Portfolio Management in Private Equity

Abstract

We study how private equity general partners (GPs) make portfolio management decisions. Using a novel deal-level dataset of 5,925 global investments from 1999 to 2016, we take a portfolio view of private equity funds to reconcile various findings in the PE literature. We document several new findings consistent with GPs trading off the benefits of focusing their skills on a relatively small number of portfolio companies with high levels of (especially idiosyncratic) risk. First, we find that a higher degree of industry or geographic concentration is associated with both higher fund returns as well as a higher fund risk. Second, we find that instead of betting more on high-returning “best ideas” as is common in portfolios of public companies (Antón, Cohen, and Polk, 2021), the largest investments in PE portfolios have the lowest returns on average but are also the lowest risk in the fund. Third, extending the Bayesian estimation of hierarchical models in Korteweg and Sorensen (2017), we find that skill only accounts for 4%-6% of the total return variation of a typical investment and that almost all remaining variation is idiosyncratic. However, GP skill accounts for more than 40% of the return variation at the fund level (i.e., after accounting for diversifiable deal-level risk) consistent with fund structures reducing skill-related information asymmetry between LPs and GPs. Overall, our findings suggest that GPs carefully consider how they construct their portfolios in an attempt to generate high risk-adjusted fund-level returns that maximize their long-run franchise value.

Authors

Greg Brown, UNC Kenan-Flagler Business School
Celine Yue Fei, UNC Kenan-Flagler Business School
David T. Robinson, Duke University Fuqua School of Business and NBER