Robust demand for leveraged loans has outstripped supply for most of the past few years and shifted negotiating power in borrowers' favor. In this environment, many companies have taken the opportunity to recapitalize their balance sheets, lower loan spreads, revise loan terms, and push out debt maturities. Not surprisingly, this has produced weaker investment protections for lenders as debt leverage has increased and loan terms have loosened and become more bond-like. Key lender concerns include:

• Covenant-lite loan structures,
• Growing EBITDA add-backs,
• Larger and more flexible incremental loan baskets, and
• Flexible language around covenants such as restricted payments, investment baskets, and asset sales.

Some of these issues are easier to directly address in our ratings at inception. For example, we believe that covenant-lite loans have a higher risk of loss given default than loans with financial maintenance covenants, and our recovery methodology will typically produce somewhat lower recovery outcomes for companies with covenant-lite loan structures, all else equal. This is consistent with our research showing lower observed actual recovery rates for covenant-lite loans versus non-covenant-lite loans for firms emerging from bankruptcy in the U.S. from 2014 through 2017. Similarly, we believe that EBITDA add-backs artificially inflate EBITDA and have conducted research showing that management-adjusted EBITDA (including add-backs) does not provide a good proxy for future profitability or leverage. Accordingly, our ratings reflect our independent views of a company's EBITDA, leverage, and cash flow, regardless of how they are presented for marketing purposes or defined in debt agreements. We also note that rising debt leverage and heavy debt-financed acquisition activity in recent years has translated into low issuer credit ratings, as illustrated by a sizable increase in the number and percentage of corporate issuers rated 'B' or lower. In addition, a shrinking cushion of junior debt and high leverage have resulted in lower recovery ratings on first-lien debt. (For more detail on these topics see the Related Research section).

Looking beyond these issues, we think looser loan terms can pose other risks that may be less obvious and that are difficult to predict and quantify. We answered questions about some of the key risks and how we approach them analytically, as well as about our recovery methodology and analysis and how this can provide a framework for assessing risks that may not be directly factored into the recovery ratings.

Frequently Asked Questions

Beyond the risk that EBITDA add-backs may overstate EBITDA and understate debt leverage, what other risks can EBITDA add-backs present?

To start, EBITDA add-backs can increase the potential for future event risk because "adjusted EBITDA" (as defined in debt agreements) is typically used to set "basket" capacity under various negative covenants, such as restricted payments. So significant EBITDA add-backs will likely increase basket sizes and the flexibility provided under various covenants. As a result, borrowers may have more room to take future actions that could increase credit risk by adding incremental debt, distributing value to equity holders, investing in third parties, or transferring assets to unrestricted subsidiaries.

Adjusted EBITDA also often includes add-backs for items that are ultimately cash operating costs and may be recurring (management fees, restructuring charges, or software development costs, for example). As such, we view the basis for adding these items back as dubious, since EBITDA should be a proxy for profitability and cash flow.

Further, many add-backs--such as those covering expected cost savings and synergies--are subjective and based on borrower estimates. As we see it, this can present two problems. First, these estimates may be overly ambitious and not ultimately realized--but even unrealized estimates may still provide the borrower with additional flexibility under various covenant baskets. Second, the borrower typically does not disclose the estimates regularly, so it can be difficult for investors to quantify adjusted EBITDA and the resulting flexibility provided under various covenants in its debt agreements. As a result, these adjustments can become a heated source of disagreement between the borrower and its creditors over a transaction's life.

The ongoing litigation between PetSmart Inc. and its lender group over the company's transfer of some of the equity in its Chewy Inc. subsidiary is an illuminating example of how the interpretation of adjusted EBITDA, the sizing of restricted payments and investment baskets, and other key terms can differ between the borrower and lenders. In PetSmart's view, the equity transfers were permitted under the negative covenants in its credit agreement and bond indentures. The transfers distributed 20% of the equity in its Chewy subsidiary to PetSmart's shareholders--which triggered the release of the guarantees and asset pledges from its Chewy subsidiary--and another 16.5% of Chewy's stock to a restricted subsidiary, which also removed this portion of Chewy's equity from the collateral package. PetSmart's lenders, however, contended that the equity transfers violated the terms of its debt agreements, and thus the Chewy guarantee and the liens on Chewy's assets and all of its stock remain in place. Following PetSmart's announcement that it completed the Chewy equity transfers in June 2018, we revised our recovery ratings on PetSmart's secured debt, although these ratings may be updated if the lenders successfully reverse these transactions. The PetSmart case highlights the potential risks and disagreements that may result from EBITDA add-backs (including how they are quantified, whether the add-backs are appropriate, and when the benefits must be realized) and the flexibility this may provide borrowers to undertake transactions that may increase credit risk to lenders (for more details, see our case study "The PetSmart Case: A Deep Dive Into Its Equity Transfer Of Chewy Inc.," published Nov. 8, 2018).

What are incremental loans and why are they a concern to leveraged loan investors?

Incremental loans refer to the capacity provided to borrowers to obtain additional loans on a pari passu basis with existing loans. These loans are uncommitted, so existing lenders don't need to provide the funding. At the same time, existing lenders cannot block the loans if other investors are willing to provide the capital. Credit agreements typically allow the borrower to obtain incremental loans up to a specified dollar amount without being subject to incurrence covenants (so-called "free and clear" baskets), plus an unquantified "additional amount" if the borrower meets relevant incurrence tests on a pro forma basis (for example maximum total or senior leverage). These incurrence tests are typically set at, or inside of, the ratio levels at loan inception.

The flexibility to add incremental loans has been a standard feature in the loan market since before the global financial crisis, but lately these loans have become a bigger concern to lenders for a few reasons. To start, incremental loan capacity is generally sized based on some percentage (often 50%-100%) of the greater of EBITDA at inception and EBITDA on a trailing 12-month basis. Because the definition of EBITDA in loan documents has expanded, the size of incremental loan facilities has increased accordingly. In addition, it's less likely that any incremental loans will be restricted by financial maintenance covenants in the loan documents now that the vast majority of term loans are covenant lite, and financial maintenance covenants that still exist under most revolving loan tranches are generally springing covenants. With springing covenants, these tests are only "run" if certain conditions are met--typically a cash draw of at least a certain percentage (often 25%-40%).

Further, "most favored nation" (MFN) pricing terms, which restrict borrowers from paying incremental lenders more than the original lenders, have also weakened. From the standpoint of an existing lender, the dilution of MFN-type protections--such as carve-outs or short sunset provisions--will make it easier for borrowers to obtain incremental loans by offering richer pricing to the new lenders without increasing the pricing on the existing loans, which suggests existing lenders may not be appropriately compensated for the risk.

How do incremental loan facilities affect S&P Global Ratings' views on credit risk and how is this factored into its ratings?

While the ability to add incremental loans is a standard feature of leveraged loans, predicting if, when, and how this flexibility might be used--if it used at all--is very difficult. Further, the amount of any borrowings and the funding purpose are important factors for our recovery analysis. Tapping incremental loan capacity to purchase EBITDA-generating assets or operations (which may increase an entity's enterprise value in a default scenario) would affect recovery expectations differently than using the proceeds to pay a shareholder dividend or to repay junior debt. In addition, whether incremental loan borrowings are accompanied by additional proceeds from new junior debt offerings or an equity injection could also significantly affect recovery expectations. As a result, we generally do not directly factor this "event risk" into our recovery analysis without specific or compelling information. Rather, we address this as part of our ongoing surveillance and update our ratings as appropriate.

From a default risk (or issuer credit rating) standpoint, our financial risk profile assessment of a borrower may capture the risk of potential incremental debt, for example when we believe a company will likely pursue aggressive financial strategies that will increase leverage and financial risk. Our financial policy assessment typically caps our financial risk assessment of private equity-owned firms at highly leveraged (our riskiest classification), even if current leverage and cash flow measures indicate lower financial risk.

Is there empirical evidence showing that the use of incremental loan capacity results in higher credit risk?

As suggested above, a borrower's use of incremental loan capacity does not necessarily translate into higher credit risk from a default risk or a loss given default perspective. We don't currently have statistics on how frequently the use of incremental loan capacity has led us to lower issuer credit ratings, recovery ratings, or both (although we hope to do some research to shed more light on this issue). However, some statistics on add-on loans from S&P Global Market Intelligence's Leveraged Commentary & Data unit provide useful context on incremental loan usage. To be clear, the statistics on add-on loans cited below include all increases to exiting loan facilities, not necessarily just those resulting from using the incremental loan capacity under existing loan agreements. While we suspect most of these add-on loans were the result of using incremental loan capacity under existing agreements, some of them likely required lender approval.

As shown below, term loan B add-on volumes represent a modest to moderate portion of term loan financing. Annually, they represented 5%-15% of total loan volumes from 2007 through 2018 (chart 1). From a before and after standpoint (chart 2), they represented just under a 20% increase in the size of the existing term loans in 2018. Lastly, the increases in 2018 were somewhat larger as a percent of the original term loan size (chart 3), at least in the fourth quarter when several large add-ons (relative to original term loan balances) were completed. Some of the more significant increases as a percentage of the original term loan amount suggest that lender approval of the add-on loans may have been required.

From a funding purpose perspective (chart 4), 62% of incremental loan proceeds in 2018 were used to fund mergers and acquisitions (M&A) activity, with debt-financed dividends representing another 16% of volumes and refinancing a similar proportion (refinancing includes repaying existing first-lien and junior debt).

Chart 1

Chart 2

Chart 3

Chart 4

While these statistics on add-on loans can't be directly extrapolated to explain the impact on default risk or recovery risk by company or debt instrument, we think they still provide some useful perspective:

• Dividends: We can assume that the 16% of deals related to dividends in 2018 increased total debt and first-lien debt, as well as total and first-lien debt leverage. As such, these transactions may have increased default risk and recovery risk.
• Refinancings: We can assume that the 16% of deals for refinancings did not materially change debt levels or default risk, but if incremental senior debt is used to repay junior debt it may increase recovery risk for the senior lenders.
• M&A: For these transactions, the greater loan size was at least accompanied by an in increase in assets or EBITDA. Even so, the impact on default and recovery risk is unclear without knowing if there were any other funding sources (junior debt, equity, or cash) or if total debt leverage or first-lien debt leverage increased.

As part of our ongoing ratings surveillance, we review each transaction and update our ratings if our views on credit risk (default risk as embodied by the issuer credit rating, or recovery ratings on debt instrument basis) changes. One example is The Hillman Cos. Inc. Hillman's leverage was expected to remain high after it completed a sizable acquisition using a $365 million add-on term loan. As a result we lowered our issuer rating credit rating to 'B-' from 'B' and revised our recovery rating on its first lien bank facility to '3' (rounded recovery percentage 60%) from '2' (70%).

Another is Equian Buyer Corp., which also used an add-on term loan (of $315 million) to fund an acquisition. In this case we affirmed the 'B' issuer credit rating because we viewed the increase in leverage as more moderate and short-term. Even so, we did lower our rounded recovery estimate on the bank debt to 50% from 60% while affirming our '3' recovery rating.

A third example is Bass Pro Group LLC, which used an $800 million add-on term loan plus a $650 million draw on its (unrated) asset-based lending (ABL) facility and a small amount of cash to redeem $1.3 billion in preferred stock and distribute $127 million to shareholders. We affirmed the 'B+' issuer credit rating because leverage was largely unchanged since we were already treating the preferred stock as debt for leverage ratio purposes. However, we still revised our recovery rating on the first-lien term loan to '3' (60%) from '2' (75%) based on the increase in the term loan and first-lien leverage.

How can flexible language in loan documents on negative covenants, restricted payments, and similar baskets affect credit risk? How does you assess these risks?

The J. Crew and PetSmart collateral transfer cases illustrate the potentially significant event risks that lenders may face as a result of more flexible negative covenant and restricted payment terms (see related research for details on both cases). At the same time, these cases also illustrate the challenges in assessing these risks and factoring them into our ratings on a prospective basis. In both cases, the borrowers took a complex series of actions in reliance on various baskets under their negative covenant and restricted payment terms to effectuate collateral transfers and other capital structure changes. In addition, these transactions were quite controversial and resulted in litigation with the lender groups that disputed the legality of the actions.

In the case of J. Crew, the company transferred its intellectual property (IP) to a non-credit party subsidiary using two exceptions in the negative covenants section of the secured loan agreement, and then borrowed against the IP to refinance junior debt. This weakened term loan lenders' collateral position, and led to litigation between lenders and the company. The case was eventually settled out of court as part of a broader restructuring agreement that included a modest repayment of the bank debt at par, a tightening of various document provisions, and an extension of certain debt maturities.

Ultimately, these cases are examples of event risks that in our view are relatively uncommon at this point and, perhaps more to the point, extremely difficult to reasonably foresee and quantify in our recovery ratings. Consequently, we typically don't directly factor these risks into our ratings on a forward-looking basis, but capture them in our ongoing ratings surveillance process. After they transferred important parts of the collateral packages, we revised our recovery ratings on J. Crew's first-lien debt to '5' (15%) from '4' (40%) and on PetSmart's first-lien debt to '4' (45%) from '3' (60%). Beyond specific rating actions like these, we continue to monitor these issues and consider when and how it may be appropriate to address these issues in our ratings (e.g. pervasive utilization of such provisions, patterns in sponsor behavior).

Following the J. Crew and PetSmart cases, we believe investors seem more focused on pushing back on certain provisions, such as blocking or restricting the sale or transfer of key assets or limiting the borrower's ability to subsequently re-designate guarantor or restricted subsidiaries as non-guarantor or unrestricted subsidiaries. Further, following the credit and equity market disruption in late 2018, it appears investors have had more success in pushing back on aggressive structuring proposals, as was the case during the tighter debt market conditions in early 2016. Even so, we think the ability to successfully push back on these terms is likely to ebb and flow with supply and demand, and it appears that the longer-term trend has been towards weaker loan terms that are more bond-like.

What other loan terms may be weaker than they were traditionally and may increase a borrower's ability to take actions that could increase credit risk?

Other notable terms that could increase credit risk include more flexibility to increase incremental debt or restricted payment capacity by stacking various baskets (e.g. separate baskets for incremental loans, general debt, receivables financing, etc.). For example, we noted that the loan documents for Refinitiv provided specific carve-outs that would allow the company to increase debt by $6 billion, or more than 40% from the amount outstanding at inception.

Similarly, the ability to subsequently reclassify incremental debt borrowed under a fixed debt basket is another way that some loan documents are becoming more bond-like and may allow borrowers to maximize debt capacity. Under these provisions, borrowers can replenish fixed debt incurrence capacity by reclassifying prior borrowings as being permitted under an unlimited debt incurrence basket (that is subject to a pro forma ratio test) if the company's financial performance later improves to the point where the requisite incurrence test would be satisfied.

Some loan documents also provide borrowers with considerable flexibility to sell assets without necessarily needing to reinvest the proceeds in new collateral or to prepay the loans. Terms that may limit lender protections and maximize borrower flexibility around asset sales include long reinvestment periods, the ability to sell capital assets and reinvest in working capital assets or unrestricted subsidiaries, step downs in required repayments or reinvestments tied to a financial ratio test, or carve-outs for certain assets or a asset sales below a designated threshold. Ultimately, weaker asset sale terms increase the possibility that material assets will be sold without a corresponding reinvestment in the business or decrease in debt.

What else is factored into your recovery ratings and analysis? How can S&P Global Ratings' recovery analysis provide a framework for assessing risks that may not be directly factored into the recovery ratings?

While we cannot realistically capture all of the potential risks to a lender's leveraged loan exposure in a way that is quantifiable and reasonable, we also believe that our recovery ratings and analysis provide a context for gaining insight into the potential impact of the types of borrower-driven events that aren't directly factored into our ratings.

To set the stage, our methodology centers on the two fundamental credit risks for corporate debt: the relative probability of default, and loss given default (or its inverse, recovery given default). Together, these two risk measures produce expected loss, which is the mathematical quantification of credit risk. For leveraged corporates, we provide a separate indicator for each--with our issuer credit or entity ratings providing an indication of relative default risk and our recovery ratings providing a debt instrument-specific estimate of recovery prospects. Because an issue rating is a blend of these two opinions, providing separate indicators for each of these risks enhances transparency by unpacking the inputs to a specific issue rating.

Our recovery methodology uses a fundamental approach to assess recovery prospects that includes factoring in our company- and sector-specific knowledge and how a company's debt and organizational structure will likely affect relative priorities and the distribution of value in a default scenario. A key part of this is evaluating where the value lies within a company's organizational structure and whether there may be:

• Material subsidiaries that do not directly provide credit support under the proposed financing terms (non-guarantors);
• Important assets that are not pledged to secured creditors; and
• Other debt instruments or liabilities that have a priority claim--for example, because of structural seniority or a priority lien--to important sources of value (assets or operating subsidiaries).

These important details are often not clearly discussed or quantified in term sheets and marketing documents. In our view, understanding where within an enterprise the EBITDA-generating assets reside, and whether lenders have recourse to those assets, enables assessments of recovery prospects that go beyond debt instrument "labels."

We strive for transparency in our recovery analysis so that investors can better understand it and draw their own conclusions. This includes summarizing our key valuation assumptions in a default scenario and showing how we'd expect this value to be distributed in a default scenario. In our view, this is important because valuation is generally quite subjective and potentially volatile. Moreover, we provide rounded recovery percentage estimates that indicate in rounded 5% increments where our recovery estimate sits within our recovery rating bands. This provides transparency around how sensitive the recovery ratings may be to potential changes in the debt structure or the recovery valuation. For example, a recovery rating on a loan that is at the low end of one of our recovery ranges may be at risk of being revised if a company does a debt-financed dividend that increases first-lien leverage but does not affect its intrinsic value in a default scenario. Further, our recovery estimates may provide investors with a benchmark for considering a distressed exchange offer or selling a debt instrument for a troubled entity. Lastly, we monitor our recovery ratings over time to try to capture events and new risks into our analysis as they arise.

Given all of these factors, what are your expectations for first-lien loan recoveries going forward?

We expect future first-lien loan recoveries to be well below the strong actual recovery rates that have been exhibited historically. Historical actual average recoveries for first-lien loans in the U.S. are generally cited in the 75%-80% area, with median recoveries of roughly 100%. In contrast, our recovery ratings on new first-lien debt in the U.S. have declined into the mid-60% area over the past two years due to a combination of rising total and first-lien debt leverage, shrinking junior debt cushions, and the predominance of covenant-lite loan structures. The trends are reflected in statistics for leveraged loans to European corporates as well. European loans show historical average and median actual recoveries in the mid-70% and mid-80% area, respectively, compared to average estimated recoveries in the high-50% area as indicated by our recovery ratings. The somewhat lower recovery percentages for European corporates primarily reflect more first-lien-only loan structures and smaller junior debt cushions.

As detailed above, however, our recovery ratings generally do not factor in potentially significant event risks related to other looser and more flexible loan terms. These terms may allow aggressive borrowers to undertake actions that could harm lender recovery prospects (for example, by distributing value to shareholders, transferring collateral to unrestricted subsidiaries, raising additional debt, repaying junior debt, or selling assets without reinvesting the proceeds or repaying loan debt). Even though we can't forecast or quantify when these actions will be taken, we certainly expect that some issuers will exploit these terms, but still default. This should put further downward pressure on first-lien loan recoveries, and may substantially increase the dispersion of actual recoveries on first-lien loans. If these expectations come true, it will be interesting to see if this makes loan investors demand tighter loan terms in the future. Ultimately, stronger traditional loan terms and structures helped support the strong historical recovery rates that made first-lien loans an attractive investment in the first place.

Related Research

• Leveraged Finance: U.S. Leveraged Loan Q4 Update, Jan. 22, 2019
• Credit FAQ: A Closer Look At The Drawbacks Of EBITDA Add-Backs, Dec. 4, 2018
• The PetSmart Case: A Deep Dive Into Its Equity Transfer Of Chewy Inc., Nov. 8, 2018
• Credit FAQ: A Closer Look At The Debt Financing In Refinitiv's Acquisition Of Thomson Reuters' Finance & Risk Business, Oct. 3, 2018
• When The Credit Cycle Turns: The EBITDA Add-Back Fallacy, Sept. 24, 2018
• Credit FAQ: A Closer Look At How Covenant-Lite Structures Affect Recoveries For Institutional Loans, July 10, 2018
• Lenders Blinded By Cov-Lite? Highlighting Data On Loan Covenants And Ultimate Recovery Rates, April 12, 2018
• J. Crew's Intellectual Property Transfer: An Update For Distressed Lenders, Sept. 12, 2017
• Lean Senior Debt Cushion Threatens Recovery Prospects For U.S. Leveraged Loans, Nov. 30, 2017