Do they always hedge? Applications of quantile regressions to risk measurement across hedge fund strategies

Abstract: Extending previous work on quantile regression analysis of hedge fund returns, this thesis explores risk factor exposures across performance quantiles of 22 hedge fund strategies from 1994 to 2016. Specific characteristics of hedge funds are taken into consideration. Factor exposures are found to vary considerably across performance quantiles, implying that common conditional mean regression models fail to capture factor dependencies across hedge fund performance periods. Potential applications of quantile regression analysis to risk measurement and management are illustrated with two examples of conditional quantiles and conditional stress tests.