On June 7, 2023, S&P Global Ratings held its Tokyo Global Structured Finance Seminar at the Four Seasons Tokyo Otemachi hotel. The seminar included several presentations on topics of interest to the Japanese market, including S&P Global Ratings' global economic outlook, the U.S. leveraged finance and broadly syndicated loan (BSL) collateralized loan obligation (CLO) markets, private-credit and middle-market CLOs, and Japanese and Australian residential mortgage-backed securities (RMBS). Below is a lightly edited transcript of our comments during the "U.S. Leveraged Finance and BSL CLO Update" portion of the seminar for those who were unable to attend or would like to review the content.

The slides used during this portion of the seminar are available here. The slide numbers in the text below refer to this version of the slides: https://disclosure.spglobal.com/ratings/en/regulatory/article/-/view/sourceId/101578868.

Speaker: Steve Wilkinson, Senior Director And Sector Lead, Leveraged Finance And Recoveries

As you've probably heard regularly over the past year or so, most forecasters expect high and persistent inflation and high interest rates to eventually push the U.S. into a recession. That said, the U.S. economy has generally been slowing over the past 12 months but continues to be surprisingly resilient and outpace the expectations of most economists (including ours). Even so, economic strains are beginning to show in various sectors and affect credit quality.

Slide 4: Growth Of The Rated Universe At 'B' And Lower Is A Long-Term Trend; A Mild Recession Will Pressure 'B' And Lower Ratings

I'm going to start with an overview of the speculative-grade (spec-grade) universe and how it's evolved over time. On this slide, we show the composition of spec-grade ratings in the U.S. and Canada by rating category and how it has grown and changed since 2004. The first point to highlight is that the number of rated spec-grade companies grew substantially--by more than 50%--from the end of the Global Financial Crisis through early 2022. Even after some contraction over the past year, it is still roughly 40% larger than in mid-2009.

The second point to highlight is that the growth has predominantly been in lower-rated categories, specifically 'B' and 'B-'. Combined, these categories now represent slightly more than 50% of all spec-grade ratings. For the most part, the growth in these rating categories (on a combined basis) is due to the influx of newly rated companies during a period of strong mergers and acquisition (M&A) activity amid low interest rates and steady economic growth.

Most of these newly rated companies are financial-sponsor-owned, with private equity firms now owning more than two-thirds of all companies rated 'B' and 'B-'. Over this time, valuations and purchase price multiples were high, and this required a lot of debt financing, even with large equity checks being written by the sponsors. This resulted in high debt leverage and lots of 'B' and 'B-' rated companies.

While downgrades partly contribute to the size of the 'B-' cohort today, more than half of all 'B-' companies were initially rated at that level. This is an important distinction that Steve Anderberg will build on later in the CLO section of the presentation. Downgrades (of course) are also why ratings of 'CCC+' and lower spiked during the COVID-19 pandemic, as well as more recently.

Over the past few months, the percent of spec-grade companies rated 'CCC+' or lower have climbed to about 12%. Looking forward, using the historical average downgrade rates for 'B' and 'B-' companies, the number of companies rated 'CCC+' or lower could expand by 5%-6% over the course of 2023. This would push this category to 15%-17% of all spec-grade ratings, which is below the nearly 19% peak reached during the COVID-19 pandemic, but above the roughly 14% peak during the Global Financial Crisis.

Slide 5: Rising Rates Weigh On Free Operating Cash Flow (FOCF) Of Lower-Rated Firms

Turning now to interest rates and cash flow, these are key items that we're watching. On the prior slide, I had emphasized ratings of 'B' and lower because these are the companies most affected by the steep rise in interest rates. Most highly leveraged companies could generate positive cash flow while interest rates were at historically low levels. However, the ability to do so in a period of significantly higher interest rates is less certain.

In the chart on the left, we show the rise in benchmark rates (specifically the secured overnight financing rate [SOFR]) going back to early 2022, when the FED's rate hikes began, with the blue bars representing trailing 12-month SOFR rates and the red bars representing forecasted trailing 12-month SOFR rates using the forward curve. The key point here is that the substantial rate hikes to date (from near zero at the beginning of 2022 to over 5% now) have yet to fully run through company financial statements, where the full impact won't hit until mid-to-late 2023; although the magnitude of the additional increases is lessening.

In the chart on the right, we see that on a trailing 12-month basis nearly 45% of 'B' companies are generating FOCF deficits (with FOCF representing operating cash flow after capex) and this proportion increasing to 55% for 'B-' companies.

One positive factor contributing to the cash flow deficits is that even lower-rated companies continue to invest in capex to support growth and improve efficiency. Even so, high debt servicing costs will continue to weigh on FOCF, and this is a key focus of the ratings surveillance being done by the corporate rating sector teams.

Slide 6: Credit Trends Turn Negative As Economic Tailwinds Flag And Headwinds Mount

Turning to rating trends, as the economy began to weaken, corporate rating downgrades started outpacing upgrades in May of 2022, with the pace of downgrades increasing in August of last year. This is generally true for spec-grade ratings overall (in the chart on the left) and for ratings "into" and "out of" the 'CCC' categories (in the chart on the right), with upgrades in the blue bars and downgrades in the gold bars. In general, we expect downgrades to continue to outpace upgrades since our base-case economic forecast currently calls for a shallow recession later this year and higher-for-longer interest rates.

Now I want to focus on the 'CCC' category ('CCC+'/'CCC'/'CCC-') for a minute, because this has important implications for both default risk and for CLO structures. At the bottom of the slide, to provide some context, we reference some key credit metrics for companies downgraded to or upgraded from the 'CCC' category using data from our first-quarter 2023 leveraged finance article (see "U.S. Leveraged Finance Q1 2023 Update: Ch-Ch-Ch-Changes -- Material Shifts In Key Credit Stats Drove Downgrades To 'B-' And 'CCC', And Upgrades To 'B-'," published May 4, 2023).

As you can see, there is a sharp distinction in credit quality for companies downgraded to this category, with median leverage of 15.5x for downgraded firms versus 8.0x leverage for companies that remained 'B-'. Similarly, cash flows metrics were also dramatically worse for downgraded firms. This reflects our view that companies rated in the 'CCC' category are more likely to default than not, which I'll cover in more detail later.

Slide 7: Spec-Grade Negative Bias Has Been Rising And Is Now At The Post-Global Financial Crisis (GFC) Average

But first, I want to talk about more about future rating expectations. On this slide, we show the negative bias for spec-grade companies over time, in the jagged yellow line from 1995 forward, with recessionary periods in the blue bars. This is the proportion of spec-grade companies with either a negative outlook, which indicates at least a one-in-three chance of a downgrade over the next year or a CreditWatch negative listing, which indicates a risk of downgrade of 50% or more (generally over a shorter time-period). Negative bias is generally the best indicator of future downgrades.

The main point is that negative bias has been rising steadily since early 2022, when it bottomed out at about 12.5%. It now sits at 22%, which is just above the post-GFC average (purple line) and rising towards the longer-term average of about 25% (green line). And we expect this upward trend to continue, with economic and financial pressures expected to increase and persist.

Slide 8: Increased Negative Ratings Bias Highlights Growing Downgrade Risk

Now, let's look at downgrade risk and rating changes by sector. In the chart on the left, we see the mix of outlooks and CreditWatches by sector, with negatives in the colors to the left of each bar, stable outlooks in red in the center, and positives in the colors to the right. And for context, the bottom-most bar shows the overall spec-grade negative bias as of May 2023 at about 22%.

Looking at the mix by sector in the bars above, the sector with the highest negative bias by far is consumer products, at more than 40% of the spec-grade companies in the sector. There are also another seven sectors with a negative bias of 20% or more, including three of the larger sectors, namely restaurants and retailing; healthcare; and media, entertainment, and leisure.

It's also useful to note the sectors with the largest increases in negative bias over the past year-and-a-half. These include consumer products, which increased roughly 20%; telecom; restaurants and retail; real estate, which increased 15% or more; and healthcare, which increased by 10%-15%.

Shifting to the chart on the right, we see upgrades and downgrades by sector, with downgrades in the yellow bars and upgrades in the blue bars, both spanning out from the center line.

A few things to highlight: first, you can see that the consumer products sector had the most downgrades, although it had quite a few upgrades too, since some subsectors are doing better (like packaged foods). Media, entertainment, and leisure also has a large, and almost even, mix of downgrades to upgrades, where trends for entertainment and leisure are largely positive while media is largely negative and under pressure.

Another point to highlight is that rating actions are mostly negative in restaurants and retail and healthcare, where labor costs are a common problem, although healthcare pressures are largely expected to level off over the course of 2023.

Slide 9: Sector Size And Credit Quality Varies, But The Largest Sectors Generally Have High Concentrations of Firms Rated 'B-' or Lower

So, let's look at rating quality by sector more closely. The purpose of this slide is to show that sector size and credit quality varies significantly. The bars in the chart show the number of rated companies by sector; so the taller the bar, the more spec-grade companies are rated. Within each bar, the colors indicate differences in the spec-grade ratings mix with 'CCC+' and lower at the bottom in blue, 'B-' in orange, 'B' in gray, and 'B+' to 'BB+' in gold.

To highlight sectors with higher (or lower) concentrations of low ratings (defined as 'B-' or lower), there are blue dots for each bar showing the percentage of firms rated 'B-' or lower (corresponding to the right-hand scale). If the sector dot is above the dashed line, there is a higher mix of lower-rated firms in that sector; and if the dot is below the line, there is a lower mix of lower-rated firms in that sector.

As you might expect, the largest sectors tend to have the highest number of companies with ratings of 'B-' or lower, since this is where rating counts have grown the most. In terms of count, the sectors with the most ratings of 'B-' or lower are healthcare; technology; and media, entertainment, and leisure—the three largest sectors—followed by capital goods; business and consumer services; and consumer products.

As you can see, five of these six sectors have a concentration of companies rated 'B-' or lower that is above the overall speculative-grade average. So, let's look at 'B-' rated firms more closely, where investors are most concerned about downgrade risk.

Slide 10: 'B-' Downgrade Risk Can Vary Widely By Sector

Digging a little deeper, one question is whether having a large number of companies rated 'B-' means that a sector has higher downgrade risk?

On this slide, we show companies rated 'B-' by sector. Again, the height of each bar shows the number of companies rated 'B-', but here, the colors show the mix of outlooks and CreditWatches by sector, with negatives at the bottom of each bar, stable outlooks in purple in the middle of each bar, and positives at the top of each bar. As a reminder, negative outlooks and CreditWatch listings are the best indicator of downgrade risk, with at least a one-in-three chance of downgrade for a negative outlook and an even higher chance of a downgrade for a CreditWatch negative listing.

Also, on this slide, are dots for each sector indicating the negative bias (percentage) by sector corresponding to the scale on the right axis. If the dot is above the horizontal line, it has a higher-than-average negative bias; and if the dot is below the line, it has a lower-than-average negative bias. Of the six sectors with the largest 'B-' populations, only consumer products has a higher-than-average negative bias. And most of sectors with the highest downgrade risk are much smaller; so simply having a large number of 'B-' companies does not make a sector "riskier" per se.

Even within the consumer products sector (as I mentioned earlier), some companies and subsectors are doing well in this environment while others are challenged. So, it is important to evaluate each credit, rather than just sectors as a whole.

Overall, the negative bias for companies rated 'B-' is just under 19%, almost 3% lower than the negative bias for all spec-grade companies, and this is because we have a different fundamental assessment of credit risk for companies rated 'CCC+' or lower. In fact, we even have separate criteria for evaluating whether a rating belongs in the 'CCC' category.

Given this, the 'B-' category can show surprisingly good rating stability and reflect a fairly wide-band of credit quality with better and worse 'B-' companies.

So, let's get into ratings of 'CCC+' or lower on the next slide.

Slide 11: 'CCC' Rated Companies Have Higher Default Risk

Taking a closer look at 'CCC' category ratings, the bottom line is that companies in the 'CCC' category are viewed as more likely to default than not. Essentially, these ratings say that we think a firm's debt structure is unsustainable, and that a default is likely absent unexpectedly favorable business, financial, and economic conditions, with the level of the rating within the 'CCC' category' dependent on the expected timing to a "visible" default scenario as shown in the table on the left.

Turning to the chart on the right, you can see that historical default rates are dramatically higher for 'CCC' category companies than for those with other ratings—even 'B-', which is only one-notch higher using average cumulative default rates from 1981 through 2021. As you can see, the dark blue bars for the 'CCC' category to the far right are substantially taller indicating a much higher default rate, with the cumulative defaults over time somewhat overstated due to companies that withdrew their ratings prior to default, which we track for one year after withdrawal.

Slide 12: Spec.-Grade Defaults Could Reach 4.25% And LSTA Leveraged Loan Index Defaults Could Reach 2.5% (Both By March 2024)

Turning now to default expectations, on this slide, we have two measures of default. The jagged gold line shows the historical spec-grade default rate over time, and the jagged blue line shows the Morningstar-LSTA Leveraged Loan Index default rate over time, with recessionary periods indicated in the gray bars.

At the right end of each line, you can see that both default rates are beginning to increase from historically low levels. As of May, the default rate for all of spec-grade is about 2.5%, and the Leveraged Loan (LL) Index rate is about 1.6%. Both are measured on a "issuer count basis".

Through the first quarter of next year, we expect both default rates to continue to increase under our base-case scenario, with the spec-grade default rate rising to 4.25%--just above its long-term average of 4.1%--and the LL Index default rate to rise to 2.5%--right at its long-term average. One obvious difference between these default measures is that the spec-grade default rate is higher. However, a significant driver of this gap is "definitional" since the Leveraged Loan Index definition of default does not include selective defaults (SDs), even if lenders take a haircut to par.

For context on materiality (as shown in the bullets), SDs are becoming increasingly common and represented about 50% of all defaults in 2020 and on a year-to-date (YTD) basis, and over 60% of all defaults in 2021 and 2022.

At one point, the risk of a selective default was almost exclusively a bondholder risk, but that is not the case anymore as selective defaults and out-of-court restructurings of loans have become more common in recent years.

Slide 13: First-Lien Recovery Expectations Are Now Well Below Historical Averages

Now turning to expectations for recovery given default. On this slide, we show that our expectations for recovery given default have declined significantly over time. In the chart, the jagged blue line shows average first-lien recovery expectations for newly rated debt each quarter, and the jagged gold line shows average first-lien recovery expectations for the whole stock of rated first-lien debt. Both measures use the rounded recovery percentages that are part of our recovery ratings.

On average, we now expect first-lien debt recovery rates will be in the mid-60% area, which is down materially from historical estimates of actual recovery ratings of 75%-80%. The reasons for the decline are fundamental and are primarily reflecting higher total debt leverage, higher first-lien debt leverage, more first-lien-only debt structures, and thinner cushions of junior debt where it still does exist. All of those factors are inversely correlated with first-lien recovery expectations.

For the most part, this reflects the growth in the spec-grade universe with new ratings concentrated at 'B' and lower, where recovery expectations are lower with high leverage pushing average recovery expectations to the low-60% to high-50% area.

Slide 14: Aggressive Loan Restructurings Significantly Impair Recovery Prospects And Don't Usually Resolve Financial Problems

As I mentioned earlier, selective defaults on loans from aggressive out-of-court restructurings have increased in recent years and are a factor for both higher total default levels and lower recovery expectations on the restructured debt.

On this slide, we show a running list of out-of-court loan restructurings, with collateral transfers in the table on the right, and priming loan exchanges in the table on the left. For both data sets, we also show the impact on the recovery ratings on the "1L" debt for the lenders that did not participate in the transactions.

There are a few interesting things that the chart highlights. The number of restructurings is fairly small given the size of the BSL market with roughly 1,500 loan issuers in the Leveraged Loan Index. However, the number of defaults has been increasing in recent years, and they tend to spike during periods of credit stress.

And most importantly, the impact on creditors that don't participate in the restructurings can be dramatic; although it can vary significantly depending on the specifics of the restructuring--which can vary significantly--with the impact the most severe when it is 51% of the lenders restructuring at the expense of the 49% who aren't asked to participate.

Given the increase in these transactions and the potentially severe negative impact on loan quality, CLO investors are focused on CLO managers having sufficient flexibility in their indentures to be able to participate in such restructurings if they have the opportunity and the restructuring can't be blocked. My personal view is that having the option to participate is critical so that you are not forced into an outcome that is worse. This is especially important because most of the restructurings didn't resolve the companies' problems. There are a total of 20 restructurings shown on the slide, undertaken by 17 firms since there are three repeat offenders. Of these 17 companies, 10 subsequently filed for bankruptcy, and six are rated in the 'CCC' category, indicating that we think their capital structures remain unsustainable and that they are likely to default yet again.

The last point I want to make on this topic is that--while the risk of such restructurings is real--we don't factor this into our recovery ratings on a forward-looking basis because they are unpredictable and unquantifiable as evidenced by the significant variation in outcomes shown on the slide.

Speaker: Stephen Anderberg, Managing Director And Sector Lead, U.S. CLOs

Steve Wilkinson covered the leveraged finance and the broadly syndicated loan space. I'll provide an update on U.S. BSL CLOs, including issuance trends, the impact of corporate rating actions on CLO collateral pools, and an update on research we've done on different topics. These include the impact of rising 'CCC' baskets on CLO overcollateralization (O/C) tests, the value of active collateral management during times of economic turbulence, and rating stress tests showing how our CLO ratings might perform under different levels of corporate loan defaults and downgrades.

Slide 16: New CLO Issuance Slows, While Resets And Refinancings Halt

CLO issuance is down this year, but there is one bright spot: middle-market CLOs. Issuance year to date has been about $50 billion, with middle-market CLOs making up an increasingly large part of that. This year, middle-market CLOs have represented almost 22% of new issuance volume, compared to a more typical 9%-12% in previous years.

Overall, U.S. CLO new issuance this year has been running at about $10 billion per month. That's down a bit from $10.75 billion of issuance per month last year, and way down from $15.6 billion per month in 2021, which was a record year for CLO issuance. The reason that new issuance is down for BSL CLOs is, quite simply, transaction economics. At the start of 2022, BSL CLO 'AAA' tranche spreads were SOFR + 130, but by the end of 2022, that had increased to SOFR + 230 due to inflation, rising interest rates, and economic uncertainty.

In the current market, BSL CLO spreads have come in somewhat, currently sitting a bit below SOFR + 200, but that's still at a level that limits the appeal of new issue CLO equity. Almost none of the new issue CLO transactions done this year have had third-party equity investors. Virtually all have equity held by the CLO manager, or by a captive equity fund controlled by the manager.

There's also a large gap between the levels that tier-one managers can price at versus the rest of the market. Recent new issue BSL CLO 'AAA' note spreads have ranged from as tight as the mid-170s to as wide as the mid-220s.

Our forecast for overall U.S. CLO new issue volume this year is $110 billion. It's possible that U.S. bank treasury departments who have historically been large CLO 'AAA' investors--Citibank, Wells Fargo, and JPMorgan--could potentially increase purchases later this year, or next year. This could help CLO senior tranche spreads come in and boost new CLO issuance. Large Japanese CLO investors have been quite active in the new issue U.S. CLO market this year.

On CLO resets and refinancings, there have been virtually none over the past year. With CLO tranche spreads this much wider than when most of the outstanding CLO transactions were originated, they don't make economic sense. This could change later this year, for a couple of reasons. First, there are some CLOs that were issued last year with high spreads and a short, one-year, non-call period. Given that spreads are somewhat tighter now than when these CLOs were issued, it would make economic sense for these transactions to be refinanced or reset.

Secondly, with so few CLO resets being done over the past year, there is a large volume of CLOs exiting their reinvestment periods this year. We'll talk about that next.

Slide 17: More CLOs Are Exiting Their Reinvestment Period, But The Pace Of Senior Tranche Amortization Has Slowed

Currently, about 25% of U.S. CLOs are outside their reinvestment period, and we expect this to increase to almost 40% of transactions by end of this year. This has a couple of implications. First, it means we might see a lot of CLO resets later this year, even for transactions for which it doesn't make sense on a purely economic basis. We have spoken with several CLO managers who are contemplating resetting their CLO transactions even if it would widen the spreads on the CLO, simply because they want to extend the dates on their CLOs rather than see them amortize.

It's worth noting that CLO reset and refi volume can be volatile and could increase significantly if CLO tranche spreads tighten even modestly.

The lack of CLO new issuance is linked to a lack of activity in the corporate loan market, which has not been growing this year but actually shrinking for the first time in 10 years. That's because CLOs buy more than 70% of broadly syndicated loans in the U.S. market, so fewer new issue CLOs means less activity in the corporate loan market.

This lack of activity in the corporate loan market has also slowed down the amortization rate of CLOs that are outside their reinvestment period. The chart at the bottom shows the paydown rate for CLO senior tranches by exit year, the year in which they exited their reinvestment period.

You can see that senior tranches of CLOs that exited their reinvestment periods in 2021 or 2022 have paid down much more slowly than the senior tranches of CLOs exiting their reinvestment periods in 2020 or before.

Slide 18: Corporate Downgrades Weigh On CLO Portfolios, Again

Turning to credit, this chart shows the proportion of assets within BSL CLO collateral pools that have been affected by downgrades by quarter, from 2018 through Q1 2023. The various colors on the histogram bars show how many of the corporate rating downgrades were in the 'CCC' range, how many were below 'CCC-', and so forth.

During the onset of the pandemic in Q1 and Q2 2020, with the sudden shutdown of the economy, there was a dramatic and very rapid increase in corporate rating downgrades. BSL CLO 'CCC' baskets swelled to 12% as managers were unable to manage around the rapid pace of downgrades. Then, for the following two years (from Q3 2020 through Q2 2022), there was a period of relative calm. With the pandemic recovery, corporate ratings saw more upgrades than downgrades each month. BSL CLO collateral pool metrics generally improved.

This changed in May 2022, when corporate ratings downgrades began to exceed corporate upgrades for the first time since Q3 2020, and then again in August 2022 when corporate downgrades accelerated given the impact of rising rates and slowing growth.

Since January 2022, a total of 108 obligors in BSL CLO collateral pools have seen ratings lowered into the 'CCC' range, mostly from a rating of 'B-', and 65 obligors have seen ratings lowered below 'CCC-', which CLOs define as non-performing. So, the credit environment has definitely shifted.

Slide 19: Most CLO Metrics Show Gradual Deterioration

All of this has had an impact on CLO collateral metrics that had been improving in the aftermath of the pandemic. What you're looking at here is an index of more than 500 BSL CLO transactions that are currently within their reinvestment periods.

BSL CLO exposure to 'CCC' rated assets had been declining each month since Q3 2020 and reached a post-pandemic low of less than 3.7% last August. But with increase in corporate rating downgrades since last May, the average 'CCC' basket has been gradually increasing and now sits at just over 6.0%. This is considerably less than the proportion of 'CCC' rated assets in the overall loan market, which sits at almost 12%. So, the managers have been actively managing exposure to 'CCC' assets and reducing them.

Exposure to non-performing, or defaulted, assets is also increasing, but remains modest at a current level of 1%.

The managers have also done a good job at building par since the pandemic. The average junior O/C ratio reached a post-pandemic high of 4.7% in August last year. It's been gradually declining since then, but still sits at a healthy 4.26%.

I should point out that these numbers are averages. There is a wide distribution of weaker and stronger transactions around the average, with CLOs originated after the arrival of the pandemic in Q1 2020 generally having stronger collateral metrics than the pre-pandemic CLOs.

Finally, one important thing we keep an eye on is the proportion of CLO assets with a negative rating bias. That is, the percentage of assets from obligors that have a rating either on Outlook negative or CreditWatch negative. This is an important forward-looking indicator of potential changes in corporate credit quality. The proportion of CLO assets with a negative rating bias now sits at 16.8%. That's a lot higher than it was a year ago, when it was at less than 11%, but still lower than historical averages.

Slide 20: Assets From 'B-' Obligors Continue To Increase

One theme that Steve Wilkinson mentioned earlier is the rise of assets from 'B-' obligors in the corporate loan market. As CLOs hold more than 70% of the loans in the loan market, these have also increased within CLO collateral pools.

The upper chart here shows the ratings distribution of BSL CLO collateral over the past six years, from year-end 2017 to May 2023. You can see that CLO assets from 'B-' rated obligors have increased significantly over that time, from less than 13% at year-end 2017 to more than 30% now.

I said earlier that CLO collateral pools had fewer 'CCC' assets than the loan market at a whole. For 'B-' rated assets, CLOs hold more than the overall loan market, but not by much. In part this is because when an obligor sees its rating lowered into the 'CCC' range, it's common that the managers will sell those loans and buy new loans from 'B-' rated obligors.

The bottom chart shows recovery ratings within CLO collateral pools over the same six-year period. Recovery ratings have also drifted downward over the period, which isn't surprising. Loans from 'B-' rated obligors tend to have lower recovery ratings, in part because many of these issuers have "loan-only" capital structures, and because 'B-' issuers tend to be more highly leveraged.

The good news here is that our CLO rating methodology accounts for this. All else being equal, a new issue CLO with lower-rated assets and lower recovery assets will require more subordination and credit protection than a CLO with a higher quality collateral pool.

Slide 21: Majority Of Current 'B-' Assets Were Born That Way

So, 'B-' exposures across reinvesting U.S. BSL CLOs are at record levels, ending the year at just over 30%, up from 26% at the start of 2022. Historically, 'B-' exposure across U.S. BSL CLOs was much smaller and was typically made up of issuers that had downgraded to 'B-' from a higher rating, rather than issuers who were assigned a 'B-' rating from the start.

The chart on this slide shows the total proportion of assets in BSL CLOs from 2007 through Q1 2023. The yellow portion of the bars shows loans from 'B-' obligors that got to 'B-' by way of a downgrade. The blue portion of the bars shows 'B-' obligors that were rated 'B-' from the start.

During periods of stress, 'B-' exposure increased as issuers within CLO collateral pools experienced downgrades to 'B-'. This is shown by the growth in the yellow portion of the bars during periods of economic stress like 2008-2010, 2015-2016, and in 2020. However, since 2017, the growth in CLO exposure to 'B-' issuers has come primarily from companies originally rated 'B-' as highlighted in the blue portion of the bars.

This is a result of changes in the loan market. Over the past six years, loan investors have become much more accepting of new issue loans from companies rated 'B-' and the proportion of these loans has grown in the market. As of year-end 2022, a majority (two-thirds) of the 'B-' assets in BSL CLOs came from issuers that were originally rated 'B-'.

This may actually be good news, because not all 'B-' companies have the same likelihood of seeing a downgrade. Companies rated B-' from the start tend to be less likely to see a downgrade than companies that got to their 'B-' rating by way of a downgrade from a higher rating level.

You can see this in the table at the bottom of the slide, which breaks out 'B-' downgrades in CLO collateral pools during 2022. Overall, 8.9% of 'B-' rated companies in CLO collateral pools saw a downgrade during the year. But this wasn't split equally between the two cohorts. Only 4.9% of companies originally rated 'B-' saw a downgrade during 2022, compared to 16.7% of companies that were 'B-' in 2022 but were initially rated higher.

So, there is a greater proportion of CLO assets coming from 'B-' issuers, but most of these issuers are less likely to see a downgrade than historical 'B-' corporate ratings.

Slide 22: Rising 'CCC' Assets Could Affect CLO O/C Test Cushions

So, a lot of companies with ratings of 'B-' means there are a lot of companies one notch away from a rating in the 'CCC' range. With more than 30% of assets in BSL CLOs rated 'B-', if even a modest proportion of those were to see a downgrade, it could swell CLO 'CCC' baskets significantly. This would have an impact on CLO overcollateralization tests.

We wanted to see how much of an increase in 'CCC' assets CLO OC tests could handle before they started to fail, so we published a stress test on them (see "How Rising U.S. BSL CLO 'CCC' Baskets Could Affect Junior Overcollateralization Test Cushions," published April 28, 2023). On this slide, two of the four scenarios from the article are shown. In scenario A, on the left, we envision a modest increase in 'CCC' assets from current levels, with a total of 10% of CLO assets rated 'CCC+' or lower. Scenario B, on the right, is much more punitive, with 20% of CLO assets being rated 'CCC+' or lower.

The two charts underneath each scenario show the impact of these hypothetical stresses on O/C tests at each CLO tranche level. The top chart shows the average O/C test cushion after the stress is applied, and the lower chart shows the proportion of CLO O/C tests failing after the stress is applied.

BSL CLO O/C tests typically haircut the value of 'CCC' assets above 7.5% of total assets, with the "excess" 'CCC' assets being carried in the O/C tests at market value rather than at par. So, we also stressed the market value of the excess 'CCC' proportion. Since an environment that would produce many corporate rating downgrades into the 'CCC' range would likely also see significant price declines for 'CCC' assets, we tested six price decline scenarios ranging from no decline in price current levels to a 50% decline in 'CCC' asset prices.

So, for example, if we look at the impact of a 20% 'CCC' exposure on 'BB' tranche CLO O/C tests, we can see that the impact depends on how much 'CCC' asset prices decline.

There's a lot to unpack here, but the key finding from the article is that BSL CLO 'CCC' baskets can increase into the mid-teens before the average junior O/C test begins to fail. There is a significant distribution around the mean though, with some individual CLOs being stronger or weaker. Generally, the pre-pandemic CLOs were weaker and the post-pandemic CLOs were stronger.

Slide 23: Value Of Active Management During A Turbulent 2022

During times of economic stress and uncertainty, CLO managers can differentiate the performance of their transactions and add value by actively managing their portfolios. 2022 was a period of considerable uncertainty, with rising interest rates, slowing economic growth, and shifting geopolitical conditions. So, we did a study to assess the impact of active portfolio management on CLO collateral metrics during the year.

Turnover of assets in BSL CLO collateral pools in 2022 was 30.5%, meaning that just over 30% of the loans that had been in CLO collateral pools at the start of the year were no longer in the collateral pools at the end of the year. To assess the impact this portfolio turnover had on CLO credit metrics, we looked at the actual change in BSL CLO credit metrics during 2022, including portfolio turnover. These credit metrics are shown in table one (the blue table) on the left side of the slide.

Then, we made an assumption that none of the trades done during 2022 had taken place, with the same assets in the CLO collateral pools at the end of the year as at the start of the year. The CLO credit metrics for this hypothetic static CLO scenario are shown in the middle (yellow) table. We then looked at the difference in these two sets of credit metrics to assess the value that active management of CLO collateral had during the year. This is shown in table 3 on the left.

On average, the trades increased the proportion of loans from 'B-' companies, because, when a company saw its rating lowered to the 'CCC' range, a manager would often sell loans from that company and purchase loans from a 'B-' rated company.

All other CLO credit metrics benefitted from the trading activity: exposure to 'CCC' assets and defaulted assets was lowered, the S&P Global Ratings' weighted average rating factor (SPWARF) was lower (indicating higher average portfolio ratings), the par value of the assets was greater, and the junior O/C test cushion was greater.

Slide 24: No U.S. CLO 'AAA' Tranche Ratings Lowered Since 2011

We don't need to spend a lot of time on this, but I wanted to touch upon CLO tranche rating actions. The chart here shows CLO tranche rating upgrades and downgrades from 2008 through Q1 2023. On the lefthand side of the chart, you can see many downgrades at the end of 2009 and start of 2010, in the aftermath of the Global Financial Crisis.

Some of these downgrades were related to transaction performance, but most were due to CLO criteria updates S&P Global Ratings implemented in late 2009. Many of those ratings were subsequently upgraded in the following years as CLO collateral improved and senior tranches paid down.

Looking over to the right on the chart, the second period of significant CLO downgrades was of course 2020, when the pandemic arrived, along with economic shutdowns and corporate rating downgrades. It's worth noting that none of the CLO tranche rating downgrades since 2011 have affected a 'AAA' rated tranche.

Slide 25: 56 U.S. CLO Tranche Defaults From 17,000+ Rated Tranches

The best measure of ratings performance over long periods of time is defaults, and CLOs have done very well here. S&P Global Ratings has been rating CLOs for more than 25 years, and during that time, has rated more than 17,000 U.S. CLO tranches. To date, 56 of these tranches have defaulted, of which 40 of them were from CLO 1.0 transactions originated prior to the Global Financial Crisis and 16 from CLO 2.0 transactions originated after the Global Financial Crisis.

The CLO 2.0 transactions have significantly more par subordination than the CLO 1.0 transactions, along with cleaner collateral pools and more restrictive CLO document provisions.

There has never been a 'AAA' CLO tranche default in the history of the market.

All of the CLO 2.0 defaults so far have been of 'BB' rated tranches. There is also a single CLO 2.0 tranche initially rated 'BBB' that is currently rated 'CCC-' and could default.

Slide 26: Rating Stress Scenarios (May 2023 Update)--1 Of 2

Finally, I wanted to close with the results of our BSL CLO rating stress scenarios we recently generated. We update these each year, and the results here are from last month.

We ran our four standard stresses, each of which envisions a proportion of assets within CLO collateral pools defaulting, and then a proportion of the remaining (non-defaulted) assets being downgraded to 'CCC'. We then see what the impact of the scenario on CLO ratings is.

So, for example, in the first stress, the "5/10" scenario, we assume that 5% of the loans held within CLOs default, with a recovery of 45%. Then, we assume that 10% of the remaining loans are rated 'CCC'.

The stresses are applied immediately, because it's easier to model and produces a greater stress.

The "5/10" scenario, with 5% of loans defaulting and 10% CLO 'CCC' baskets, is relatively close to the loan defaults and downgrades seen during the pandemic in 2020.

The stress level experienced by CLOs during the Global Financial Crisis is a bit more stressful than the "5/10" scenario, but not by too much. The GFC saw CLO collateral defaults reach about 6% and CLO 'CCC' baskets peak at about 12%.

The graphic here shows the average notch movement for each CLO tranche under the various scenarios. So, for example, under the "5/10" scenario the average CLO 'AAA' tranche rating barely moves at all, while the 'BB' tranche ratings move downward by just under one notch.

Slide 27: Rating Stress Scenarios (May 2023 Update)--2 Of 2

This slide shows the detailed results of the four stress scenarios. The columns show the number of notches that tranches move off their current rating. A zero-notch movement means a rating was affirmed at its current level, while the one-notch column indicates ratings that were lowered by one notch, and so forth.

So for the "5/10" scenario, in the table at the top with 5% defaults and 10% 'CCC' baskets, we see that 99.3% of 'AAA' ratings under this stress are affirmed at their current rating level, and 0.70% of ratings go down by one notch, to 'AA+'. Further down the CLO capital stack, the impact is greater. For example, for 'BBB' CLO tranches under the "5/10" scenario, about 80% of the ratings are affirmed. while the remaining ratings get downgraded, mostly by one notch.

It's worth noting that the stress test results are based solely on model results, without qualitative factors being taken into account. In reality, with qualitative factors taken into account during a committee discussion, the 'AAA' tranche ratings would be a little bit less likely to be downgraded than shown here because the committee would look at senior tranche paydowns.

But aside from that, the "5/10" scenario lines up pretty closely with the CLO tranche rating downgrades seen during 2020.

Overall, the rating stress tests highlight the resilience of the CLO structure and the way it protects senior tranche investors during times of economic stress with high levels of loan defaults and elevated CLO 'CCC' baskets.