Asset diversification is a key concept in collateralized loan obligation (CLO) portfolios, as it is elsewhere in finance, and it has played an important role in the strong performance history of U.S. CLOs. Diversification spreads out risks across loan portfolios, which is important given the relatively long lifespan of CLO transactions. A typical new issue broadly syndicated loan (BSL) CLO transaction in the U.S. starts with a five-year reinvestment period and a roughly nine-year weighted average life (WAL) limit for its assets, which might step down to something like five years by the end of the reinvestment period. As such, it could be 10 years or more before the senior notes get paid down in full (even if, historically, it's happened more quickly, and many CLO investors get taken out by a reset or refinancing long before that). A lot can happen in a decade: speculative-grade companies can default, entire sectors can be disrupted, or there may be a full-blown recession. Portfolio diversity can mitigate the idiosyncratic risks among these, but too much diversity can also guarantee some exposure to the next troubled sector.

In a prior study we published on how different types of diversity might affect U.S. CLO performance, we reviewed the impact the Global Financial Crisis (GFC) and 2015-16 energy slowdown had on CLO portfolios with different obligor and industry diversity profiles (see "Can Too Much Diversity Have Negative Effects On CLO Portfolios?" published April 23, 2018). In this follow up study, we will revisit obligor and industry diversification again, looking at it through the lens of two more recent periods of economic dislocation and slowed economic growth: the 2020 pandemic-driven recession, and the period of rising interest rates and slowing economic growth that followed in 2023.

CLO Obligor Diversity Has Doubled Over The Past 20 Years

At the end of 2004, the median reinvesting U.S. BSL CLO 1.0 transaction (those originated from the inception of the CLO market through 2009) had about 150 obligors in its portfolio (see chart 1). By the end of 2023, the median count of obligors for reinvesting U.S. BSL CLO 2.0 portfolios had doubled to about 300 obligors. Much of this growth in obligor diversity has taken place within the past 10 years. The median U.S. BSL CLO portfolio today would have been considered one of the most diverse (top 10th percentile) as recently as eight years ago, in 2015. More recently, however, we may have reached "peak" diversity. Since the end of 2021, after a year of heavy CLO issuance, the median number of obligors across reinvesting transactions has held steady at around 300.

Chart 1

Despite the increase in individual CLO portfolio obligor diversity, the aggregate number of obligors (excluding tranches from other CLOs held as assets in CLO 1.0 transactions) across rated U.S. BSL CLO portfolios is lower now than it was prior to the GFC. The aggregate number of unique obligors in CLO 1.0 transactions increased sharply in the years before the GFC, to more than 1,900 by the end of 2007 from fewer than 1,500 at the end of 2004. It then declined by the end of 2008, as CLO managers traded out of about 100 lesser-held, smaller obligors and rotated into larger, more widely held obligors (which typically had higher ratings) as part of a flight to quality during uncertain economic times. By the end of 2009, CLO 1.0s reduced their exposure to the larger top 250 obligors and rotated back into some of the lesser-held, smaller obligors. As they traded through the GFC, some CLO 1.0 transactions were able to regain par that had been lost during the recession, including some CLOs that were able to get back above their target par.

After the GFC, the aggregate number of unique obligors across BSL CLOs declined sharply from 2011 through 2015, before gradually picking up again by 2018 as new CLO 2.0 transactions gained exposure to a new batch of smaller obligors, many of which were rated 'B-' (see chart two). By the end of 2022, exposure to loans from 'B-' rated companies had more than tripled relative to the CLO 1.0s issued before the GFC, changing the overall credit profile of the portfolios notably. By end of 2023, we noted the first year-end decline in 'B-' exposure in about a decade, a decline to about 26% from just over 30% in 2022 year end.

Chart 2

How Exposure To The Most Widely Held CLO Obligors Changed Over The Years

Between 2005 and 2012, over two-thirds of BSL CLO assets came from the largest 250 CLO obligors (see chart three). As CLO 1.0 transactions exited their reinvestment periods, the percentage exposures to the largest 250 obligors across CLOs gradually declined to the low 50% range as post-GFC CLO 2.0 transactions invested in smaller, less widely held companies.

What hasn't changed is the reaction of CLO managers during periods of economic stress. During the pandemic-driven recession in 2020, there was a flight to quality, which resulted in the uptick in CLO exposure to the "top 250" obligors, just as there had been at the end of 2008 during the GFC. From a manager perspective, this is a rational response, as, historically, the largest 250 most widely held CLO obligors have experienced lower rating volatility compared to smaller, less widely held BSL CLO obligors.

Chart 3

In our study on diversity from 2018, we found that portfolios with lower obligor diversity and higher industry diversity were more likely to have experienced greater credit deterioration during the GFC and the energy slowdown. In this article, we use a similar framework to assess the impact of the 2020 pandemic and the period of lower economic growth that followed in 2023 on our index of reinvesting U.S. BSL CLO transactions.

In this study, we use different ways to measure diversity depending upon the context. In chart 1, we referenced the count of obligors in a CLO portfolio or across CLO portfolios, regardless of the size of the various positions held. The second measure is an effective count of obligors (or industries) in a CLO portfolio or across CLO portfolios, adjusted for the respective size of each obligor exposure (for example, a portfolio of 100 equally weighted obligors is more diverse than a portfolio with 100 obligors where one obligor is worth 99% of the portfolio). For this, we borrow from our CDO Monitor framework (see "All You Need To Know About CDO Monitor," published March 24, 2020), which has diversity measures based on a modified Herfindahl–Hirschman index calculation. Obligor diversity measure (ODM) is an effective count of unique issuers in a CLO portfolio, and the industry diversity measure (IDM) is an effective count of unique industries, both normalized to account for position sizes.

Obligor diversity and industry diversity are both typically incorporated in various ways in CLO indentures through concentration limitations and other means, as well as the CDO Monitor framework for S&P Global Ratings-rated transactions. Based on median ODM and IDM values across our rated universe of CLOs, we've bucketed the portfolios into the following four groups:

• Low obligor/low industry (low obligor diversity across a narrow array of industries);
• Low obligor/high industry (low obligor diversity across a wide array of industries);
• High obligor/low industry (high obligor diversity across a narrow array of industries); and
• High obligor/high industry (high obligor diversity across a wide array of industries).

In addition to grouping the CLO transactions using the categories above, for purposes of this study, we've provided some average figures across mangers with larger, medium, or lower assets under management (by issuance volumes since the second half of 2020 as reported within the CLO Global Databank as reported by Leveraged Commentary & Data at the start of 2024):

• Larger assets under management (AUM) managers (managers that have issued greater than $2.9 billion post-pandemic transactions);
• Medium AUM managers (managers that have issued between $1.2 billion and $2.9 billion of post-pandemic transactions); and
• Lower AUM managers (managers that have issued less than $1.2 billion post-pandemic transactions, if at all).

The Impact Of Diversity During The 2020 Pandemic

With our index of over 400 reinvesting U.S. BSL CLOs at the start of 2020, larger AUM CLO managers tended to have higher obligor diversity in their portfolios, while lower AUM managers tended to have lower obligor diversity. Industry diversity on the other hand, was fairly even across the three manager size cohorts, with larger managers having slightly lower industry diversity. Meanwhile, larger AUM managers had lower exposure to assets rated 'CCC+' or lower, and the loans held in their CLOs were more liquid and had higher market prices, primarily because these CLOs had greater exposure to top 250 CLO obligor names. Smaller AUM managers, on the other hand, had greater exposure to less widely held companies, along with more exposure to 'CCC+' and below assets and loans that were less liquid with lower market prices.

There were a mix of diversity profiles across all three cohorts of managers by size, but relatively speaking, at the start of 2020:

• Larger AUM managers were more likely to have portfolios with high obligor diversity and either low or high industry diversity.
• Medium AUM managers were more likely to have low obligor and high industry portfolios.
• Lower AUM managers were more likely to have portfolios with low obligor diversity and either low or high industry diversity.

Table 1

At the start of 2020, the low obligor/high industry transactions (which were more likely to come from tier 2 and tier 3 managers) had weaker credit metrics (higher default exposure, lower junior overcollateralization (O/C) cushion and higher par loss relative to original target par). No deal was exempt from the credit impact of the pandemic, despite manager intervention to de-risk the portfolios (rotating out of weaker credits into larger obligors with higher ratings), and all CLOs experienced some level of credit deterioration. Most transactions experienced a dramatic deterioration in 'CCC' asset exposures and junior O/C cushions during the second quarter of 2020, followed by a gradual improvement for the remainder of the year. Exposure to defaulted assets rose sharply and remained at elevated levels for the remainder of the year. Portfolio par balances experienced a gradual decline since the onset of the pandemic, particularly for the low obligor/high industry transactions towards the end of 2020. Much of the par loss can be attributed to asset sales at distressed prices (de-risking activity) as well as defaults when assets were written down or converted to equity.

Charts 4-7 show how the four cohorts of CLOs based on ODM and IDM values performed during 2020 through the lens of four performance metrics: 'CCC' asset exposure, defaulted asset exposure, junior O/C test cushions, and changes in CLO par relative to target par.

Chart 4

Chart 5

Chart 6

Chart 7

When defaults in the leveraged loan market occurred during the pandemic, the low obligor/high industry portfolios experienced a larger negative impact across their CLO metrics: default exposures ticked up faster and remained at higher levels; and O/C cushions experienced a sharper decline, while portfolio par balances experienced a steeper decline for the rest of the year. Across the four cohorts of different diversity types, the low obligor/high industry transactions ended 2020 with the highest defaulted asset exposures, the lowest junior O/C test cushions, and the lowest portfolio par balances.

The Impact Of Diversity In 2023

We repeated the analysis across our U.S. BSL CLO Index in 2023 (over 490 transactions). By 2023, many leveraged loan issuers have already experienced stress from the high interest rate environment, resulting in elevated downgrade-to-upgrade ratios as well as a notable uptick in defaults. Some issuers also experienced sector-specific stresses (for example, across healthcare-, consumers-, telecom-, and chemicals-related issuers). We also noted a slight uptick in the top 250 exposures across reinvesting transactions by the end of 2023, although this was much less significant than during the GFC or pandemic (see chart 8).

At the start of 2023:

• Larger AUM managers were more likely to have portfolios with high obligor diversity and either low or high industry diversity.
• Medium AUM managers were more likely to have portfolios with low obligor diversity and either low or high industry diversity.
• Lower AUM managers were more likely to have low obligor and high industry diversity portfolios.

Table 2

There was a slight uptick in CLO diversity after the pandemic (median obligor count at the start of 2023 was 297, up from 268 at the start of 2020; see chart 1), mainly from the transactions formed in 2021, which was a high issuance year (median obligor count remained flat for two years after 2021). In table 2 above, we find part of the increase in diversity is driven by newer CLOs issued by smaller AUM managers, which saw average ODMs increase to 190 at the start of 2023 from 159 back at the start of 2020. Meanwhile, the average ODMs of larger and medium AUM managers have not changed materially during this time. Low obligor/high industry transactions, again, started 2023 off with weaker credit metrics (higher exposure to 'CCC' and below) and lower junior O/C cushion.

Charts 8-11 show how the four cohorts of CLOs based on ODM and IDM values performed during 2023 through the lens of four performance metrics: 'CCC' asset exposure, defaulted asset exposure, junior O/C test cushions, and changes in CLO par relative to target par.

Chart 8

Chart 9

Chart 10

Chart 11

In 2023, low obligor/high industry transactions started and ended the year with weaker portfolio metrics. At the start of the year, the low obligor/high industry transactions experienced the highest percentage exposure to obligors that would have experienced a downgrade in 2023. Of note, the average number of obligors that experienced a downgrade in 2023 was lower within low obligor/high industry CLO transactions; however, because these deals had less obligor diversity, the proportion downgraded was larger. We examine the differences in exposure across these four transaction cohorts based on ODM and IDM values at the start of the year to explore what led to the different trajectories. Interestingly, the least diverse portfolios--the low obligor/low industry transactions--experienced the smallest proportion of 2023 downgrades at the start of the year. Perhaps these managers were more deliberate in their credit selection, only diversifying into obligors and industries they were most familiar with.

Table 3

At the start of 2023, relative to the overall sample, the high industry transactions (both low and high obligor) tended to have:

• Less exposure to IT;
• Less exposure to healthcare;
• Less exposure to financial;
• More exposure to materials (including chemicals); and
• More exposure to consumer staples sectors.

Interestingly, the low obligor/high industry transactions had more exposure to consumer discretionary and, to a lesser extent, more exposure to communication services (which includes telecom), both sectors that have weaker credit rating distributions and a high proportion with a negative rating outlook.

Chart 12

In 2023, the low obligor/high industry transactions also stood out as they had notably less exposure to the top 250 obligors; conversely, this means they had greater exposure to thinly held obligors (750+), which also had a weaker credit rating distribution (see chart 13). With a portfolio that was concentrated in fewer obligors and larger exposure to thinly held obligors, in hindsight, its unsurprising these transactions experienced greater deterioration within their CLO metrics.

Chart 13

High industry diversity and low obligor diversity portfolios showed weaker performance relative to other CLOs in stressful periods

U.S. BSL CLO portfolios with low obligor diversity and high industry diversity showed weaker performance relative to other CLOs during the GFC, energy slowdown, pandemic, and in 2023. It appears the level of exposure to larger widely held issuers with higher ratings (i.e., top 250) may have a material impact on CLO performance during periods of economic stress.

Related Research

• CLO Insights 2024 U.S. BSL Index: Some Thoughts On The Value Of CLO Diversity, Feb. 1, 2024
• All You Need To Know About CDO Monitor, March 24, 2020
• Can Too Much Diversity Have Negative Effects On CLO Portfolios? April 23, 2018