Title: Nonparametric quantile regression for double censored data with application to stock markets with price limits
Authors: Chi-Yang Chu - National Taipei University (Taiwan) [presenting]
Abstract: Quantile regression is often used in the analysis of stock return-volume relations. Many countries impose upper and lower limits on stock returns and losses, respectively, in order to reduce price volatility. However, this double censored property appears to be ignored in the literature. To analyze stock markets with price limits in a proper fashion, a nonparametric quantile regression model for double censored data is proposed. The proposed estimator performs well in simulations. In the application to Taiwanese stock markets, the proposed approach seems to alleviate some potential biases arising from double censored data. Specifically, the proposed estimator suggests larger estimated losses via conditional value at risk.