Title: A semiparametric estimation of value-at-risk and its applications
Authors: Shih-Feng Huang - National Central University (Taiwan) [presenting]
Abstract: A semiparametric approach consisting of GARCH-type models and a generalized nearly-isotonic regression (GNIR) is proposed for Value-at-Risk (VaR) estimation. The GNIR is capable of depicting the up/down fluctuation of data automatically. An algorithm for the GNIR is proposed and its convergence property is derived. The proposed VaR estimator, denoted by NVaR, is shown to have fewer fluctuations than the VaR estimators obtained from GARCH-type methods. It also has better timely responses to market risks than the VaR estimators obtained from extreme value theory (EVT). We apply the NVaR to compute capital requirements and employ the daily indices of 13 global financial markets from 2003 to 2017 for our empirical investigation. The numerical results reveal that the NVaR is capable of satisfying the backtesting, reflecting market risks in time, and reducing the fluctuations of the VaR sequence and capital requirements simultaneously.