CMStatistics 2022: Start Registration
View Submission - CFE
Title: Which hedge funds are systemically risky, and when: A dynamic extreme value regression approach Authors:  Philippe Hubner - HEC Liege, University of Liege (Belgium) [presenting]
Julien Hambuckers - University of Liege (Belgium)
Abstract: A novel approach is introduced to measure the time-varying systemic risk contribution of hedge funds at the fund level, overcoming short reporting periods in commercial databases. To do so, we extend the extreme value systemic risk model to a regression context, where marginal tail indices of hedge funds and banks are driven by a set of covariates. This formulation makes it possible to estimate systemic risk contributions by exploiting extreme value regression methods on pooled time series of hedge funds returns - in spite of the short reporting period of the funds. It also has the advantage of identifying whether a high level of systemic risk of a given fund originates from a high risk of spillovers to the banking sector, or the high level of the fund tail risk. These measures are then used to identify funds characteristics and market conditions that indicate a high systemic threat, an information of interest for regulators. Using a large sample of funds over the period 1994-2021, we find that investment strategies are clear determinants of the hedge funds' systemic risk.