Hedge funds typically reveal some statistical properties like serial correlation, non-normality, volatility clustering, and leverage effect, which have to be considered when risk positions of hedge funds are computed. We describe an autoregressive Markov-Switching model that captures the specific features of hedge fund returns and allows especially to fit for volatility clustering and leverage effects in the data. The model is tested using publicly available hedge fund index data from different regions. We compare two different variants of the model by means of risk and performance measures. Our case study implies that if the leverage effect appears in the data, it is worth to fit for leverage in the parameter estimation process.     «

Hedge funds typically reveal some statistical properties like serial correlation, non-normality, volatility clustering, and leverage effect, which have to be considered when risk positions of hedge funds are computed. We describe an autoregressive Markov-Switching model that captures the specific features of hedge fund returns and allows especially to fit for volatility clustering and leverage effects in the data. The model is tested using publicly available hedge fund index data from different...     »