CLOs are no stranger to economic slowdowns and shocks. Over a nearly 30-year history, CLOs have weathered multiple storms, including the dot-com bubble, the global financial crisis (GFC), energy and retail slowdowns, and the COVID-19 pandemic. During all of this, the CLO structure has continued to hold up and perform as intended. Since their inception in 2010 through 2021, CLO 2.0s (issued 2010 or later) had not experienced a default, but that is no longer the case.

From the mid-1990s through second-quarter 2022, 50 S&P Global Ratings-rated U.S. CLO tranches saw their ratings lowered to 'D' out of nearly 16,000 tranches rated. Of these, 40 were CLO 1.0 transactions issued prior to the GFC (2009 and before), while the other 10 defaulting tranches were from CLO 2.0 transactions and none were initially rated higher than 'BB'.

In August 2021, we lowered the first CLO 2.0 notes to 'D', and since then, several more CLO 2.0 notes have defaulted and been downgraded to 'D' (ten tranches) or have experienced a downgrade to a non-performing rating of 'CC' where we view a default will be highly probable by the legal final maturity of the tranche (one tranche). The tranches across these 11 CLO 2.0s are listed here: "U.S. CLO Defaults As Of March 17, 2022," published March 23, 2022.

Below, we will explore what happened to understand why these junior CLO tranches defaulted. How were they different? Were the structure or portfolio exposures unique? Or were they just unlucky?

The Pre-2015 2.0 Cohort

At the start of 2015, there were about 400 outstanding S&P Global Ratings-rated CLO 2.0 transactions still within their reinvestment period (the pre-2015 CLO 2.0s). Starting in 2015, U.S. speculative-grade corporates experienced an increase in downgrades and subsequent defaults, which peaked in mid-2016. A significant proportion of these downgrades followed the significant drop in commodity prices during that time as the oil and gas and commodities industries came under significant distress due to the price collapse. Then, in 2017, several retail-related issuers experienced downgrades as well resulting in over 20 obligors being rated in the 'CCC' category (a 10-year high at the time) and contributing to the slight uptick in defaults in mid-2018. These corporate rating actions impacted the pre-2015 CLO 2.0s specifically as CLO 2.0s closing after 2015 avoided these already distressed sectors.

During this time, the CLO market evolved with the introduction of refinancings (where the spreads on a CLO's notes could be reduced without calling the transaction) and resets (where the notes were refinanced, the CLO's reinvestment period end date and legal final maturity date could be extended, and other changes could be made). Starting around 2016 and 2017, we noted an uptick in these transactions. A number of pre-2015 vintage CLOs that had maintained healthy overcollateralization (O/C) ratio cushions reset and extended their reinvestment periods, which resulted in many of these transactions being able to reinvest through the pandemic and later. Meanwhile, other pre-2015 CLO 2.0s refinanced some or all of their outstanding notes without extending the reinvestment period.

Chart 1 shows three cohorts of pre-2015 vintage CLO 2.0 transactions:

• CLOs that were able to reset, continued reinvesting, and didn't see any tranches default (yellow line);
• CLOs that did not reset, amortized during the 2020 pandemic, and didn't see any tranches default (orange line); and
• CLOs with defaulting tranches (defaulting CLOs), none of which were reset and all of which were amortizing during the pandemic (blue line).

All three cohorts experienced a decline in their junior O/C cushions during the pandemic-driven economic downturn in 2020. However, the defaulting CLOs experienced a much more notable decline during their reinvestment periods, which got noticeably worse in 2019 when the average junior O/C cushions turned negative even before the pandemic.

Chart 1

The Defaulting CLOs: How They Differed From Others

At the start of 2015, the defaulting CLOs had larger-than-average exposure to assets from 'CCC' category obligors and non-performing assets. They also had lower junior O/C cushions and about double the exposure to energy-related issuers than non-defaulting peers. Interestingly, the defaulting CLOs weighted average recovery rating (WARR) distribution was higher ('AAA' WARR of 44.1%) than that of the average pre-2015 CLO 2.0s (43.5%). This was largely due to the their higher exposure to loans from energy-related issuers, whose debt tended to have higher recovery ratings due to higher debt cushions. These energy-related companies also issued loans with high spreads, which contributed to the greater than average weighted average spreads amongst the defaulting CLOs.

The defaulting CLO portfolios were also less diversified relative to the average diversity levels of the pre-2015 CLO 2.0s as measured by the obligor diversity measure (ODM), which gives an effective count of unique issuers in a portfolio. Additionally, we noted the these CLOs' junior note subordination levels were slightly lower than the other pre-2015 CLO 2.0s.

Table 1

Reinvestment Periods

As noted above, the defaulting CLOs started 2015 on slightly weaker footing relative to the rest of the pre-2015 CLO 2.0s. As they reinvested, corporate downgrades and defaults from 2015-2018 contributed to further credit deterioration of the defaulting CLOs portfolios. We noted in a prior article (see "CLO Spotlight: How Do CLO Managers Perform In Times Of Stress?" published Sept. 6, 2016) that many managers traded during this time to reduce exposures to 'CCC'-rated and defaulted obligors at the cost of par.

For the defaulting CLOs, on average we saw that by the end of their reinvestment periods:

• Their portfolio diversity (as measured by ODM) remained less diverse than other CLOs at the time;
• They reduced energy-related exposure by approximately a third (down to 6.6% from 10.1%);
• They increased retail-related exposures by more than two-fold (up to 7.3% from 3.4%);
• They increased exposure to higher-rated issuers ('BB-' and above) by approximately 1.6%;
• Their par balances declined to an average of 98.5% of target par, due to asset defaults and par losses from trading;
• Their 'CCC' category buckets increased to well above the 7.5% threshold; and
• Their junior O/C test cushions declined to 1.5% due to the above-mentioned par loss and the market value haircuts from both excess 'CCC' and defaulted exposures.

During the reinvestment period, the managers for the defaulting CLOs rotated out of weaker performing issuers and into higher-rated issuers, some of which were rated 'BB-' and above. These trades may have offset the credit deterioration of the portfolio, but came at the price of par. Additionally, the rotation out of energy-related issuers and into retail-related issuers gave the CLOs exposure to another sector that would experience heightened defaults, and likely contributed to the deterioration they would face during their respective amortization periods.

Table 2

Only one deal amongst the defaulting CLOs failed its junior O/C test during its reinvestment period, which came back into compliance shortly thereafter. Relative to the other CLO 2.0s, these transactions had less portfolio diversification, which likely resulted in increased volatility within the 'CCC' buckets of these deals. During this time, the manager's attempts to reduce exposure to 'CCC' category issuers and defaults likely contributed to the gradual O/C decline. All the CLO 2.0s with defaulting tranches ended their reinvestment periods with lower but positive O/C cushions.

Amortization Experience Of The Defaulting CLOs

During their amortization periods, the defaulting CLOs, on average, saw:

• High levels of amortization occur early, which then slowed somewhat in the latter years;
• The 'AAA' notes paid off in full about two years after the end of the reinvestment period;
• A significant proportion of the initial amortization came from loan prepayments of higher-rated issuers ('BB-' and above) in the portfolio; and
• The remaining loans from weaker issuers ('CCC+' rated and below) become an increasing proportion of the portfolio, which would have negative implications on O/C cushions.

Chart 2

Portfolio concentration can make the portfolio increasingly sensitive to idiosyncratic events. Although the default rate for the U.S. speculative-grade corporate issuers was relatively benign in 2019, the downgrades and defaults during this year had an increasingly negative effect on the portfolio metrics for the defaulting CLOs, evidenced by the notably sharp decline in their O/C cushions in table 1.

Collectively, the defaulting CLOs had exposure to over 550 issuers at the start of 2019, 20% of which were downgraded in 2019; approximately 8.8% of these issuers saw their ratings lowered into either the 'CCC' or non-performing categories. The downgrades as well as the pay downs of stronger credits led to significant increases in the percentage exposures to 'CCC'-rated and non-performing issuers noted in chart 2 above. Since the defaulting CLOs had already exceeded the 7.5% 'CCC' threshold, O/C cushions significantly declined in 2019, with the average junior-most O/C cushion amongst the defaulting CLOs turning negative by mid-2019, though still slightly above water as junior O/C triggers were set slightly above 100%. At this point, liquidation was not an option. Given the low market values for 'CCC'-rated and non-performing assets, many of these deals were underwater with respect to liquidation values (market value of portfolio minus principal balance of rated notes), leaving the CLO equity holders little incentive to optionally redeem before the pandemic.

Conclusion

Well before the arrival of the pandemic-driven recession in March of 2020, the defaulting CLOs were showing far weaker performance (and failing more junior O/C test ratios) than other CLOs from the same era. They closed with portfolios that had slightly less obligor diversity, junior tranches with relatively less subordination, and with relatively overweight energy exposure ahead of the oil and gas and commodities downturn in 2015-2016. Not surprisingly, these deals were not selected for reset, and thus they continued to experience credit deterioration during their amortization period, only to hobble into the pandemic with already underwater portfolios. Perhaps most importantly, the CLO tranche defaults amongst the defaulting CLOs were limited to the junior notes that were originally assigned a speculative-grade rating, as structurally intended. Most of the senior notes amongst the defaulting CLOs were paid off in full well before it became evident the junior notes were at risk of default. For the junior notes of the defaulting CLOs, 2020 was the final nail in the coffin.