Modeling Private Equity: A Combined Time-Series Approach

Abstract

We introduce a flexible time-series estimator that improves upon existing approaches at recovering factor exposures of buyout funds. Ours is the first that aims to recover individual systematic fund returns while remaining agnostic to the underlying return process. Under varying auto-regressive and moving-average simulations, we recover systematic exposures with less bias and variance than competing estimators. We apply our model to a newly available commercial data set from PitchBook and find alphas that are higher than those reported in recent literature. Furthermore, we find substantial time-series model fit improvement from our optimized estimator. The large dispersion of our estimated alpha and beta coefficients highlight the benefits of fund specific analysis compared to cross-sectional estimations.

Keywords: private equity, factor modeling, long horizons, overlapping, lagged, t-stats, supervised learning

JEL Classification: C01, C22