Published in London & New York
10 Queen Street Place, London
1345 Avenue of the Americas, New York
Creditflux is an
company
© Creditflux Ltd 2024. All rights reserved. Available by subscription only.
Analysis CLOs
Relating arbitrage, issuance and equity
by Poh-Heng Tan
Day one arbitrage remains a key metric for CLO equity investors. But how does the US BSL CLO arb today compare with historical averages? And does the figure impact deals’ equity returns?
Our analysis combines the discounted spreads of the Morningstar LSTA US B-BB Ratings Loan Index with the funding costs of CLO debt to get a proxy version of CLO arbitrage (see methodology). Chart 1 (below) shows the arbitrage since 2017.
As of 26 July 2024, arbitrage was 227bps, an improvement from earlier this year due to the tightening of CLO liabilities without a corresponding compression in asset spreads.
1: US BSL CLO new issue arbitrage (bps)
Source: Pitchbook Data, CLO Research
CLO tranches are priced tightly
Recently, demand has outstripped supply, as reflected by the fact that CLO tranches are priced tightly across the capital stack compared to historical norms. For example, historically, the median top-tier single A CLO tranche was around 231bps when the four-week average loan index’s discounted spreads ranged between 360bps and 390bps. However, in today’s market, single A tranches are priced firmly inside 200bps at the time of writing.
191
bps
Our arbitrage metric was at its lowest in 2018
Given the attractiveness of floating-rate assets and the strong track record of senior and mezzanine CLO tranches, it is no surprise to see solid demand across the capital stack. As shown in table 1 (below), an arbitrage level of 227bps (as of 26 July 2024) is slightly higher than the median for deals in both 2022 and 2023. The median YTD 2024 arbitrage metric is 211bps, which aligns with the 2019 median metric.
One challenge of looking at historical data is the change in reference rate from LIBOR to SOFR. To smooth the arbitrage trend, the 2021 vintage deals’ arbitrage metrics have been adjusted for the basis between LIBOR and SOFR. As shown in chart 1, the arbitrage line is smoother as it moves from 2021 into 2022. For example, two deals were priced within two months: King Park CLO in December 2021 and Boyce Park CLO in January 2022, both managed by the same manager. The arbitrage metrics for the two deals were 214bps and 217bps, respectively, after adjusting for the SOFR transition.
Low arb deals have not performed well
Our data shows that the arbitrage metric was lowest (at 191bps) in 2018, despite deals from that year having the tightest liability prints, because asset spreads were tight, too.
As more deals from this vintage are being called, it remains to be seen if these deals will perform well. So far, equity IRRs are not looking good. For example, based on over twenty 2018-vintage deals redeemed in Q2 2024, the median equity IRR was estimated to be in the mid-single-digit range — far from the targeted low-teens IRRs. This shows that a very tight primary liability CLO print, which locked in attractive long-term, non-recourse funding costs, does not necessarily translate into attractive returns.
243
bps
2020 deals had the best arbitrage, but new issue volume was low
Nevertheless, although arbitrage offers a glimpse into a deal’s initial position, it doesn’t promise success. As we’ve discussed in previous articles, countless factors shape a deal’s final equity IRR. For example, based on recent observations, redeemed deals from the 2020 and 2022 vintages experienced much better average IRRs than those from other vintages. These deals were issued during periods of volatility, allowing loans to be purchased at a good discount. Subsequently, the loan market rallied, and the pull-to-par effect translated to a solid final equity NAV, resulting in impressive IRRs for many.
Another observation is that the arbitrage picture and new CLO issuance do not always go hand in hand. Chart 2 highlights that the arbitrage metric and issuance volume are not correlated. This year’s median arbitrage metric YTD is actually lower than in both 2022 and 2023, but this year’s new issue volume has already far exceeded that of the past two years (on an annualised basis).
2: US BSL CLO arbitrage vs issuance
* Based on deals with 3-year reinvestment period ** Annualised based on new issue issuance in the first half of 2024
Source for all charts and tables: Pitchbook Data, CLO Research
Poor arb in a record-breaking year
Similarly, 2018-vintage deals had the lowest median arbitrage metric at 191bps, yet 2018 was a record-breaking year for new issue CLO volume. On the other hand, 2020-vintage deals had the best median arbitrage metric at 243bps, but new issue volume was low.
This highlights that the arbitrage picture is only one of many factors driving new issuance. The strength of the loan market, investor confidence, loan supply, CLO amortisation rate and redemptions, demand for CLO paper, captive CLO equity dry powder and outstanding CLO warehouses all play a role in the timing and volume of new issuance.
Methodology
CLO arbitrage is the difference between the returns on the loans owned by a CLO and the funding costs of its debt.
For our analysis, we calculated a version of arbitrage using the leveraged loan index’s discounted spread net of the cost of funding, based on discount margins (of triple A to double B tranches) rather than spreads. The discounted spread metric is used for the calculation of the arbitrage metrics as it captures the discounted prices of loans.
The loan index’s moving 4-week average discounted spreads are used as a proxy for US BSL CLO portfolios’ discounted spreads. The loan index used for this analysis is the Morningstar LSTA US B-BB Ratings Loan Index.
Based on our study, US CLO managers have largely tracked this index’s performance, indicating that it is not easy to outperform it. This consistency in tracking the index’s performance is one reason we believe this index is a suitable proxy for the US BSL CLO portfolios’ discounted spreads at issuance.
CLO tranches issued at a discount are a cost to equity investors. For a like-for-like comparison, new issue upfront costs and management fees are not accounted for in the calculation of the arbitrage figures.
