At the heart of the global financial crisis of 2007-09 was an obscure credit derivative called the collateralised debt obligation (CDO). CDOs were financial products based on debts – most notoriously, residential mortgages –which were sold by banks to other banks and institutional investors.
The profitability of these CDOs largely depended upon homeowners’ ability to repay their mortgages. When people began to default, the CDO market collapsed. And because CDOs were interwoven with other financial and insurance markets, their collapse bankrupted many banks and left others requiring government and central bank support.
Many thought this would put an end to the market for complex structured credit derivatives, but it didn’t. As of 2021, a close cousin of the CDO known as the collateralized loan obligation or CLO was approaching the equivalent value of the CDO market at its peak. A record number of CLOs were issued in August, and the market as a whole is approaching USD 1trn in value. Many within the financial services industry say that there is nothing to worry about, but there are good reasons why they could be wrong.
How CLOs differ from CDOs
Collateralized loan obligations are underpinned not by mortgages but by so-called leveraged loans. These are corporate loans from syndicates of banks that are taken out, for example, by private-equity firms to pay for takeovers.
Proponents of CLOs argue that leveraged loans have a lower record of defaults than subprime mortgages, and that CLOs have less complex structures than CDOs. They also argue that CLOs are better regulated, and carry weightier buffers against default through a more conservative product design.
None of this is untrue, but this does not mean risk has disappeared. Mortgages, for example, had low rate of defaults in the 1990s and early 2000s. But since CDOs enabled banks to sell on their mortgages to free up their balance sheets for more lending, they began lending to riskier customers in their search for more business.
This relaxation of lending standards into subprime mortgages – mortgages issued to borrowers with a poor credit rating – increased the eventual default rate of CDOs as people who could ill afford their mortgages stopped repaying them. The danger is that the same appetite for CLOs may similarly reduce standards in leveraged lending.
In one respect, CLOs may even be worse than CDOs. When homeowners failed to repay their mortgages and banks repossessed and sold their houses, they could recover substantial amounts that could be passed through to CDO investors. However, companies are rather different from houses – their assets are not just bricks and mortar, but also intangible things like brands and reputation, which may be worthless in a default situation. This may reduce the amount that can be recovered and passed on to CLO investors.
In a recent paper, we examined the similarities between CDOs and CLOs, but rather than comparing their design, we examined legal documents which reveal the networks of professionals involved in this industry. Actors working together over a number of years build trust and shared understandings, which can reduce costs. But the mundane sociology of repeat exchanges can have a dark side if companies grant concessions to each other or become too interdependent. This can drive standards down, pointing to a different kind of risk inherent in these products.
The US-appointed Financial Crisis Inquiry Commission (FCIC) found evidence of this dark side in its 2011 report into the CDO market collapse, underlining the corrosive effects of repeat relationships between credit-rating agencies, banks, mortgage suppliers, insurers and others. The FCIC concluded that complacency set in as the industry readily accepted mortgages and other assets of increasingly inferior quality to put into CDOs.
Unsurprisingly, creating CLOs requires many of the same skill sets as CDOs. Our paper found that the key actors in the CDO networks in the early 2000s were often the same ones who went on to develop CLOs after 2007-09. This raises the possibility that the same industry complacency might have set in again.
Sure enough, the quality of leveraged loans has deteriorated. The proportion of US-dollar-denominated loans known as covenant-light or cov-lite – meaning there are fewer creditor protections – rose from 17% in 2010 to 84% in 2020. And in Europe, the percentage of cov-lite loans is believed to be higher.
The proportion of US dollar loans given to firms that are over six times levered – meaning they have been able to borrow more than six times their earnings before interest, tax, depreciation and amortisation (EBITDA) – also rose from 14% in 2011 to 30% in 2018.
Before the pandemic, there were alarming signs of borrowers exploiting looser lending standards in leveraged loans to move assets into subsidiaries where the restrictions imposed by loan covenants would not apply. In the event of a default, this limits creditors’ ability to seize those assets. In some cases, those unrestricted subsidiaries were able to borrow more money, meaning the overall company owed more in total. This has strong echoes of the financial creativity that drove riskier borrowing in 2005-07.
So how worried should we be? The CLO market is certainly very large, and corporate defaults could soar if it turns out that the extra money pumped into the economy by central banks and governments in response to the COVID crisis provides only a temporary reprieve. The major buyers of these derivatives again seem to be large, systemically important banks. On the other hand, according to some accounts, these derivatives are less interwoven with other financial and insurance markets, which may reduce their systemic risks.
Nevertheless the market is at least large enough to cause some disruption, which could cause major ructions within the global financial system. If the networks behind these products are becoming blind to the risks and allowing CLO quality to slowly erode, don’t rule out trouble ahead.