| 徐刚刚,狄千姿,石慧.基于动态R-Vine Copula的银行股指投资组合及风险度量研究[J].井冈山大学自然版,2020,41(5):10-17 |
| 基于动态R-Vine Copula的银行股指投资组合及风险度量研究 |
| RESEARCH ON BANK STOCK INDEX PORTFOLIO AND RISK MEASUREMENT BASED ON DYNAMIC R-VINE COPULA |
| 投稿时间：2020-05-06 修订日期：2020-06-17 |
| DOI：10.3969/j.issn.1674-8085.2020.05.003 |
| 中文关键词: 动态R-Vine Copula模型 Monte Carlo模拟 CVaR MST-PRIM算法 |
| 英文关键词: dynamic R-Vine Copula model Monte Carlo simulation CVaR MST-PRIM algorithm |
| 基金项目:新疆维吾尔自治区高校科研计划项目（XJEDU2018Y021）;新疆农业大学大学生创新创业计划训练项目（dxscx2020518） |
|
| 摘要点击次数: 1344 |
| 全文下载次数: 2143 |
| 中文摘要: |
| 金融资产价格之间的波动往往具有结构相依性。为了研究银行股指数据间这种复杂关系，能够进一步准确度量金融风险，本文以我国六大银行股指数据为研究对象，利用ARMA-GJR-SkT模型作为单一资产序列的边缘分布，以灵活的R-Vine Copula模型为基础，联合构建投资组合模型。通过滚动时间窗口的Monte Carlo技术及MST-PRIM算法确定各类模型的RVM结构，在此基础上结合逆变换法仿真模拟收益率序列，并利用模拟收益率进一步计算VaR与CVaR，最后经返回值检验法对模型进行验证。结果表明：建立最优的投资组合模型是精准度量金融风险的关键；在风险度量时，相同的置信水平下，CVaR模型比VaR模型更可靠。置信水平不等时，其值增大的同时，失败天数会减小。 |
| 英文摘要: |
| The fluctuation of financial asset prices often has structural dependence. In order to study the complex relationship between banking stock index data and further accurately measure the financial risk, this paper takes the six major stock index data of China's banking as the research object, Using arma-gjr-skt as the marginal distribution of a single asset sequence, Based on the flexible r-vine copula model, jointly build the portfolio model. The RVM structure of various models is determined by the Monte Carlo technique of rolling time window and mst-prim algorithm. On this basis, the return sequence is simulated by inverse transformation method, and the VaR and CVaR are further calculated by the simulated return rate. Finally, the model is verified by back-testing method. The results show that the establishment of the optimal portfolio model is the key to accurately measure the financial risk. Given the same level of confidence in risk measurement, CVaR model is more reliable than VaR model.When the confidence level is different, the number of failure days will decrease as the value increases. |
|
查看全文
查看/发表评论 下载PDF阅读器 |
| 关闭 |
