Predicting Private Equity Fund Returns

Abstract

This thesis investigates the potential of a Private Equity fund performance forecasting model, to assist Private Equity investors in their investment decision making process. Fund performance is measured by the fund’s Kaplan Schoar Public Market Equivalent and is forecasted using a binary classification approach. The top performing Machine Learning models are able to forecast Buyout fund performance with 63 % accuracy, and Venture Capital fund performance with 66 % accuracy. Therefore, the features used to train the models and selected based on the literature on Private Equity performance drivers, possess important predictive power, which can be integrated in the investment procedure.