In recent months, portfolio managers of cash flow CLO vehicles have been faced with the unenviable task of keeping their CLOs afloat in an environment in which bank loan prices in the global loan markets have declined to unprecedented levels. This task has become increasingly difficult, due in large part to certain provisions in the underlying CLO documentation that, in today's illiquid market, no longer operate as intended. At a time when it is more important than ever for managers to actively manage the credit risk in their CLO portfolios, CLO managers have suddenly found their hands tied by provisions that, while initially conceived to safeguard the credit quality of CLO portfolios, now serve as a disincentive for managers to replace credit impaired loans with stronger ones. This article will examine the dilemma confronting CLO managers who wish to improve their portfolios by trading rapidly deteriorating loans for better-performing so-called "deep discount" loans.
The OC Tests
Any discussion of the deep discount dilemma warrants an overview of one of the most fundamental protections granted to cash flow CLO investors - the overcollateralization (OC) tests. In a typical cash flow CLO, each class of notes, other than the most subordinated "equity" class, has a corresponding OC test, which will be satisfied if the "OC test ratio" is at least equal to a specified "trigger" ratio, usually expressed as a percentage. In its most basic form, the OC test ratio is determined by dividing the "principal balance" of the assets in the CLO portfolio (subject to certain exceptions discussed below) by the sum of (i) the aggregate principal balances of the CLO notes of the relevant class, and (ii) the aggregate principal balances of the CLO notes of all other classes of notes that are senior or pari passu to such class. The higher the OC test ratio, the greater the cushion between the potential available proceeds from the CLO's underlying collateral and the principal owed to a CLO investor. Stated another way, a high OC test ratio for a class of CLO notes is indicative of a strong likelihood that the CLO's asset portfolio will generate sufficient cash flow to pay such class of notes in full.
If the OC test for a particular class of CLO notes is not satisfied on a determination date, an automatic deleveraging mechanism is activated, whereby interest collections from the CLO portfolio, which normally would be allocated to the payment of interest on the more junior classes of notes and other fees, expenses and obligations, will instead generally be diverted to reduce the principal balance of the more senior classes of notes in order of seniority until the OC test failure is cured.
The diversion of interest proceeds brought about by an OC test failure can have a profoundly negative impact upon the returns of investors in the lower rungs of a CLO's capital structure, equity holders in particular. This impact is even more severe in CLOs that include a pay-in-kind (PIK) feature, pursuant to which the unpaid interest resulting from the diversion of interest proceeds brought about by an OC test failure is added to the principal balance of the affected class of notes. This capitalization of interest has the effect of magnifying the negative impact of the OC test failure on the returns of the more subordinated classes of notes.
Many CLOs also include another, separate OC test trigger, the failure of which results in an event of default. These triggers are customarily set at such a level that this OC test will only fail under conditions of severe portfolio deterioration. The consequences of an event of default can be severe. After an event of default, a specified percentage of the CLO's "controlling class" - normally a majority or supermajority of the holders of the CLO's senior-most class of liabilities - generally have the right to direct the CLO trustee to declare all outstanding CLO notes to be immediately due and payable.
After an acceleration of payment on the notes has been directed, a specified percentage of the controlling class (in many cases, together with a specified percentage of one or more additional classes of noteholders) may direct the trustee to commence a liquidation of the entire CLO portfolio, the proceeds of which are required to be used to pay, in accordance with the priority of payments specified in the related indenture, the accrued interest on and principal of each class of notes in full, generally in order of seniority, together with certain transaction costs and expenses. In today's illiquid market, where performing loans are selling at a fraction of their par values, the liquidation of a CLO portfolio could result in all but the most senior noteholders recognizing only pennies on their investment dollar.
As stated above, the numerator of the OC test ratio is generally the sum of the aggregate principal balance of each asset in the CLO portfolio. For purposes of calculating the OC test ratio numerator, the general rule is that each CLO asset's par value (i.e., the outstanding principal balance of the asset) is used for such computation. There are, however, exceptions to this general rule, including the following:
* the principal balance of any "defaulted" loan (i.e., any loan that is not paying principal and/or interest or suffers from another severe infirmity specified in the CLO documents) is generally deemed to be the lesser of its market value and its recovery rate (calculated by reference to certain rating agency imposed formulae);
* the principal balance of any "deferring" loan (i.e., any loan that has deferred payment of interest for a significant period of time, as specified in the CLO documents) is generally deemed to be the lesser of its market value and its recovery rate;
* the principal balance of any deep discount loan (i.e., any loan purchased below a designated percentage of par, typically 80% to 85%) is generally deemed to be its purchase price; and
* the principal balance of any loan deemed to be included in a CLO's "excess 'Caa'/'CCC' basket" (a concept that will be addressed at length below under the heading "Manager's Dilemma") is generally deemed to be either its market value, or the lesser of its market value and its recovery rate.
Since each of the loan types set forth above is valued at less than its par value for purposes of calculating the OC tests, the presence of any of these loans in a CLO portfolio will negatively impact the OC tests' numerator, thereby increasing the probability of triggering an OC test failure.
A primary objective of a CLO manager is to seek to improve or maintain the credit quality of the CLO portfolio while managing the portfolio in such a manner as to avoid OC test failures and events of default. Under normal market conditions, these two objectives were rarely in conflict.
As noted above, deep discount loans are valued at their purchase price for purposes of calculating the OC test ratio numerator. The most obvious consequence of using this valuation methodology in the current market environment is that it creates a disincentive for CLO portfolio managers to make certain trades that would, in the manager's reasonable judgment, improve the overall credit quality of the CLO portfolio.
To illustrate this point, we will use the example of a CLO portfolio containing a loan that did not constitute a deep discount loan as of the date of purchase (and would therefore not constitute a deep discount loan for purposes of the OC test), but is rapidly decreasing in credit quality and currently trading at 55% of par. We will also assume that the loan in question is not subject to any other exception to the general rule of valuing a loan at par for the purposes of calculating the OC test ratio numerator. If the manager's CLO documents follow the prevailing market convention of designating assets purchased at below 80% to 85% of par as deep discount loans, then the manager would be effectively discouraged from selling the deteriorating loan at its market value (55% of par) and purchasing a loan of a better credit quality at a price of greater than 55% of par but less than the minimum deep discount threshold. The reason for this disincentive is simple -making the trade would force the manager to value the substituted loan at its purchase price rather than its principal balance for the purposes of calculating the OC test ratio numerators, thereby negatively impacting the OC tests.
Substituting a credit impaired loan for a stronger deep discount loan, while not ideal, may in certain cases be in the best interest of the CLO, even if such substitution triggers an OC test failure. While such substitutions could trigger a reallocation of interest proceeds to the most senior noteholders in the short term, they are likely to improve portfolio quality in the long run. This example, however, only serves to underscore the inherent deficiencies with deep discount loan provisions in the current market environment. As this example clearly demonstrates, by meeting its objective of maintaining or improving the credit quality of the CLO portfolio, the CLO manager would simultaneously be failing its objective of maintaining compliance with the OC tests and (in circumstances of widespread credit impairment across the CLO portfolio) avoiding events of default.
The consequences of carrying deeply discounted loans at their purchase prices are further exacerbated as more and more loans in cash flow CLO portfolios are downgraded to 'Caa'/'CCC' ratings, since the majority of such CLOs contain maximum thresholds (typically 5.0-7.5% of the principal amount of the CLO portfolio collateral) for holding Caa/CCC-rated loans. To the extent that a CLO portfolio exceeds its stated threshold, loans in excess of the threshold (the "excess 'Caa'/'CCC' basket") are valued at their market value for the purposes of the related CLO's OC tests. Since CLO documents typically count the 'Caa'/'CCC'-rated loans with the lowest market values toward the excess 'Caa'/'CCC' basket, this mark-to-market feature can have drastic OC test implications, particularly in today's illiquid market environment, where as of mid-April 2009, the average concentration of 'Caa'/'CCC' loans in CLO portfolios was more than 10.0%.
An increase in rating migrations to 'Caa'/'CCC' levels, when combined with the rapid deterioration of loan values, puts CLO portfolio managers in a difficult bind. On one hand, such managers could continue to hold their excess 'Caa'/'CCC' basket loans, running the risk that further market value erosion of such loans will lead to OC test failures. Alternatively, such managers could sell the 'Caa'/'CCC'-rated loans and use the sale proceeds to purchase stronger, more highly-rated loans, which, in a loan market where average bids for performing loans have recently hovered in the mid-60s, will most likely be deemed deeply discounted pursuant to the terms of the related CLO documents. Since these deeply discounted loans will be valued at their purchase prices, by substituting an excess 'Caa'/'CCC' loan for a deeply discounted loan, such managers will essentially lock in a significant reduction in their CLO's OC tests.
Managers of CLOs with 'Caa'/'CCC' exposure below the maximum 'Caa'/'CCC' threshold in many cases have an even greater disincentive to trade their 'Caa'/'CCC' loans for stronger credits, even if such sales would benefit the long-term credit quality of the CLO. As long as the concentration of 'Caa'/'CCC' loans in a particular CLO portfolio remains within the stated threshold, all such 'Caa'/'CCC' loans will be valued at par for the purposes of the OC test ratios. In an environment where most substitute loans available for purchase are likely to be deemed deeply discounted, managers are often effectively discouraged from trading their 'Caa'/'CCC' exposure, since doing so would negatively impact the CLO's OC tests.
The dilemma described above is the direct result of provisions in CLO documentation that were drafted before the current liquidity crisis and, as a result, no longer operate as originally intended. At its origin, the notion of valuing a deep discount loan at its purchase price rather than par for purposes of the OC test ratios was aimed at protecting the credit quality of CLO portfolios by preventing managers from avoiding OC test failures by manipulating par. To use an example, if deep discount assets were not valued at their purchase prices but rather at par for purposes of the OC test, a manager could artificially inflate the numerator of the OC test ratios by using the proceeds from the sale of a strong asset trading at 100% of par to purchase two weak assets trading at 50% of par. Although such a trade would vastly improve the OC test ratios, thereby decreasing the possibility of an OC test failure, it would also significantly weaken the quality of the CLO collateral pool.
Prior to the liquidity crisis, the market value of a loan was, for the most part, a direct reflection of its credit quality. When the 80% to 85% of par threshold for deep discount loans was originally formulated, loans trading at 80% to 85% of par were considered to be distressed. In today's highly illiquid market, where performing loans are trading at abnormally low levels, loans trading at 80-85% of par are often of superior credit quality. In 2008, the Loan Syndications and Trading Association (LSTA) issued a Loan Market Advisory urging loan buyers and sellers to look beyond price as the dispositive factor in determining whether a loan should trade as "distressed." Because deep discount provisions no longer operate as intended in an illiquid market setting, these provisions have brought about the unintended consequence of providing a disincentive for CLO managers to replace weak assets with stronger ones. Moody's Investors Service, in its June 24, 2004 publication entitled Haircuts for Excess Caa Assets and Deep Discount Obligations seemed to verify this conclusion, stating: "[t]o the extent that particular assets and markets suffer from a lack of liquidity, then market value treatment may not be appropriate."
The continuation of this article will be published in the next edition focusing on the industry's call for reform and suggested solutions.
(c) 2009 Asset Securitization Report and SourceMedia, Inc. All Rights Reserved.