7. Debt tranches

and ratings

Typically, CLOs are around nine times levered, and therefore a large majority of all CLO liabilities are debt tranches.

When rating corporate bonds or loans, rating agencies look at measures such as the company’s level of indebtedness, its cashflows, its access to finance and the dynamics affecting its industry. They then apply some level of judgment to come up with an appropriate rating. In principle, rating a CLO should be more straightforward than rating a corporate bond. Assuming that the rating agency already rates all the underlying assets, it simply needs to figure out how the risk of the portfolio is shared between the different tranches.

However, reality is not so simple. For one thing, the rating agency may not rate all the corporate borrowers in the portfolio. For another, the differences in the waterfalls and other rules from one deal to another make it difficult to rate CLOs using a simple, formulaic approach.

And, more fundamentally, nobody knows (until it happens) how defaults in a portfolio feed through to losses for different CLO tranches. Most CLOs throughout the history of the market have survived without inflicting any losses on debt investors. So CLO default is an event with low probability. That makes it hard to predict, and highly dependent on the assumptions used.

Each of the rating agencies involved in rating CLOs has a different methodology for coming up with its ratings. The main rating agencies involved are Fitch Ratings, Moody’s Investors Service and S&P Global Ratings. Other rating agencies such as Kroll and DBRS Morningstar are involved to a lesser extent, mainly by giving private ratings for CLOs in specific circumstances.

The methodologies used by the rating agencies to rate CLOs have evolved over time. All publish details of their methodology, but not in such detail that an outsider could be certain of guessing what their rating on a particular tranche would be. This is deliberate. Rating agencies fear that if they make their rating methodology public as a series of mathematical rules, then the investment banks that structure CLOs will find ways to “arb the ratings”, that is, include assets or features which boost the CLO’s ratings while adding to its real risk.

Especially following their experience of mis-rating structured products such as subprime mortgage-backed securities in the lead-up to the financial crisis of 2007 and 2008, rating agencies like to retain an element of discretion in rating CLOs.

In many respects, Fitch Ratings and S&P Global Ratings take a similar approach to ratings, while Moody’s is a little different from the other two. This is true, for example, of what the agencies say a rating actually means. For Fitch and S&P, the rating of any debt is their opinion of its likelihood of suffering any kind of loss. This is referred to as a first-dollar-of-loss approach. Moody’s, by contrast rates to what it terms an expected loss. This is an opinion of the magnitude of loss that any debt can be expected to sustain during its life. The difference sounds small. But it could be important for a bond that has a high likelihood of suffering a default but a very high expected recovery rate in the event of such a default. The Moody’s rating for such a bond should be higher than the S&P or Fitch rating, assuming that they are all in agreement about the bond’s default and recovery risk.

Similarly, Moody’s takes a rather different approach to rating CLOs than the other two agencies. Moody’s general approach is first to take the proposed CLO portfolio and convert it into a series of numerical scores. One is a rating factor (weighted average ratings factor or warf) based on ratings of the underlying assets, which represents the average default risk of the portfolio (ranging from 1 for Aaa risk to 10,000 for anything below Caa3). Moody’s also comes up with a single number to address correlation within the portfolio, the diversity score.

Moody’s then puts this version of the portfolio through a model which takes account of the cashflow rules of the CLO. The model gives an expected loss calculation for each tranche under a number of different scenarios. Moody’s then compares this expected loss with a target expected loss rate which it believes is appropriate for each rating over the covenanted life of the deal. For example, a CLO tranche needs to have an expected loss over its life of below 0.0055% in order to gain a rating of Aaa. The agency also imposes additional haircuts, or stress factors, to take account of any other risks it identifies in the deal.

Both Fitch and S&P use a Monte Carlo modelling approach to rating CLOs. This is a model that uses heavy computing power to calculate the result of many different scenarios for the portfolio at each rating level, such as different levels and patterns of default, different interest rate projections and different timings of defaults.

From this, the agencies calculate how many defaults it would take for each tranche to suffer a loss. They compare this to what S&P and Fitch call a break-even default rate for each tranche. These are a number of defaults compatible with a desired rating that a given tranche should be able to sustain. In other words, a tranche rated AAA would need to be able to withstand a much greater number of defaults than one rated BB, for example.

For their efforts, rating agencies generally take an upfront fee in the order of seven basis points (0.07%) of the size of all the tranches they rate. In pre-crisis CLOs it was common for all tranches above the equity to carry two or more ratings. Since the crisis, it has become rare to see a tranche with three ratings, and many carry only one. Even so, a $500 million CLO with dual ratings on at least some of its debt can easily pay $500,000 to rating agencies in upfront fees.

Rating agency fees are higher when there are shadow ratings (see box below).

Of course, many investors do want to keep a close eye on how their CLOs are performing. CLOs differ from investments such as hedge funds in that they provide very detailed reports to all their investors. Generally, all classes of investors, including debt and equity, receive the same regular reports. They are popularly known as “trustee reports” but in reality are prepared by the portfolio administrator (which is often the same company as the trustee). In addition to these number-based reports required under the rules of the CLO, the manager may choose to send out its own more descriptive commentary on the performance of the CLO and the state of the market.

There are two main types of report. The most frequent is a monthly report which lists the CLO’s current assets, gives the size and price of all purchases and sales in the month, details its cash balances and reports whether it is in compliance with the various performance tests (see chapter 6: Understanding the tests).

Investors receive an additional report at each payment date (either quarterly or every six months depending on the payment frequency of the CLO). This is known as the note valuation report, and it shows how much cash the CLO has received in that period and details all the payments the CLO has made. This report shows how the CLO’s interest waterfall and principal waterfall are working in practice (see chapter 5: Understanding the cashflows), including how much money is left over to pay equity investors after all the other classes of investors have received their payments.

CLO notes are generally issued in a form that allows investors to switch between two different types of note known as 144A and Regulation S (Reg S).

Broadly, US investors buy 144A notes and non-US investors buy Reg S notes. A 144A is a private placement, meaning that the issuer is exempt from the onerous marketing stipulations of US securities laws. Reg S bonds are bonds issued to international investors (and therefore also exempt from the same laws).

In practical terms, what this means is that 144A bonds are cleared through the US DTCC and carry a cusip identifier number. Reg S bonds, meanwhile, clear through either Clearstream or Euroclear and are identified with an isin identifier number.

One may imagine that having different formats would impede trading of CLOs in the secondary market. But this is not the case. If a holder of a 144A CLO note wants to sell to an investor who requires a Reg S note, or vice versa, the format of the note is simply changed by a clearing house.

The legal format of a CLO investment makes no difference to the payment that investors receive. However, there are several kinds of CLO debt that have unconventional payment profiles.