The combination of historically low interest and default rates and growing institutional investor demand has fueled an increase in covenant-lite loans. According to S&P Global Ratings' Leveraged Commentary & Data (LCD), covenant-lite loans made up 75% of new institutional loans in 2017. Perhaps more notable, through early February 2018, covenant-lite deals accounted for record high 85% of 'B' rated institutional issuance.

For years, there was much discussion about the covenant-lite phenomenon, and in particular, what it means for ultimate recovery in a default. The data on recovery rates for covenant-lite loans in the past downturn did not suggest incremental recovery risk, although the data was viewed as inconclusive given limited number of data points and a general perception that companies with covenant-lite structures at that time were generally better credits.

In this report, we look at empirically observed recovery levels for covenant-lite and noncovenant-lite institutional loans issued by companies that recently exited bankruptcy. We also highlight how our recovery methodology analyzes covenant-lite and noncovenant-lite loans. We examine a sample of recently reviewed ratings on institutional loans, both with and without covenants, and our estimated recoveries across the two sets.

Here are the key findings and other highlights:

• Our review of the entities that emerged from bankruptcy between 2014 and 2017 indicates that covenant-lite first-lien loans recovered an average of 72%, compared to 82% for noncovenant-lite loans.
• Among S&P Global Ratings recovery ratings newly assigned or reviewed in the fourth quarter of 2017, we estimated that recoveries for rated first-lien institutional loans average over 75% for those with maintenance covenants and about 66% for those without such covenants.

However, it is important to note that the new data set is limited and to emphasize that recoveries are complex and driven by a myriad of factors. Covenants alone do not tell the full recovery story. Other material factors that influence recovery outcomes include the nature and value of a company's assets, debt structure and mix, and the timing of exit from bankruptcy among other variables. An additional caveat is that some market participants note that in the current market only weaker credits tend to have more covenant protections.

As Loan Terms Loosen, More Risks Loom

The low-interest-rate and default environment in the recent years has resulted in an increasingly borrower-friendly leveraged finance conditions as investors seek higher yield. An outcome of this market dynamic is weaker protection for lenders in various ways that bode ominously for recoveries in the next credit downturn. The weaker protections include:

• Covenant-lite loan structures as the predominant loan structure, which is generally the most widely discussed development.
• Aggressive EBITDA add-backs that allow borrowers to take on more leverage than might otherwise be possible. While our methodology calculates EBITDA and leverage independently, and factors the appropriate incremental risk into our ratings ("Adjustments To EBITDA In Technology Leveraged Buyouts: How We Read The Story They Tell", published Dec. 21, 2016), these add-backs can distort actual ratios. As a result, the add-backs can weaken the effectiveness of financial maintenance covenants in noncovenant-lite loans. Further, the add-backs also flow through to incurrence tests and, when combined with a trend toward more generous baskets (discussed next), can further compound the risks of both covenant-lite and noncovenant-lite loans.
• More generous or flexible baskets for permitted debt or restricted payments including dividends, investments, dispositions, transactions with affiliates, etc. The risks these features can pose are illustrated by the aggressive actions taken by J. Crew Group Inc., in which the company used various baskets to carve out key assets from the collateral package and pledge them to support additional debt (see our review of the series of transactions used in J. Crew's asset transfer, "J. Crew’s Intellectual Property Transfer: An Update For Distressed Lenders", published Sept. 12, 2017). This obviously increases the risk of loss to loan investors in a default scenario and negatively affected our recovery ratings. However, such actions are hard to predict and generally not directly factored into our ratings on a prospective basis, other than perhaps through more conservative assumptions when we view a company or financial sponsor as prone to pursuing aggressive financial strategies.

The Rise Of The Covenant-Lite Phenomenon

During 2017, covenant-lite loans made up 75% of new institutional loans, according to LCD. To illustrate how fast it has grown, the average from 2007-2009 was 14% but steeply climbed to exceed 70% since 2014 (see chart 1). Meanwhile, the spread increase through covenant-relief amendments edged up on a lagging-12-month basis to 80 basis points (bps) in January from 67 bps one year ago, according to LCD data. As covenant-lite structures proliferate, one irony is that now only the credits the market deems weakest are likely to have financial maintenance covenants in their institutional bank debt. In the past, only the credits the market deemed strongest were able to obtain covenant-lite terms.

How Financial Maintenance Covenants Can Benefit Loan Investors

In general, financial maintenance covenants used to be a standard feature for loans and provided lenders with the opportunity to influence borrower behavior in ways that may protect loan investor interests if a borrower's credit quality declined below established limits (e.g., by limiting dividends, tightening collateral packages, or reducing revolver availability to minimize total lender exposure).

Further, these covenants provide lenders with the opportunity to reprice loans as compensation for the increase in credit risk as part of negotiating an amendment if financial maintenance covenants are breached. The violation of a maintenance covenant triggers a technical default and provides lenders the option to accelerate payments if such defaults are not cured. In reality, though, lenders generally take less dramatic actions; instead, they amend the agreement to either waive or reset the covenant, for which they receive a consent fee and/or an interest-margin increase. With this in mind, we believe that there is a material opportunity cost of not being able to reprice the loan risk in the event of covenant breach, which hurts covenant-lite lenders from a risk-return standpoint.

From the perspective of a borrower, the absence of maintenance covenants could provide much needed financial flexibility in times of stress. This keeps lenders at bay from negotiating better terms or, worse, potentially accelerating a loan because of a technical event-of-default. In fact, the risk of not being able to negotiate an amendment following a covenant breach on a widely syndicated loan has been offered as one of the primary justifications for excluding financial maintenance covenants from institutional loans. This argument posits that the investor base is largely similar to the bond market, in which investors do not require such protections, and that certain nontraditional loan investors (such as hedge funds, distressed debt investors) might have different motivations and be unreasonable in such circumstances.

This added time and liquidity might provide a lifeline to issuers in a period of stress, allowing them to recover without needing to renegotiate terms. However, the lack of maintenance covenants can also give aggressive management teams and sponsors the latitude to pursue shareholder-friendly or other actions that may ultimately hurt a company's credit profile. In some cases, a covenant-lite loan structure may not save a company but could simply delay an inevitable default--and potentially undermine a company's enterprise value along the way--and impair recovery rates.

Actual Recoveries In 2014-2017: Covenant-Lite Versus Noncovenant-Lite

We reviewed empirical data on recoveries between covenant-lite and noncovenant-lite institutional loans. For this purpose, we derived recovery rates from a review of bankruptcy and other documents, including Chapter 11 disclosure statements, plans of reorganization, and asset sales. The data set excluded distressed exchanges or out-of-court restructurings, as well as companies for which we were unable to obtain reliable recovery data.

The following shows the ultimate recovery rates of first-lien institutional term loans of 28 companies that exited bankruptcy between January 2014 and December 2017 (see table 1)--either through a reorganization or asset sales/liquidation. All but one filed for bankruptcy during the same period (see table 2).

Table 1

Table 2

While the sample size is small, the limited empirical evidence suggests that the absence of financial maintenance covenants appears to negatively affect the ultimate loan recovery rates in the data sets we reviewed. In particular, our analysis of the two sets finds:

• Among companies that exited bankruptcy in 2014-2017, the average ultimate nominal recovery rate of first-lien institutional term loan is 76%. The median is slightly higher at 81%. By comparison, the average 40-year historical recovery levels for senior secured first-lien term loans is 81.2% ("Default, Transition, and Recovery: Recovery Study (U.S.): Quantitative Easing, Low Yields, And Distressed Exchanges Have Boosted Bond Recoveries Since 2010", published Dec. 14, 2017). We note that the recovery statistics quoted above are not perfectly comparable because our statistics are calculated on an issuer-count basis while the 40-year recovery study statistics are calculated on an issue-count basis. Further, our statistics exclude distressed exchanges while the 40-year study does not; however, this should not meaningfully affect recovery rates on loans.
• Loans with financial maintenance covenants recovered 82.2% on average (median 84.1%). Those without such covenants recovered 71.6% (median 63.5%).
• We note that average recovery rates for noncovenant-lite loans by emergence year can be heavily influenced by the limited number of data points. For example, in 2016 there were only two entities with noncovenant-lite loans that emerged, and both had nearly full recovery (100% by Colt Defense LLC and 98.7% by Aspect Software Inc.) Further, in 2017 the high average recovery achieved by noncovenant-lite loans was over-represented by two separate loans to different entities under Caesars Entertainment hotel and casino group (Caesars Entertainment Corp. and Caesars Entertainment Resort Properties LLC), which emerged from bankruptcy after nearly three years of Chapter 11 proceedings and provided full recovery to term loan lenders.
• We are mindful that factors other than maintenance covenants can drive the recovery rates observed for the two sets of loans. These include debt structure, the value of company's assets, management quality, earnings potential, sector stability, leverage, legal issues, industry dynamics, and other idiosyncratic issues.

Treatment Of Covenant-Lite In Our Recovery Ratings Methodology

Our recovery analysis simulates a hypothetical default for the borrower. As part of our recovery methodology, we define a fixed-charge proxy as a level below which the borrower will have insufficient funds to meet its fixed 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 maintenance covenants would breach them and lenders would subsequently demand a higher interest margin to offset the greater default risk to which they are now exposed. These 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. This earlier default expectation prevents further enterprise value deterioration and thus drives relatively higher recovery prospects for lenders (all else being equal).

As an example, call center software provider Aspect Software Inc. went through a series of covenant amendments as it struggled to fend off bankruptcy. The first was in late 2012, when the company after a few quarters of disappointing results netted covenant relief in exchange for an uptick in both the term loan spread (by 75 bps to L+525) and amortization (to about 1% per quarter in 2013 and 2014 before reverting to 1% per annum, as lenders sought to gradually reduce potential exposure). Over the period, each such incremental borrowing cost raised the hurdle for debt servicing, as Aspect needed to divert more cash flow to satisfy the rising interest and/or principal payment, reducing its liquidity and contributing to its eventual default. The increased fixed charge that led to default likely helped preserve the enterprise value and contributed to the lender's strong recovery of 98.7%.

S&P Global Ratings' Estimated Recoveries In The Fourth Quarter Of 2017: Covenant-Lite Versus Noncovenant-Lite

The trend observed in empirical data is consistent with our recovery methodology and reflected in our forward-looking recovery estimates.

Indeed, a closer look at our recovery analyses performed in the fourth quarter of 2017 reveals that we expect covenant-lite loans to result in lower recovery rates on an average. We reviewed our recovery ratings on 301 senior secured first-lien institutional loans in the U.S., all of which were newly assigned or re-reviewed (either an event driven review or as part of ongoing surveillance) in the fourth quarter of 2017. In selecting this timeframe, we believe it provides a meaningful sample size and the more recent history may be the most relevant (see table 3).

Table 3

• Approximately two-thirds of the loans in our review are not governed by a financial maintenance covenant--indicative of broader market trends--with 'B' rated issuers accounting for the bulk.
• On average, we expect senior secured covenant-lite loans to recover about 65.8%. This compares to a recovery expectation of 75.5% for a noncovenant-lite loan.
• The difference between the two sets is slightly less pronounced among the most represented 'B' rated issuers, with the gap narrowing to 8.5%. We notice that with 'CCC' rated companies, the difference is further blurred, again highlighting that factors other than covenants can influence recovery ratings.
• Going by recovery rating (see chart 2), covenant-lite loans are more concentrated in '3' recovery ratings (58% of all covenant-lite; the comparable figure for noncovenant-lite is 39%) with a marginally lower estimated average recovery of 58% (versus 61.2% for noncovenant-lite).

Chart 2

We want to reiterate the two important caveats around this review:

• Our sample size of historical recovery rates is small and may not be applicable for the entire market or for all time periods. As it happens, the covenant-lite borrower base has expanded significantly over the past few years. Today, lenders tend to insist only on covenants for certain small and low-rated borrowers. They were once demanded of all but prestigious, large, and established borrowers.
• A number of factors can cause variations in average recovery rates. One primary factor is the debt mix. The average proportion of senior secured debt increased meaningfully in recent years, while the average share of subordinated debt has shrunk over the same period (see "Lean Senior Debt Cushion Threatens Recovery Prospects For U.S. Leveraged Loans," published Nov. 30, 2017). Structure and level of subordination also have a bearing on the average recovery rates.

While some hypotheses are easier to substantiate, others are more nuanced and less clear cut. It is possible to argue that a tighter leverage and tightly defined EBITDA-based incurrence covenants might be more effective than a maintenance covenant with massive headroom. Similarly, a loan secured by liquid and easily sold collateral may yield a higher recovery than one backed by assets that are difficult to value or illiquid, regardless of covenants.

Conclusion

The structuring of loans continues to evolve with the changing dynamics between lender appetites and borrower demands. When we first tackled the challenge of assessing the covenant-lite impact ("Credit FAQ: Investor Questions On U.S. Leveraged Finance And Recovery", published June 25, 2013), we observed no material difference in recoveries in the last cycle, a result that partly reflects the quality of the companies able to obtain covenant-lite financing in the previous cycle. Today, as covenant-lite financing spreads more widely in the market, our review of the 2014-2017 bankruptcies indicates that covenant-lite first-lien institutional loans recovered on average 10% less than fully covenanted loans. While our findings are in line with general market expectations, we caution that with individual credits, there is more to review than just the covenant label and encourage investors to look into the fundamental aspects of company's credit profile and debt structure as we do when performing our recovery analysis.