Guide to CLOs

9. How are CLOs priced?

How are CLOs priced? It’s simple. The price of a CLO depends on what someone is prepared to pay for it. As in any other market, the balance between supply and demand determines the price.
Understanding the value
But it isn’t quite this straightforward. Those who buy a CLO usually want to know that the price they are prepared to pay reflects its true value – the return they can reasonably expect to receive in future. As with any investment, a CLO bond or equity can be thought of as a stream of payments. To know the value of the bond you need to figure out what all the cashflows will be. Then you need to discount them back to today’s value using an appropriate discount rate (which, among other things, takes account of the probability of receiving those cashflows).
Doing this with a corporate bond is hard enough, since the investor needs to make assumptions about how likely the bond is to default and its recovery rate if default happens. To do this with a CLO debt security is harder still. The level of the note’s coupons may be predictable, but the probability of receiving those interest payments and the size of the final principal payment (if any) depend on the level and the timing of defaults in the portfolio.
The timing of the final cashflow is particularly hard to estimate. The weighted average life of a CLO note is determined by the speed at which the CLO pays down its debt, by the number and timing of defaults and by when the CLO equity investors choose to exercise their call option.
A widely used approach to CLO analysis is to “run the cashflows”. This is an exercise in which the would-be investor looks at the rules governing the CLO’s payments, such as the order of the waterfalls and the trigger levels of the overcollateralisation and interest coverage tests, and then projects all the payments that the CLO will make to each of its tranches. This forecast requires the investor to make various assumptions about what happens to the CLO portfolio (see Tweaking the variables, below).
In the early days of the CLO market, investors made these forecasts using spreadsheets they created. Such spreadsheets were cumbersome to manage and prone to errors. However, there are now many software tools which allow investors to run forecasts, such as those from Intex or Moody’s Analytics.
However the calculation is done, the cashflow forecast will depend on the rules of the CLO having been interpreted correctly. The forecast will also vary greatly according to which assumptions the analyst uses when running the model.

A weigh in the life

Weighted average life is a key concept in CLO valuation.
The weighted average life of any bond is the average amount of time until the principal is repaid. The principal may be repaid in full on one future date. Or it may be paid gradually through amortisation.
In the case of a CLO, a note may be repaid as a result of the CLO winding down, either because it has reached the end of its reinvestment period or because it is failing coverage tests. Alternatively, the note may be repaid in full when the equity investor calls the deal. An investor’s estimate of weighted average life needs to take account of both of these possible routes to repayment.
Pricing comparisons
Forecasting weighted average life is important for another reason. When people ask how a CLO is priced, what they often want to know is how they can make sense of the price an investor is prepared to pay and how that price compares with the price of other CLOs and other, non-CLO, investments.
Share buyers know that a company’s share price means nothing in isolation. To compare it with the shares of any other company, the investor needs to look at how the share price relates to the company’s earnings and balance sheet.
Similarly, CLO prices mean little in isolation. In order to make comparisons between CLOs, investors need to convert prices into yield terms. Because CLOs are usually floating rate instruments, the standard measure of spread is a discount margin.
Converting prices into discount margins and vice versa is relatively simple bond mathematics (see Making the numbers work on the previous pages) provided the investor knows the weighted average life of the bond. But, as mentioned before, this requires cashflow forecasting.
How pricing in the primary market changes over time: US CLO triple A 5Y benchmark vs European CLO triple A 4Y+ benchmark (bp)
Pinch to zoom in
Tweaking the variables
The main assumptions that the CLO analyst needs to make are about defaults, recoveries, prepayments, reinvestments and the extent to which the manager is able to make investments after the end of the reinvestment period.
The number and severity of defaults in the portfolio (in other words, the default rate and the recovery rate) have a big effect on the projected returns on all the tranches. So too do the timing of those defaults. Most obviously, higher defaults mean there is less principal and interest income available to equity investors. Ultimately, defaults cause losses to the debt tranches. Perhaps less obviously, defaults can shorten the expected life of debt tranches because they cause the CLO to fail one or more overcollateralisation tests, leading to early wind-down of the CLO and amortisation of the notes.

This is where CLO cashflow modelling becomes very complex. Each development in the life of a CLO – a default, a test failure, a test coming back into compliance – has knock-on effects. Many different scenarios are possible for the CLO’s life story. In quant terms, CLO modelling is path dependent, a phenomenon which makes calculations difficult and time consuming, even for the most powerful computers.
One particular level of defaults could trip the double B OC test, for example, causing the triple B notes to get paid back earlier than expected. A slightly higher default rate (or lower recovery rate) could cause both the double B and triple B test to fail, cutting off payments to the triple B notes for a time. The outcome for triple B investors would be hugely different in these two scenarios.
Many investors project CLO cashflows using a flat assumption about defaults. That is, they assume that defaults will be at the same level each year. A flat 2% default rate is a traditional yardstick for CLO modelling. Clearly, no investor believes that a 2% a year default rate is actually going to happen. But it is an assumption which makes the modelling process more manageable and helps an investor to compare one CLO to another.

Making the numbers work

Numerical comparison of bonds in the CLO market is typically done using the concept of discount margin (DM).
The DM is the spread (typically expressed in basis points) over a risk free discount curve like Euribor which makes the combined “Euribor plus spread discount curve” discount the bond’s cashflows back to the initial purchase outflow. Formally, DM is the solution to:
co = ∑ni=1 ci di
where co is price times par amount (the initial purchase outflow); ci are cashflows (interest and principal) at time i; and di is the discount factor at time i.
Forward discount factor
di + 1
1 + [forward Euribor from time i to i+1] + DM
As most CLO debt carries a floating rate of interest, DM can be used to compare prices across bonds, as it adjusts for maturity and coupon differences. An important point is that the DM of a bond priced at par (and with no accrued interest) is the spread on that bond.
As the formula above is complex, there are back-of-envelope calculations to compute DMs. For debt where impairment or write-down is a remote risk, we have the classic price-yield-duration formula:
% change in price = duration × –change in yield (in %) Recast for DMs, we have: % change in price = duration × –change in DM (in %)
Modified duration (as opposed to Macaulay duration) is the correct measure to use here. The duration for a zero coupon bond is its average life. For coupon bonds, the duration is less than the weighted average life (WAL).
This leads to an interesting second approximation, where we use the WAL as a proxy for the duration with a minor adjustment. The following table gives an example of the adjustment:
Pinch to zoom in
So finally, we have:
% change in price = WAL × adjustment × –change in DM (in %)
Note that while the change in DM is an absolute percentage change (for example, a change from 2% to 3% is a change of 1%), the price change is a relative percentage. In other words, a change of price from 50% to 55% is a 10% (5/50%) change.
Bringing all this together in an example, let’s say a bond has a coupon of six-month Euribor plus 1.45% and a WAL of five years.
Price 100%
→ DM of 145 basis points
DM of 135 basis points
→ –0.1% change in DM → % change in price ≈ 5 x 0.96 x + 0.10 → Price ≈ 100.48%
The reality of the loan market is that most loans do not continue until their maturity date. Because loan borrowers conventionally have a right to pay back the loan with little or no penalty, companies tend to refinance them (pay them back and take on a new loan or bond) when they can borrow on more favourable terms than those provided by the existing debt. Borrowers also like to avoid letting the loans run close to maturity since this could leave them in the position of being forced to borrow at a punitive rate. For both reasons, companies tend to repay their loans some time well before they reach maturity.
One class of loans, known as the revolver and A tranche, do commonly receive some principal back well ahead of their maturity date. These loans, sometimes known as the pro-rata tranches, are usually bought by banks rather than CLOs and other institutional investors. CLOs typically own only a small amount of these loans (usually less than 5% of their portfolios). They therefore often receive only a small amount of loan repayments during the life of the CLO from these kinds of assets. However, most of the loans owned by CLOs are the B and C loans, which have a longer maturity profile (but the same seniority) as the pro-rata tranches. These “institutional” classes of loan do not usually amortise until close to their legal maturity. Typically, the company that issues the loan will either choose to pay it back well before maturity or will default.
The CLO analyst needs to make an assumption about the amount of loans in the portfolio that will be refinanced every year – in other words, the speed of prepayments. The rate of prepayments partly reflects the prevailing borrowing rate for high yield corporate borrowers. When they can borrow cheaply, companies will tend to prepay faster. If many companies are facing difficulties or if borrowing costs are high, they are less likely to repay their loans.
The CLO analyst could forecast prepayment levels by simply projecting today’s rate of repayments forward. Alternatively, he or she could look at average historical repayment rates. Often, as with defaults, analysts use a flat prepayment rate for the life of the deal.
A related assumption is about reinvestment spreads. If, for example, 20% of the portfolio prepays each year, those repayments will need to be reinvested in new assets. The analyst needs to make an assumption about what yields or spreads those new loans will provide.
A typical approach is to assume that future proceeds will be invested at some level of spreads similar to those prevailing today.
Many CLOs, especially older deals, allow the manager some leeway to make investments after the end of the reinvestment period. So, the analyst needs to make an assumption about what proportion of repayments the manager is able to deploy into new assets post-reinvestment.
A common simplistic assumption is that the manager will be able to re-invest all the prepayment proceeds while the CLO is within its reinvestment period and 50% of them after the reinvestment end date.
Modelling CLOs using flat, across-the-board assumptions has the benefits of speed and simplicity. But many investors prefer to use what they regard as more realistic projections of defaults, recovery rates, prepayment rates and so on. Using their own assumptions may give them insights into whether a particular CLO bond is underpriced or overpriced compared to other assets.
Many investors run cashflows using something other than a flat projection of defaults. A popular approach is to assume that there will be a spike in defaults at some point in the life of the deal. This comes a little closer to the real world, where defaults tend to remain low for long periods of time and then jump higher during recessions.
Similarly, investors may choose to model the cashflows using the assumption that prepayment rates and prevailing loan spreads will fall and rise depending on where we are in the economic cycle.
Two other areas of modelling require a different kind of analysis.

Hitting the floor

One thing that has made CLO valuation more complicated since 2008 has been the appearance of a phenomenon known as the Libor floor (named for the old reference rate used for loans issued in dollars). This has been a feature of many loans, particularly in the US, since Libor fell to below 1% in early 2009.
As described in chapter 1: What is a CLO?, loans pay an interest rate which is composed of a reference rate of interest, and a spread reflecting the individual borrower’s creditworthiness. If a loan pays quarterly and has a spread of 200 basis points (or 2%) over the three-month reference rate, this means that on any given payment date it will pay interest calculated on the loan’s outstanding principal amount multiplied by one quarter of three-month refence rate plus one quarter of 2%.
Libor floors change this calculation by simply stating that the “Libor” part of the calculation will never fall below a certain amount. So, a loan with a 2% reference rate floor will pay 2% plus the spread at any time whenever the reference rate is below 2%. If the refence rate goes back above 2% during the lifetime of the loan, the reference rate will revert to the current rate.
Libor floors are good for CLO investors in the sense that the loans pay more than they would without the floor. But they mean that investors need to form an opinion about if and when interest rates (of which Libor is an example) will rise in order to make sense of the price of the CLO.
Beyond standard assumptions
One is post-reinvestment trading. The investor needs to take a view on the extent that the CLO manager is allowed by its documentation to buy new assets after the end of the reinvestment period. This requires close reading of the CLO indenture. The investor also needs to guess how aggressively the CLO manager will choose to interpret the rules of the deal. Both of these factors can make a big difference to the weighted average life of a particular CLO tranche.
Standard CLO modelling assumes that the CLO will remain outstanding until it receives and deploys enough principal to pay down all its notes. In practice this is unlikely to happen. Instead, equity investors will choose to call the deal well before the most junior notes start to amortise.
Timing the call
It is difficult to forecast the date of a CLO call, but it can have an even greater impact on the CLO note’s weighted average life than any other factor.

If they are acting rationally, equity investors will call the CLO at the point at which they could raise a new CLO with cheaper funding, taking into account the transaction costs of issuing a new CLO (such as arranger and rating agency fees). As a CLO starts to amortise, paying down its most senior (lowest coupon) tranches first, its funding costs rise. In order to model the call date, CLO investors need to forecast not only the performance of the CLO but also future CLO spreads.
CLO equity investors do not always act rationally. CLOs commonly remain outstanding beyond the point at which they could be refinanced more cheaply. The most common reason for this is that the equity is held by a large number of different investors and it proves impossible to muster enough votes to trigger a call.
Therefore, CLO analysts need to gain an understanding of who the equity investors are in order to forecast the timing of a call. This information is not usually available in any public source. But investors may be able to glean some information on who the holders are from contacts who were involved in selling the deal in the first place, or who have traded the bonds in the secondary market.
Pricing in practice
Perhaps the best way to know what a CLO is worth is to look at the price and spread of comparable CLOs. There are two markets that provide price signals: the primary market (for new CLOs) and the secondary market (the trading of existing CLOs).
Pricing in the two markets is closely connected. But there are differences. Partly, they reflect the fact that the CLOs being issued today aren’t identical to the deals that change hands in the secondary market (see chapter 10: How the secondary market works). Partly, the variations result from the different dynamics of the two markets, and the different ways that investors use these markets.
In the primary market, CLOs may price at par or at a discount, depending mainly on the preference of investors. Often pricing at a discount is a way for debt investors to reduce the risk of the CLO being called early.
For example, if the market price of a new CLO note is 600 basis points over the risk-free rate, the arranger and investors may choose to have the bond pay a 550 basis point coupon but price at less than par. This amounts to the same thing. But it reduces the incentive for the equity investors to call the CLO. It only becomes economical for them to call it once spreads fall to 550 basis points rather than when they are at 600 basis points.
Pricing equity
Valuing equity presents similar but bigger challenges for investors than valuing CLO debt. Whereas debt has a promised stream of payments and its cashflows will never be greater than those that are promised, equity, on the other hand, has variable payments which depend on the performance and spreads on all the individual loans in the CLO portfolio.
So, whereas CLO debt is usually priced in terms of its spread over a low-risk investment such as US dollar Libor or Euribor, this approach does not work for equity. Rather, CLO equity investors try to estimate the absolute return they will receive over the life of their investment, just as investors in a hedge fund or private equity fund would.
Future returns are converted into today’s money using a discount rate which is itself the investment’s expected return. In other words, the investor conventionally assumes that any money received is reinvested in an instrument that provides an identical return. This calculation is called an internal rate of return or IRR.
Besides calculating the expected IRR of the investment, CLO equity investors have a second way to think about valuation. The value of the equity tranche should be the total market value of all the CLO’s assets minus the total market value of its debt (that is, all the tranches above equity in the capital structure). This is referred to as the net asset value of the CLO or sometimes as the market value overcollateralisation.
There are various practical problems in calculating a CLO’s net asset value. For one thing, there may be no visible market price for the CLO’s debt. On the other side of the balance sheet, the investor may well feel confident about the value of most of the CLO’s assets. But there are usually at least a handful of assets in each CLO that are illiquid or obscure and for which it is hard to estimate a price.
One of the biggest uncertainties for CLO equity investors is that they do not know what future assets the manager will buy. When they are marketing the deal to prospective investors, the manager and arranger will often show a list of loans that they expect to buy. However, savvy investors know that this modelled portfolio can vary markedly from the one which emerges in practice.
Another big uncertainty for equity investors, as for debt holders, is that they do not know how long the CLO will last. Although equity investors as a class control the call option, an individual equity investor does not, unless it owns a majority of the tranche.
Therefore, equity investors need to make assumptions about when the CLO will be called. The identity of other investors in the equity tranche could be critical to the timing of a call. Their willingness to call a CLO may depend on the kind of institution they are, their investment mandate, and the price at which they bought the position. Therefore analysing the equity investor base is an important part of valuing the equity.