Trends for loan and bond recovery rates in the U.S. have diverged since 2021. Loan recoveries remain well below their long-term average amid tough financing conditions for weaker credits and a weakening of loan structures in recent years. Bond recoveries, meanwhile, are holding solidly above their long-term average. This could, in part, represent a sample-size bias. The recent pool of bonds emerging from default in our study includes no subordinated bonds, and most of these emergences have followed distressed exchanges; both appear to be lifting recent averages.

Higher-for-longer interest rates also loom over credit in the U.S. While higher rates should contribute to lower valuations, and while they could challenge ultimate recoveries, recovery rate trends in rising rate environments have historically been mixed.

Loan Recoveries Remained Below Average In 2023 Even As Bond Recoveries Stayed Elevated

Discounted recovery rates for bonds and loans converged in the U.S. over the past two years (with both of those rates in the mid-60s) as recovery trends have diverged. Bond recoveries remain well above their historical average even as loan recoveries remain below theirs.

Loan recoveries improved somewhat this year through September, rising seven percentage points to 65.2% but staying well below the long-term average of 73% (see chart 1). Higher-for-longer interest rates may be part of the drag on loan recoveries as challenging financing conditions strain weaker borrowers. Leveraged loan issuance declined in 2023 to its lowest level since 2010 (according to Leveraged Commentary and Data from PitchBook, a Morningstar company), with most of the new issuance going toward refinancing. Additionally, weakening debt structures--including the prevalence of covenant-lite loans, permissive loan documents, and shrinking debt cushions--are also weighing on loan recoveries.

Bond recoveries, by contrast, continue to be solidly above their long-term average of 40%, staying elevated at 65.9% even though they fell this year (through September) by nearly four percentage points. For the first time in our data series, recent average bond recovery levels (from instruments emerging between January 2022 and September 2023) exceeded loan recoveries.

The pool of bonds recently emerging from default largely consists of instruments that entered default through a distressed exchange. Further, recent bond defaults include a high share of secured bonds (and a lack of subordinated bonds) relative to historical levels. These factors are reflected in higher recent average recoveries for bonds.

Chart 1

The number of defaults in the U.S. has more than doubled in 2023 (through September). Distressed exchanges and out-of-court restructurings have accounted for more than half of U.S. defaults so far this year, and more than half of emergences as well.

The elevated level of bond recoveries that we have currently likely reflects the prevalence of distressed exchanges. We've observed that bonds tend to offer higher average recoveries after a distressed exchange or nonbankruptcy restructuring than after a bankruptcy. However, while the recovery immediately following a distressed exchange may be higher, we've also observed a higher propensity for repeat defaults following a distressed exchange. Subsequent defaults of an issuer could lead to lower recoveries.

Higher Interest Rates Can Strain Valuations

While the higher interest rates that issuers face can be expected to reduce the valuations of restructured companies (contributing, in turn, to lower recoveries), the divergence of bond and loan recovery trends suggests that there's more to the story.

Since the beginning of 2022, nominal loan recovery values have fallen as interest rates have risen (see chart 2). But historically, the behavior of loan recovery rates during periods of rising interest rates has been more mixed. For instance, we saw a steep increase in loan recoveries during the rate tightening cycle in 2004-2006 that contrasts with the recent decline.

Notably, rising interest rates do not occur in a vacuum: They often occur during periods of economic growth. During these periods, the benefits of economic expansion may outweigh the costs of higher interest rates for struggling companies. Also, higher rates might cause companies to default with less profit deterioration, which may partially offset valuation pressure from higher rates.

Chart 2

Recoveries Vary By Instrument Type

Loan recoveries for 2022-2023 (through this past September) are down across loan types, and this could reflect weakening loan structures. Recoveries for revolving credit facilities, first-lien term loans, and second-lien (or lower) term loans during this period are all below their long-term averages (see chart 3).

By contrast, recoveries of senior secured bonds and senior unsecured bonds are both up for 2022-2023 (through this past September). Furthermore, there are no subordinated bonds in the bond recovery data for this period. This could change as we receive more ultimate recovery information, but for now, overall bond recoveries have been higher without subordinated bonds in the pool of recent bond emergences. Meanwhile, the lack of subordinated bonds among recent emergences could also reflect the shrinking debt cushions we've observed among first-lien term loans.

Chart 3

We based our analysis in this study on data from S&P Global's LossStats, which is available through S&P Global Market Intelligence's CreditPro. The database provides ultimate recovery values for over 4,600 defaulted instruments from over 1,100 U.S. issuers that emerged from default between 1987 and September 2023. Except where otherwise noted, the recovery values we cite refer to the discounted recovery, where the discount rate applied is the instrument's effective interest rate. This approach to calculating recoveries differs from that of S&P Global Ratings' recovery ratings methodology (and recovery studies based on that methodology). S&P Global Ratings' recovery ratings indicate the recovery prospects of a given loan or bond from a speculative-grade issuer, calculated on a nominal basis, based on future hypothetical default scenarios. (For more details on the approaches, see Appendix I: Data Approach and Appendix II: Comparison With Recovery Ratings.)

The Pool Of Loans Emerging From Default In Recent Years Shows Weaker Structures

In several respects, the structures of the loans emerging from default in 2022 and 2023 display structures with characteristics that are weaker than those found in earlier pools of loans.

Unsurprisingly, covenant-lite loans--that is, loans without financial maintenance covenants--now account for a growing share of defaulted term loans as cov-lite loan structures have become predominant in broadly syndicated loan issuance. About 45% of the first-lien term loans emerging in 2022-2023 have been cov-lite, up from 39% in 2016-2021 and 14% in 2010-2015.

Cov-lite loans were more uncommon before 2010, and the few examples we have of recoveries of cov-lite loans before 2010 displayed higher-than-average recovery. At the time, as cov-lite loans were the exception rather than the rule, lenders were more likely to agree to cov-lite structures only for issuers with stronger credit quality. Now that cov-lite loans have grown to encompass most new broadly syndicated loan issuance, they're no longer reserved for issuers with above-average credit strength.

For emergences since 2010, we've observed lower average recoveries for cov-lite term loans than for non-covenant-lite loans. First-lien term loans that were cov-lite averaged a 61% discounted recovery between 2010 and September 2023, nearly 11 percentage points below the average recovery of non-covenant-lite first-lien term loans (see chart 4). For 2022-2023 (through this past September), the difference between cov-lite and non-covenant-lite recoveries was less pronounced, with cov-lite recoveries averaging just a couple of percentage points less than first-lien term loans with covenants.

Chart 4

The lack of financial covenants, coupled with more flexible loan documents, has also triggered a rise in out-of-court loan restructurings, which can weaken loan recoveries by allowing the transfer of collateral or by permitting more priority debt.

In addition to the weakening of covenants, the debt cushions--the percentage of the debt capital structure that is junior to the defaulting instrument--of recently emerged loans have been notably smaller than in prior periods. In a default, the more junior instruments take the first loss, providing some cushion to the more senior instruments. Larger debt cushions are commonly associated with higher recoveries for an instrument.

None of the first-lien term loans that have emerged from default in 2022-2023 had a debt cushion of more than 50%, and most had a debt cushion of less than 25%. By contrast, of the first-lien term loans that emerged from default between 1987 and 2021, more than a third had a debt cushion of 50% or more (see charts 5a and 5b). Historically, we've observed that higher average recoveries correspond to larger debt cushions for first-lien term loans (see chart 6).

Chart 5a

Chart 5b

Chart 6

Bond Recoveries Remain Elevated

We find that recovery rates generally tend to be inversely correlated with default rates. Recoveries tend to be higher during periods of lower default rates as the conditions that support lower default rates--for example, improving economic conditions--also tend to support higher recoveries. In 2022, bond recoveries were elevated (averaging near 70%) at a time when the U.S. speculative-grade default rate was exceptionally low (at 1.65%). Recoveries have remained high for the bonds that have emerged this year through September (near 66%) even though the default rate has more than doubled (to 4.1%). While the increase in the speculative-grade default rate has been rapid, it remains just below its long-term average of 4.2%.

Chart 7

The majority of the bonds emerging from default in 2022-2023 (through this past September) defaulted through either a distressed exchange or another nonbankruptcy restructuring. Bonds often experience higher recoveries after a distressed exchange or nonbankruptcy restructuring than after a more-costly bankruptcy restructuring, where less of the company's residual value may accrue to bondholders and junior creditors.

Distressed exchanges have risen as a share of defaults since 2014, and in most years during this period, distressed exchanges have lifted average annual bond recoveries (see chart 8). For the 12 months ended Sept. 30, 2023, bond recoveries overall averaged almost 63%, nearly 13 percentage points higher than the average recovery for bonds excluding distressed exchanges and nonbankruptcy restructurings. The population of bonds emerging from default in 2022-2023 (through this past September) was relatively small, with just 20 instruments. Most emerged after distressed exchanges; just four of these bonds emerged following a bankruptcy.

Chart 8

The recovery following a distressed exchange may be higher, but if the restructuring doesn't sufficiently address the secular or operational challenges the company faces, then a repeat default may follow. In our study "Credit Trends: A Rise In Selective Defaults Presents A Slippery Slope," published June 26, 2023, we found that 34.9% of issuers that defaulted through a selective default experienced a repeat default within the subsequent 48 months.

Chart 9

The pool of bonds emerging in 2022-2023 (through this past September) is also tilted toward bonds that are more senior. Nearly half are senior secured, and none are subordinated. By contrast, between 1987 and 2021, less than 15% of the bonds emerging from default were senior secured, and over 35% were subordinated. All of this is likely also helping to make recent bond recoveries appear higher.

The media and entertainment sector and the retail and restaurant sector account for larger shares of the bond recoveries in 2022-2023 (through this past September) than any other sector. Together, they account for half of the bond recoveries during this period. In each sector, recent recoveries for bonds have come in higher than the sector's long-term average, and these recoveries are largely recoveries following distressed exchanges.

Meanwhile, the pool of loans emerging in 2022-2023 (through this past September) includes a large share of instruments from the media and entertainment, health care, and consumer products sectors. These three sectors combined accounted for more than half of the loan recoveries during the period. For each of these sectors, loan recoveries were below their historical average.

Notably, there have been several high-profile bankruptcies of rated issuers in 2023--including those of Bed Bath & Beyond Inc., Party City Holdings Inc., and Yellow Corp., where ultimate recovery is not yet available for this study. Because we focus on ultimate recovery values in this report, inclusion of recovery data for an issuer will lag the eventual emergence from default and will also depend on the availability and timing of sufficient recovery data. While recent bond recoveries have been above average, this trend may not last, especially with an increase in bankruptcies.

Table 1

Recovery By Instrument Type

Our historical data shows how recoveries vary by debt instrument--with more senior instruments exhibiting higher recoveries with lower variance than more junior debt.

• Historically, revolvers show the highest recoveries, with a mean recovery of 79.2% and a median recovery of 95.2%.
• First-lien term loans follow, with a mean recovery of 70.9% and a median of 79.6%.
• Recoveries for second-lien (and unsecured) term loans are considerably lower, with an average recovery of 43.9% and a median of 31.2%.
• Second-lien (and unsecured) term loans show the highest standard deviation, at 40.2% (reflecting, in part, the relatively small sample size), and they often display a bimodal recovery distribution--either fully paid or not paid at all.

Bonds overall have lower average recoveries than loans, though recoveries vary widely by bond type (see table 2).

• Among bonds, senior secured bonds have the highest mean recovery, at 58.1% (with a median of 58.7%).
• Senior unsecured bonds show a mean recovery of 44.8% (with a median of 41.9%).

Since subordinated debt typically accounts for a small share of a company's debt structure, there's often no value left to provide recoveries for these instruments in a bankruptcy after the senior debtholders are paid.

• Subordinated bonds (including senior and nonsenior) show a modal recovery value of zero and a median recovery of 14.7%.
• Senior subordinated bonds show an average recovery of 29.9%.
• Nonsenior subordinated bonds show an average recovery of 22.8%.

Table 2

On a nominal basis, where the ultimate recovery value has not been discounted to account for the time between default and emergence, recoveries across the debt structure are notably higher: loan and revolver recoveries average 81.3%, while bond and note recoveries average 47.3%.

In addition to these issue-weighted recovery values, we calculate recoveries on a dollar-weighted basis for both the discounted and the nominal rates (see table 2). For these dollar-weighted recoveries, we calculate the discounted (or nominal) sum of debt recovered and divide it by the total amount of defaulted debt in the sample for that instrument type.

Instruments with lower seniority tend to have a wider range of average recoveries than higher-priority instruments. The coefficient of variation, or the standard deviation scaled by the mean, rises for instruments lower in the capital structure. While the coefficient of variation for loans and revolvers is 42.7%, it rises to 84.6% for bonds overall.

Secured debt tends to show higher recoveries than unsecured debt, and among secured debt, we see variance by collateral type. Debt secured by inventories or receivables shows the highest average recovery (91.3%), with a lower standard deviation than other collateral types. Among secured debt, instruments secured by a second lien (or below) show the lowest average recoveries with the highest relative standard deviation (see table 3).

Table 3

Loans From Middle-Market Companies Continue To Exhibit Higher Recoveries Than Loans From Larger Entities

As in previous periods, term loans and revolvers from small or middle-market companies (those with $350 million or less in debt outstanding at the time of default) continue to show higher recovery rates than loans from larger companies, even as loan recoveries have fallen for both middle-market and larger companies. Loans and revolvers from middle-market companies averaged a 66.3% recovery in 2022-2023 (through this past September), above the 59.4% average for larger companies (see chart 10).

For loans that both emerged from default in 2022-2023 (through this past September) and are from middle-market companies, nearly two-thirds had defaulted through a distressed exchange--a higher share than for large corporate loans that emerged from default during the same time period. Among the recent recoveries of loans from middle-market companies, the media and entertainment sector and health care sector were the most represented.

Chart 10

Higher recoveries for middle-market companies can reflect several factors, including simpler debt structures (such as unitranche) with fewer lenders than are involved in the debt structures of larger issuers. This can reduce the potential for conflicting lender groups. Furthermore, financial maintenance covenants remain more common among loans for middle-market companies than among broadly syndicated loans for large companies. Further, more protective loan documents also reduce the flexibility issuers may have to transfer collateral or layer in priority debt.

Recovery rates for first-lien term loans from middle-market issuers have averaged 77.6% on a discounted basis since 1987, higher than the 67.2% average recovery for first-lien term loans from larger corporate issuers (see table 4).

Table 4

One notable caveat about this data is that we only capture ultimate recoveries from publicly available data sources. With less data available for smaller middle-market and private-credit companies, the trends we observe in the pool of companies we capture in this study may not necessarily be representative of broader trends within the middle and private-credit markets.

The Distribution Of Recoveries Varies By Instrument

Because of their seniority in the debt structure, term loans and revolving credit facilities tend to show higher recoveries than bonds and notes. In fact, loans and revolvers show recoveries of par or greater much more often than bonds or notes. Some 55.5% of defaulted loans and revolvers in our study experienced elevated recoveries of 80% or higher, and 25.7% of loans and revolvers recovered par or greater (see chart 11).

Recovery rates are not quite as high among all first-lien debt. Including first-lien senior secured bonds and first-lien revolving credit facilities and term loans, 54.0% of the defaulted instruments recovered 80% or more, and 24.9% of first-lien instruments recovered par or greater (see chart 12).

Chart 11

Chart 12

In contrast, the distribution of bond and note recoveries skews much lower. Just 5.2% of defaulted bonds and notes had recoveries at par or greater, while 27.4% of defaulted bonds and notes experienced negligible recoveries of 10% or less (see chart 13).

Chart 13

Appendix I: Data Approach

We define recoveries as the ultimate recovery rates following emergence from three types of default: bankruptcy filings, distressed exchanges, and nonbankruptcy restructurings. Unless we specify otherwise, we base recoveries at the instrument level, and we discount them by using each instrument's effective interest rate. Instruments that did not default are excluded from this study.

In S&P Global's LossStats database, the coupon rate at the time the last coupon was paid is the effective interest rate used for the discount factor.

We calculate the discounted recovery values by discounting instruments or cash received in the final settlement on the valuation date, back to the last date that a cash payment was made on the prepetition instrument.

The last cash-pay date represents the true starting point for the interest accrual, which is why we use this date as the starting point for the discounting rather than the default date of the instrument or the bankruptcy date of the company.

For fixed-coupon instruments, this is the fixed rate, and for floating-rate instruments, it is the floating rate used at the time of default.

Nominal recovery rates, which are the nondiscounted values received at settlement, are also reported.

We prefer discounted rates in this study because they allow us to better compare bankruptcies of different lengths. For example, the nominal rate for a distressed exchange could be the same as that for a bankruptcy case that takes two years. However, investors in the bankruptcy case are significantly worse off because they could lose significant time value while waiting for the final settlement.

On the other hand, a distressed exchange could take only a day. In a historical study, discounted recovery rates offer the major benefit of making different periods more comparable by preventing any major bias that could arise if times between default and emergence differ greatly. S&P Global Ratings provides recovery ratings that map to nominal values.

Recovery

Recovery is the value creditors receive on defaulted debt. Companies that have defaulted and moved into bankruptcy will usually either emerge from the bankruptcy or be liquidated. On emergence from bankruptcy, creditors often receive a cash settlement, new instruments (possibly debt or equity), assets or proceeds from the sale of assets, or some combination of these.

Ultimate recovery

Ultimate recovery is the value of the settlement a lender receives by holding an instrument through its emergence from default. The recovery is based on the amount received in the settlement divided by the principal default amount. Within S&P Global's LossStats database, three recovery valuation methods are used to calculate ultimate recovery:

Trading price at emergence.  We can determine the recovery value of an instrument by using the trading price or market value of the prepetition debt instruments upon emergence from bankruptcy. Of the three valuation methods, this one is the most readily available because most debt instruments continue to trade during bankruptcy proceedings. Note that this approach differs from the commonly used "30 days after default" method, which measures recovery estimates shortly after default rather than at emergence.

Settlement pricing.  The settlement pricing includes the earliest public market values of the new instruments that a debtholder receives in exchange for the prepetition instruments. This method is similar to the trading price method, except that it is applied to the new (settlement) instrument instead of the old (prepetition) instrument.

Liquidity-event pricing.  The liquidity-event price is the final cash value of the new instruments or cash from the sale of assets that the lender acquires in exchange for the prepetition instrument.

Appendix II: Comparison With Recovery Ratings

The approach we use for assessing recovery in this study is different from how S&P Global Ratings determines its recovery ratings. The recovery ratings are estimates of recovery for debt instruments from entities rated speculative grade ('BB+' or lower). These recovery ratings indicate recovery prospects calculated on a nominal basis, based on future hypothetical default scenarios and reflecting the expected recovery following an entity's emergence from bankruptcy via a going concern or liquidation. These recovery ratings are issue-specific and range from '1+' (high expectations for a full [100%] recovery) to '6' (0%-10% recovery).

The approach to determining recovery values that we use in this report also differs from the approach we use in our study "North American Debt Recoveries May Trend Down For Longer," published Dec. 11, 2023. That study excluded recoveries following distressed exchanges, and it used the implied recoveries from the bankruptcy plans as the basis for the value received, rather than trading prices, because the bankruptcy plan information was more consistently available for the issuers included in the study. In addition, that study only covers companies going through U.S. bankruptcy, and it uses actual recovery estimates on a nominal, rather than discounted, basis.

Related Research

• Credit Trends: A Rise In Selective Defaults Presents A Slippery Slope, June 26, 2023