| Summary: | This dissertation explores the ability of risk measures to explain cross-sectional differences in future hedge fund returns from 1994 to 2012, using two databases: Lipper TASS and Barclay Hedge. We analyse the predictive ability of total risk, skewness and kurtosis and then, using factor models, decompose total risk into its systematic and unsystematic components, testing how well these perform in explaining cross-sectional differences in future hedge fund returns. Particular attention is given to the performance of risk measures under stressful conditions, namely periods with systemic shocks and periods of poor hedge fund sector performance. We find that results vary significantly during periods affected by systemic shocks and in periods in which the sector performs poorly, thus arguing that the predictive ability of risk measures might be limited. A possible explanation can lie in significant cross-fund differences in managers’ ability to properly time and adjust their risk exposures.
|
|---|