CLOs: Still No Reason To Be Scared of the Initials

KEY POINTS
• Although gross issuance of CLOs (collateralized loan obligations) and bank loans has been running high this year, net new issuance has been subdued — allaying concerns about overheating.
• Structural features built into CLOs are a buffer against potential capital losses as underwriting for bank loans loosens amid strong investor demand.
• With the economic upcycle entering its later stages, we are closely monitoring pressures on lending standards and maintaining an emphasis on upper-tier managers.
• We think that fears of overleveraging among buyers of CLOs are overblown.

Recent months have seen growing concerns about potential overheating in the bank-loan sector and the possible impact on CLOs, whose managers are major purchasers of bank loans. Weakening loan terms and less disciplined lending are often cited as evidence of such overheating. Also, the low-yield environment has driven robust demand for CLOs’ relatively high yields, spurring a rise in issuance. A potential rollback in risk-retention rules could reinforce this rise.

While acknowledging these concerns, we remain positive on CLOs for many reasons.1

Net new issuance has been subdued
We see no signs that the bank-loan market is overheating. Importantly, the share of the market made up of lower-quality loans and leveraged-buyout deals remains small. Meanwhile, the portion held by CLOs has held steady around 60% – 65%, consistent with its long-term average. Even though total CLO issuance (including new issues, refinancings, and resets2) is on pace to reach a record high this year, it has been made up mostly of refinancing and reset deals. For such deals, CLO managers do not need to buy new bank-loan collateral; instead, they are refinancing/resetting existing deals to take advantage of lower coupons. In contrast, new CLO creation, which requires new bank-loan collateral, has been subdued (Figure 1). We have observed a similar trend in the bank-loan market: Total issuance has been quite high but mostly composed of refinancings, while new bank-loan creation has been much closer to its historical average.

Robust CLO structures can protect against increased bank-loan losses
Backed by a tailwind of brisk growth in demand, terms on new bank loans are becoming increasingly issuer friendly as manifested in the rising frequency of “covenant-light” loan agreements. (Covenants are commitments made by borrowers to lenders, typically featuring lender-protective measures such as maintenance of minimum cash-flow to debt service ratios. Covenant-light agreements do not contain the usual lender protections.)

We believe covenant erosion has reached a peak, as many investors are starting to push back against this trend. Nevertheless, we acknowledge that higher levels of covenant-light loans in the market could translate into higher defaults and lower recoveries in an economic down cycle, with negative knock-on effects for CLOs. However, bank-loan defaults remain near historical lows (1.4% as of 31 October 2017 according to JPMorgan) and we do not expect a material pickup in the near term given our constructive outlook on loan fundamentals.

Although an economic downturn is not our base-case scenario, bank-loan defaults could potentially reach the high single digits in such a regime. As perspective, for several years after the bursting of the tech bubble in the early 2000s, the trailing 12-month bank-loan default rate ran in the 5% – 8% range, marking the worst prolonged period in the sector’s history. (The default rate briefly spiked during the global financial crisis, to roughly 11%, but the upsurge was short-lived.) Applying the same loss levels experienced during the 2000 – 2003 default cycle to CLOs currently in the market, we estimate that AAA rated bonds could withstand about four times the defaults experienced in this worst-case scenario before incurring capital losses, while BBB tranches could survive about two times that scenario. These multiples would be even higher if we were to apply the bank-loan loss experience of the global financial crisis. This resilience speaks to the strong structural protections built into CLOs.

Some critics draw parallels between today’s CLO market and the non-agency residential mortgage-backed securities (RMBS) market in 2008. We believe such comparisons are inaccurate. Pre-crisis nonagency RMBS were loosely underwritten on the assumption that home prices would keep rising. As a result, loans were repackaged into deals with very little investor protections. Leverage was then applied throughout the structure, which amplified losses during the housing crisis.

In contrast, there is no layering of risks within CLO structures (that is, there are no CLO deals backed by other CLOs). Furthermore, the average maturity of loans in pre-crisis RMBS was 30 to 40 years; the average maturity of loans in CLOs is just three years. Of course, there is still the risk of maturity extension and defaults if loans can’t be refinanced. But unlike RMBS and as outlined above, CLOs have strong structures with the ability to withstand such defaults and extensions.

Easy financial conditions could encourage bad behavior
Economic recovery and a favorable credit environment may be propping up some companies that would cease to exist under less-supportive conditions. When the cycle eventually turns, credit conditions will tighten, default risk will rise, and lenders will be less inclined to lend to companies with less-stable balance sheets. Likewise, tightening bank-loan spreads may be encouraging some CLO managers to stretch for riskier loans with higher yields. So, we think it’s important to assess the resilience in an economic downturn of both the CLO managers and the underlying collateral.

Credit bonds issued in 2017 could underperform relative to deals under-written to more strict guidelines in earlier years. As a result, we favor credit bonds of slightly more seasoned CLOs, although we remain comfortable with newly issued senior bonds given their credit support. Within the overall sector, we tend to favor upper-tier CLO managers who have demonstrated the ability to navigate through an economic cycle and who are less apt to loosen their credit standards.

Potential rollback of US risk-retention rules
The US Dodd-Frank Act requires CLO managers to retain at least 5% of the credit risk of a CLO. However, a group of CLO managers is challenging this rule, and the US Treasury Department recently released a report calling the rule overly punitive and recommending that it be dialed back. US CLO managers could be required to hold less than 5% of the deal; or they could end up being exempt from risk retention altogether.

The risk-retention rule has kept a lid on issuance and has limited credit availability, so relaxing it would likely boost CLO issuance. Loosening would also benefit smaller CLO managers hampered by the increased capital requirements that come with the rule. From an investor’s standpoint, rolling back the rule could improve manager diversification, which has narrowed in recent years. With more managers in the market, CLO issuance could rise, and consequently heightened demand for bank loans could exacerbate current pressures to lighten lender protections. We will continue to closely monitor this risk.

Buyers not using excessive amounts of leverage to buy CLOs
Investors’ thirst for income-producing assets has drawn an increasing number of buyers into the market for CLOs, based on these instruments’ attractive yields versus other credit assets. This has raised concerns that some buyers of CLOs could be using leverage to boost income-earning potential in the current low-interest-rate environment. Yet we have seen no evidence of outsized leverage being employed across the CLO market. In fact, we estimate that less than 10% of AAA CLO purchases involve leverage. Hedge funds typically use leverage, but they make up a very small share of new-issue buyers.

Much more prominent in the buyer base are longer-term asset owners such as banks, insurance companies, and corporate pensions (Figure 2). This buyer-base profile should help dampen price volatility — especially in a downturn, since there should be limited forced selling.

In conclusion: Positive outlook on CLOs, but monitoring lender and manager behavior
We believe the behavior currently exhibited by lenders and CLO managers is indicative of the later stages of the economic cycle and not something more systemic. There are no signs of excesses building in the sector that would put higher-quality senior bonds, and even many subordinate bonds, at risk of loss. In the current environment, we are maintaining a higher-quality portfolio, focusing on upper-tier managers. When we invest down-in-credit (A rated and below), we favor refinanced bonds with shorter weighted-average lives; such bonds have similar credit support/ratings as their longer-dated new issue peers but with lower risk.

As always, we benefit from our CLO and bank loan analysts’ assessment of the quality of the underlying loan pool, and their evaluation of CLO managers’ strength, to identify security-specific opportunities and avoid potential underperformers.