Back to Insights

Private Debt Fund Returns, Persistence, and Market Conditions

Feb 27, 2023

Private Debt Fund Returns, Persistence, and Market Conditions

Abstract
This paper examines net-of-fees private debt fund performance, performance persistence across funds managed by the same general partner and a general partner’s ability to time the market. We document that private debt funds outperform bond and equity market benchmarks in the cross-section, with high performance dispersion across strategies and performance quartiles. Lagged performance significantly affects current fund performance. While ex ante and ex post credit market conditions strongly affect fund performance, general partners can only partially time them.

Introduction
We investigate private debt (PD) fund performance and determinants thereof. PD funds represent an important segment of the private capital industry, which soared on the boom in unlisted assets and tripled their market capitalization since the COVID-19 pandemic induced market sell-off.Footnote1 PD funds emerged as an asset class in the late 1990s and exceeded $1.1 trillion assets under management in 2020 (Preqin Pro Citation2021). As of today, PD funds’ assets under management represent some important 12.3% of the aggregate value of private capital funds. They approximately match the size of real-estate funds ($1.15 trillion) and have outgrown infrastructure ($0.8 trillion) and natural resources ($0.2 trillion) funds (Preqin Pro Citation2021). This growth has been driven by a surge in the demand for non-bank private debt, as banks retrenched from cash-flow-based lending to the middle market after the Global Financial Crisis due to increased bank regulation and the resulting reduction in risk appetite on the part of the banks (see, for example, Langfield and Pagano Citation2016; van der Veer and Hoeberichts Citation2016; Bordo and Duca Citation2018; Cortés et al. Citation2020). Also, PD fund growth was spurred by an increase in the supply of capital by yield-seeking institutional investors challenged by a low-yield environment in traditional credit markets.

Despite the growing importance of PD funds, which have reached average fund sizes exceeding $1.3 billion (in 2018 US dollars), our understanding of PD fund returns to limited partners (LPs) is limited to date. While there is a large literature on the returns to private equity investing, this study is one of the first to investigate thoroughly the returns to PD fund investing, and especially their outperformance of traded fixed-income benchmarks. Moreover, we analyze whether there is performance persistence over subsequent funds managed by the same general partners (GPs), and whether superior GPs can time credit market conditions.

In a typical private debt transaction, credit funds lend capital to an existing corporation. Each PD fund follows an investment strategy, such as direct lending, distressed debt lending, mezzanine lending, special situations lending and venture debt lending.Footnote2 Comparable to private equity funds,Footnote3 PD funds are organized as partnerships, in which the investor becomes a LP and the asset manager, which also invests in the partnership, is the GP. We investigate closed-end PD funds. Such funds imply that the investor cannot withdraw funds until the fund is liquidated, typically eight to ten years after inception. GPs are tasked with the selection of attractive credit opportunities, the negotiation of lending contracts, their execution, the active monitoring of investments and investee companies, sometimes combined with the execution of advisory roles at the management or board level of the borrower, the renegotiation of lending agreements in case of covenant breaches, the execution of credit workouts and the realization of secondary market transactions. For its services, the GP receives a management fee of 1.5% to 2% and typically also a success fee in the amount of 15% to 20%, the latter paid at the end of the lifetime of a fund and calculated on a return exceeding a preferred return to LPs, which is usually around 6% to 8% per annum.

The LP is a passive investor, does not obtain any decision control over investments and therefore has no influence on the selection and implementation of investments or on the general investment strategy. Both elements, however, are laid out in a detailed limited partnership agreement (LPA) in principle. Given the limited life of a fund, GPs must raise new funds. In general, such new funds are launched when either the existing fund comes to the end of its life or when 75% or more of the committed capital of the current fund has been called from investors. The life of a PD fund hence consists of (i) a fundraising period, in which investors commit capital throughout a series of closings and in which the GP is tasked with deal sourcing, deal evaluation and executing first investments; and (ii) an investment period, in which the GP continues its activity, which lasts two to four years from the final closing in the fundraising period. During the investment period, the GP may recycle its capital, i.e., reinvest the proceeds of early exits. The investment period is followed by (iii) the harvesting period, in which the GP exits investments and distributes all proceeds, net-of-fees, to LPs.

PD funds generate returns from various sources. First, funds earn regular cash coupons paid on the loans. Second, GPs may structure payments in kind (PIK), accruing interest paid out at maturity. Third, they may cash in early repayment penalties (Cumming et al. Citation2019) and recycle or reinvest capital from early repayments. Fourth, portfolio company’s fees may include advisory fees, transaction and deal fees, directors’ fees, monitoring fees, capital market fees, organization cost compensations, placement fees and others (CalPERS Citation2015; ILPA Citation2016). These are paid to the GP, but typically paid back to the LPs by means of fee-offset provisions, that is, a reduction in management fees, thereby boosting fund performance. Finally, funds may actively exit some of their investments on the secondary markets if valuations are good.

The first aim of this paper is to provide systematic evidence related to the performance a LP may expect from a PD fund. We first focus on the absolute and relative returns to PD funds, using relevant benchmarks. Thereafter, we analyze whether GPs are skilled in managing PD funds in two dimensions: First, we evaluate whether GPs provide performance persistence in subsequent PD funds. Persistence exists if some GPs possess specific skills that allow them consistently to perform better than their peers. Second, we analyze the ability of GPs to time the market, thereby providing more fine-grained insights into the GPs’ skillset. Borrowing from Kacperczyk, Nieuwerburgh, and Veldkamp (Citation2014, 1455), we define GP skill as “a general cognitive ability to pick [stocks] or time the market”.

We collect data on 448 PD funds with vintage years 1986 to 2018, and calculate PD funds’ net-of-fees Internal Rate of Return (IRR) and net multiples. We further calculate the excess return of PD funds compared to public benchmarks, using the public market equivalent (PME) method (Kaplan and Schoar Citation2005).

We find that the average PD fund renders a 9.19% net-of-fees IRR to LPs. There is a large dispersion between top quartile funds, with an IRR of 23.3%, as compared to the bottom-quartile funds, with an IRR of −3.6%. PD funds achieve a net investment multiple of 1.3 in the cross-section, again with large performance dispersion between top quartile funds (1.76X) and bottom quartile funds (0.98X). PD funds outperform the investment-grade (IG) bond market benchmark with 8%, the high-yield (HY) bond market benchmark with 6% and the S&P500 equity market benchmark with 6%, with relatively equal outperformance across different investment strategies. Against these same benchmarks, top quartile funds reach a market outperformance of 38%, 33%, and 42%, while bottom quartile funds underperform the market by −18%, −19%, and −21%. These results echo Munday et al. (Citation2018), who find an average IRR of 8.1% and a market outperformance of 6.2% to 9.8% in the cross-section.

Multivariate analyses suggest that a GP’s prior fund performance is a significant and economically important predictor of its future fund performance: A one standard deviation increase in IRR (net multiple, PME IG, PME HY, PME S&P500) of the previous fund increases the performance of the current fund by 3.42% (0.11X, 5.10%, 4.88%/8.73%). However, our persistence results are largely driven by mature predecessor funds with at least 75% of capital called. Past performance of early-stage funds should thus be considered with more caution when considering an investment in a new PD fund.

Moreover, we find that a higher ex ante level of funding illiquidity, as proxied by the level of the Treasury-EuroDollar rate (TED) spread, significantly and negatively affects the outperformance of a PD fund against the IG and HY benchmark. On the contrary, a higher ex ante level of credit risk spreads and equity market volatility are positively related to PD fund multiples and their outperformance against the IG or HY benchmark. These findings are in line with prior research, which shows that ex ante credit market conditions affect the performance of debt investments (Cumming and Fleming Citation2013; Cumming et al. Citation2019).

We extend our analysis and, to the best of our knowledge, are the first to test whether and how ex post credit market conditions affect fund performance. Do GP’s have market timing skill and anticipate changes in ex post credit market conditions? We find mixed evidence for market timing skills. On the one hand, PD funds that are initiated in periods with ex post improvements in funding illiquidity (or a TED spread contraction) perform better. On the other hand, improvements in credit spreads affect fund performance negatively. Ex post changing equity market volatility, as proxied by the VIX, does not impact PD fund returns or outperformance. The effect size of changes in ex ante funding illiquidity is approximately more than twice as large as that of performance persistence, while the effect of a change in ex post funding illiquidity is about as important as past performance.

This finding is of particular interest to investors committing capital to PD funds in their fundraising period, which can take two years or more after the inception of the fund. As funds typically have a series of closings, diligent analysis of changes to funding illiquidity and credit spreads during the fundraising and the investment period may importantly improve an LP’s investment decision. According to our estimation, performance is highest when credit spreads expand ex ante and ex post, whilst TED spreads contract, i.e., when funding liquidity improves.

We contribute to the literature in various dimensions. First, we extend the previously sparse empirical evidence that PD funds offer attractive returns (Munday et al. Citation2018). Second, we are, to the best of our knowledge, the first to investigate performance persistence of PD funds. Third, previous literature remains quiet on the question of market-timing skill. We present a model to analyze skill and find that it significantly affects PD fund performance.

Literature
PD fund performance has not received a lot of attention in the academic literature.Footnote4 Studying loan portfolios of major US life insurance companies, Carey (Citation1998) found that private corporate loans have lower default and higher recovery rates than public bonds. Cumming and Fleming (Citation2013) study the performance of 311 loans used by private firms across 25 countries between 2001 and 2010 and find that performance depends on the portfolio size per manager, highlighting the role of time allocation for due diligence and monitoring. Cumming et al. (Citation2019) study more than 400 loans acquired by PD funds in 13 Asia-Pacific markets between 2001 and 2015. They find that trading private debt delivers higher returns than buying and holding a primary issuance.

In contrast to these studies focusing on gross returns on individual investments, we focus on PD fund net-of-fee investor returns, following Munday et al. (Citation2018). Using the Burgiss database, they analyze net returns to LPs of 476 private credit funds and 155 direct lending funds and find positive IRRs for the top three quartiles across all investment strategies.

Persistence in fund returns is defined as performance persistence across funds of the same GP. In competitive financial markets, GPs shall capture the returns to their skill by either growing fund size or increasing fees, thereby eliminating persistence in net-of-fee performance (Berk and Green Citation2004). Contrary to the Berk and Green (Citation2004) model, various empirical studies find persistence in PE or VC fund performance (Kaplan and Schoar Citation2005; Kaplan and Sensoy Citation2015; Nanda, Samila, and Sorenson Citation2020; Korteweg and Sorensen Citation2017). Individual partners employed by the GP also show persistent returns (Ewens and Rhodes-Kropf Citation2015).

A skill that might explain persistent returns is market timing. Previous literature finds mixed results with respect to the market-timing ability of asset managers. While some studies find evidence for market-timing ability (Ball, Chiu, and Smith Citation2011; Kim and In Citation2012; Cao, Simin, and Wang Citation2013; Chen, Adams, and Taffler 2013; Kacperczyk, Nieuwerburgh, and Veldkamp Citation2014; Yi et al. Citation2018; Jenkinson, Morkoetter, and Wetzer Citation2018), others find mixed or negative evidence (Carhart Citation1997; Elton, Gruber, and Blake Citation2012; Ferson and Schadt Citation1996; Andreu, Matallín-Sáez, and Sarto Citation2018; Bodson, Cavenaile, and Sougné Citation2013; Tchamyou and Asongu Citation2017). More recent studies show that corporates can time the markets when issuing bonds (Frank and Nezafat Citation2019) or equity (Wadhwa and Syamala Citation2019). It is thus an empirical question whether GPs of PD funds do possess market-timing skills.

Sample and Data

Sample Description

We use a worldwide PD dataset obtained from Preqin, which contains fund-level dataFootnote5 based on the Freedom of Information Act (FOIA), or their equivalent outside the US. Cash-flow data from Preqin are increasingly used in academic research on PE funds and found to be reliable (Ang et al. Citation2018; Barber and Yasuda Citation2017; Phalippou Citation2014; Kaplan and Waldrop Citation2016).

Our sample consists of 448 PD funds raised between 1996 and 2018; cash-flow data extend to December 2020.Footnote6 The sample size of this study is comparable to early research on the performance of private equity funds. Kaplan and Schoar (Citation2005), for example, draw conclusions on the performance of buyout (VC funds) using a sample of 169 (577) funds. Table 1 provides summary statistics of our sample.