Competing Risks in Hedge Fund Lifetimes

Abstract

Recent evidence suggests that institutional investors wish to commit to hedge funds on a long-term basis, but they are worried that their investments may liquidate prematurely, leading to large capital losses. Providing them with rigorous models to predict liquidation is important, yet current studies of hedge fund lifetimes and mortality have treated all funds that exit the database as liquidated, when many of these funds are not liquidated. This has likely produced imprecise estimates of hedge fund lifetimes, attrition, and survivorship bias. This study aims to reconcile these important issues by applying a competing risks model to account for the different types of exits hedge funds can experience. By treating exit types separately, we avoid blurring the effect of predictor variables on survival because the variables are not attempting to predict a heterogeneous aggregation of exit types. Predictor variables whose value changes over time, such as returns or volatility, are modeled as time dependent variables rather than as fixed variables. This provides better warning signals to investors about possible fund liquidation, since the impact of the predictor variables is measured at every instant of hedge fund lifetimes, rather than during the last months of their life as is done in existing studies. Hedge funds that exit the database for reasons other than liquidation often have very good returns, so the direction of survivorship bias is not always evident. We avoid underestimating survivorship bias because funds that are not liquidated are not counted as dead. We find that separating exit types leads to estimates of annual attrition that are not as high as previously thought. The results of this study will allow investors to better estimate the expected lifetime of a hedge fund before allocating any money to it, and provide better warning signals to investors already committed to a hedge fund about possible fund liquidation, thereby helping all investors avoid the large capital losses that are often associated with liquidation.