Stress Testing Private Credit

Abstract

Private credit is an exponentially growing market that is frequently described as mysterious and opaque. In this paper, I peel back some of that opacity, creating new data sources on private credit and making many of them publicly available. Like open-source software, the value of this information is not in that it is inherently perfect, but that it is transparent, meaning that errors that exist can be identified and corrected, improving its quality and reliability over time.

First, I create a new publicly available data set that provides detailed information on the size and strategy of every US private credit fund. This is based on extensive enhancements I perform, building off a base of regulatorily mandated disclosures that private funds file with the SEC. I show that the US private credit market is substantially larger than generally recognized, and I show how the data enables anyone to precisely calibrate their measure of private credit market size based on their preferred definition of private credit.

Second, I create a publicly available data set detailing the leverage used by private credit funds, with borrowing supplied primarily via banks. I build this by piecing together data from diverse sources such as regulatory filings, press releases, investor presentations, sponsor annual reports, and more. I use this data to estimate total bank lending to private credit, and I validate it by showing how yields earned by insurance company investments in private credit funds vary predictably based on underlying fund leverage.

Finally, I combine my newly developed public data sources with proprietary insurance data (which unfortunately cannot be publicly released) to precisely measure the exposure of individual insurance companies to different types of private credit funds, and I perform stress test analyses measuring expected losses to insurance companies in different hypothetical stress scenarios.