B1601
Title: Efficient estimation for EV regression models of tail risks
Authors: Julien Hambuckers - University of Liege (Belgium)
Marie Kratz - ESSEC Business School, CREAR (France) [presenting]
Antoine Usseglio-Carleve - Avignon Université (France)
Abstract: A method is introduced to estimate simultaneously the tail and the threshold parameters of an extreme value regression model. The standard model finds its use in finance to assess the effect of market variables on extreme loss distributions of investment vehicles such as hedge funds. However, a major limitation is the need to select an ex-ante threshold below which data are discarded, leading to estimation inefficiencies. To solve these issues, the tail regression model is extended to non-tail observations with an auxiliary splicing density, enabling the threshold to be selected automatically. An artificial censoring mechanism of the likelihood contributions is then applied in the bulk of the data to decrease specification issues at the estimation stage. The superiority of the approach is illustrated for inference over classical peaks-over-threshold methods in a simulation study. Empirically, the determinants of hedge fund tail risks are investigated over time, using pooled returns of 1,484 hedge funds. A significant link between tail risks and factors such as equity momentum, financial stability index, and credit spreads is found. Moreover, sorting funds, along with exposure to the tail risk measure, discriminate between high and low alpha funds, supporting the existence of a fear premium.