THE JOURNAL OF FINANCE •VOL. LXXVIII, NO. 3 •JUNE 2023
CLO Performance
LARRY CORDELL, MICHAEL R. ROBERTS, and MICHAEL SCHWERT*
ABSTRACT
We study the performance of collateralized loan obligations (CLOs) to understand the
market imperfections giving rise to these vehicles and their corresponding economic
costs. CLO equity tranches earn positive abnormal returns from the risk-adjusted
price differential between leveraged loans and CLO debt tranches. Debt tranches of-
fer higher returns than similarly rated corporate bonds, making them attractive to
banks and insurers that face risk-based capital requirements. Temporal variation in
equity performance highlights the resilience of CLOs to market volatility due to their
closed-end structure, long-term funding, and embedded options to reinvest princi-
pal proceeds.
COLLATERALIZED LOAN OBLIGATIONS (CLOS)HAVE RECEIVED a great deal
of attention in recent years because of their rapid growth and broad reach.
Standard & Poor’s (2020b) reports that two-thirds, or $2.1 trillion, of lever-
aged loan issuance since the 2008 financial crisis has been funded by CLOs. A
broad array of financial institutions invest in CLOs, including banks, insurers,
*Larry Cordell is with the Federal Reserve Bank of Philadelphia. Michael R. Roberts is with
the Wharton School, University of Pennsylvania and the National Bureau of Economic Research.
Michael Schwert is with AQR Arbitrage, LLC. Schwert was with the Wharton School during the
drafting of this paper. Wethank Jeremy Brizzi; Alan Huang; Yilin Huang; Akhtar Shah; and the
customer support team at Intex Solutions for their invaluable assistance in constructing the data
set for this project; Stefan Nagel (the Editor); an anonymous Associate Editor; two anonymous ref-
erees; Xudong An; Bo Becker; Darrell Duffie; Daniel Green; Fotis Grigoris; Fred Hoffman; Chris
James; Arthur Korteweg; Mark Mitchell; Taylor Nadauld; Jordan Nickerson; Greg Nini; Yoshio
Nozawa; Matt Plosser; Todd Pulvino; Bill Schwert; Serhan Secmen; Rob Stambaugh; René Stulz;
Fabrice Tourre; Stephane Verani; John Wright; seminar participants at the Corporate Finance
Virtual Seminar series, Dartmouth College, Federal Reserve Bank of Chicago, Federal Reserve
Bank of Philadelphia, Frankfurt School of Finance, London Business School, New York Univer-
sity, Ohio State University, Rutgers University, University of Florida, University of Rochester,
Wharton; and conference participants at the SFS Cavalcade, WFA, and NBER Summer Institute
for helpful comments. We gratefully acknowledge financial support from the Jacobs Levy Equity
Management Center. The views expressed in this paper are those of the authors and do not nec-
essarily reflect the position of the Federal Reserve Bank of Philadelphia or the Federal Reserve
System. AQR Capital Management is a global investment management firm, which may or may
not apply similar investment techniques or methods of analysis as described herein. The views
expressed here are those of the authors and not necessarily those of AQR. The authors have read
The Journal of Finance disclosure policy and have no conflicts of interest to disclose.
Correspondence: Michael R. Roberts, The Wharton School, University of Pennsylvania, 3620
Locust Walk, Suite 2400, Philadelphia, PA19104; e-mail: mrrobert@wharton.upenn.edu.
DOI: 10.1111/jofi.13224
© 2023 the American Finance Association.
1235
1236 The Journal of Finance®
pension funds, mutual funds, and hedge funds. As a result, U.S. and European
regulators have expressed concerns about the growth of the CLO market and
the financial system’s exposure to these vehicles (MarketWatch (2019), Stan-
dard & Poor’s (2020a)).
In this paper, we address two questions arising from the growth of CLOs.
What market imperfections are CLOs designed to address, and how large are
the economic costs of these imperfections? With perfect capital markets, there
is no role for CLOs, or securitization more broadly, because economic agents
can costlessly transform cash flows. Therefore, CLOs exist because of market
imperfections. We test the implications of different imperfections for the per-
formance of CLO assets and liabilities. In doing so, we provide the first large-
sample evidence on CLO performance, shedding light on the risks and rewards
of these vehicles.
We begin by constructing a novel data set that offers a near-comprehensive
view of the CLO market. The data include the full history of cash distribu-
tions to every CLO tranche, as well as information on contract terms, collateral
holdings, and trading activity. The sample period, August 1997 to March 2021,
encompasses three distinct business cycles including the financial crisis and
the first year of the COVID-19 pandemic.
Our central finding is that CLO equity tranches provide statistically and eco-
nomically significant abnormal returns, or “alpha,” against a variety of public
benchmarks. Using the generalized public market equivalent (GPME) frame-
work of Korteweg and Nagel (2016), we find that the average completed CLO
equity investment offers a net present value (NPV) of $0.66 per dollar invested,
net of fees. This estimate equates to approximately $33 million, or 6.6% of
total assets, for the typical deal. The exact magnitude of the NPV estimate
depends on the stochastic discount factor (SDF) specification, but all of the es-
timates are economically large and statistically significant. Before-fee GPME
estimates imply that managers capture approximately 40% of the before-fee
surplus. Because many managers fund a portion of the equity tranche in the
CLOs they manage, this fraction is a conservative estimate of the total com-
pensation of CLO managers.
Since equity investors receive the residual cash flow from the collateral
pool after debt tranches are paid, these abnormal returns must be due to
risk-adjusted price differentials between the leveraged loans in the collateral
pool and the debt tranches issued to finance the vehicle. We confirm this re-
lation by applying the GPME framework to the cash flows produced by CLO
collateral and debt tranches. The point estimates show that debt tranches offer
lower risk-adjusted returns than loan collateral, but the difference is statisti-
cally insignificant due to the low volatility of these cash flow streams and the
relatively short time series.
Closer inspection reveals that the abnormal equity returns are concentrated
among CLOs originated before 2010, so-called “CLO 1.0” transactions, and in
particular by CLOs issued just before the onset of the financial crisis. CLOs
issued in 2006 and 2007 locked in low-cost financing prior to the crisis and
reinvested in high-yielding loans during and after the crisis. The result was
CLO Performance 1237
a windfall of excess interest and principal for CLO equity investors as the
economy recovered. CLOs have also proven resilient to the COVID-19 crisis,
which has thus far had a negligible effect on equity distributions.
This resilience is attributable to several structural features of CLOs. First,
CLOs are closed-end vehicles in which capital inflows and outflows are limited.
Second, coverage tests are based on par values and credit ratings instead of
market prices. Consequently, market volatility does not cause the diversion of
cash flows to pay down debt tranches unless the volatility coincides with rating
downgrades and defaults. Third, embedded options to reinvest collateral and
reissue debt after a noncall period enable opportunistic trading and refinanc-
ing by CLO managers. Finally, CLOs employ a long-term funding structure
known as “term leverage” that insulates the vehicle from rollover risk. Unlike
most levered investment vehicles that use short-term debt (e.g., hedge funds),
CLOs issue long-term debt with maturities in excess of seven years and fixed
credit spreads.
We investigate several potential economic mechanisms behind the perfor-
mance results. We find that the pool of leveraged loans comprising CLO assets
generates gross returns that are economically indistinguishable from a broad-
based index of leveraged loans. Net-of-fee returns are similar to those gener-
ated by a diversified portfolio of loan mutual funds. Thus, the average CLO
manager does not exhibit skill in selecting leveraged loans relative to other
market participants, though some managers do produce persistent outperfor-
mance relative to their peers.
Informational frictions that form the basis of traditional securitization the-
ories (e.g., Glaeser and Kallal (1997), Riddiough (1997), DeMarzo and Duffie
(1999), DeMarzo (2005)) are also unlikely responsible for our findings. The
vast majority of CLOs are “open-market” transactions in which collateral is
acquired through participation in broadly syndicated loans (BSLs) arranged
by banks. This setting leaves relatively little scope for adverse selection, a con-
jecture confirmed by Benmelech, Dlugosz, and Ivashina (2012), who show that
securitized loans perform no differently than nonsecuritized loans.1Moral haz-
ard has little role in explaining our results, as CLO managers do not expropri-
ate value from debt investors by trading into riskier credits after issuance.
Thus, the typical CLO does not appear to capitalize on imperfect information
about the quality of loans in the collateral pool.
What CLOs do appear to capitalize on are the incentives of regulated fi-
nancial intermediaries. Like CLO equity, debt tranches offer higher returns
than public benchmarks. Discounting CLO debt cash flows using the returns
of rating- and duration-matched corporate bonds, we find public market equiv-
alent (Kaplan and Schoar (2005)), or PME, estimates for debt tranches that are
1We do find that the subset of middle-market (MM) CLOs, for which collateral comprises loans
arranged by the CLO manager,earn higher risk-adjusted returns on collateral and equity tranches
than BSL CLOs. However, this difference is statistically insignificant due to the small number of
MM deals, it suggests a role for CLOs in mitigating the costs of adverse selection where informa-
tion asymmetry is most pronounced.
