have not yet included any potential sources of additional liquidity that have not been committed so far in our analysis, such as government aid, although we expect that airlines may be able to access some funding, which could prevent what would otherwise be further deterioration in creditworthiness.

How have COVID-19-related rating actions affected specific sectors?

S&P Global Ratings is publishing details every few days on the breadth and depth of COVID-19 and oil price-related rating actions. Consumer-facing sectors have been the most widely affected sectors by social distancing, which our rating actions reflect. For example, automotive, hotel and gaming, retail, and transportation sectors rank at the top for the percentage of issuers affected by rating actions. As of April 10, 2020, about 27% of North American corporate and sovereign issuers (based on 639 rating actions) had COVID-19 and oil price-related rating actions.

For a regularly updated list of rating actions we have taken globally on corporations and sovereigns, as well as summary table and support charts, see the "COVID-19- And Oil Price-Related Public Rating Actions To Date" link on our Coronavirus microsite: https://www.spglobal.com/ratings/en/research-insights/topics/coronavirus-special-report.

In particular, the technology sector has a high industry proportion rated 'B-'. What is your rating outlook for that sector?

Software is the largest single industry category within U.S. broadly syndicated (BSL) CLOs, with loans from this sector comprising nearly 9% of total CLO assets.

Our sector team has written that the coronavirus will hurt enterprise and consumer IT spending, with bleak ramifications for certain credits in the hardware and semiconductor segments (see "Global IT Spending Set To Slide As Coronavirus Hits Hardware Sales," published March 19, 2020). We expect global IT spending to decline year over year (down 3%, versus our previous forecast of 2%-3% growth), with smartphones and PCs particularly hit. Lower hardware spending lowers semiconductor industry revenues in 2020 (versus the previous forecast for growth). We expect significant negative rating actions due to revenue deferral or, in some cases, revenue destruction.

As of March 31, 2020, all of our technology sector rating downgrades have been one notch. Of all rating actions in technology, 43% have been downgrades, while 57% have been CreditWatch negative placements.

What are your thoughts about healthcare companies in this environment? Can this sector be considered a less vulnerable (i.e., "defensive") one?

Companies from various healthcare-related sectors (healthcare providers and services, pharmaceuticals, etc.) represent about 11% of CLO assets in aggregate.

Our healthcare team has written that they currently expect the impact of the coronavirus and related mitigation strategy to result in more moderate rating actions for companies in the broad healthcare universe, compared to other sectors. As of April 10, 21% of healthcare sector credits have had a negative rating action, compared to an average of 48% for the 10 most affected sectors. However, the situation is changing quickly; the longer and more widespread the pandemic is, the higher the potential for more negative ratings actions. Of particular concern to us are subsectors that rely on discretionary, low-acuity procedure volume, such as physical therapy, dentistry, shoulder surgery, and hip and knee replacements. (For more details, see "The Health Care Credit Beat: Growing Health Care Ratings Contagion From COVID-19," published April 15, 2020.)

The drop in patient and procedure volumes during the pandemic could lead to a decline in EBITDA and cash flows. While we believe the procedures are only delayed, not canceled (some procedures can only be delayed for so long before they become non-discretionary), a prolonged decline in procedure volume could lead to liquidity issues for the typically more highly-leveraged companies held in CLOs. Subsectors identified as most at risk for negative rating actions include hospitals, physician groups, outpatient surgical centers, physical therapy companies, and dental service organizations.

It's also worth noting that our rated healthcare universe is concentrated in the lower speculative-grade area, with 65% of all companies rated in the 'B' category (131 out of 201 companies). 'B' and 'B-' rated companies make up 59% of the overall universe.

What kinds of negative impact are you seeing on companies in the telecommunications sector?

Our sector analysts have said that the vast majority of U.S. telecom and cable providers can withstand the effects of a surge in COVID-19 cases and a sinking stock market with limited impact to credit quality near term (see "As COVID-19 Cases Surge, Pockets Of Risk Emerge For Certain U.S. Telecom And Cable Providers," published March 17, 2020). Longer-term credit implications will depend on the severity and duration of COVID-19 outbreaks and their impact on the U.S. economy. A handful of issuers' operations have direct exposure to the coronavirus, some of which have cushion at current ratings, while others are in the 'CCC' rating category already.

As of April 10, 2020, about 7.6% of the telecom sector had experienced a negative rating action, and most of those were in speculative grade.

How are recovery ratings likely to be affected by the significant economic disruption resulting from COVID-19 and sweeping mitigation measures?

Because recovery ratings already look through to a default scenario, they don't necessarily need to change as a result of the current economic disruption resulting from COVID-19. We shouldn't estimate recovery valuations based on a worst-case scenario or a trough-level EBITDA. Our recovery valuations are forward looking and factor in a run-rate EBITDA and enterprise valuation that we view as appropriate after the severe near term disruption eases. Consequently, we don't expect widespread or dramatic changes to our recovery ratings at this time. That said, some recovery ratings are likely to be lowered in certain situations:

• Where changes in capital structure will materially increase debt to boost liquidity or re-order the relative priorities of existing debt, especially for low speculative-grade credits; or
• Where we expect the impact on the business to be more severe and longer lasting, with less realizable value after a hypothetical default.

Are you changing anything about your approach to recovery analysis given the increase in fully drawn revolvers?

Yes and no. We are exercising the flexibility already inherent in our criteria (see "Recovery Rating Criteria For Speculative-Grade Corporate Issuers," published Dec. 7, 2016) to make revolver usage assumptions that differ from our standard recovery assumptions. Our standard assumptions for revolving credit facilities, based on empirical data from our recovery rating performance studies, is:

• Cash flow revolving facilities will be 85% utilized (cash draws plus material standby letters of credit usage) at the point of default.
• Asset-based revolving loan facilities (ABLs) will be 60% utilized.

The criteria allows for alternative assumptions (see paragraph 64 of the criteria) if we have more specific information regarding current or projected usage, if there are drawdown restrictions due to covenants, or if there is limited availability under a borrowing base formula.

In practice, it is relatively uncommon that we'd have specific information to make alternative assumptions for most credits. Under the current conditions, we view a company's proactive usage of revolving facilities above our normal recovery expectations as a clear signal that they need the liquidity and that we should factor this into our recovery analysis. For credits near the recovery percentage threshold of one recovery rating category and another, a higher debt balance at default has the potential to change our recovery rating. That said, we would typically avoid changing recovery ratings without a strong analytical rationale for doing so given the various uncertainties and assumptions underlying estimating recovery rates.

How could possible government intervention affect your recovery analysis for companies such as airlines, cruise lines, and casinos?

The simple answer is that it will depend on the specifics of each situation, which are not known at this time. To the extent the funding materially adds to a company's debt burden, this could negatively affect our existing recovery ratings. Similarly, the type of debt--secured versus unsecured--could also affect our existing recovery ratings. In contrast, forgivable grants or subordinated loans would be less likely to affect our recovery ratings on existing debt. Beyond the form of the support, the magnitude of support for each credit will also likely be unique for each credit and would need to be factored into our recovery analysis on a case-by-case basis.

CLO-Related Questions

What percentage of outstanding U.S. CLO notes are held by banks?

The U.S. CLO investor base is relatively diverse and includes different types of institutions who tend to focus on different parts of the CLO capital structure (i.e., CLO tranches with differing levels of risk and reward). U.S. and Japanese banks are the largest investors in U.S. CLO tranches by dollar amount, but their investments focus almost entirely on the 'AAA' CLO tranches, which carry the lowest level of risk. CLOs also represent a relatively small proportion of the assets banks invest in. For example, regulatory filings show that less than 1% of total assets in the U.S. banking system are invested in CLO tranches.

If U.S. CLOs continue to see an increase in assets from 'CCC' rated companies, could they be forced to sell these loans?

U.S. CLOs are generally not subject to a forced sale of any assets, including excess 'CCC' loans. Most U.S. BSL CLOs have an allowable basket for loans from 'CCC' rated obligors of 7.5% of total collateral. Beyond this amount, the excess 'CCC' loans above the 7.5% threshold--typically those loans with the lowest market values--are carried at market value rather than par value for purposes of calculating the CLO coverage test ratios. This can pressure the CLO's overcollateralization (or O/C) ratio tests, which, if they fail, may cause interest payments to be diverted from the CLO equity (and possibly also the junior CLO tranches) to instead be used to reduce the balance of the senior CLO notes outstanding. When CLO transactions see collateral stress, these mechanisms can serve to provide protection for the CLO senior notes. However, the CLO manager is not forced to sell assets because of the test failure. For more information on how CLO structural features work, including O/C ratio tests, see "S&P Global Ratings' CLO Primer," published on Sept. 21, 2018.

What is the U.S. CLO exposure to Canadian loans?

U.S. BSL CLO collateral pools are relatively diverse from an obligor and industry category perspective, with (as of fourth-quarter 2019) an average exposure to loans from 202 companies operating across about 25 industries (see "Sector Averages Of Reinvesting U.S. BSL CLO Assets: Credit Quality Deteriorated In Fourth-Quarter 2019 As Loan Prices And Spreads Increased," published Feb. 4, 2020). From a regional perspective, a large majority of the loans in U.S. CLOs--more than 88%--come from U.S. companies. There are a couple of widely held Canadian obligors in U.S. BSL CLOs (see "The Most Widely Referenced Corporate Obligors In Rated U.S. BSL CLOs: Fourth-Quarter 2019," published Jan. 10, 2020), but on the whole, Canadian companies represent only about 3.2% of total U.S. BSL CLO collateral.

What are the similarities and differences between the global financial crisis (GFC) in 2008-2009 and the current downturn in terms of CLO performance and rating actions?

The current economic downturn appears sharper and more severe than the 2008-2009 downturn, and this has been reflected in the pace of rating actions on corporate loan issuers found within CLOs. The shift in CLO collateral credit quality during the global financial crisis took place over quarters, with corporate rating downgrades (and ultimately defaults) playing out over a much more extended timeframe than what we're seeing now.

With respect to CLO ratings, the GFC saw downgrades driven in part by negative credit migration of the underlying corporate collateral, but also by rating agency methodology changes. For subordinate and mezzanine CLO tranches, rating actions were driven in large part by negative credit performance (corporate ratings downgrades, defaults, etc.), while for the more senior CLO tranches, the change in our rating methodology drove most of the rating changes. Absent the change in our CLO rating methodology, relatively few 'AAA' CLO ratings would have seen a rating action.

Partly due to our CLO rating methodology changes, post-GFC CLOs (aka CLO 2.0 transactions) have more subordination beneath most rated CLO tranches, especially at the top of the CLO capital stack. All else being equal, this should provide additional protection against CLO tranche defaults. However, U.S. BSL CLO collateral quality is lower today than it was ahead of the GFC in 2008. In fourth-quarter 2019, before the current market stress began, loans from 'B-' rated companies comprised more than 20% of the assets in U.S. BSL CLOs, compared to less than 10% before the start of the GFC in 2008 (see "To 'B-' Or Not To 'B-'? A CLO Scenario Analysis In Three Acts (UPDATE)," published March 26, 2020). This increase reflected changes in the broader leveraged loan market over the years following the GFC. In 2010, following the GFC, loans from 'B-' rated companies peaked at 12.75% of total U.S. BSL CLO assets, while loans from 'CCC' obligors peaked at just under 15%.

Where do the major CLO tests (e.g., O/C tests, 'CCC' buckets, and others) currently stand?

We publish this information periodically as rating actions occur on the underlying CLO collateral. Our most recent published update (see "U.S. CLO Exposure To Negative Corporate Rating Actions (As Of April 12, 2020)," published April 14, 2020) includes the following table showing the evolution of various U.S. BSL CLO credit metrics in 2020 year to date.

Table 1

Additionally, the article also includes a link to a spreadsheet with a list of corporate ratings within U.S. BSL CLOs that have been downgraded since the beginning of March, and a list of CLOs tranche ratings on CreditWatch negative.

Assuming defaults reach 10% over the coming 12 months, do you have a broad estimate of how many CLOs are likely to trip interest diversion tests this year? How many CLOs are currently violating their junior O/C tests?

In the short term, the combination of increased loans from 'CCC' rated obligors and the drop in loan prices since early March will have the most significant impact on junior O/C ratio tests. As a reminder, most U.S. BSL CLOs have an allowable basket of 7.5% of total collateral for 'CCC' assets, above which the excess 'CCC' assets get carried at market value rather than par value in the O/C test calculations. The increase in defaults over the next 12 months will also have an impact if/when they materialize, but this will occur over a greater time.

Given that the average U.S. BSL CLO junior O/C test cushion at the end of March was about 3.73% above the minimum required threshold, a 10% default environment would likely result in many, or even most, CLO junior O/C tests failing, especially considering the increase in assets from 'CCC' companies that would likely accompany such an elevated default rate. A few caveats apply here:

• The 10% speculative-grade default forecast under our base-case scenario includes both speculative-grade bond and loan issuers, while U.S. CLOs are collateralized almost entirely by loans (which in past downturns have had a lower default rate than bonds).
• CLO portfolios (and O/C test results) will vary from transaction to transaction, with some CLO managers or vintages potentially outperforming and seeing less deterioration than others.

What portfolio stress level can a 'AAA' CLO sustain and still maintain its rating? Also, based on recent downgrades to date, do you expect 'BB'/'BBB' CLO bonds to be downgraded?

Generally, we expect CLO 'AAA' ratings to be stable absent a significant level of portfolio stress. The stress run analysis we presented for scenario one during our webinar showed 7.5% of U.S. CLO tranches failing their quantitative analysis when subjected to our hypothetical stress, including having a 'CCC' loan bucket of 17.7% and non-performing assets of 5.5%. All of the tranches that saw their 'AAA' ratings lowered in this scenario experienced a one-notch downgrade, to 'AA+', and the remaining 92.5% of 'AAA' tranches remained at a 'AAA' rating.

We note that the 'AAA' CLO downgrade rate in the real world would likely be lower. This is because rating committees take into account qualitative considerations, our expectations for future senior tranche pay downs, and other factors beyond the results of the quantitative analysis when considering potential rating actions.

For the lower rated CLO tranches ('BB', 'B', and potentially 'BBB'), the downgrades would be much higher if our stress scenario were to play out in real life. Our stress scenario analysis has some pretty conservative assumptions built in: the portfolio downgrades and defaults are assumed to occur on day one, and no credit is given for a CLO manager's ability to avoid downgrades or defaults. However, if we were to see a real world scenario in which 5.5% of a CLO's collateral were non-performing, and another 17.7% was rated 'CCC', this would leave most U.S. BSL CLO 'BB' and 'BBB' tranches covered entirely by 'CCC' assets, recoveries on non-performing assets, and excess spread; this is not a great scenario from a subordinate or mezzanine CLO tranche rating stability perspective.

When you use a "scenario default rate" in your CLO analysis, how are recoveries accounted for?

Our quantitative analysis of CLOs uses two key metrics: scenario default rates (SDRs), which are generated by CDO Evaluator based on the assets in the CLO portfolio, and breakeven default rates (BDRs), which are generated by our cash flow model, S&P Cash Flow Evaluator, based on the CLO structure and credit protections. The SDR from CDO Evaluator represents the level of portfolio defaults we would expect to see, through the lens of our criteria, under a given rating stress level. It also sets the "hurdle rate" of gross defaults that a CLO tranche must withstand in our cash flow modeling in order to be assigned a given rating level.

For example, for a typical BSL CLO portfolio, CDO Evaluator produces a 'AAA' SDR in the mid-60% range. This means that, through the lens of our CLO criteria, this is the level of gross defaults we would expect from these pools under a 'AAA' stress level commensurate with a Great Depression stress scenario; and, for a CLO tranche to be assigned (or maintain) a 'AAA' rating, it would generally have to withstand that amount of defaulting loans in the collateral pool without the CLO tranche experiencing a loss in our cash flow modeling using our criteria stresses. For CDO Evaluator, if the CLO pool in question is lower in quality or less diversified, then the resulting SDRs will increase, reflecting the greater risk, and vice versa.

Since the SDRs from CDO Evaluator represent the model's estimate of gross default rates under economic stresses commensurate with our various rating levels, our assumed recovery assumptions have to be accounted for elsewhere. These are handled in our cash flow analysis using our S&P Cash Flow Evaluator model, and are reflected in the BDRs produced by the model for each rated tranche. The BDR for a given tranche represents the level of gross defaults the tranche can withstand without missing any required payments under the stresses (interest rate increases and decreases, default timing, etc.) outlined in our CLO criteria.

How do the SDRs produced by your CDO Evaluator model compare to the cumulative default rates seen in recent recessions?

At the 'BBB' rating level, commensurate with a "moderate" economic stress under our ratings definitions, CDO Evaluator would produce an SDR in the low 40% range for a typical U.S. BSL CLO 2.0 portfolio. This means that for a CLO tranche to be rated 'BBB', it should be able to withstand this amount of cumulative defaults in our cash flow modeling without the tranche experiencing a loss.

To compare this against corporate defaults actually seen in CLO portfolios during the 2008-2009 downturn, we ran CDO Evaluator on about 500 CLO 1.0 transaction portfolios to see how many cumulative defaults the model would predict for each portfolio under a moderate ('BBB') economic stress, and then compared that to the level of defaults each portfolio actually experienced. We assumed each CLO portfolio was static, with no credit given for active management of the portfolio, and tracked the defaults in each portfolio until the last asset in each portfolio had either paid down or defaulted.

The results of this testing can be seen on the chart below, with each marker representing one of the CLO portfolios tested. The vertical axis shows the level of defaults CDO Evaluator estimated for each portfolio under a "moderate" (i.e., 'BBB') stress level, and the horizontal portfolio shows the actual cumulative default rate experienced by the portfolio in the years during and after the financial crisis. The amount a marker (each representing a CLO portfolio) is above the diagonal line indicates the extent to which CDO Evaluator's estimate of portfolio defaults exceeded the actual observed defaults during and after the downturn.

Questions Related To The CLO Stress Test Scenarios

On your stress scenarios, do you use parent corporate rating or facility rating?

In line with our criteria for rating CLOs, we use the issuer credit rating (ICR) assigned to the company that issues the loan, not the facility rating on the loan itself. In instances where we don't have an ICR assigned to a loan issuer in a CLO, our criteria provide some other means to derive an 'implied rating' for purposes of running CDO Evaluator (see "Global Methodology And Assumptions For CLOs And Corporate CDOs," published June 21, 2019).

For U.S. BSL CLO transactions, we have ratings on more than 95% of the companies that issue the loans, and recovery ratings on a scale of '1+' through '6' assigned to more than 95% of the loans.

What recoveries rates are you using in your CLO analysis?

The CLO rating stress tests presented during the webinar (and all of the previous CLO rating stress scenarios we've published since the first one in July of 2014) essentially consist of two parts. In the first part, a hypothetical stress (or series of stresses) are applied to CLO portfolios: loans are assumed to default, corporate loan issuer ratings are assumed to be lowered (and the CLO 'CCC' buckets increase in size), recoveries are stressed to lower than historical values, etc.. Then, in a separate step, we batch run the CLOs through our CDO Evaluator and S&P Cash Flow Evaluator models, using the assumptions outlined in our CLO rating methodology and including the stresses from the first step. This gives us an idea of the modeling results a surveillance committee might review for the affected transactions after the portfolios experienced the hypothetical stress. The stress tests results presented in our webinar slides are based on the CLO ratings that this quantitative analysis would infer, without the benefit of committee discussion or qualitative factors.

As such, there are two sets of recovery assumptions applied in the CLO rating stress analysis:

• First, the recovery rates for the assets we assume will default as the result of applying our stress scenario; and,
• Second, the recovery rates we use in the batch runs we do to simulate our rating analysis after the hypothetical stress is applied: that is, what would the quantitative results reviewed by a surveillance committee after the stress look like?

The second set of recovery assumptions is based on the numbers outlined in our CLO rating methodology, since these are presumably how a surveillance committee would generate the results they would review. The recovery assumptions in our criteria are also conservative: for purposes of analyzing a CLO tranche rating, we look to the '1+' through '6' recovery rating assigned to each loan, and haircut the recovery value inferred by it for purposes of analyzing CLO tranches with higher ratings. For example, a 'AAA' CLO recovery assumption for a typical BSL CLO collateral pool might be in the mid-40% range, compared to historical loan recoveries in the mid-to-high 70% range over the past several decades.

For the hypothetical stresses we applied in the scenarios presented in the webinar:

• For the first of the three scenarios, we assumed 60% recovery for the assets we defaulted.
• For the second and third scenarios, we assumed a 45% recovery for the assets we defaulted.

In the CLO scenario 1 mentioned in the webinar slide, can you please give a sense how much SPWARF has increased in this scenario?

The average SPWARF of CLO exposures under the first hypothetical stress scenario would have been about 3,487, compared with an SPWARF of about 2,678 for our rated CLOs absent the hypothetical stress at the time the analysis was done. In comparison, the SPWARF of the CLO Insights Index increased to 2,923 as of April 12, 2020, from 2,639 as of March 1, 2020.

On your CLO scenario analysis, why are you assuming 7.6% and 11.0% non-performing asset exposure?

During the credit crisis in 2008-2009, average default exposures across U.S. BSL CLO 1.0 portfolios peaked at around 7%. At the time before the webcast, we felt 7% and 11% were both significant stresses, loosely correlating to the 10% and 13% overall speculative-grade default rate projections, respectively (leveraged loan default rates are generally lower than the overall speculative grade default rate, which includes bond issuers). We will follow up with a scenario analysis that covers a wider range of default assumptions.

For a CLO manger that trades really well, how much could they affect the ratings transition in your scenarios?

We did a study back in 2016 where we compared CLO 1.0 portfolios at the peak of the credit crisis with a hypothetical static portfolio where the CLO manager did not make any trades going into the credit crisis (see "How Do CLO Managers Perform In Times Of Stress?" published Sept. 6, 2016). We found that CLO 1.0 managers turned over their portfolios during the credit crisis by about 46%. On average, they had about 11% 'CCC' and about 7% non-performing assets at the peak of the crisis; had the managers not made any trades going into the crisis, the static portfolio would have had 17% 'CCC' and 10% non-performing. It's difficult to quantify the benefits of manager intervention, and the results will obviously vary from manager to manager (and CLO to CLO), but it would seem the assets purchased during the crisis had better ratings at the peak of the crisis when compared to the assets that dropped off the portfolios.

Related Criteria

• Global Methodology And Assumptions For CLOs And Corporate CDOs, June 21, 2019
• Recovery Rating Criteria For Speculative-Grade Corporate Issuers, Dec. 7, 2016

Related Research

• "U.S. CLOs In The Time Of Coronavirus" webinar. https://event.on24.com/wcc/r/2236630/758403004151FDCC6CA659BC1B0FA316.
• The Health Care Credit Beat: Growing Health Care Ratings Contagion From COVID-19, April 15, 2020
• U.S. CLO Exposure To Negative Corporate Rating Actions (As Of April 12, 2020), April 14, 2020
• To 'B-' Or Not To 'B-'? A CLO Scenario Analysis In Three Acts (UPDATE), March 26, 2020
• Global IT Spending Set To Slide As Coronavirus Hits Hardware Sales, March 19, 2020
• As COVID-19 Cases Surge, Pockets Of Risk Emerge For Certain U.S. Telecom And Cable Providers, March 17, 2020
• Sector Averages Of Reinvesting U.S. BSL CLO Assets: Credit Quality Deteriorated In Fourth-Quarter 2019 As Loan Prices And Spreads Increased, Feb. 4, 2020
• The Most Widely Referenced Corporate Obligors In Rated U.S. BSL CLOs: Fourth-Quarter 2019, Jan. 10, 2020
• When the Cycle Turns: The Continued Attack of the EBITDA Add-Back, Sept. 19, 2019
• S&P Global Ratings' CLO Primer, Sept. 21, 2018
• How Do CLO Managers Perform In Times Of Stress? Sept. 6, 2016