Estimating of CVaR and Analysis of Leverage Effect Based on Nonlinear Quantile Regression Model

Abstract

In most articles, the leverage effect is often analyzed by ARCH type models from the angle of volatility, and VaR is assumed to have linear relationship with volatility. In this paper, a new nonlinear quantile regression model is proposed, and we argue that the constant and slope varies with the realized volatility and yield. An empirical analysis of stock market of China based on above model is presented at last section. In addition, the CVaR is estimated conditioned on realized volatility, which can also verify leverage effect in stock market.