As the U.S. collateralized loan obligation (CLO) asset class approaches its 30-year mark, it has established an enviable track record of performance through three full-blown recessions and a number of non-recessionary periods of economic stress. This is due in large part to CLO managers working to add value for investors by building a portfolio that seeks to meet return objectives while mitigating risks to the CLO noteholders. This takes many forms, but a large part of a manager's job is to reposition their portfolio exposure to downgrades and defaults, reduce the risk of downgrades to the CLO tranches themselves, as well as other types of risk during periods of stress (i.e., active management).

Over the past year-and-a-half, interest rates have gone through steep increases and economic growth has slowed, which put pressure on the highly levered, speculative-grade-rated corporate borrowers that issue most of the loans inside U.S. CLO collateral pools. Given the economic uncertainty and a gradual deterioration of U.S. broadly syndicated loan (BSL) CLO credit metrics over the past 18 months (see "SLIDES: U.S. BSL CLO And Leveraged Finance Update: Elevated Rates Expected To Keep Pressure On Low-Rated Corporates," published Nov. 3, 2023), we are now taking stock of how active management of collateral has affected CLO performance. (The high degree of transparency across BSL CLOs, a data-rich asset class, have provided lots of information to assess performance). This article builds on similar studies we have done previously to assess the value of active management during previous periods of stress, including the Global Financial Crisis; the oil, gas, and commodities slowdown in 2015 and 2016; and the 2020 pandemic (see the Related Research section at end of this article).

Study Framework

For this study, we reuse the framework used in our prior articles on manager value where we compare actual performance of CLO portfolios with a hypothetical static portfolio scenario where we assume the managers made no trades. We go into detail about the time period and the sample we selected for this study below.

Sample

Our sample consists of 485 S&P Global Ratings-rated U.S. BSL CLOs from 108 CLO managers that have been reinvesting during the entirety of this seven-quarter period. Slightly more than half (56%) of these CLOs were originated after the arrival of the pandemic in the first quarter of 2020, while the remaining 44% are pre-pandemic CLO transactions.

Given the sample includes transactions that have already gone effective by the start of 2022, our sample does not include transactions from newer CLO managers that have issued their inaugural CLO in 2022 or 2023. Some of these CLOs from the newer CLO managers have unique CLO metrics (e.g., some recent inaugural transactions have very low 'B-' exposures or very low S&P Global Ratings' weighted average rating factor [SPWARF] values).

Time period

To assess the impact of active management in the post-pandemic era, we focused on a seven-quarter period from the start of 2022 through third-quarter 2023. This period follows 2021, a year with near-record activity in the U.S. BSL CLO loan market and record issuance for U.S. CLOs. By the start of 2022, average credit metrics for pre-pandemic CLOs had recovered much of the ground they lost in 2020, while post-pandemic transactions have continued with fairly clean metrics. Average CLO metrics across the entire sample (including both pre- and post-pandemic CLOs) in January 2022 had an average 'CCC' asset exposure of about 4.6%, a non-performing asset exposure at near zero (0.13%), a junior overcollateralization (O/C) test cushion of 4.7%, and an average portfolio par balance at 99.9% of target par.

By the end of first-quarter 2022, concerns over geo-political risk and inflation began to mount, and economic growth started to slow. By May 2022, downgrades of obligors in U.S. BSL CLOs were outpacing upgrades for the first time since 2020, which has continued to be the case since. From the start of 2022 through third-quarter 2023, more than 140 U.S. CLO obligors saw their ratings lowered into the 'CCC' category, while another 90 saw downgrades to non-performing ratings ('CC', 'SD', and 'D'). Now seems like a good time to review how CLO portfolios have fared and assess the impact of manager turnover.

Chart 1

For a more detailed look at the CLO metrics during this time period, see tables 6-8 in the Appendix section.

What Could Have Happened Without Active Management

Using the same framework from our prior manager value studies, we imagined a hypothetical static CLO scenario where CLO managers made no trades after the start of 2022. Given portfolio exposures at the start of 2022, we recalculated hypothetical CLO metrics with Oct. 1, 2023, corporate ratings and pricing data.

Hypothetical deterioration of U.S. BSL CLO credit metrics

Absent active portfolio management from CLO managers between the start of 2022 through third-quarter 2023, the average hypothetical change in metrics across our sample could have been:

• 'B-' exposure increase by 0.67% to 27.22%;
• 'CCC' exposure increase by 5.45% to 10.07%;
• Non-performing exposure increase by 1.17% to 1.30%; and
• Junior O/C cushion decrease by 1.68% to 2.97% (over 6% of our sample failed their junior O/C test).

Table 1 shows the hypothetical change in BSL CLO portfolio metrics absent manager trading from first-quarter 2022 through third-quarter 2023.

Table 1

Had the managers done nothing over the seven-quarter period, the average SPWARF of the portfolio could have increased to 2890 (instead of the actual 2768), and 'CCC' buckets could have broken into the low double digits, levels not seen since 2020. The average CLO would have exceeded the 7.5% 'CCC' threshold by 2.57%, exposing transaction O/C ratios to market value haircuts and potentially reducing the junior test O/C cushions by 1.68% to 2.97% instead of 4.66%.

What Actually Happened: Deterioration Of BSL CLO Portfolio Metrics Mitigated

After this seven-quarter period, on average, just over half of the assets that had been in the CLOs at the start of 2022 still remained in the portfolios, yielding a 48% portfolio turnover. The turnover volume of 48% over seven quarters was about the same as the turnover during 2021, when reinvesting U.S. BSL CLOs saw a 49% turnover on average over just four quarters.

Minimal-to-modest deterioration of U.S. BSL CLO credit metrics

Given the predominantly negative corporate rating actions over the period, a number of the CLO credit metrics have deteriorated across our sample:

• Average exposure to 'CCC' assets increased by 2.69% to 7.31%;
• Non-performing asset exposure increased by 0.47% to 0.60%;
• Junior O/C test cushions decreased by 0.69% to 3.96%; and
• Portfolio par balance decreased by 0.11% to 99.76%.

In addition, the average weighted average life (WAL) of the CLOs declined by 0.73 years after 1.75 years had elapsed.

The buildup of 'CCC' buckets across CLO portfolios was minimal-to-modest in full-year 2022, with an average increase of just 0.33%, before accelerating another 2.36% in the first three quarters of 2023, particularly during the second and third quarters.

Table 2 shows the change in U.S. BSL CLO portfolio metrics from first-quarter 2022 through third-quarter 2023.

Table 2

Purchase of fixed-rate assets from higher-rated issuers helped to mitigate some collateral credit deterioration

One development in this seven-quarter period that may have mitigated collateral credit deterioration for some CLOs is the purchase of fixed-rate assets from higher-rated issuers. Given the increase in interest rates, these were (and are) trading at a discount, which helped to build par, and which, in turn, combined with the higher ratings, improved portfolio credit metrics. By third-quarter 2023, average exposure to fixed-rate assets in CLO portfolios increased by 1.21%, to 1.52% from 0.31% since the start of 2022. In full-year 2022, we found that managers were able to build par by about 19 basis points (bps), partially due to this strategy; however, three quarters later, many transactions had lost most of the par they'd built in 2022 as CLO 'CCC' buckets and exposure to defaulted assets began to build up in 2023 and managers stepped up their de-risking efforts. 'B-' exposures experienced a steady decline in second- and third-quarter 023, a first in several years, partly due to a slower pace of CLO obligors being downgraded to 'B-' from 'B' and higher, and also from managers attempts to reduce weak 'B-' exposures (or potential future 'CCCs').

The Value Of CLO Manager Intervention

In the seven quarters since the start of 2022, by turning over the portfolio by 48% and at a cost of 11 bps of par, on average, managers were able to:

• Reduce the SPWARF by 121;
• Reduce their 'CCC' and non-performing exposures by 3.46%; and
• Preserve 0.99% of junior O/C cushion.

We found these transactions increased their exposure to the more widely held issuers (the top 250), which as a group, tend to have higher ratings and historically greater price stability (link to top obligor update). As a result, the trades into higher-yielding and more liquid widely held issuers helped to preserve the market value of the portfolio by Oct. 1, 2023, while the junior O/C cushions ended up about 1% higher had the managers done nothing.

Table 3 shows the difference between actual and hypothetical third-quarter 2023 portfolio metrics.

Table 3

Manager trades could save a ratings notch in the next downturn similar to 2020

In the second half of 2020, at the height of the pandemic, we lowered our ratings on hundreds of BSL CLO tranches, representing about 13% of the outstanding ratings. A large majority of these tranches were from the tranches originally rated within the 'BB' and 'B' categories. When we compare the credit metrics of the CLOs that experienced a 'BB' category tranche downgrade in 2020 with the transactions that didn't see a downgrade, we find some key differences. The CLOs that experienced a downgrade, on average, had 'CCC' exposures that were 1.63% larger, non-performing asset exposures that were 0.57% larger, and junior O/C test cushions that were 1.6% lower (see table 4).

Table 4

In reviewing the manager value metrics, across our sample, we find 56 transactions (11.5% of our sample) that experienced manager trades reduced 'CCC' exposure by greater than 1.63%, reduced the non-performing asset exposures by greater than 0.57%, and reduced the drop in estimated junior O/C test cushions by greater than 1.62%. Put simply, if we were to enter a period of economic stress similar to what we saw in 2020 in the future, the trades done by these managers across these 56 transactions between first-quarter 2022 and third-quarter 2023 may have been enough to save the junior CLO tranche from a downgrade.

Vintage And Manager Size Had Impacts On Different Outcomes Of Our Study

Vintage and manager size had impacts coming to different outcomes in our study.

By vintage

We found vintage had a large impact on the outcomes of our study. At the start of 2022, the pre-pandemic transactions had weaker relative CLO metrics, given they had already reinvested through a downturn in 2020. Relative to the post-pandemic cohort, the pre-pandemic cohort had (pre-pandemic vs. post-pandemic):

• Higher SPWARF values (2720 vs. 2676);
• Higher 'CCC' buckets (6.07% vs. 3.51%);
• Higher non-performing buckets (0.22% vs. 0.07%); and
• Lower junior O/C cushions (3.94% vs. 5.21%).

Pre-pandemic transactions had weaker credit metrics, and were closer to triggering their 'CCC' threshold and junior O/C limits. We found these older transactions experienced a higher turnover volume of 49.03%, vs. 47.44% for the post-pandemic cohort.

Relative to post-pandemic transactions, the pre-pandemic transactions experienced a (pre-pandemic vs. post-pandemic):

• Larger increase in 'B-' exposure (+2.24% vs. +1.68%);
• Lower increase in 'CCC+' and below exposures (+2.19% vs. +3.91%); and
• Larger decline in par balance (-0.31% vs. +0.04%).

This highlights the dynamic nature of managers as they tend to de-risk more when needed and trade less if the transaction is not at risk of failing their tests (i.e., managers may be less willing to give up par if there is still ample 'CCC'/O/C cushion). The pre-pandemic deals wound up trading more, moving out of 'CCC' category and below assets and into 'B-' rated assets, even at the cost of par, resulting in a larger decline in junior test O/C cushion.

The post-pandemic transactions experienced slightly less turnover/trading, a larger increase in 'CCC', and less par loss (in fact, many post-pandemic deals experienced a par gain). Because most post-pandemic deals are still well below their 7.5% 'CCC' bucket thresholds, these deals experienced a smaller decline in junior O/C cushion.

Some correlation between manager size and our study results

We found that there may be some correlation between the size of the manager and the results of our study. We created cohorts of our sample based on how active the managers have been at issuing CLOs after the pandemic (a rudimentary indicator of the manager's size), based on the issuance amount BSL CLOs priced since the pandemic. We grouped our sample by:

• Group 1: managers that have issued greater than $2.9 billion worth of post-pandemic U.S. BSL CLOs;
• Group 2: managers that have issued greater than $1.2 billion worth of post pandemic U.S. BSL CLOs; and
• Group 3: managers that have issued less than $1.2 billion worth of post pandemic U.S. BSL CLOs.

Even within each of the three manager groups (by manager size), we found the older pre-pandemic sub-cohort of transactions tend to trade more as they are closer to failing their various tests, likely for the same reasons as noted above (see table 5).

Table 5

In January 2022, the CLOs from group 1 managers started off with (group 1 vs. group 2 vs. group 3):

• Higher SPWARF values (2715 vs. 2615 vs. 2686);
• Higher 'CCC' bucket and average non-performing exposures;
• Higher junior O/C cushions; and
• Higher exposure to the top 250 obligors.

During the seven quarters, relative to the three manager groups, transactions issued by group 1 managers experienced:

• The least amount of turnover (47.87%);
• The smallest increase in 'CCC' and non-performing;
• The smallest decline in junior O/C cushion;
• The smallest decline in par (as % of target par); and
• The largest increase in fixed-rate debt.

Group 1 managers started with the largest exposure to the most widely-held loan issuers in BSL CLO collateral pools. These "top 250" obligors tended to experience greater rating and price stability during the study period than did smaller, less widely-held obligors. Because of the tendency to hold loans from the larger top 250 exposures, even if the managers had done nothing during this period, group 1 managers would have experienced less of an increase in the SPWARF, 'CCC', and non-performing exposures, as well as smaller declines in market values and junior O/C cushions (see hypothetical group 1 results).

Group 1 managers were also more likely to purchase fixed-rate assets, most of which were from higher-rated issuers that and were purchased at a discount in 2022 after the interest rates started to rise. This helped to further preserve CLO metrics. Despite starting the period with weaker credit metrics (higher SPWARF, 'CCC', and non-performing exposures), transactions issued by group 1 managers ended the seven-quarter period with the lowest 'CCC' and non-performing exposures, even though turnover was the lowest across the three manager cohorts. Junior O/C cushion and market values also declined by the least.

Group 2 and 3 managers started off with lower 'CCC' buckets, but they had greater exposure to smaller issuers, which historically have experienced less rating stability. During the seven quarters, these managers turned over their portfolios more than the group 1 managers, reducing the hypothetical 'CCC' bucket by a greater amount (trading out of exposures that would eventually experience downgrades to 'CCC-' or below). For example, had the group 2 managers done nothing, their 'CCC' buckets could have increased to 10.81% from 4.26%. After the turnover of 48.96% for group 2 managers, the actual 'CCC' exposures only increased to 7.64%, a reduction of 3.18%, which is higher than the 2.61% 'CCC' reduction for group 1 managers. The de-risking trades came at a cost of par as the group 2 and 3 managers experienced greater par loss (-0.18% and -0.19%, respectively; compared to par loss of -0.08% for group 1), which also led to larger declines in junior O/C cushion. We also found that group 2 and 3 managers were less likely to trade into fixed-rate assets, which could have helped slightly with the par loss and credit metrics. By the end, most of the CLO metrics across the group 2 and 3 transactions were a bit worse than group 1 transactions; though, through a higher volume of active trading, these managers were able to avoid double-digit 'CCC' buckets.

CLO Managers' Active Management Supports An Impeccable 30-Year Performance Track Record

Portfolios built during benign times and trades done during periods of stress are decisions that clearly have a meaningful impact on CLO performance. The decision to give up spread to reduce risk is never an easy one, given the push and pull between 'AAA' investors and equity holders inherent in the BSL CLO structure, but managers have shown through almost 30 years and multiple recessions that they can balance these incentives effectively for an impeccable track record of performance.

Appendix

Table 6

Table 7

Table 8

Related Research

• SLIDES: U.S. BSL CLO And Leveraged Finance Update: Elevated Rates Expected To Keep Pressure On Low-Rated Corporates, Nov. 3, 2023
• Under Stress: Assessing CLO Manager Performance During COVID-19, June 1, 2020
• How Do CLO Managers Perform In Times Of Stress? Sept. 6, 2016