【摘要】 在金融市场中，对金融资产的尾部风险的精准度量一直是研究者关注的焦点。本文提出了一个新的二元时间序列模型，用来计算和预测金融资产的在险价值（Va R）和条件在险价值(Co Va R)。该模型可以同时捕捉二元时间序列中存在的序列相关性和横截面相关性，从而提高估计和预测的精度。本文在模型推导中给出了该二元时间序列模型的参数估计值，并且基于plug-in方法给出了Va R和Co Va R的估计值。还建立了Dvine模型估计量的渐近性质。对金融股价的实证分析表明我们的模型在风险度量和预测方面表现良好。更多还原

【Abstract】 Accurate measurements of the tail risk of financial assets are major interest in financial markets. The main objective of our paper is to measure and forecast the value-at-risk(Va R) and the conditional value-at-risk(Co Va R) of financial assets using a new bivariate time series model. The proposed model can simultaneously capture serial dependence and cross-sectional dependence that exist in bivariate time series to improve the accuracy of estimation and prediction. In the process of model inference, we provide the parameter estimators of our bivariate time series model and give the estimators of Va R and Co Va R via the plug-in principle. We also establish the asymptotic properties of the Dvine model estimators.Real applications for financial stock price show that our model performs well in risk measurement and prediction.更多还原