When a private equity firm buys a business, they often load up the business with debt. This works out great for the private equity firm because they can use the debt to increase the return on their investment in the short term, while in the long term the business will be stuck with the debt and the private equity firm will have already made their money. But the practice of leveraged buyouts creates a problem: no one actually wants to lend to a company that is about to be stripped for spare parts (or at least they don’t want to do so at interest rates that would make said stripping profitable). To get around this, private equity gives their loans to a very complicated financial structure called a CLO. In normal times, CLOs provide private equity with cheap debt financing while granting their investors relatively high performing interest rates.
CLOs, or Collateralized Loan Obligations, are a type of Asset Backed Security that make leveraged buyouts possible. Like other ABS, CLOs work by pooling leveraged loans and dividing the security into tranches. Payments are made to the owners of each tranche according to a waterfall system, so the AAA tranche (the tranche at the top of the waterfall) gets paid first, while the equity tranche (the tranche at the bottom) gets paid last. Because AAAs get paid first, they are considered more secure, while equity investors have the riskiest investment, so they are compensated by receiving a higher return on their investment (assuming the cash makes its way down to the equity tranche in the first place).
In addition to the waterfall system, CLO investors are protected by overcollateralization (OC) tests. The basic idea behind overcollateralization is if investors put $500M into the CLO, that CLO needs to be backed by $510M of leveraged loans. The extra $10M gives investors a cushion so that even if some of the loans perform poorly, there will theoretically still be enough loans to pay all of the investors. Each individual tranche actually has its own OC test, i.e. the OC test for the AAA tranche will be more stringent than the OC tests for the lower tranches. If a particular tranche fails its OC test, payments from the loans are diverted from that particular tranche and are used to (1) pay back investors higher up in the waterfall, and (2) to try and buy new loans to try to get the OC test back into compliance.
For OC purposes, the loans that make up CLOs portfolios are normally valued at par. (If I lend someone $100 dollars, $100 is the par value of the loan, while if I sell the loan to the bank and the bank buys it for $90, $90 is the market value of the loan). There are three exceptions to this where loans aren’t valued at par. The first is if a loan defaults (in which case it’s valued at the lower of the market value and the expected recovery value). The second is if the market value of a loan goes below 80%, in which case the loan is valued at market, not at par. The third is CLOs have a CCC bucket, which means they are only allowed to hold a small percentage of the worst rated corporate debt (CCC rated debt). If the amount of CCC debt exceeds some percentage (typically 7.5%) then the CCC debt is valued at market instead of at par.
CLOs are also protected by the correlation assumptions used to build their portfolio. The basic idea is that CLOs tend to be diversified and hold debt from multiple industries. Theoretically, downturns will impact some industries more than others, so diversification theoretically means that only some, rather than all, of a CLO’s portfolio will be adversely impacted in a downturn. Unfortunately, the correlation assumptions failed to take into account the possibility of an event like Covid-19, which impacts every industry and completely blew up whatever protection the correlation assumptions may have given CLOs in a normal recession.
In downturns OC tests that normally protect CLOs end up putting CLOs in a vice. After a wave of Covid-19 induced downgrades in the corporate debt market, many CLOs triggered their OC tests in April due to exceeding their CCC buckets. Now, CLO managers needed to dump the CCCs while purchasing higher rated debt. In part because of the federal reserve’s actions to prop up the corporate debt market, and in part because every CLO needed to purchase higher grade debt, the price of BBs went up and the price of CCCs went down, which made attempts to cure the OC test by selling CCCs and buying BBs much more expensive.
The older a CLO, the less flexibility it has to cure its OC test failures. CLOs go through a lifecycle with several distinct stages. First, there’s a period where they acquire all of the debt that forms their collateral. Then, CLOs have a (2-4 year) reinvestment period where they make payments to their investors while also actively trading their portfolio of collateral. Finally, CLOs have an amortization period where the portfolio is liquidated, investors are paid off, and the CLO is slowly winded down. In general, the more time left in the reinvestment period, the more flexibility CLOs have to trade for new securities and thus avoid or cure their OC tests.
CLOs actually have a bit of flexibility in determining the length of the reinvestment period. At the end of the reinvestment period, the holders of the equity tranche (i.e. the lowest level of the waterfall) have three call options: they can begin amortization, they can “reset” the CLO, or they can refinance the CLO. A reset just means that the length of the reinvestment period is extended for additional years. Refinancing means that all of the tranches are reissued; because equity gets paid after all of the other tranches, refinancing is an opportunity to find investors who will accept less interest payments which leaves more money for the equity tranches.
All of the flexibility CLOs had was not enough to stop a sell off in March. (The graph below displays the yields of different tranches, if the yields go up then that means the prices went down). The sell off was particularly extreme in the lower tranches because these tranches are the most at risk of taking losses due to OC test failures. However, the sell off has since largely subsided. The question now is whether the asset class will suffer any sort of long term damage.
In the fall, there will likely be a wave of bankruptcies and defaults on corporate debt. CLO managers will face a choice: do they sell their debt early, or do they hold on in the hopes of getting their money back in bankruptcy proceedings. Managers that hold on to their debt may be in for a rude surprise. Hedge funds and distressed investors figured out that rules that protect CLO investors leave them vulnerable to certain types of bankruptcy deals, such as when a company converts its debt to equity (CLOs are designed to hold debt not equity). Distressed investors are well aware of these restrictions and have structured the terms of bankruptcies like Acosta to be particularly unfavorable for CLOs. CLOs may have to write off large swaths of their portfolios as bankruptcies pile up. (CLOs would have had to write off bankrupted loans anyway because they can’t hold bankrupt debt at par, which would have caused more problems with their OC tests.)
The winners of CLOs’ Covid struggles are distressed investors who are going to have a lot of opportunities to purchase bankrupt corporate debt for cheap. In some cases like Acosta, distressed investors will be able to force CLOs to give up their debt for pennies on the dollar. In other cases, CLOs will further liquidate their holdings to try to cure their OC tests and distressed investors will still have opportunities to pick up debt. CLOs themselves may become distressed investors. One of the more interesting tweaks to the CLO model are “enhanced” CLOs, which are just CLOs with a much larger CCC bucket (as much as 50%). It’s a pretty simple solution; if the CCC bucket is triggering OC tests, just increase the size of the bucket so CLOs can hold more poorly performing debt. For existing CLOs, it’s too late to become an enhanced CLO: their documents are already agreed upon and it’s too late to change them. But in some cases, CLOs will become distressed investors and profit from the dynamics that will tear other CLOs apart.
There’s a bit of controversy about who would ultimately bear the pain if CLOs collapse. In a widely criticized Atlantic piece, Frank Partnoy (well known for his writings about derivatives and the 2008 crisis) described CLOs as 2020’s version of 2008’s CDOs, arguing that CLOs could place systemically important institutions (i.e. banks) at risk. One of the stronger rebuttals of Partnoy’s piece came from Nathan Tankus, who argued that banks’s holdings of CLOs weren’t that large, that banks tend to hold the AAA tranches which are at the top of the waterfall and relatively save, and that there isn’t evidence that banks are taking on the same forms of synthetic leverage that caused CDOs to blow up in 2008. On balance, Tankus is probably right about the risk to banks. The AAA tranches are probably going to be safe.
However, everything below the AAA tranches (i.e. the mezzanine) is at risk, which will have consequences for important elements of the economy. Lately, insurance companies have been investing in the mezzanine as they search for higher interest rates in a low rate environment. If CLOs fail their OC tests, the payments to lower tranches get diverted to the AAA tranche, which protects the banks but puts everyone else at risk. There’s a decent chance that mezzanine holders may not get all of their principal back, let alone interest payments. Even if mezzanine holders still get paid, they may be forced to sell anyway if their CLO tranches get downgraded and their internal risk management rules mandate that they sell the tranches. Either way, the market for mezzanine tranches will likely be severely depressed.
If no one buys the mezzanine tranches at reasonable prices, then it’s unlikely that any new CLOs will be issued or that existing CLOs can refinance. This will have huge consequences for private equity firms’ ability to conduct leveraged buyouts. It’s difficult to feel sympathy for private equity. But private equity owns or has owned a lot of companies that have a lot of leveraged debt, all of that leveraged debt will eventually need to be refinanced, and removing the largest source of financing for leveraged loans is going to have consequences for the broader economy. In the long run, less leveraged buyouts may be a good thing. In the short term, the loss of CLOs may cause a lot of pain for a lot of different businesses, making any recovery from a Covid recession much more difficult.
It’s going to be interesting watching this all play out in the coming months. Subscribe to stay up to date!