有偏t分布在金融资产波动率和相关性研究中的应用.docVIP

有偏t分布在金融资产波动率与相关性研究中的应用 The Application of the Skewed t Distribution in The Research of the Volatility and Correlation of Financial Assets 尹莹 指导教师姓名：牛霖琳 助理教授 专 业 名 称：数量经济学 摘 要 金融资产收益率的波动率和相关性一直是金融理论和实践中的重要课题，一直受到国内外研究者的关注。在金融市场的风险管理中，估计和预测单个资产的风险和多组资产之间收益率的相关性对投资者的决策发挥着重要的作用。我们已经逐步认识到，金融资产收益率的分布在统计特征上表现了尖峰厚尾的现象，并非服从传统的正态分布，而在现阶段研究金融市场资产的波动率与相关性的大部分文献主要是基于资产收益率为正态分布的假设下进行。这篇文章的研究重点在于，在拒绝传统的金融资产收益率服从正态分布的假设的基础上，采用一种更能刻画收益率序列的统计分布——有偏t分布来刻画真实数据。我们抽取了中国股票市场上的三组资产和上证综合指数，通过观察这些资产收益率的统计特征表现，验证了资产收益率服从有偏t分布的假设更加符合真实数据的推想。在这样的基础上，我们将有偏t分布应用在估计单个资产波动率的模型以及估计多组资产收益协方差矩阵的模型中。无论是经济动态模型和静态模型中，我们的研究结果表明，有偏t分布模型比正态分布模型的估计效果都要好。此外，在研究多组资产收益率相关性的部分，我们是在传统的极大似然估计方法的基础上对其改进，并加入稳定的条件约束，提高估计的有效性，运用有约束的极大似然估计方法对资产收益相关性进行研究。 关键字：有偏t分布；资产收益率相关性；GARCH模型；有约束的极大似然估计方法 Abstract Research of assets return’s volatility and the covariance matrix between financial assets has been a very important issue, it has long attracted researchers attention. In the risk management, the estimates of individual assets risk and multiple asset returns covariance matrix model play an important role when investors make invest decision. Besides, we have realized that, the distribution of financial assets return has shown a statistical characteristics, called the heavy tail phenomenon, that means financial assets return is not subject to the traditional normal distribution. But recent research of the financial markets volatility and multiple assets covariance matrix was mainly under the assumption of normal distribution. To solve this problem, this article focus on a better distribution assumption which can depicts the statistical characteristics of assets return more precisely. The distribution assumption we used is called the skewed t distribution. Compared with normal distribution, skewed t distribution can describe the real data more flexible and more exactly. On the Chinese stock market , we selected three stocks and the Shanghai Stock Index. By observing the statistical characteristics of those assets return