Most new issue collateralized loan obligations (CLOs) have call provisions. The most common (benchmark) deals feature a two-year non-call period, during which the notes issued by the CLO cannot be refinanced, reset, or called. After that period ends, CLO equity holders have the option to refinance or reset the CLO notes, or to redeem them altogether via an optional redemption. In a refinance (or a partial refinance), typically, the only change is to lower the coupons paid on the notes. A reset is more complex and can include other changes to the deal documentation and terms, including extending the CLO's reinvestment period end date and maturity date, as well as adding a new non-call period. In most cases, however, the pool of underlying assets backing the CLO remains largely in place (or can be added to). In this article, we consider the impact that market CLO spreads could have on the potential volume of CLO refinances and resets.

The Economic Incentive To Refinance/Reset A CLO

The motivations to issue a new CLO--or to refinance or reset an existing transaction--are many and varied. In the case of new issues, the decision depends mostly on the spread between the underlying loans and the cost of capital for the CLO, based on the current level at which CLO liabilities can price. The decision in the case of refinances and resets, on the other hand, depends mostly (but not exclusively) on the economic incentive provided by the prevailing level of market CLO spreads relative to the spreads on the outstanding CLO notes.

To illustrate the distribution of spreads that might influence individual call decisions, we have created a dot plot, by vintage, of every pre-2024 broadly syndicated loan (BSL) CLO that is within its reinvestment period or within two years of its reinvestment period end date (see chart 1). The dependent axis represents the current spread on the 'AAA' notes, and the independent axis represents the end of the CLO's non-call period.

Chart 1

While the overall cost of capital for the CLO's debt stack will drive the decision to refinance or reset the CLO (other than for a partial refinancing), we (consistent with the broader CLO market) often use the 'AAA' tranche spread as a proxy because the 'AAA' notes make up the largest portion of the BSL CLO capital stack, typically representing nearly two-thirds of the CLO notes and equity by par amount. The solid horizontal line in the chart above approximates the average level of 'AAA' spreads for U.S. BSL CLOs priced in January 2024, and can be thought of as a bar that raises or lowers the pool of transactions outstanding in the market that are incentivized to refinance or reset.

With some 1,822 observations, one can see that if spreads continue to tighten, numerous 2018 and 2021 transactions will be "in-the-money." In other words, managers will have an economic incentive to refinance or reset these CLOs.

Investors watch this (potential) source of supply for multiple reasons. First, a call will result in a return of capital to allocate to new investments (including CLOs) and a decision to either accept a lower spread or search for an alternative investment. Second, the amount of new supply impacts market "technicals" (i.e., supply/demand dynamics that can affect spread and market sentiment more broadly). The chart above also illustrates the rise and fall of 'AAA' credit spreads over the past few years. A cluster of transactions issued in early 2022 priced at lower spreads, similar to those of the 2021 vintage, but the level and dispersion of 'AAA' spreads increased throughout the year amid Fed rate hikes and heightened interest rate volatility. During 2023, spreads descended to around 175 basis points (bps) from historically wide levels.

Our Framework For Estimating Possible Future Refinance/Reset Volume

To approximate the number of refinance/reset-eligible and incentivized CLOs each month, we developed a program that applies a set of criteria to the individual CLOs in our issuance database. This framework assumes a CLO is eligible for a refinance/reset for any month under examination if it meets three criteria:

• The CLO must be within its reinvestment period, or less than two years outside its reinvestment period during the examination month;
• The CLO must be outside its non-call period during or prior to the examination month; and
• The CLO must be the latest iteration of the transaction.

The third condition eliminates the possibility of double counting. Suppose, for example, that CLO "X" has reset twice during its lifecycle, extending its reinvestment period at each reset date. Assume further that initial new issuance occurred in February 2018 (CLOX-2018), the first reset occurred in February 2020 (CLOX-2018-R1), and the second reset occurred in February 2022 (CLOX-2018-R2). From March 2020 to February 2022, CLOX-2018-R1 is the only eligible iteration of the deal. Looking only to the current iteration of the CLO (the third refinance/reset eligibility condition in the example) ensures that, up to the examination month, CLO X is not counted more than once.

Our refinance/reset framework presumes a CLO is incentivized for a refinance/reset in a particular month if the CLO's 'AAA' tranche spread to three-month SOFR exceeds the average 'AAA' spread of CLOs that priced during that month. Put differently, a CLO is incentivized for a refinance/reset if its "moneyness" (CLO 'AAA' spread minus monthly market average 'AAA' spread) is positive. The moneyness of MM CLOs is calculated with respect to monthly average middle-market (MM) CLO 'AAA' spreads. We note that various CLO managers achieve different pricing depending on investor preference, so comparing a CLO's spread to an overall market average may not always paint an accurate picture of moneyness. For this analysis, we base refinance/reset incentives on the 'AAA' spread alone, but a CLO's weighted average cost of capital (WACC) can also be employed. All CLO tranches in our database that referenced LIBOR before its final cessation in mid-2023 and haven't reset or refinanced since received a +26 bps adjustment to their spreads.

Scenarios For Potential CLO Refinance/Reset Volume

We present four 'AAA' spread scenarios with estimates of potential refinance/reset volume through the end of 2024. For the four scenarios under consideration, the monthly estimate of eligible and incentivized CLOs is cumulative. If, for example, there are 200 eligible and incentivized CLOs in month one and 300 in month two, this means 100 additional CLOs "came online" from month one to month two. We do not attempt to forecast the actual number of future refinances and resets in this analysis. Rather, we are forecasting the number of CLOs that are both eligible and incentivized to refinance or reset. Therefore, the cumulative count of eligible and incentivized CLOs does not decrease due to refinances and resets occurring, but the count may decrease as older CLO vintages fail to meet eligibility criteria, or as market spreads increase.

Historically, 8%-10% of cumulative refinance/reset contenders go through with a refinance/reset during each month. This rate varies greatly depending on current spreads and market conditions, and we do not factor its effect into our analysis. In practice, however, it means that if there are 100 eligible and incentivized CLOs in month one and the refinance/reset rate is 10%, then 10 CLOs would refinance/reset in month one. Assuming market spreads do not shift from month one to month two, no new CLOs become eligible or ineligible in month two, and the refinance/reset rate remains 10%, then nine CLOs (i.e., 10% of the remaining 90) would refinance/reset in month two.

We present four scenarios, each corresponding to a different spread environment. Each is characterized by a path of BSL CLO 'AAA' spreads over SOFR. Note that we set MM 'AAA' spreads 60 bps higher than BSL 'AAA' spreads.

Scenario 1: BSL CLO 'AAA' spreads constant at SOFR + 160 bps

Scenario 1 assumes average market BSL 'AAA' spreads remain constant at 160 bps from February to December 2024, resulting in a cumulative total of about 240 eligible and incentivized deals by year end. The count of refinance/reset contenders ticks up in July and October as groups of CLOs from the 2022 vintage exit their non-call periods, but the count of CLOs from other vintages remains stable.

Chart 2

Scenario 2: BSL CLO 'AAA' spreads gradually tighten to SOFR + 140 bps

For scenario 2, we reduce BSL 'AAA' spreads by two basis points per month to a terminal spread of 140 bps by fourth quarter 2024. The 2021 CLO vintage accounts for the largest portion of eligible and incentivized CLOs by a wide margin (nearly 50% of the year-end total). New issue and refinance/reset CLO volume reached record highs in 2021, and assuming the entire 2021 vintage received upward credit spread adjustments (CSAs) of 26 bps post-LIBOR transition, it is no surprise that 2021 deals represent the bulk of potential refinance/reset issuance. Unlike scenario 1, the increase in refinance/reset contenders is driven by spread compression to the 140 bps range as opposed to cohorts of widely priced CLOs becoming eligible. Most 2021 CLOs exited their non-call periods prior to 2024 and become incentivized for a refinance/reset as spreads approach 140 bps.

Chart 3

Scenario 3: BSL CLO 'AAA' spreads rapidly tighten to SOFR + 125 bps

Scenario 3 calls for rapid 'AAA' spread tightening to 125 bps, with the pace of spread compression accelerating, and then leveling off toward the end of the year. The green line in the chart below reflects the average 'AAA' spread of CLOs that priced in 2021 (adjusted up by 26 bps). When the prevailing market spread line (purple) intersects the green line, potential refinance/reset issuance surges as the 2021 vintage becomes "in the money". Under these spread assumptions, the count of incentivized and eligible deals reaches nearly 1,500, which is a sizeable proportion of the total outstanding CLO universe.

Chart 4

Scenario 4: BSL CLO 'AAA' spreads gradually widen to SOFR + 180 bps

Based on current pricing trends, significant spread widening appears less likely than stability or gradual tightening. Bullish market sentiment, a growing CLO investor base, and the prospect of renewed CLO demand from bank treasuries could keep CLO liability costs capped. Forecasting spreads is a complex exercise, however, and market conditions can shift suddenly. In our fourth scenario, we assume BSL 'AAA' spreads increase by two basis points per month to a terminal spread of 180 bps by November 2024. Refinance/reset potential falls gradually until July, when CLOs with wider spreads from the 2022 vintage exit their non-call periods. This dynamic is also easy to visualize in scenario 1 (constant at 160 bps).

Chart 5

Areas For Further Research In Projecting Future Refinance/Reset Volume

Applying our framework to the data set we've compiled allows us to estimate the volume of refinances and resets that might arise in any given period, assuming (among other things) that managers exercise their options as soon as their deals are in the money. However, there are several reasons why this simplifying assumption isn't entirely realistic. For example, if a manager believes that spreads will continue to tighten, it may be better to wait before calling the deal even if it is in-the-money. That is, the spread incentive is not necessarily the only factor that determines the likelihood of a refinance or a reset at a particular point in time.

Other considerations can also influence the decision to refinance/reset a given CLO. These include par loss and the credit quality of the existing CLO collateral pool, which could necessitate a contribution (typically from the CLO's equity holders) of collateral. The addition of new collateral ensures the CLO will meet its target par level (required for the rating process) and have sufficient cushion for its collateral quality tests. A reset or refinance might also become less likely once a CLO has amortized a significant portion of its senior notes after exiting the reinvestment period.

We could address this uncertainty in the timing of refinances and resets by treating the time until a CLO (that meets the eligibility criteria outlined above) refinances or resets as a random variable. Such a random variable could in turn depend on observable predictors, such as the extent to which a deal is in the money, the size of the deal, whether it is a MM or BSL CLO, etc. This refinement could take the form of a survival analysis, calibrated on historical data at the deal level. A survival analysis is the topic of future research and will enhance the current framework by allowing us to incorporate relevant information about individual CLOs (e.g., whether asset par value is less than its target and whether the deal is underperforming), as well as information about the manager (e.g., whether the manager typically achieves strong pricing relative to the market and whether they have frequently reset in the past and have been able to access the market with ease).