Estimating Hedge Fund Leverage: A Three‐Step Estimation Protocol

ABSTRACT Utilizing a micro‐level hedge fund dataset, we propose a methodology for estimating hedge fund leverage. Initially, we perform a Principal Component Analysis on a set of 49 risk factors for dimension deduction purposes. After acquiring 10 Principal Components, we deploy the Least Absolute Shrinkage and Selection Operator regression (Lasso) per fund by seven 3‐year monthly non‐overlapping intervals in order to select which Principal Components affect each fund's return. As a last step, we execute a regression in the same manner as previously, with only the non‐zero Principal Components. By aggregating , we estimate an average sectorial leverage of 3.3 with an average of 58.2%. Moreover, we observe an analogous degree of Deleveraging in 2007–2009 that includes the 2008 financial crisis as in 2019–2021 that includes the COVID‐19 stress period.