CHAPTER 23
Collateralized Loan Obligations (CLOs)
Demystifying a Versatile Asset Class
Mendel Starkman
WHAT IS A COLLATERALIZED LOAN OBLIGATION (CLO)?
CLOs are a type of investment fund. Similar to a mutual fund, a CLO is a professionally managed fund that invests in a pool of financial assets, specifically corporate loans. Unlike a mutual fund, however, a CLO is a structured fund. Instead of each investor owning a share with the same risk and return as all other investors, the CLO offers various investment tiers, each with a different risk‐and‐return profile. This chapter discusses the characteristics of a CLO, including its structure, underlying loan pool, and collateral manager.
At its core, the fundamental purpose of a CLO is to provide an efficient source of financing to below‐investment‐grade corporate borrowers. Corporations require funding for various purposes, including general business and growth opportunities, capital planning, or to finance acquisition activity. CLOs provide a meaningful source of funds from which these companies can borrow.
On the other side of the equation are institutional investors, who find that CLO tranches provide an attractive return for their level of risk. Investors may not be willing or able to lend directly to speculative‐grade companies. Instead, a CLO provides investors with a professionally managed, well‐diversified pool of such loans, with the added protection of loss shock absorbers that are inherent to the CLO structure.
The CLO market connects the equation. Investors who have available funds are able to provide them to the CLO, which in turn lends them to companies that want to borrow. The innovation of CLO technology enables such an interaction, opening a sizeable source of funds to borrowers while crafting a risk‐and‐return profile that is attractive to investors.
CLOs have a relatively long and impressive track record. The CLO market began to emerge in the late 1990s and early 2000s. Issuance picked up significantly from 2005 through 2007, paused for several years during and after the financial crisis, and then roared back to life in 2012. Moody's Investors Service calculated that only 0.8% of the CLO tranches that they have rated since 1996 have incurred (or were expected to incur) principal losses.1 While certain improvements have been made to CLOs due to lessons learned during the crisis, their resilient performance through market cycles provides meaningful perspective for potential future performance (Figure 23.1).
Figure 23.1: Annual issuance of U.S. broadly syndicated loan CLOs, 1996–2015
Source: Wells Fargo
It should be noted that CLOs come in several different flavors. This chapter focuses on U.S. cash flow CLOs that are backed by broadly syndicated leveraged loans (these transactions are also referred to as arbitrage CLOs). However, the concepts described still broadly apply to other types of CLOs, which include:
- Balance sheet CLOs—CLOs constructed as a means for a financial institution to transfer an existing loan portfolio into a new off‐balance‐sheet entity.
- European CLOs—CLOs managed by European collateral managers and primarily backed by loans to European companies.
- Middle‐market CLOs—CLOs backed by loans to smaller companies (the European equivalent of a middle‐market CLO is referred to as a small‐ and medium‐enterprise (SME) CLO).
STRUCTURES AND STRUDELS: AN INTRODUCTION TO STRUCTURED FINANCE
CLOs are card‐carrying members of the structured finance family. But before examining the specific characteristics of CLOs, it is worth beginning with a brief introduction to structured finance in general (which we will slant toward an application to CLOs). To explain the basic concept of structured credit, one can pay a visit to their local bakery shop.
A typical investment fund, such as a mutual fund, can be thought of as a pie. Each investor owns a slice of the pie proportional to his investment. For example, let's say a fund had a total value of $100, and five investors each invested $20 into the fund. Each of the five investors in this example owns a proportionate 20% slice of the fund. Should the fund experience losses, the entire pie would shrink, and all investors would sustain a loss proportional to their investment. In our example, if the fund value declined by $20, and now had a remaining value of only $80, any one of the investors would still own a 20% slice of the fund. But since the value of the entire fund has decreased, an original investment of $20 would now be worth only $16. Each investor sustained an equivalent proportional loss in their investment.
Continuing with the analogy, structured finance can be thought of as a layer cake, in which each investor owns a different horizontal layer of the cake. For example, imagine that a cake had five layers, in which each layer represented 20% of the height of the cake. If someone were to take a knife and slice off the entire bottom‐most layer of the cake, 80% of the cake would remain perfectly intact. However, now it would be a four‐layer cake instead of a five‐layer cake.
This analogy illustrates the broad concept behind structured finance. Similar to the layers in the cake, imagine an investment fund that is worth $100, but in which five investors each invested $20 into different, equally sized layers of the fund. If the fund were to sustain $20 of losses, it would be comparable to the bottom‐most layer being sliced off the cake. The investor in the bottom‐most layer of the fund would absorb the full $20 of loss, thereby wiping out his entire investment. In contrast, the four other investors would not sustain any losses at all. (See Figure 23.2.)
Figure 23.2: Illustration of investment fund ownership and loss allocation
The key difference is that the pie‐like nature of typical investment funds attributes losses proportionately, while the layer‐cake‐like nature of structured finance attributes losses sequentially. Specifically, losses in structured finance are allocated from the bottom up, otherwise referred to as being in reverse‐sequential order.
Why, then, would anyone want to invest in the bottom‐most layer of a structured fund? Wouldn't everyone want to invest in a higher layer and have the protective buffer that the lower layers provide against losses?
The answer relates to the classic balance of risk and reward. Clearly, the lowest layer of a structured fund is the most risky, because it is the first to absorb any losses. But correspondingly, it is also designed to yield the most. On the opposite extreme, the highest layer of the fund is the least risky, because it does not absorb any losses until the value of all of the other layers has been exhausted. It is therefore correspondingly designed to yield the least. The layers in between, with varying degrees of risk, are issued with yields that are appropriate for their respective levels of risk.
In an actual structured finance transaction, each investment layer is referred to as a tranche (based on an Old French word meaning “slice”). The existence of junior tranches provides a greater buffer to protect the senior tranches from loss. Another way of saying this is that the more senior the tranche, the greater the subordination that exists beneath it to protect against losses. The junior‐most tranche has no subordination (i.e., no tranches exist beneath it to buffer from losses) while each more‐senior tranche is protected by progressively greater amounts of subordination.
BON APPETIT: REENGINEERING FOR RISK APPETITE
One aspect of structured finance that is often misunderstood is how a structure can include tranches that are rated investment grade (including a large proportion with the highest possible rating of AAA), when in the case of a CLO all of the assets that act as collateral are speculative‐grade loans. Don't the ratings of the tranches need to more closely reflect the risk of the asset pool that supports them?
The answer relates to a fundamental difference between the risk of the loans and the tranches. The ratings of the loans in the pool consider the risk, as measured by probability of loss of each individual corporate loan. Because the loans are speculative‐grade credits, the probability of loss for any given loan is relatively high. Recall, however, that the tranches have an added buffer to protect against losses, due to the subordination of more junior tranches. Consequently, a tranche's risk is not whether there will be any losses in the loan pool, but rather, whether losses on the loan pool will exceed the tranche's entire loss buffer. In this way, a tranche's risk largely becomes a matter of statistics. What is the likelihood that losses in the pool, net of any recovery value, will erode all of a given tranche's available subordination and result in a loss to the tranche?
This can be illustrated with a simplified example. Let's assume that the senior‐most tranche of a CLO has 35% subordination. Let's further assume that after a loan defaults, it recovers, on average, 50 cents on the dollar. In order to erode all of the subordination to the senior‐most tranche, 70% of the entire loan pool would need to default (and then recover at 50%) before the 35% of subordination is eroded and the senior‐most tranche begins to sustain any losses. Based on historical data, the statistical likelihood that 70% of a diversified corporate loan pool will default is very remote, indicating that an investment in this senior tranche is expected to be much less risky than a direct investment in the underlying loan pool.
Due to the alchemy of financial engineering, the structure is able to issue tranches whose risk is significantly lower than the risk of the collateral pool itself. In essence, it is possible to issue tranches that have AAA‐grade risk, even though they are backed by a pool of single‐B‐rated loans, because there is a very low probability that the losses on those loans will exceed all of the tranche's available subordination. For this reason, the tranches are able to achieve credit ratings that can be meaningfully higher than the credit ratings of the underlying loans.
CHASING WATERFALLS
Until now, our focus has been on the inherent seniority order of a structured credit product, and the protection afforded to each tranche through the subordination of more‐junior tranches. Underlying the structure, however, is a portfolio of assets that serve as collateral. Structured products are each collateralized by a single type of asset class, but that asset class differs based on the type of securitization. Structured finance collateral pools can range from residential or commercial mortgages, to credit card receivables, auto loans, or student loans, to other more esoteric forms of loans (for example, aircraft leases or railcar receivables). For CLOs, the pool generally consists of loans to relatively large companies, with a single CLO typically investing in loans issued by 100 to 300 different obligors. Some examples of household names that have had loans represented in CLO pools include: American Airlines, Chrysler, Dell Computers, Hilton Hotels, MGM Resorts, PetSmart, and Sea World.
With any fixed‐income product, the lender (or investor) receives payments of interest on a regular basis, and payments of principal that are based either on a preset payment schedule or on the instrument's maturity date (similar to a mortgage or a student loan). CLOs are no different. They invest in a pool of loans, and thereby play the role of lender to a range of corporate borrowers. In turn, these corporate obligors make periodic payments of interest and principal back to the CLO.
A CLO's payment structure follows a sequential tranche system, but as opposed to loss allocation (which is implemented from the bottom up), payments follow a top‐down sequential order. The senior‐most tranche receives payments first, followed by the second‐priority tranche, and so on. If on any payment date, available funds are insufficient to pay all of the tranches, it is the junior‐most tranche that is left holding the bag. This sequential prioritization of the tranche payments is referred to as the payment waterfall.
The CLO maintains a ledger that separately tracks interest and principal receipts. Interest payments received from the loan pool are used for the CLO's interest waterfall to sequentially pay the periodic interest due to each of the tranches. Principal payments received are used for the CLO's principal waterfall to sequentially repay the principal balance of each of the tranches (Figure 23.3). In normal operation, principal and interest waterfalls are almost always kept separate. (However, there can be circumstances, particularly in times of distress, when interest proceeds would be used to cover shortfalls in principal payments, or vice versa.)
| Priority | Payable to: | Amount |
| Available Interest Proceeds: | 5,783,840.36 | |
| (A) | to the payment of accrued and unpaid taxes and governmental fees: | — |
| (B) | to the payment of accrued and unpaid Administrative Expenses: | — |
| (B)(i)(a) | to the Bank under each of the transaction documents: | 21,097.93 |
| (B)(i)(b) | to the payment of any Petition Expenses | — |
| (B)(i)(c) | to Moody's and S&P for fees and expenses in connection with any rating of Rated Notes pro rata: | — |
| (B)(i)(d) | to the payment of all other Administrative Expenses as directed by the Collateral Manager on their respective amounts: | 50,893.00 |
| (B)(ii) | upon discretion of Collateral Manager, deposit to the Expense Reserve Account lesser of: OERS or OEEA: | — |
| (C) | to the payment of any accrued and unpaid, current and prior, Senior Collateral Management Fee: | 191,799.34 |
| (D) | to the payment of any Interest Rate Hedges, including any termination payments to a Hedge Counterparty: | — |
| (E)(i) | to the payment of accrued and unpaid interest on the Class A‐1a Senior Notes inc. Defaulted interest, pro rata: | 1,578,740.80 |
| (E)(ii) | to the payment of accrued and unpaid Interest on the Class A‐1b Senior Notes inc. Defaulted interest, pro rata: | 195,000.00 |
| (F) | to the payment of accrued and unpaid interest on the Class A‐2 Senior Notes, inc. Defaulted Interest: | 465,451.35 |
| (G) | if either Senior Coverage Test is not satisfied | — |
| (G)(i) | to the payment of principal of the Class A‐1a Notes (pro rata): | — |
| (G)(ii) | to the payment of principal of the Class A‐1b Notes (pro rata): | — |
| (G)(iii) | to the payment of principal of the Class A‐2b Notes: | — |
| (H)(i) | to the payment of acrrued interest on the Class B Mezzanine Notes…: | 370,771.66 |
| (H)(ii) | and inc. Defaulted Interest…: | — |
| (H)(iii) | and Interest on Deferred Interest: | — |
| (I) | in the event that any Class B Coverage Test is not satisfied: | — |
| (I)(i) | to the payment of principal of the Class A‐1a Notes (pro rata): | — |
| (I)(ii) | to the payment of principal of the Class A‐1b Notes (pro rata): | — |
| (I)(iii) | to the payment of principal of the Class A‐2b Notes: | — |
| (I)(iv) | to the payment of principal of the Class B Mezzanine Notes: | — |
| (J) | to the payment of any Class B Mezzanine Deferred Interest: | — |
| (K)(i) | to the payment of acrrued interest on the Class C Mezzanine Notes…: | 284,506.95 |
Figure 23.3: Interest waterfall excerpt from sample CLO payment date report
Source: U.S. Bank
THE CAPITAL STRUCTURE: THE INTERSECTION OF STRUCTURED AND CORPORATE FINANCE
Earlier, we illustrated the concept of subordination thorough the example of a layer cake with equally sized horizontal slices. In fact, the tranches of a structured fund are not equally sized. The senior‐most tranche is by far the thickest. In a modern CLO, it can comprise 60–65% of the total structure (meaning that 35–40% of subordination is provided the combined value of all the more junior tranches). Each of a CLO's other tranches typically represent 5–10% of the structure. Altogether, CLO structure will often have a total of 6 or 7 tranches.2
The tranches of a CLO are typically structured so that each one is designated with a different rating category by the major credit rating agencies. The rating for the senior‐most tranche is AAA, followed by the second‐priority tranche at AA, and so on. At the lower‐end of the capital structure, the rated tranches typically dip into below‐investment‐grade territory, with ratings as low as double‐B, or even single‐B. The junior‐most tranche, however, is known as the equity tranche and is left unrated.
CLO tranches are generally issued as floating‐rate notes, with the stated coupon rate being represented as a spread over three‐month LIBOR. The first‐priority tranche, being the least risky, earns the lowest spread over LIBOR. (Because the senior‐most tranche is the thickest tranche, it also has the greatest influence on the CLO's overall interest costs.) Each lower‐ranking tranche is progressively more risky and therefore earns a somewhat higher stated spread over LIBOR.
The only exception is the junior‐most tranche, which does not have a stated payment rate. Rather, the junior‐most tranche represents the equity of the CLO's capital structure and it earns the residual interest and principal proceeds that remain after all other tranches have received their payments. (Another term used to describe the residual cash flow is excess spread, which refers to the excess proceeds available within the CLO after all payments due to the more senior tranches.)
In a way, a CLO structure resembles a corporate balance sheet, in which assets equal liabilities plus shareholder's equity (Table 23.1). The corporate loans held by a CLO are its assets while the interest‐bearing tranches represent the debt and the junior‐most tranche represents the equity. A company uses its revenues to service the periodic interest payments on its outstanding debt, and the remaining proceeds belong to the shareholders (either through retained earnings or dividends). In the same way, the proceeds received from the CLO's loan pool are its revenues, from which payments are made to the debt tranches, with any remaining residual cash flow generally distributed to the CLO's equity tranche.
Table 23.1: Sample CLO Capital Structure
Sources: Bloomberg, Intex
| Galaxy XX CLO | |||||
| Manager: | PineBridge Investments, LLC | ||||
| Underwriter: | Goldman, Sachs & Co. | ||||
| Trustee: | U.S. Bank National Association | ||||
| Initial Ratings | |||||
| Security | Principal Amount ($) | Original Subordination | Coupon | Moody's | Fitch |
| Class A Notes | 352,000,000 | 36.0% | 3mL + 1.45% | Aaa | AAA |
| Class B Notes | 63,250,000 | 24.5% | 3mL + 1.95% | Aa2 | NR* |
| Class C Notes | 33,000,000 | 18.5% | 3mL + 2.60% | A2 | NR* |
| Class D Notes | 31,350,000 | 12.8% | 3mL + 3.50% | Baa3 | NR* |
| Class E Notes | 26,950,000 | 7.9% | 3mL + 5.50% | Ba3 | NR* |
| Subordinated Notes (Equity) | 48,950,000 | N/A | Residual | NR* | NR* |
| Total Initial Pool Balance: | 550,000,000 | ||||
| Pricing Date: | 5/19/15 | ||||
| Closing Date: | 6/25/15 | ||||
| End of Non‐Call Period: | 7/20/17 | ||||
| End of Reinvestment Period: | 7/20/19 | ||||
| Stated Maturity Date: | 7/20/27 | ||||
*NR indicates not rated.
BEING FRIENDLY: DEBT VERSUS EQUITY
Since the debt tranches of a CLO are only entitled to receive their stated interest rate (i.e., the tranche's spread over three‐month LIBOR) and the ultimate return of their principal, there is no potential for upside in payments. If the CLO performs well, then the debtholder will earn the quarterly coupon payments to which they are entitled and ultimately receive back their principal balance. If a CLO underperforms, the debt investor would have downside risk in the form of lost interest or principal payments.
A CLO's equity tranche, on the other hand, is not paid a stated coupon, but rather earns all of the cash flow that remains after the debt tranches are paid. Therefore, while the equity tranche has the greatest risk of payment variability and downside, it also has the potential for meaningful upside.
This dichotomy of risk profiles between debt and equity creates an implicit divergence of interests within the CLO's investor base. For example, a skew toward loans that are more risky and higher yielding would be of greater benefit to the equity investors, while not directly benefitting the debtholders.
When analyzing a CLO's structure and the investment style of the CLO manager, certain aspects may be considered more “debt‐friendly,” as they would maximize the protection of the debt tranches at the possible expense of extra yield for the equity tranche. Other aspects may be considered more “equity‐friendly,” as they would involve some additional degree of risk in an effort to achieve a somewhat higher return. The role of a CLO manager includes a perspective of both elements, ensuring that the interests of all investors are appropriately balanced throughout the life of the transaction. An investor needs to access the potential sources of risk and opportunity in each CLO and ensure that the transaction meets the intended risk profile of their investment.
CLO COVERAGE TESTS: TAKE YOUR PIK
Another noteworthy aspect of a CLO structure relates to the ongoing measurement of coverage tests. These are ratios for each debt tranche that measure the extent to which it is supported by the loan collateral. There are two types of coverage tests:
The objective of an overcollateralization test (also referred to as an OC or par coverage test) is to ensure that sufficient collateral exists to repay the principal balance of the CLO tranches. (See Figure 23.4.) The test measures the par balance of the loan pool relative to the par balance of each debt tranche. A ratio result of less than 100% for a given tranche level means that the par balance of the loans does not fully support the par balance of that tranche.
Figure 23.4: Illustration of hypothetical OC test curing calculation
The objective of an interest coverage test (also referred to as an IC test) is to ensure that sufficient interest cash flow exists to support the ongoing interest payments due to the CLO tranches. The test measures the interest expected to be received from the loan pool relative to the interest payments due to be paid on each debt tranche. A ratio result of less than 100% for a given tranche means that the interest payments to be received from the loans are insufficient to fully service the upcoming payments due to that tranche.
In order to ensure that the CLO's debt tranches are sufficiently supported by the loan pool, the OC and IC ratios are required to be maintained at or above certain predetermined levels. Failure to maintain these levels for a given tranche causes the temporary diversion of cash flows away from more junior tranches to prepay the senior‐most debt tranche. This cash flow diversion continues until the failed test is brought back into compliance with the required predetermined level.
Mathematically, failing ratios can be brought back into compliance either by reducing the denominator or by increasing the numerator. In the case of a failing OC test, for example, the test can be cured either by paying down a portion of the debt tranches, which reduces the denominator, or by adding additional amounts of collateral, which increases the numerator.
In the event of a failed OC or IC test, the tranches that are junior to the tranche level with the failing test would not receive any current payments (as their interest payment would instead be diverted to prepay the senior tranche). Rather, those tranches would pay‐in‐kind (PIK), which essentially provides them with an IOU. The outstanding balance of the non‐paid tranche would be increased by the amount of the missed payment, and that amount would be due (to the extent available) on a later payment date.
The CLO's debt investors receive meaningful protection from the OC and IC tests, and the possible diversion of more‐junior cash flows. It also acts as a somewhat self‐curing mechanism in the event of a failed OC test, because it introduces funds from the CLO's interest waterfall to be used instead to support the transaction's principal waterfall.
It should be noted that many CLOs also have an additional overcollateralization test, often referred to as the interest diversion test. This test is typically calculated only at the junior‐most debt tranche level, and acts as an early warning before any of the actual OC tests are breached. The failure of this test has two key differences within the CLO structure from a failed OC or IC test. First, instead of diverting all of the more‐junior cash flows, this test typically only diverts half of them. Second, the diverted funds are used by the CLO to invest in extra loans, rather than to prepay the debt tranches.
THE LIFE AND TIMES OF A CLO
Chronologically, there are several important milestones in the lifecycle of a CLO (Figure 23.5).
Figure 23.5: CLO lifecycle illustration
First is the marketing period. In the weeks prior to the issuance of a new CLO, the manager and the CLO underwriter meet with prospective investors and discuss the transaction's intended structure, tranche spreads, documentation, and collateral composition. Investors provide their feedback, upon which the manager and the underwriter may choose to adjust the intended transaction to satisfy the investor requests and secure orders for the marketed tranches. Often, a loan warehouse is opened prior to the marketing period, and the CLO manager is provided with a budget to begin purchasing loans that will ultimately be included in the CLO's loan pool. The existence of a warehouse provides investors with a degree of certainty as to which specific corporate obligors will be included in the CLO, and at what price the manager is able to source the loans.
Next is the pricing date. On this day, coupon rates are set for each of the CLO tranches, and allocations are made to investors who placed orders for investment in the transaction. Depending upon the demand for particular tranches, investors may be filled on their entire order, or the CLO may be oversubscribed, causing investors to be allocated less than they had ordered.
Following the pricing date is the closing date, also referred to as the issue date. This is the day on which the CLO legally begins to exist and the CLO tranches are issued as securities. The CLO tranche trades that were allocated to investors on the pricing date all settle on the closing date. Investors provide funds and formally invest in the CLO tranches. In turn, the CLO uses the funds to purchase any loans that had been warehoused for the CLO prior to that point and continue investing in other loans.
The next key milestone for a CLO is its effective date, often up to six months or longer after the closing date. This period represents the time it takes for the CLO manager to fully select and acquire its complete initial loan pool. The entire period during which the manager purchases the pool is known as the ramp‐up period, which begins when loans begin to be purchased into a warehouse (often during the marketing period), and continues until the effective date, when the loan pool is fully invested. Prior to becoming effective, the quality and characteristics of the CLO's loan pool must meet or exceed certain predetermined metrics and the expected ratings of each debt tranche must be confirmed by the relevant credit rating agencies.
Another important date in the lifecycle of a CLO is the end of the non‐call period, the period during which the tranches are not subject to optional redemption, refinancing, or repricing. At times, the equity investors may benefit from redeeming the CLO or from amending the stated coupon of the CLO's debt tranches. To protect the interests of debtholders, there is no ability for the equity investors to elect to do so until after the end of the non‐call period, typically about two‐years after the CLO's closing date.
One of the most significant dates for a CLO is the end of the reinvestment period (also referred to as the revolving period). During the reinvestment period, which is typically at least four years following the CLO's closing date, the CLO manager ensures that any cash proceeds (received from loan maturities, prepayments, sales, or recoveries) are reinvested into new loans. In contrast, after the end of the reinvestment period, cash proceeds are generally not reinvested, and are rather used to amortize the CLO tranches. It is important to note that even after the end of the reinvestment period, CLO managers generally maintain some ability to make substitutions into new loans and do not always amortize the CLO tranches. However, there are significant restrictions on such abilities, particularly in circumstances where the transaction is underperforming.
The final milestone in the CLO lifecycle is the stated maturity date, also referred to as the legal final maturity date, which is the date upon which the CLO tranches are due to be repaid in full. This date typically occurs at least 12 years after the closing date. Due to the structure of a CLO, which begins to pay down the tranches after the end of the reinvestment period, most tranches of a CLO are no longer outstanding at the time of the stated maturity date. However, any remaining loans in the CLO pool would be sold at such time and the sale proceeds would be used to pay down any remaining outstanding debt tranches (with any residual value belonging to the equity). In order to prevent the CLO from being required to sell loans, and thereby subject to risk from the market value of the loans at such time, CLOs are generally restricted from investing in any loans that mature after the stated maturity date.
THE LOAN MARKET
With a background in the creation and structure of a CLO, we now turn our attention to the underlying collateral pool. While the CLO's structure regulates the risk, return, and payment timing of the tranches, it is ultimately the performance of the underlying loan collateral that drives the CLO.
The collateral pool of a CLO is comprised of loans to speculative‐grade corporate borrowers. CLOs predominately invest in single‐B rated loans (which are between four and six notches below investment grade), although CLO pools also often include some proportion of double‐B and triple‐C‐rated loans. The loans are frequently referred to as leveraged loans, due to the fact that the issuing companies operate with a relatively large proportion of debt in their corporate capital structures. Companies often issue loans at the same time as they issue high‐yield debt, and may use the entire new financing package to fund a leveraged buyout (LBO) or for other purposes.
Loans are issued by a diverse range of obligors, incorporating virtually every industry sector and domestic geographic location. Most loans are floating‐rate instruments, which accrue on a quarterly basis and pay a stated rate (also referred to as a spread or margin) over three‐month LIBOR.
With limited exception, CLOs generally make investments in a company's senior‐secured, first‐lien loans. This means that in the event of a default, the lenders (in this case, the CLO, which acts as a lender) have a senior claim to the assets of the company and should therefore expect a higher recovery value as compared to the recovery prospects of the company's more subordinate, second‐lien loans or unsecured debt. Recovery prospects may be based on the liquidation value of the company's hard assets or on its enterprise value as it continues operations.
The size of the loan market has grown significantly over time as loans have gradually grown in acceptance as a mainstream asset class. (See Figure 23.6.) Demand for loans is due in part to the fact that they are secured instruments with floating‐rate coupons. This stands in contrast to other corporate debt instruments that are often unsecured or issued with fixed‐rate coupons. Loan issuance picked up after the financial crisis, as the economic recovery spurred companies to issue additional loans to fund M&A activity or to optimize their corporate liability structures. In addition, there has been meaningful demand for loans from open‐ended mutual and exchange‐traded funds (ETFs) in addition to CLOs and privately managed accounts.
Figure 23.6: Institutional loan outstanding amount, 2004–2015 ($ billions)
Source: Thomson Reuters, S&P/LSTA Leveraged Loan Index
The loan market is sizeable and an active secondary market exists. Loan trading, however, is more operationally intensive, and trade settlement times tend to be longer than for corporate bonds or other non‐loan securities.
Historically, loan agreements have included positive or negative covenants to which the borrower is required to adhere, with breaches of such covenants generally constituting a default. There are two broad types of covenants. Incurrence covenants restrict the borrower from taking certain actions that could dilute the value, control, or protection of the existing debtholders. These may include limitation on the borrower assuming additional debt, paying dividends, or repurchasing shares. Maintenance covenants require the borrower to continually uphold a specified degree of healthy performance, such as maintaining earnings by at least a certain proportion above their debt expenses.
Over the past few years, fewer covenants have been included within loan agreements, resulting in the advent of loans that are so‐called covenant‐lite (or cov‐lite). A primary risk of covenant‐lite loans is the prospect for lower recovery values, as lenders would need to wait for the borrowing company to deteriorate to the point of missed debt payments rather than being able to react to the earlier occurrence of a breached maintenance covenant. The jury is still out on the fundamental prospects for covenant‐lite loans. Empirical evidence suggests that covenant‐lite loans have not historically experienced lower post‐default recovery values.3 However, historical data may not be predictive of future recovery prospects for cov‐lite loans (Figure 23.7). There were relatively few historical observations of covenant‐lite loan defaults, resulting in a limited dataset. In addition, the unique government policy intervention that existed during the previous recovery cycle may have influenced the historical observed recoveries while future recovery cycles may perform differently.
Figure 23.7: Percentage of annual primary first‐lien loan issuance by covenant‐lite loans, 2004–2015
Source: Wells Fargo, S&P LCD
The growth of the loan market, and the rise of covenant‐lite loans in particular, has attracted the attention of banking regulators. In 2013, U.S. banking regulators issued guidance on leveraged lending practices, which defined certain expected characteristics for bank‐originated leveraged loans. Essentially, underwriting banks are expected to issue prudent loans that are able to repay over time. While the specific balance of characteristics is generally left to the discretion of the underwriting bank, certain broad examples were provided. These included a maximum total leverage multiple (including total debt to EBITDA), and an expectation of the borrower's ability to repay the debt within a specified time horizon. The extent to which this guidance changes underwriting standards for future loan issuance has yet to be fully realized.
CLO REPORT CARDS: PERFORMANCE METRICS
How can an investor tell how a CLO is performing? Fortunately, CLO reporting is very transparent. Investors receive monthly trustee reports as well as quarterly payment date reports. These reports show the CLO's current tranche balances, the identity of all loans held in the pool, and various metrics that represent the quality of the loans. The reports also show the metrics relative to each of the required portfolio guidelines (which are technically referred to as the concentration limitations), as well as certain high‐level averages that represent the quality of the underlying loan pool (which are technically referred to as the collateral quality tests). If one were to look at the summary page of a monthly trustee report, one would typically see the following data and metrics:
- Current tranche balances
- Current coverage test results (i.e., overcollateralization, interest diversion, and interest coverage tests)
- Collateral characteristics and limits (e.g., the percentage of loans that are second lien or covenant‐lite, presently in default, or rated CCC+ or below)
In addition, the report contains several other key metrics, which fluctuate somewhat each month based on the then‐current composition, performance, and active‐management of the underlying loan pool.
Several of the primary metrics that are reported in a CLO's trustee report are described in the following:
The weighted average rating factor (WARF) is a standard Moody's measure of the CLO pool's average credit quality. Each rating subcategory, from AAA to D, is given a numeric value (on a scale of 1 to 10,000).4 The scale is applied to each individual loan in the pool, from which a weighted‐average pool rating is calculated (Table 23.2). Typically, CLO WARF measures have tended to be in a range from 2500 to 3000, representing a weighted‐average loan pool credit quality between single‐B and single‐B‐minus.
Table 23.2: WARF: Moody's Default Probability Ratings versus Moody's Rating Factors
Source: Moody's Investors Service
| Moody's Default Probability Rating | Moody's Rating Factor |
| Aaa | 1 |
| Aa1 | 10 |
| Aa2 | 20 |
| Aa3 | 40 |
| A1 | 70 |
| A2 | 120 |
| A3 | 180 |
| Baa1 | 260 |
| Baa2 | 360 |
| Baa3 | 610 |
| Ba1 | 940 |
| Ba2 | 1350 |
| Ba3 | 1766 |
| B1 | 2220 |
| B2 | 2720 |
| B3 | 3490 |
| Caa1 | 4770 |
| Caa2 | 6500 |
| Caa3 | 8070 |
| Ca, C | 10000 |
The weighted average life (WAL) is a measure of the CLO pool's tenor. It is calculated by measuring the average time until maturity for each loan, weighted by the balance of the loan.
The weighted average spread (WAS) is a measure of the CLO pool's rate of return, calculating the weighted‐average floating spread over LIBOR for each loan in the pool. (For example, if a loan pays three‐month LIBOR + 3.50%, then the 3.50% spread would be included in the WAS calculation.)
The diversity score, as implied, it is a measure of the CLO pool's diversification. The value of the metric is based on a Moody's model that generates a numeric value for the pool, based upon inputs such as each loan's issuer, industry, and geographic location (Figure 23.8).
| Coverage and Collateral Quality Tests | ||||
| Coverage | ||||
| Test Name | Current Threshold | Current | Result | Prior |
| Senior Interest Coverage Test | 120.00% | 244.77% | Pass | 282.89% |
| Class C Interest Coverage Test | 112.50% | 220.94% | Pass | 254.33% |
| Class D Interest Coverage Test | 107.50% | 197.57% | Pass | 226.00% |
| Class E Interest Coverage Test | 102.50% | 174.06% | Pass | 197.25% |
| Senior Overcollateralization Test | 121.40% | 132.47% | Pass | 132.38% |
| Class C Overcollateralization Test | 114.10% | 122.72% | Pass | 122.63% |
| Class D Overcollateralization Test | 108.10% | 114.70% | Pass | 114.62% |
| Class E Overcollateralization Test | 103.70% | 108.59% | Pass | 108.52% |
| Class E Reinvestment Test | 105.10% | 108.59% | Pass | 108.52% |
| 5.1(c) Test | 102.50% | 156.27% | Pass | 156.17% |
| Quality | ||||
| Test Name | Current Threshold | Current | Result | Prior |
| Moody's Weighted Average Recovery Rate Test | 47.5% | 50.0% | Pass | 50.0% |
| Weighted Average Life Test | 7.50 | 5.07 | Pass | 5.11 |
| Diversity Test | 65 | 74 | Pass | 73 |
| Maximum Moody's Rating Factor Test | 2,963 | 2,729 | Pass | 2,664 |
| Minimum Weighted Average Spread Test | 3.75% | 4.01% | Pass | 4.27% |
Figure 23.8: Summary metrics from a sample CLO trustee report
Source: U.S. Bank
The probability of defaults occurring in a CLO's loan pool is in part a function of time. The longer a loan is outstanding, the greater the probability that the obligor may become distressed. Due to this, one can consider how the metrics described earlier may balance against one another. For example, while a lower‐credit‐quality portfolio has a greater probability of default, there may be offsetting factors, such as a shorter remaining life or greater diversification, which could each act to offset the loan pool's overall probability of default.
THE CLO MANAGER
A CLO's underlying loan pool is actively managed by a professional collateral manager. Several different types of investment firms act as managers to CLO funds. These include traditional asset or fund managers, such as BlackRock or Neuberger Berman, as well as insurance companies, like Prudential and New York Life. Some CLOs are managed by the credit investment arms of private equity firms, such as Apollo, Carlyle, or GSO/Blackstone, while others are managed by credit teams within hedge funds, such as Och‐Ziff. CLOs may also be managed by standalone investment companies whose primary business is to manage credit and/or loan funds, either in the form of CLO portfolios or as privately managed accounts.
The CLO manager plays an integral role in the performance of the transaction. Essentially, their mandate is to prudently invest the CLO's available cash balances into loans in accordance with all of the CLO's investment guidelines and within the limits required by the CLO's various metrics (such as constraints on OC ratios, triple‐C percentages, or WARF values). In order to do this, the manager employs a credit team of portfolio managers, analysts, and traders, who evaluate, transact, and conduct ongoing surveillance on corporate credits that are appropriate for investment by the CLO.
Typically, the CLO manager's analyst team is divided according to individual sector expertise. For example, one credit analyst may cover health‐care companies while another specializes in energy credits and a third focuses on names in technology, media, and telecom. The CLO management company may rely on their analyst team to support their CLO effort in addition to their other funds that invest in speculative‐grade corporate credit.
CLOs represent an opportunity for managers to add fee‐generating assets under management (AUM). With each new CLO, the manager adds several hundred million dollars of AUM, and is often able to tap into an investor base that does not otherwise invest in the manager's non‐CLO funds. CLO managers typically earn a total management fee of 0.50% per annum on the outstanding balance of the CLO's assets. (The total fee is split between a “senior” fee that is paid prior to any of the CLO tranches, and a “subordinate” fee that is paid just before the distribution to equity investors.) Later in the life of the CLO, there is also the potential for the manager to receive an incentive fee to the extent that the realized equity return exceeds a pre‐specified hurdle.
QUALITATIVE ANALYSIS OF A CLO
When analyzing a CLO, there are two main qualitative aspects that warrant consideration: due diligence of the CLO's collateral manager and thorough review of the transaction documentation.
Meet and Greet: Collateral Manager Due Diligence
An investment in a CLO is often intended to be held for the medium or long term, so the investor is effectively entering into a long‐term relationship with the CLO manager. The manager's expertise, strategy, and actions can have a significant impact on both the performance and tradability of the CLO's tranches throughout the life of the investment. On the positive side, a manager can select strong credits, maintain consistent equity returns, manage steady transaction amortization, and maintain open communication with the CLO's investor base. On the negative side, a manager can be slow to act on deteriorating problem credits or try to enhance equity returns by dipping into more yieldy loans that represent weaker credits. They may also take actions that do not fully balance the interests of all investors in the structure.
Prior to making an investment, due diligence should be conducted to assess the extent to which the manager has the requisite expertise, staffing, systems, and processes to properly manage the CLO. In addition, the investor should consider the manager's investment style and track record, and their views regarding the loan market. The qualities of different CLO managers can vary widely in these areas, which are briefly touched upon in what follows.
Expertise and Staffing
An investor should assess the experience of the manager's credit analysts, traders, and portfolio managers, and any recent turnover within the research team. It is also worth considering the size of the team relative to the number of credits that are followed, the depth of analysis that is conducted, as well as whether the analysts' time is shared with the manager's other funds or asset classes.
Investment Style and Track Record
Getting a sense of the manager's investment style and track record involves discussion, analysis, and ongoing review. Some styles are more readily apparent. For example, some managers select loan portfolios with fewer credits while others construct granular portfolios with many credits. Other investment styles may be more difficult to detect, such as an affinity or aversion to loans from particular industry sectors, or the regular use of allowed risk buckets (such as those for triple‐C‐rated or second‐lien loans).
The manager's historical track record can be difficult to ascertain from data alone. It is possible that key performers may have shifted positions since the track record was established and past equity returns or loss experience may not tell the whole story. For example, strong historic equity returns may have resulted in part from investments in more risky loans, heavy trading activity, or structural features within individual CLOs. The investor should carefully evaluate historic performance data to determine whether it represents how the manager would be expected to perform in the future.
Systems and Processes
CLO management is a data‐intensive business. Each manager actively follows hundreds of corporate loan investments, public and/or private corporate financial data, loan prices, and CLO transaction metrics. An investor should assess the manager's approach to data management and how seamlessly the data is incorporated within their investment and surveillance processes.
The investor should also evaluate the manager's normal business processes. An understanding of the manager's credit selection process is of foremost importance. Other processes worthy of discussion include the manager's approach to credit surveillance, particularly how they address troubled credits, and how purchases or sales are allocated across funds.
The CLO investor should meet with the manager and request sufficient information from which they can assess the entirety of the manager's platform. Investors who are interested in stable, consistent future performance, and a more readily‐understood risk profile, would need to consider these qualitative factors, and not simply take the information in the CLO's marketing materials at face value.
The Devil in the Details: CLO Documentation Review
The other key aspect of qualitative CLO analysis relates to a review of the transaction's documentation. While the general structure of a CLO's tranches, payment waterfall, and loan investment guidelines have all become more consistent across post‐crisis CLOs, a thorough review of the documentation will still reveal meaningful differences between transactions. The degree of flexibility that the documents provide to the manager can affect the composition and risks of the loan portfolio, the pace at which debt tranches are ultimately repaid, and the benefit to investors from the CLO's inherent structural protections. Overly restrictive documents, on the other hand, could prevent the CLO manager from taking investment actions that would be beneficial to the CLO's performance. Differences in documentation would likely have the most pronounced effect on performance in circumstances where the underlying loan pool became distressed.
Two main documents delineate the terms of a CLO: the offering circular and the indenture. The offering circular is a marketing document that describes key provisions of the CLO in relatively easy‐to‐read summary sections and also includes the disclosures necessary for the sale of securities. The indenture is the formal governing document, which includes all of the legal terms and definitions that are relevant to the transaction.
The following is a brief guided tour through some of the key considerations that pertain to CLO documentation.
Investment Guidelines
The documents prescribe investment guidelines that define the parameters of the instruments into which the CLO pool is allowed to invest. Certain investments are never permitted. For example, the pool may be required to consist entirely of loans, and not contain corporate bonds, structured finance securities, or other debt instruments. Other investments may be permitted, but only within certain constraints. Typical limits might include a requirement that at least 90% of the loan pool consist of senior secured loans, that no more than 60% of the pool can consist of covenant‐lite loans, or that no more than 2% of the pool be exposed to any individual borrower. These sections of the document place parameters around the expected constitution and quality of the collateral pool.
Trading Activity
The documents also describe the circumstances under which the CLO manager is allowed to invest (or reinvest) available cash balances, and when they would be required to stop investment activities and allow the loan pool to naturally pay down. Broadly speaking, the manager is allowed to invest during the reinvestment period, and not invest after the end of the reinvestment period. However, there may be restrictions on investment activities even during the reinvestment period, particularly if the transaction is underperforming. Likewise, limited investment activities may still be permitted even after the end of the reinvestment period. The document may also set parameters for how the manager may address the workout of distressed loans, and under what circumstances a loan's maturity date can be amended and extended. These sections of the document can have a meaningful influence on the total lifespan of the CLO's collateral pool, and by extension, on the duration of the CLO's tranches.
Priority of Payments
An important aspect of the CLO documentation defines the step‐by‐step order of payments that are made to each of the tranches. This section prescribes how available interest and principal proceeds are disbursed to pay the CLO's management and administrative fees to each of the debt tranches and to the equity investors. It also identifies the points within the payment waterfall when the overcollateralization (OC) and interest coverage (IC) tests are applied, and how the failure of such tests would redirect cash flows away from subsequent steps in the payment order. This section also describes whether the payment sequence would change after the occurrence of extreme events, such as an event of default.
OC Test Calculations
Several areas of the document describe adjustments made when calculating the CLO's OC tests. Recall that the OC tests measure the par value of the loans relative to the par value of the CLO tranches. However, there can be circumstances in which certain loans are included in the calculation at less than their stated par value. These typically include loans that have already defaulted, loans that were originally purchased at deep discounts from par, or loans that contribute to an excessive amount of triple‐C‐rated collateral within the CLO. In circumstances like these, the loans are given a haircut, meaning that they are considered to have a value that is less than their full par value when being considered within the OC test calculations. These haircuts lead to a reduction in the resulting ratio. The application of OC haircuts is important to understand, because they influence how quickly the tests may fail, which in turn can cause the diversion of junior tranche cash flows for the purpose of supporting the senior tranches.
Deal Document Amendments
Another section of the document defines the required process that must be followed in order to make future changes to the CLO documentation. Unforeseen market events or other circumstances, such as changes in government regulation, may create the need to make amendments to the original terms of the CLO documents. In such cases, the documents describe the notices that must be provided to investors, credit rating agencies, or other parties to the transaction, and who must provide consent prior to the execution of a proposed amendment. It is important for an investor to understand these aspects of the document, because it is possible for certain terms of the CLO to be materially altered even if it were against the will of some investors in the transaction.
Extreme Events
Some sections of the document specify what will occur in the event that certain extreme circumstances impact the CLO. For example, the document describes what would constitute an event of default for the CLO, as well as how the transaction would operate and who would retain primary control after such an event occurred. Other sections relate to what would occur if the CLO manager were to resign (or in some cases, if specific key personnel left the collateral manager's firm).
QUANTITATIVE ANALYSIS OF A CLO
Quantitative analysis of a CLO tranche involves projecting future expected cash flows and measuring risk or return under a variety of possible scenarios. The assumptions used to project future cash flows are the heart and soul of the analysis. The exact scenarios that should be projected depend on both the tranche level and the risks that are being assessed. For example, an investor in a senior tranche may be less concerned about default risk and more concerned about tranche extension risk while an equity investor may have the opposite perspective. Assumptions can be further refined, based upon the quality of the underlying loan pool and the analyst's perspective of what investment actions may be taken by the CLO manager.
As a practical matter, it is feasible to model CLO cash flows in a spreadsheet. However, most market participants utilize third‐party software, due to the complexity of CLO modeling and the data requirements necessary to maintain current collateral information for an actively managed underlying loan pool.
The key assumptions that are input into a CLO cash flow model relate to default rates, recovery rates for defaulted loans, prepayment rates, and the characteristics of loans into which the CLO is expected to invest in the future. Each of these assumptions is briefly touched upon in the following.
Default Rates
One of the most significant cash flow assumptions—but arguably the most difficult to estimate—is the portfolio default rate. The relatively long‐term nature of CLOs requires a multiyear perspective of the future default experience of speculative‐grade corporate credit. Three approaches to modeling future defaults are described ahead: conditional default rates (CDRs), cumulative default rates, and Monte Carlo simulation. All three approaches intentionally focus on the loan pool as a whole, as opposed to an in‐depth understanding of each specific underlying credit in the pool. A CLO typically has very limited exposure to any individual loan issuer, so there is generally greater utility in understanding how many loans are expected to default over time and the impact they will have on the CLO's payment structure, rather than which ones they are likely to be.
Conditional Default Rates (CDRs)
The easiest and most common default rate approach is to assume a constant percentage of annual portfolio defaults, commonly referred to as a CDR. Market participants have traditionally assumed a constant rate of 2–3%, which is loosely based upon historical experience. In practice, this means that 2–3% of each year's projected outstanding loan pool is assumed to default.
Benefits of this flat CDR approach relate to it being both simple and uniform. Differences in future expected performance can be addressed separately through the discount rate that is applied to the resulting cash flow projections. One drawback is the fact that it does not directly account for the current position in the corporate credit cycle. Corporate defaults tend to fluctuate between periods of higher and lower defaults. Modeling a constant, average rate is therefore less realistic—and could have a different effect on the CLO structure—than more specific default timing assumptions.
A variation of the flat CDR approach is to use a vector of different CDR values. The CLO's total outstanding life can be divided into several time horizons, each with a different assumed CDR. While this approach may be somewhat more realistic than the previous flat CDR approach, it is still not customized to the characteristics of the specific loan pool.
As an additional variation to both the flat and vector CDR approaches, it may be useful to identify specific risky portfolio characteristics, such as particularly concentrated loan exposures, loans that pose an imminent default risk, or allocations to weaker industry groups. (These loans may possibly be identified on the basis of their current market price, as low marks on the loans may at times be used as a proxy for weaker credit quality.) In such cases, asset‐specific default assumptions may be applied to particular loans within these identified risk buckets. The general CDR assumption for the remainder of the loan pool would then be overlaid in addition to the asset‐specific assumptions.
Cumulative Default Rates
A second approach is to assume that a certain cumulative proportion of the CLO's loan pool will default over the next several years, and then account for the credit cycle by staggering the proportion of that cumulative rate over the course of the projection period. For example, if the analyst wanted to apply a five‐year cumulative default rate of 15%, and assume that the greatest stress occurs in the middle of that period, they might apply a cumulative annual vector such as the following: 1% / 2% / 6% / 4% / 2%, which together sums to the intended rate of 15%.
One approach to determine a cumulative lifetime default rate assumption is to rely on historical precedent—by looking at the performance of similar loans over average or specific past time periods. Historical default data, both long term as well as for specific cohort years, is often available from credit rating agencies or other data providers. The challenge is to determine which historical data is appropriate to apply to a given CLO. Current credit ratings on the CLO's underlying loans may be a good starting point to inform a cumulative default rate assumption, but adjustments may be necessary to account for disproportionate exposures to particular loan types, sectors, or other risks.
The main drawback of historical data is that it may not reliably predict future performance. This is especially true when considering the severe performance of loans during the financial crisis, as well as the unprecedented global monetary stimulus that contributed to the loans recovering so rapidly thereafter. Neither of these factors will necessarily be repeated in the future. In addition, loan underwriting standards may have changed due to a combination of market forces and regulatory pressures. For these reasons, there may have been a regime shift between the available historical data sets and the expected future default (and recovery) experience for loans. Effectively using historical data may therefore be limited in its ability to project future performance.
An alternative to using specific historical data is to project future default rates based upon one's own macroeconomic outlook. An investor may have a view on the future default rate for speculative‐grade corporate debt, and they can overlay that perspective onto the expected performance of the specific CLO portfolio. Credit rating agencies and sell‐side Wall Street research teams also periodically publish default rate projections, which can be useful for this type of analysis.
There are also challenges to practically applying a broad macroeconomic perspective to a specific CLO portfolio. It is important to consider how the macroeconomic‐based default rate was determined, and if there are differences that would limit its application to the CLO. Some macroeconomic rates are inferred from prior recession cycles while others are calculated by regressing variables against historical default rates and then extrapolating those rates into the future. Still others are based on the idea that current loan market prices imply an expected default rate (given an assumed recovery rate). Properly calibrating a macroeconomic default rate with the CLO's collateral pool requires an evaluation of the source and methodology of the default rate projection.
Monte Carlo Simulation
A third approach, although clearly the most computationally advanced, is to simulate future default rates by modeling the performance of each loan in the CLO's portfolio to its maturity or default. A rating transition matrix can be used to simulate the probability of each future period's change in credit quality and possible jump to default. This calculation is performed for every individual asset within the CLO and then aggregated to generate different paths of lifetime portfolio defaults. This process is repeated thousands of times and modeled in conjunction with the structural rules of the CLO's payment waterfall to simulate all possible future outcomes for the CLO. A simulation approach can also apply dynamic recovery or prepayment rates that vary as a function of each path's simulated default rate.
To a certain extent, simulations provide for more realistic modeling of CLO cash flows. By predicting defaults on an asset‐level basis, the simulation can capture the disproportional shock to the CLO structure from the loss of specific, large, high‐yielding, or low‐recovery assets. This stands in contrast to the other default rate approaches, which implicitly assume that defaulted assets have average characteristics. (In fact, relative to lower‐yielding loans, the higher‐yielding assets may actually have a greater probability of default, and their loss may have a more adverse impact on the CLO structure.) Further, simulation results can be used for risk management purposes to quantify more extreme scenarios or specific confidence intervals or to evaluate a tranche's sensitivity to particular cash‐flow assumptions.
While simulations provide for superior structural analysis, they are somewhat limited by the data sets that are used as model inputs. Historical rating transition matrices are necessary to conduct the simulation, as are corporate industry and asset correlation rates. These historical relationships tend to change over time, however, and may not be fully relevant for future periods. In addition, from a practical perspective, subscription‐based third‐party CLO models with Monte Carlo capability can be difficult to come by.
In summary, there are three general approaches to modeling CLO default rates: conditional default rates (CDRs), cumulative default rates, and Monte Carlo simulations. All three approaches can be applied in practice, each with its own benefits and drawbacks. At times, an analyst may find it beneficial to apply more than one approach to the same transaction. Investors should first assess the qualities of the particular CLO structure and loan pool to determine what risks or qualities are most important to evaluate. Then, they can select an analytical approach that best addresses their modeling and valuation requirements.
Recovery Rates
The seniority of leveraged loans within the capital structure of the issuing corporation has historically resulted in high post‐default recovery rates. It should be noted that the opposite of a recovery rate is a loss severity rate. (The recovery rate measures the percentage of par that is repaid on a defaulted loan while, conversely, the severity rate measures the percentage of par that is not repaid. Severity is also referred to as loss‐given‐default, or LGD.) Market participants have generally assumed flat, average recovery rates of 70% or higher.
Assumed recovery rates should be adjusted to account for loans that are not senior‐secured, such as second‐lien or unsecured loans. These loans have historically realized lower recovery rates, although they do comprise only a small minority of a CLO's loan pool.
Recovery rates are influenced by various factors such as the company's enterprise value at the time of default, or the value of its assets in the event of liquidation. The recovery value is also affected by the company's capital structure, including any other obligations of the defaulted company that would be paid prior to, or together with, the loan in question. Timing can also play a role in recovery value. With certain loans, it may take longer before a company formally triggers a default (which may be the case in the future with covenant‐lite loans). Likewise, the workout process for a defaulted obligor may take a long time to process through the legal system, particularly in periods when many other companies have also defaulted. The longer the delay in realizing the recovery, the greater the chance that the company or its assets decline in value, which in turn could lead to lower recovery rates.
Similar to default rates, there are also several ways to implement recovery rate assumptions. One is to apply flat, average recovery rates. Another is to use a multiyear vector, with lower recovery rates in certain years, but higher rates in other years, to account for changes in the business cycle. Finally, recovery rates can also be modeled within a Monte Carlo simulation as a function of each path's projected default rate.
To account for the longer expected timing of the workout process in times of higher corporate stress, the cash‐flow analysis can incorporate recovery lag assumptions. These assumptions can have a notable impact on the CLO, as investors may face extension risk from delays in realizing recovery values.
When reviewing historical recovery rate data, it is important to note the method used to measure the recovery rate (Figure 23.9). Often, the loan's recovery value is considered to be the trading price of the loan 30 days after the date of default. However, some data services capture the trading price of the loan on the day of the default itself while others measure the ultimate cash recovery received on the loan regardless of how long it took. Each of these approaches has different implications for the way one may view future recovery prospects. In addition, it is important to note that actual recoveries are sometimes realized in the form of equity or warrants (or other securities) in a reorganized company, rather than as a lump‐sum cash payment to the CLO. In contrast, CLO models typically assume that recoveries are received as cash that is then deployed to reinvest or to amortize the CLO's senior notes. Non‐cash recoveries have a different impact on the CLO structure and may require adjustments to the cash flow assumptions in order to achieve the intended effect.
Figure 23.9: Historical leveraged loan default and recovery rates, 1998–2015
Note: Annual default rates are par‐weighted. Annual recovery rates are issuer‐weighted, first‐lien recovery rates and based on the loan price 30 days after the default date.
Sources: J.P. Morgan, Moody's Investors Service, S&P LCD, Markit
Prepayment Rates
Loan prepayments occur when the corporate borrower repays a loan prior to its stated maturity date. Loans typically have minimal call protection, which provides the borrower with the option to elect for an early redemption. The company's decision to do so is largely based upon corporate finance considerations. If current market interest rates are lower than the rate being paid on the loan, a new loan would result in a lower interest expense. The borrower may therefore elect to refinance their loan, using the proceeds of the new loan to prepay the original lenders. Alternatively, the borrower may wish to extend the maturity of outstanding obligations in order to secure greater certainty of future funding or to show only long‐term debt on their financial statements. At times, a loan prepayment is caused by merger activity. In a prepayment, the CLO receives cash earlier than previously expected, rather than continuing to hold an ongoing exposure to the loan.
Depending on the phase in the CLO's lifecycle, there can be different implications of a prepayment. During the reinvestment period, prepayments can generally always be reinvested, provided that the CLO's metrics are satisfied or at least not made worse as a result of such reinvestment. After the reinvestment period, prepayments can typically be substituted into new loans with similar credit and duration characteristics as the prepaid loan. In circumstances where prepayments are not allowed to be reinvested, the proceeds are used to sequentially amortize the CLO tranches. In general, faster prepayments benefit investors in the CLO's debt tranches, because they receive their money back faster and therefore have less exposure to potential future credit issues. In contrast, slower prepayments generally benefit the equity investors, because the inherent leverage of the CLO structure remains outstanding for longer.
Market participants often assume conditional prepayment rates (CPRs) of 15%–20% per annum, which is broadly consistent with average historical experience (Figure 23.10). However, CPRs can fluctuate, often inversely to the state of the market. When market conditions have been favorable, actual annualized CPRs have been in excess of 50%, largely due to heavy refinancing activity. During weaker periods, when capital markets were not readily open to speculative‐grade borrowers, annualized prepayment rates have dipped below 10%. CLO pools can also suffer from adverse selection, in which the stronger credits are redeemed from the pool while the weaker credits remain outstanding.
Figure 23.10: Historical loan repayment rate, 2002–2015 (trailing 12‐month annualized rate)
Source: Morgan Stanley
Reinvestment Rates
One other important CLO cash‐flow assumption is often given less focus, and it relates to the reinvestment of available cash. Recall that CLOs receive cash through both maturities and prepayments within the loan pool, and as actively managed vehicles, the CLO manager is often able to reinvest those funds into new loans. The characteristics of those new loans (such as the assumed maturity, spread, and purchase price) will have a material impact on both the duration of the debt tranches and the rate of return to the equity investors. For example, an assumption that the manager will reinvest into longer‐duration assets would keep the debt tranches outstanding for longer and enhance the equity returns. Assumed reinvestment into higher‐yielding loans or into loans at discounted purchase prices can likewise increase the modeled equity returns. When setting reinvestment assumptions, an investor should consider the manager's track record and style, as well as the constraints within which the manager is operating. Typically, one would expect the manager to optimize the use of available constraints in order to realize the best results for all investors.
CLO Valuation
Once CLO cash flows have been projected, those projections can be used to value an investment. This is typically accomplished through a standard discounting process. However, since the debt tranches are floating rate, they would normally be discounted at a floating spread over LIBOR, also referred to as a discount margin (DM), rather than a fixed internal rate of return (IRR). A fixed‐rate tranche, or the equity tranche, would commonly be discounted at a fixed rate.
When discounting the cash flows, it is important to consider the relative stress that was incorporated into the cash flow projections themselves. If the cash flows are already risk adjusted and thereby already represent a conservative or stressful projection, then the discount rate should be lower. On the other hand, if the projections are not risk adjusted, then the risk would need to be captured by the use of a higher discount rate. In addition, for equity tranches, whose cash flows may be very sensitive to specific assumptions, a range of scenarios and sensitivity analysis is appropriate rather than just applying a discount rate to a single projected scenario.
Another approach that is sometimes used to estimate a tranche's value is by calculating its net asset value (NAV) coverage. This approach considers the market value coverage of a CLO tranche by assuming that the underlying loan portfolio is liquidated at its current market value and sale proceeds are used to repay the CLO's liabilities. NAV coverage is a useful data point, particularly for more distressed transactions in which some tranches are not fully collateralized on a market value basis. However, it is an incomplete proxy of actual tranche value. The NAV does not account for the CLO's structural features or the fact that the transaction most likely cannot be redeemed or liquidated for many years to come.
A WHOLE NEW WORLD: CLOs IN THE POST‐CRISIS REGULATORY ENVIRONMENT
Lessons learned through the financial crisis and the different financial and regulatory atmosphere in the era following the crisis have each contributed to changes in the structure of the markets for loans and CLOs. Two regulations in particular that were part of the Dodd‐Frank Wall Street Reform and Consumer Protection Act have had a particularly direct impact on the CLO market.
The Volcker Rule
One of the sections of the Dodd‐Frank Wall Street Reform and Consumer Protection Act is commonly referred to as the Volcker Rule. It is named after Paul Volcker, the former Federal Reserve Bank chairman whose ideas inspired that section of the Act. One of the goals of the Volcker Rule was to prohibit banks from investing in hedge funds or private equity funds, which were deemed to be overly risky for the safe‐and‐sound investment practices that are typically expected of banks.
At first blush, it would seem that this rule is unrelated to CLOs. After all, CLOs are not hedge funds or private equity funds. However, the issue arises from the fact that it is difficult to clearly define what constitutes a “hedge fund” or a “private equity fund.”
The definition that the legislators opted to use to define prohibited funds also ended up including most CLOs. However, the regulators did provide an allowance for CLOs, provided that they qualify as so‐called loan securitizations. This requires, among other things, that the collateral of the CLO be comprised exclusively of loans with no exposure to bonds or other securities.
Historically, banks have tended to be significant investors in CLO tranches. Since the Volcker Rule effectively requires that any CLO in which a bank invests qualifies as a loan securitization, the entire CLO market has generally shifted to be 100% loan‐only. (Previously, there had been certain benefits, such as diversification, to the allowance of small proportions of bonds or other non‐loans within the underlying CLO portfolio. However, due to the Volcker Rule, CLOs with such allowances are generally no longer issued in practice.)
Skin‐in‐the‐Game: Risk Retention Requirements
Another section within the Dodd‐Frank Act set new requirements for securitization sponsors to retain risk in the funds that they issue and essentially requires a CLO manager to retain 5% of the balance of any new CLO issuance. The concept is that if the manager is exposed to investment risk with some of their own skin‐in‐the‐game, they are more likely to select loans that are well‐underwritten and expected to perform.
The risk retention requirement makes it much more onerous for managers to issue new CLOs. In the past, managers had sometimes co‐invested in their own transactions, but there was no regulatory requirement for them to do so. Issuing more CLOs simply added to the manager's assets under management (AUM) and enhanced the manager's fee revenue. The manager's interests were aligned with investors through both their general reputation as well as the CLO's fee structure of senior, subordinate, and incentive management fees. In contrast, with the new requirements, a typical CLO of $500 million will require the manager to make a $25 million cash investment in the transaction. Consequently, CLO issuance will therefore be transformed into a very capital‐intensive business, changing the manager's economic benefit of new issuance and potentially placing CLO management out‐of‐reach for some managers.
The outcome of the risk retention requirements may be that the universe of active CLO managers will shrink. Existing CLO managers will need to have an adequate source of capital to be able to fund new issuance, and they may need to reevaluate the economic benefits of continuing to issue additional transactions. New CLO managers will find that these requirements further raise the barriers to enter the CLO market. In addition to the staffing, analytical, operational, and marketing issues that new managers have always needed to address, they must now also find a significant source of capital to enable them to bring their CLOs to market. The CLO manager and issuance landscape is therefore likely to evolve as a result of this regulation.
CONCLUSION
CLOs present investors with the ability to gain exposure to leveraged loans in a way that buffers the risk for senior investors and enhances the return for junior investors. CLOs boast a relatively long and impressive track record, which demonstrates their resilience as an asset class. However, past performance should not lead one to underestimate the potential risks. The performance of a CLO ultimately depends upon the performance of speculative‐grade loans and the active management of the collateral manager. Investors should do their homework prior to making an investment, including manager due diligence, document review, and quantitative cash‐flow analysis, in addition to conducting ongoing surveillance throughout the life of the transaction. With a deeper understanding of a CLO's construction, features, considerations, and risks, an investor can make informed decisions about this attractive and resilient asset class.