论文部分内容阅读
以我国黄金现货市场主力品种——Au99.95的日数据为研究对象,运用滚动时间窗法进行了多种波动率模型和不同条件收益分布假定下的样本外动态VaR预测。进一步,运用更加严谨和稳健的KupicLR检验以及动态分位数回归检验法,对不同模型得到的VaR预测精度进行了深入的后验分析。主要实证结果显示,我国黄金现货市场的波动不存在显著的杠杆效应,但却同时具有明显的条件“有偏”和“尖峰胖尾”特征。另外,假定条件收益服从有偏学生分布的EGARCH模型具有最好的样本外极端风险预测精度。
Taking the daily data of Au99.95 as the research object, this paper uses the rolling time window method to conduct a variety of volatility models and the dynamic VaR forecasting under the assumption of different conditional income distribution. Further, using more rigorous and robust KupicLR test and dynamic quantile regression test, the VaR prediction accuracy obtained by different models is deeply analyzed. The main empirical results show that there is no significant leverage effect on the fluctuation of the gold spot market in our country, but at the same time, it has obvious characteristics of “partial” and “peak fat tail”. In addition, it is assumed that the conditional return obeys the EGARCH model with deviant distribution and has the best accuracy of exogenous risk prediction.