| 3,296 | 114 | 157 |
| 下载次数 | 被引频次 | 阅读次数 |
Copula函数在金融分析和风险管理中有广泛的应用,利用Copula函数可以构建组合风险资产的联合收益分布和资产之间的相关性。在构建Copula模型时,一个关键的问题就是如何选择最佳的Copula来拟合实际的金融数据。文章分析了Copula函数选择困难的原因,指出了现有的似然准则选择方法的不足,提出了基于参数Bootstrap技术的对数似然准则检验方法,考虑了更大范围的Copula函数族群,利用模拟实验检验了该方法的选择能力,模拟结果表明对于没有尾部相关性的Copula函数和具有较小的尾部相关性的Copula函数可以较好地进行区分,而且也能区分大部分的具有较大尾部相关系数的Copula函数。同现有的只能区分常见的几类Copula的似然准则选择方法相比,文章提出的方法可以在更大范围内识别不同的Copula函数。
Abstract：The copula function is applied extensively in the financial analysis and the risk management,which is typically used to model the joint distribution of the portfolio risk assets and the correlation among the risk assets. When the copula model is constructed,it is a crucial problem how to choose the fitted-best copula to fit the actual financial data. This paper analyzes the reason why the choice of the optimal copula is so difficult and shows the deficiency of the current method based on the likelihood approach to choose the optimal copula. This paper then provides a parametric bootstrap-based log likelihood approach. With the approach,this paper allows for a wider range of the copula functions,and uses the simulation trial to test the choosing power of the approach. The result shows that the approach can distinguish the copulas with or without the tail correlations. Compared to the likelihood criterion method which can just identify a few common copulas,the approach can identify different copula functions in a wider range.
[1]Deheuvels P.The function and its dependence empirical properties:a non parametric test of independence[J].Royal Academy of Belgium.Builetin of Science Class,1979,65(6):274-292.
[2]Genest C.,Rivest L.P.Statistical inference procedures for bivariate Archimedean Copulas[J].Journal of the American Statistical Association,1993(88):1034-1043.
[3]Malavergne.Y,D.Sornette.Quantitative testing the Gaussian copula hypothesis for financial asset dependencies[J].Finance,2003(3):231-50.
[4]Genest,C.Goodness-of-fit procedures for Copula models based on the probability integral transformation[J].Scandinavian Journal of Statistics,2006(33):337-366.
[5]Jeremy Berkowitz.Testing density forecasts with applications to risk management[J].Journal of Business and Economic Statistics,2001(4):465-474.
[6]Chen X.,Fan Y.Pseudo-likelihood ratio tests for semi-parametric multivariate Copula model selection[J].The Canadian Journal of Statistics,2005(33):389-414.
[7]王沁,王璐,何平.基于截断tau的copula模型选择及应用[J].数理统计与管理,2008,27(1):118-123.
[8]王沁,王璐,袁代林.基于Spearman的rho的Copula参数模型的选择[J].数学的实践与认识,2011,41(15):145-150.
[9]李根,邹国华,张新雨.高维模型选择方法综述[J].数理统计与管理,2012,31(004):640-658.
[10]王博,孟生旺.极值Copula在商业银行操作风险度量中的应用[J].统计教育,2010(8):55-60.
[11]吴庆晓,刘海龙,龚世民.基于极值Copula的投资组合集成风险度量方法[J].统计研究,2011,28(7):84-91.
[12]汪朋.资产组合风险的极值Copula模型与实证研究[J].武汉金融,2013(7):24-26.
[13]Nelsen R.B.An introduction to Copula[M].Springer-Verlag,New York,2006:32-34.
[14]Jeffreys H.Theory of Probability[M].The Clarendon Press Oxford University Press,New York,1998:245-248.
[15]Efron B.Bootstraping Method:Another Look at the Jackknife[J].Annals of Sttistics.1979(7):1-26.
[16]Mardia K.,Kent J.,Bibby J.Multivariate Analysis[M].Academic Press,London,1979:320-322.
基本信息:
DOI：10.19343/j.cnki.11-1302/c.2014.10.015
中图分类号:F830.91
引用信息:
[1]吴建华,王新军,张颖.相关性分析中Copula函数的选择[J].统计研究,2014,31(10):99-107.DOI:10.19343/j.cnki.11-1302/c.2014.10.015.
基金信息:
教育部人文社科规划基金项目“基于非线性分析方法的金融市场波动与信用风险控制研究”(13YJAZH091);; 济南大学科研基金(青年项目)“基于Copula函数和非参数估计的尾部相关性研究”(XKY1315)的资助
2014-10-15
2014-10-15