Risk quantification and validation for Bitcoin

This paper introduces a semi-nonparametric approach for modeling Bitcoin risk relatively to other parametric distributions and volatility models. Model performance is assessed through different backtesting techniques, including multinomial test, for three risk measures: Value-at-Risk, Expected Shortfall and Median Shortfall. Our results show that the 'large' semi-nonparametric expansion is a good alternative to measure Bitcoin risk according to recommendations of Basel Committee on Banking Supervision, but also that 99%-Median Shortfall seems to be an accurate and robust risk measure for Bitcoin. (C) 2020 Elsevier B.V. All rights reserved.