We plan to publish more on this soon, but this month we take a quick look at how different types of diversity might affect U.S. collateralized loan obligation (CLO) performance. But first, let's look at the metrics across the CLO Insights Index.

After gradually trending downward over the past year, some of the CLO metrics have improved slightly in recent months. In particular, exposure to 'B-' and 'CCC' category obligors has improved slightly in recent months, and the proportion of obligors with ratings with a negative outlook has declined to 17.46% from 18.21% two months ago (see table 1).

The downward trend in the average junior overcollateralization (O/C) test cushion has also moderated somewhat, declining just a few basis points over the past two months. Average 'CCC' exposure across the index continues to hover around the 7.5% limit, with about 45% of reinvesting CLOs in the index (492 rated transactions this month) above the excess 'CCC' threshold.

Since December, there have been six widely held assets from the top 500 U.S. BSL CLO obligors that have either been downgraded into the 'CCC' category or to a nonperforming rating ('CC', 'SD' or 'D'). Three of these are software companies (see table 2).

In addition to these, since the start of the year, another five issuers have seen their ratings lowered to nonperforming ratings following a restructuring, distressed exchange, payment-in-kind (PIK), or seeking bankruptcy protection (see "U.S. BSL CLO Obligors: Corporate Rating Actions Tracker 2024 (As Of Jan. 19)," published Jan. 24, 2024). Exposures to these issuers were less widely held across the index (i.e., not within the top 500 obligors), and the impact on the index numbers was relatively muted.

Table 1

Table 2

Assessing The Benefits Of Portfolio Diversity

In our prior study on CLO portfolio diversity (see "Can Too Much Diversity Have Negative Effects On CLO Portfolios?" published April 23, 2018), we found that CLO portfolios with low obligor diversity but high industry diversity experienced a higher level of credit deterioration during the Great Financial Crisis (GFC) as well as the energy slowdow in 2016. These CLO portfolios were more likely than average to have material exposure to the next industry that would experience stress (for example, energy and retail during the 2016 slowdown). Doing the same analysis on our CLOs in 2023, we see pressure across several industries (e.g., healthcare, consumer-related, telecom, chemicals, etc.), likely resulting in an outsized negative impact to some CLO portfolios.

Using the data sample in table 1 above (with 492 outstanding U.S. BSL CLOs that have been reinvesting since January 2023) and applying a similar framework to the one we used for our 2018 diversity study, we found notable differences in CLO performance given the level and type of CLO diversity. To quantify levels of diversity for CLO collateral pools, we borrowed two diversity measures from our CDO Monitor framework (see "All You Need To Know About CDO Monitor," published March 24, 2020) to differentiate our sample of 492 transactions into four cohorts. The obligor diversity metric (ODM) provides an effective count of obligors within the portfolio, normalizing for the CLO's exposure to each obligor; the industry diversity metric (IDM) does the same for industry categories. Across our sample, ODM ranges from 113 to 330, while IDM ranges from 13 to 30.

We differentiate our sample of 492 reinvesting CLOs into four cohorts:

• High obligor and high industry diversity (141 transactions);
• High obligor and low industry diversity (105 transactions);
• Low obligor and high industry diversity (105 transactions); and
• Low obligor and low industry diversity (141 transactions).

We find the larger managers (by amount of CLOs issued since the pandemic) tend to have higher levels of diversity, in particular for obligor diversity. Meanwhile, smaller managers (managers that have issued fewer CLOs since the pandemic) have notably lower levels of obligor diversity.

During 2023, many transactions across the index experienced a decline in junior O/C cushion, mostly from default haircuts and, to a lesser degree, excess 'CCC' haircuts (see chart 1). We find the cohort of transactions with low obligor diversity and high industry diversity experienced notably higher levels of junior O/C cushion decline. We find that CLOs with low obligor diversity are more likely to experience defaults and O/C test volatility, as the impact from any single default is larger as it represents a larger proportion of the portfolio (relative to a transaction with high obligor diversity).

Chart 1

By the middle of 2023, there was a notable decline in 'B-' exposure, partially due to de-risking efforts from the manager. As managers trade out of weaker credits and into stronger credits (or in some cases where impaired credits recover less or get converted into equity), par balances tend to decline (see chart 2). Again, we find the transactions with low obligor diversity and high industry diversity tend to experience above-average levels of par loss during this rolling one-year period.

This makes sense as the impact from each trade from a portfolio with fewer obligors can have a larger impact on the portfolio par balance, relative to a portfolio with a higher obligor count. Again, in 2023, we find CLO portfolios with fewer obligors spread across more industries tend to underperform, just like they did during the GFC and the energy slowdown in 2016.

Chart 2