We recently reviewed ultimate recovery rates for covenant-lite (loan tranches without financial maintenance covenants) and non-covenant-lite first-lien term loans for over 67 entities that emerged from Chapter 11 bankruptcies between 2014 and the first half of 2020. Our findings are similar with those from our previous study of 2014-2017 bankruptcies covering 28 firms.

Our updated dataset has more than doubled to 67 companies. We added 32 companies that emerged from bankruptcy from 2018 through the first half of 2020, and another seven credits that emerged before 2018 for which we were able to obtain sufficient information. In addition to covering empirical evidence on recoveries for cov-lite and non-cov-lite term loans, we also looked at:

• The dominance of covenant-lite term loan structures in institutional leveraged loans and how financial maintenance covenants can help protect term lenders from pricing and event risks;
• Why covenant-lite loans loan exist and persist, and how they can delay or mitigate default risk, at least during periods of short-lived or moderate stress followed by a rebound; and
• How our recovery rating methodology factors in the presence of debt with and without financial maintenance covenants.

To be clear, covenant-lite loans are not devoid of all forms of covenants. Loan terms still require the borrower to meet incurrence covenants before taking various actions (such as adding debt, pledging assets, or making restricted payments), and thus provide limits on what a borrower can and cannot do. Even so, these limits may be less restrictive than they appear because there are usually a variety of "free and clear" baskets that may provide borrowers with additional flexibility to incur debt or make restricted payments without meeting incurrence covenant tests.

Ultimate Recovery Rates From 2014 Through The First Half Of 2020

Data overview

We derived actual recovery rates (on a nominal, as opposed to discounted basis) based on the implied valuations and recoveries indicated in our review of bankruptcy documents, such as Chapter 11 disclosure statements, plans of reorganization, and asset sale disclosures. Our universe excludes distressed exchanges or out-of-court restructurings, as well as companies for which we couldn't obtain reliable recovery data. This approach is consistent with our approach for measuring the performance of our recovery ratings relative to actual recoveries. Specifically, we tracked the ultimate recovery rates of first-lien term loans issued by 67 rated companies that exited bankruptcy between January 2014 and June 2020, either through a reorganization or asset sales/liquidation.

Approximately two-thirds of the term loans in our study have no financial maintenance covenants--directionally consistent with the broader market trend. The loans in the dataset are also predominately institutional term loans with only two pro rata term loans, both of which had maintenance covenants. The dataset covers a wide range of industries, with an outsized (though not dominating) representation from the oil and gas sector (17 entities, or about a quarter of the entities), followed by restaurants/retailing (13 entities, mostly retailers). The oil and gas companies covered in our dataset are not primarily exploration and production (E&P) companies, which one might expect considering how many E&P companies have defaulted in recent years. Instead, there are more than twice as many oil field services companies. That is because E&P companies predominantly rely on reserve-based lending (RBL) revolving credit facilities as their main source of senior secured financing, as opposed to borrowing through the more common first-lien institutional term loans that we focus on in this report. About a quarter of the constituents emerged from bankruptcy in 2019, followed by 2017 and 2016 when a wave of oil and gas companies completed their restructuring and exited Chapter 11.

Chart 1a

Chart 1b

Cov-lite versus non-cov-lite

Tables 1 and 2 detail the nominal recovery rates for first-lien term loans both in aggregate and broken down by the presence of financial covenants, most in the form of a leverage test or fixed-charge coverage test. While our dataset is not exhaustive of all the Chapter 11 filings during the observation period, the empirical evidence suggests that the absence of financial maintenance covenants appears to hurt ultimate term loan recovery rates. When calculating averages, we did not weight individual recovery by debt size, meaning our recovery averages do not skew toward borrowers with larger amounts of debt, nor toward ones that were the most highly levered. Both tables also show the overall statistics for the full dataset, as well as the results excluding the oil and gas credits and/or retail/restaurant credits.

Table 1

Among the 67 companies that exited bankruptcy between 2014 and the first half of 2020, the average first-lien recovery was 72% on an issuer count basis; the median is slightly higher at 74%. Oil and gas outperformed other sectors by both average and median measure, with more than half of the sector's loans repaid in full. Helping to vault oil and gas borrowers to the top is the generous amount of junior subordination seen in their debt mix. The sector's median first-lien cushion (defined as the percentage of a company's total debt claims that are junior to the first-lien, as outstanding at default) was 68%, double the 34% median of the full dataset. As expected, oil and gas borrowers' going-concern business values, which were at the core of the restructuring process, were highly sensitive to expectations of future crude oil and natural gas prices, especially when they were struggling to reorganize in economically challenging times.

Conversely, recoveries for the retailing/restaurant sector were notably weaker in the low- to mid-50% area on an average and median basis because the sector remains under considerable pressure due to significant and ongoing secular changes and because it's generally not asset-rich. Looking at the sector's asset mix, most retailers in our dataset did not own a substantial amount of real estate properties, which would have been a potential source of recovery value. When forced to liquidate, retailers with weak brands had to clear inventories and hard assets at hefty discounts, resulting in poor recovery rates for creditors. Regarding debt mix, most retailers heavily relied on asset-based revolvers for funding. These asset-based lending facilities, given their large size and tight grip on working capital assets, diminished the value of what was left and available for term loan lenders. We've also recently seen more retailers seeking bankruptcy protection based on a dual-track plan, either a stand-alone reorganization or an asset sale.

Table 2

When looking at recoveries based on the presence or absence of financial maintenance covenants, our comparison of actual recovery rates between the two loan types reveals:

• Cov-lite first-lien term loans recovered on average about 68% of par, well below the 79% average realized by non-covenant-lite first-lien term loans (close to the 72% and 82%, respectively, in our previous review).
• The gap is larger on a median basis, with cov-lite term loan recoveries trailing those of non-cov-lite term loans by a sizable 34-percentage-point margin (up from 20% in our prior review).
• While the number of deals with covenants is limited, the oil and gas sector had an average gap of 22 percentage points, nearly double the difference of the overall dataset, while the remaining sectors average out to a gap of 14%.
• The updated dataset has two pro rata term loans that we included given the dearth of syndicated term loans with financial maintenance covenants. Pro rata term loans are held by banks and almost always still have financial maintenance covenants. Both loans were to oil and gas companies and their inclusion does not significantly shift the statistics.

The outperformance in recovery for non-cov-lite term loans is broad as opposed to sector-specific---the only exception is the retail/restaurant sector, where, oddly, cov lite recovered more than non-cov-lite loans, rather than the reverse. The opposite direction of the gap reflects below-average recoveries for all four of the non-cov-lite loans to this sector, and in particular extremely low recoveries for jewelry and accessories retailer Charming Charlie LLC (2017) and gift retailer Things Remembered Inc. (2019). Both saw exceptionally low single-digit recoveries on their covenanted loans, dragging down the average for non-cov-lite. For Charming Charlie, our data only capture its first Chapter 11 filing, where at emergence term lenders received 25% of the reorganized company's equity. The company has since filed for a second bankruptcy to liquidate its business

Outside of the oil and gas and retail/restaurant sectors, loans with maintenance covenants outperformed those without 26% on average. Measured by median, the divergence is wider at 40%. We did not show the pair-wise comparison of other sectors in the table because the small size of these subgroups would limit the value of the data.

Even with covenant-lite loans, term lenders can still indirectly benefit from a financial covenant embedded in the cash flow revolver, issued by the same borrower, as the revolver lenders (mostly banks) can still influence the borrower's actions in ways that may help all lenders (such as by limiting debt incurrence or restricted payments). However, financial maintenance covenants for cash flow revolvers are generally springing covenants that are only tested when a prescribed borrowing threshold (often 25%-40% of the commitment) is met, which offers less protection than a covenant that is tested regularly every quarter. (To be clear, non-cov-lite term lenders do not need to hope that revolving lenders take actions to protect their interests since they would have a seat at the negotiating table as a result of their own covenants.)

Factors other than financial covenants can also influence recovery outcomes. As mentioned above, one such factor is subordination or the first-lien cushion. The proportion of senior secured debt in the debt mix has increased significantly in recent years, while the share of subordinated debt has shrunk over the same period. Although cushion size positively correlates with recovery, it did not contribute to the performance gap observed in our study; the median cushion of cov-lite loans is 35%, actually exceeding the 32% of non-cov-lite by a thin margin.

Likewise, although we believe that a prolonged restructuring process or a very complex capital structure can saddle companies with larger legal and financial advisory fees, and leave less value on the table for all creditors (as these expenses are paid before all other liabilities), in our dataset we did not observe a close tie between actual recoveries and the time companies spent in bankruptcy. Ultimately, lower recoveries for longer restructurings likely affect overall debt recovery rates rather first-lien recoveries, and also likely feature idiosyncratic challenges such as resolving complex litigation or legacy liabilities, which cloud the ability to draw broad conclusions.

Table 3

We also compared the recovery rates and differences realized by cov-lite and non-cov-lite first-lien term loans by emergence year. Average and median recoveries for cov-lite and non-cov-lite term loans vary from year to year, as do the differences in average and median statistics between these loan types. We think these data highlight the folly of trying to draw definitive conclusions regarding absolute recovery expectations or differences between cov-lite and non-cov-lite term loans. Undoubtedly, the underlying macroeconomic and sector conditions will materially influence recovery outcomes, as will fundamental factors such as total debt leverage, debt mix, asset composition and quality, industry dynamics, and idiosyncratic issues.

Even with the noise in the underlying data and limited data points for various years and categories, the numbers are directionally consistent with covenant-lite loans generally exhibiting lower average and median recovery prospects across the individual years.

Cov-Lite Structures Dominate In Institutional Term Loans And Weaken Lender Protections

One reason that cov-lite term loan structures concern institutional loan investors is because they morphed from a concession that was only granted to so-called "better" speculative-grade credits before the global financial crisis, to one that has been afforded to most speculative-grade borrowers since 2013. Since 2018, cov-lite loan structures accounted for roughly 85% of institutional term loans (see chart 2).

Chart 2

Further, financial maintenance covenants used to be a hallmark of the loan market and were a key point of distinction from speculative-grade bond structures. From a lender's perspective, financial maintenance covenants provide a mechanism to renegotiate loan terms if a company's credit deteriorates below the limits established at loan inception or--in extreme cases--to accelerate repayment by calling a default. In practice, covenant violations tend to result in repricing the loan (higher margins, fees, or both) to reflect higher credit risk, rather than triggering an event of default. This thesis was upheld in 2020's sharp COVID-19-related downturn as more lenders were willing to accommodate a covenant holiday (suspend or relax the maintenance test (primarily under revolving loan facilities) until borrowers were expected to return to pre-pandemic performance levels) than elected to accelerate the loan.

Somewhat less frequently, covenant amendment negotiations may allow lenders to revise other terms to help reduce potential losses or limit further credit deterioration. For example, lenders may seek to tighten collateral terms, increase amortization payments, limit capital expenditures or revolver availability, or restrict the company's ability to pay dividends or make investments outside of the credit group.

Of course, lenders holding loan tranches without financial maintenance covenants do not have a mechanism to reprice their loans when companies violate the financial maintenance covenants on other loan tranches, so they bear opportunity costs or pricing risk because they may no longer be adequately compensated for the risk of lending to a stressed borrower. Further, cov-lite loan structures, coupled with a revolving loan with springing financial maintenance covenants, may have higher event risk. This is because the number and size of various free and clear baskets have increased over time and now provide borrowers with significant flexibility to take actions under various buckets (e.g. to add debt of various types or make distributions or investments). This flexibility is free and clear of needing to comply with incurrence covenant tests related to these actions, and are also not subject to compliance with springing financial maintenance covenants under a revolver if borrowing levels remain below the thresholds that activate covenant testing. This can provide distressed issuers with more flexibility, especially if they can secure access to liquidity from sources other than the revolving facility (for example, a separate asset-based loan, receivables factoring, or sale leasebacks).

Why Do Covenant-Lite Structures Exist And Persist?

To start, cov-lite structures exist because borrowers (and financial sponsors) like them! From a borrower's perspective, having no maintenance covenants can provide much-needed financial flexibility in times of distress The inability of term lenders to call a technical default due to a covenant breach keeps them from negotiating better pricing, protective restrictions, or, worse, potentially accelerating a loan. In fact, the risk of not being able to negotiate an amendment with the diverse group of holders after a covenant breach has been offered as one of the primary justifications for excluding financial maintenance covenants from broadly syndicated institutional loans. This argument, which has some merit in our view, posits that the investor base today is largely similar to the bond market, in which investors do not require such protections, and that certain nontraditional loan investors--such as distressed debt investors and hedge funds--might have different motivations and be unreasonable in such circumstances.

The flexibility provided to borrowers by cov-lite term loan structures is enhanced by the springing nature of the financial maintenance covenants that exist under revolving loan facilities that are part of the same credit agreement. These covenants are not tested unless a predetermined borrowing threshold is met (typically 25%-40% of the revolving commitment). Further, even when such thresholds are met and there is a risk of breaching financial covenants, the borrower only needs to negotiate with a relatively small group of pro rata lenders (predominantly banks) that have relatively modest loan commitments at risk (compared to term lenders). There was a surge in covenant amendments and waivers in early 2020 following a flood of revolver draws after the COVID-19 crisis emerged, but since that time, many companies have tapped the bond market to repay revolver borrowings, rendering the springing revolver covenants inactive again.

Cov-lite structures also persist because they are marketed to buyers individually, providing each institutional loan investor with very limited power to influence the proposed terms, especially with these structures firmly established in the marketplace since 2013.

On the other hand, cov-lite term lenders' inability to call a default and or reprice risk in periods of stress may, as we see it, modestly reduce default risk by alleviating cash flow and liquidity pressure. This is especially true if the magnitude and duration of the distress are not too severe and followed by a rebound. Of course, if these firms default eventually and the cov-lite terms merely delay an inevitable restructuring, this may undermine a company's enterprise value and impair recovery rates.

Treatment Of Covenant-Lite In Our Recovery Ratings Methodology

As noted earlier, we factor the presence or absence of financial maintenance covenants into our recovery ratings, which generally results in somewhat higher recovery rates for debt structures with non-cov-lite term loans versus cov lite. Our recovery analysis simulates a hypothetical default that includes estimating a fixed-charge proxy at the level below which a company will have insufficient cash flow to meet its fixed payment obligations. Our calculation of the fixed-charge proxy captures the interest expense and principal amortization due in the default year plus minimum capital spending needs. We use the fixed-charge proxy as a measure of the required stress to cause a default.

Along the path to default, we assume borrowers with traditional financial maintenance covenants breach them and lenders demand a higher interest margin to offset the greater default risk to which they are now exposed. This is consistent with our observations of market behavior. The incremental borrowing costs are factored into our fixed-charge proxy as an additional fixed obligation, implying that a company would default somewhat earlier and at higher profitability than it would have without a covenant breach. Holding all else constant, this earlier default expectation prevents further enterprise value deterioration and thus drives relatively higher recovery prospects for lenders. In our 2018 study we also compared our recovery estimates for cov-lite and non-cov-lite first-lien loans that we rated during fourth-quarter 2017. The results were similar to the empirical outcomes observed for actual recoveries, showing average recoveries for first-lien cov-lite term loans of 65.8%, compared to 75.5% for a non-cov-lite loans.

For further discussion of our views on how weak loan terms may hurt loan recovery rates, as well as what is and is not factored into our recovery ratings on a prospective basis (and why), please refer to "Credit FAQ: When The Cycle Turns: Assessing How Weak Loan Terms Threaten Recoveries," published Feb. 19, 2019.