Private Credit, Setting the Record Straight
A recent academic study claiming private credit “barely outperforms” public market benchmarks has sparked debate in the investment community. However, the study’s methodology raises serious questions. By using an inappropriate performance metric (TVPI instead of industry-standard internal-rate-of-return or time-weighted returns) and comparing private drawdown funds against ETFs without accounting for leverage, the research overlooks critical factors that explain private credit’s value proposition. When analyzed correctly using time-weighted returns and adjusting for leverage, private credit funds demonstrate a consistent 2.83% annual illiquidity premium over comparable public market investments, with lower default rates than high-yield bonds and 20 years of data supporting sustained outperformance.
The Controversy
Bloomberg columnists Rene Ismail and Kat Hidalgo generated significant industry discussion with their October 16, 2025 article “Private Credit on the Defensive Again Over ‘Mark-to-Myth’ Study.” The piece highlighted the study “Residual Risk: Benchmarking the Boom in Private Credit,” which was published in the Fall 2025 issue. The authors Jeff Hooke, Xiaohua Hu, and Michael Imerman compared private debt fund performance against publicly traded ETFs holding broadly syndicated loans (BSL)—using live funds as benchmarks instead of indices.
The study examined vintage years 2015 to 2020, focusing on senior and mezzanine private debt funds. The authors’ conclusion was stark: “private credit funds barely outperform, or in some cases underperform, the benchmarks.” In his interview with Ismail and Hidalgo, Hooke declared, “The two main marketing points of the industry seem to be illusory.”
The Methodological Flaw
The study’s central problem lies in its choice of performance metric. Rather than using industry-standard measures like internal rate of return (IRR) or time-weighted return (TWR), the authors employed Total-Value-to-Paid-In (TVPI) capital. This is a critical error.