基于半参数Logistic回归模型的我国财产保险公司偿付能力研究
|
摘要
保险业在现代金融体系中占据着重要地位。在现代社会中,保险已经成为保障经济正常运行和社会稳定发展的不可缺少的组成部分。如果保险公司偿付能力不足,不仅会影响投保人的利益、保险公司的稳定性、甚至影响到中国保险业的健康发展。保险公司偿付能力是指保险公司对所承担的风险责任在发生赔偿和给付时所具有的经济补偿能力。保险偿付能力监管已经成为各个国家监管机构对保险业进行监管的核心部分,研究偿付能力有着十分重要的意义。
本文主要研究的是影响财产保险公司的因素。本文首先对国内外偿付能力的研究现状进行了阐述,介绍了研究偿付能力常用的方法和当前国内保险监管和偿付能力的一些情况；然后对偿付能力监管的涵义和意义作了细致说明,列举了世界上主要国家的三种偿付能力监管模式并且评价了这三种模式对我国偿付能力监管的启示；随后着重分析影响财产保险公司偿付能力的因素,由于2008年发布的《保险公司偿付能力管理规定》没有涉及监管指标,只是进行一些额度监管,本文主要是针对2003年中国保险监督委员会《关于保险公司偿付能力额度及监管指标管理规定》的11项与偿付能力关系密切的因素作为主要指标进行分析。最后,建立Logistic回归模型和半参数的Logistic回归模型。
本文数据来自《中国保险年鉴》(2005-2008)。本文利用Logistic回归模型分析各个指标的显著程度以及对偿付能力的影响大小,得到包含六个指标的Logistic回归模型。为了进一步分析,建立半参数Logistic回归模型,我们将影响显著的六个变量作为半参数Logistic回归模型中的线性部分进行分析,影响不显著的五个变量列为模型中的非线性部分,最终得到半参数Logistic回归模型。
通过两种方法对比,两个模型的共同点是得到的指标与偿付能力的相关性一致,但是,由于半参数Logistic回归模型的系数与Logistic回归模型不一致,通过分析,半参数Logistic回归模型研究财险公司偿付能力问题比普通的Logistic回归模型更加优越。但是,这只是对于偿付能力方法研究的一次尝试,还有许多问题需要进一步研究和探讨。
The insurance industry plays an important role in the modern financial system, in modern society, insurance has become an indispensable part for protecting the normal economic operation and social stability. Lack of solvency will not only affect the interest of the insured and affect the operational stability of insurance company, but also affect the healthy development of China's insurance industry. Insurance Solvency refers to the risks borne by insurers in the event of responsibility for payment of compensation and when it has the financial ability to compensate. Insurance solvency regulation has become the core of every national regulatory agency to monitor the insurance industry. It is of great significance.
This Dissertation mainly researches the influence factors of Property Insurance Company Solvency. First it elaborated the solvency status at home and abroad, introduced the commonly ways used to study solvency and the current situation of domestic insurance regulation; Then it explained the meaning and significance of solvency supervision, listed the three important kinds of solvency regulation mode in the world and Assessed their Enlightenment to our country. Subsequently our analysis focused on the factors which affect the insurance company solvency, Since the "insurance solvency regulations" in 2008 does not involve monitoring indicators, but some amount of supervision.We selected 11 important factors closely related to solvency as the main indicators reference to Insurance Regulatory Commission on the insurance solvency margin and regulatory indicators of the provisions of regulations in China 2003. Finally we give a Logistic regression model and semi-parameter Logistic regression model.
Data comes from the "China Insurance Yearbook" (2005-2008),We use the Logistic regression model to analyze the significant degree of individual indicators, as well as the impact on the solvency. Then we get the Logistic regression model of six indicators.In order to further analyze, we give a semi-parameter Logistic regression models. The six variables which significantly affect solvency are as the linear part of the influence of factors of the semi-parametric logistic regression model. Factors which are not significant are as the non-linear part of the semi-parametric model. Finally, we get the semi-parametric logistic regression model.
Comparison of the results of two methods, Both models have in common is the relevance of the solvency and indicators, But the coefficients of two methods is inconsisten. Semi-parametric Logistic regression model is advantageous than normal Logistic regression model in analying solvency. However, this method is only an attempt for analying the solvency, There are many issues that need further study and discussion.
本文主要研究的是影响财产保险公司的因素。本文首先对国内外偿付能力的研究现状进行了阐述,介绍了研究偿付能力常用的方法和当前国内保险监管和偿付能力的一些情况；然后对偿付能力监管的涵义和意义作了细致说明,列举了世界上主要国家的三种偿付能力监管模式并且评价了这三种模式对我国偿付能力监管的启示；随后着重分析影响财产保险公司偿付能力的因素,由于2008年发布的《保险公司偿付能力管理规定》没有涉及监管指标,只是进行一些额度监管,本文主要是针对2003年中国保险监督委员会《关于保险公司偿付能力额度及监管指标管理规定》的11项与偿付能力关系密切的因素作为主要指标进行分析。最后,建立Logistic回归模型和半参数的Logistic回归模型。
本文数据来自《中国保险年鉴》(2005-2008)。本文利用Logistic回归模型分析各个指标的显著程度以及对偿付能力的影响大小,得到包含六个指标的Logistic回归模型。为了进一步分析,建立半参数Logistic回归模型,我们将影响显著的六个变量作为半参数Logistic回归模型中的线性部分进行分析,影响不显著的五个变量列为模型中的非线性部分,最终得到半参数Logistic回归模型。
通过两种方法对比,两个模型的共同点是得到的指标与偿付能力的相关性一致,但是,由于半参数Logistic回归模型的系数与Logistic回归模型不一致,通过分析,半参数Logistic回归模型研究财险公司偿付能力问题比普通的Logistic回归模型更加优越。但是,这只是对于偿付能力方法研究的一次尝试,还有许多问题需要进一步研究和探讨。
The insurance industry plays an important role in the modern financial system, in modern society, insurance has become an indispensable part for protecting the normal economic operation and social stability. Lack of solvency will not only affect the interest of the insured and affect the operational stability of insurance company, but also affect the healthy development of China's insurance industry. Insurance Solvency refers to the risks borne by insurers in the event of responsibility for payment of compensation and when it has the financial ability to compensate. Insurance solvency regulation has become the core of every national regulatory agency to monitor the insurance industry. It is of great significance.
This Dissertation mainly researches the influence factors of Property Insurance Company Solvency. First it elaborated the solvency status at home and abroad, introduced the commonly ways used to study solvency and the current situation of domestic insurance regulation; Then it explained the meaning and significance of solvency supervision, listed the three important kinds of solvency regulation mode in the world and Assessed their Enlightenment to our country. Subsequently our analysis focused on the factors which affect the insurance company solvency, Since the "insurance solvency regulations" in 2008 does not involve monitoring indicators, but some amount of supervision.We selected 11 important factors closely related to solvency as the main indicators reference to Insurance Regulatory Commission on the insurance solvency margin and regulatory indicators of the provisions of regulations in China 2003. Finally we give a Logistic regression model and semi-parameter Logistic regression model.
Data comes from the "China Insurance Yearbook" (2005-2008),We use the Logistic regression model to analyze the significant degree of individual indicators, as well as the impact on the solvency. Then we get the Logistic regression model of six indicators.In order to further analyze, we give a semi-parameter Logistic regression models. The six variables which significantly affect solvency are as the linear part of the influence of factors of the semi-parametric logistic regression model. Factors which are not significant are as the non-linear part of the semi-parametric model. Finally, we get the semi-parametric logistic regression model.
Comparison of the results of two methods, Both models have in common is the relevance of the solvency and indicators, But the coefficients of two methods is inconsisten. Semi-parametric Logistic regression model is advantageous than normal Logistic regression model in analying solvency. However, this method is only an attempt for analying the solvency, There are many issues that need further study and discussion.
引文
[1]高虹.论保险公司偿付能力监管[J].市场周刊.管理探索.2005(7)：121—122。
[2]刘云海,张琳.论保险业监管模式的现实选择[J].保险研究.2005(10)：67—69。
[3]粟芳.中国非寿险保险公司的偿付能力研究[M].上海.复旦大学出版社,2002：106-107。
[4]谢志刚、韩天雄.《风险理论与非寿险精算》[M].天津：南开大学出版社.2000.
[5]俞自白.《建立中国保险公司偿付能力监管指标体系的建议》.世界华人论坛专题.2002.9.3-9.4。
[6]周力生.欧盟偿付能力监管体系改革及意义[J].保险研究,2005(2),90—92。
[7]李建标,杜传忠.试析保险企业的偿付能力及监管对策,经济学研究,2001(3)：10-14。
[8]刘茂山.保险经济学,天津：南开大学出版社,2000。
[9]伍超标.保险精算学基础,北京：中国统计出版社,1999。
[10]毛宏.保险企业偿付能力的监控,运筹与管理,2001(2)：144-148。
[11]高虹.论保险公司偿付能力监管,市场周刊,管理探索,200,5年第7期。
[12]刘畅.因子分析法在我国寿险公司偿付能力监测中的应用,统计与决策,2005年第5期。
[13]黄琦.西方偿付能力监管对我国保险监管的启示,科技创业月刊,2005年第2期。
[14]李欣霞.浅析我国非寿险保险公司偿付能力现状。特区经济,2004年第12期。
[15]孙祁祥.保险学.北京：北京大学出版社,1999：168-197。
[16]闫春,赵明清,张彦梅.非寿险保险公司偿付能力影响因素的灰色关联分析[J].山东科技大学学报(自然科学版),2003第4期：115-118。
[17]占梦雅.中国非寿险业法定偿付能力额度标准的合理性分析与实证分析[J].财经理论与实践,2006,第2期：25-28。
[18]张伟,邱长溶.财产保险公司偿付能力实证分析[J].江西则经大学学报,2004 第4期：16-18。
[19]周晶晗.从欧美财险公司破产看我国非寿险偿付能力监管[J].金融教学与研究,2007第2期：46-48。
[20]刘栋.保险企业偿付能力的理论与使用模型.保险研究,1991第2期。
[21]柴根象,洪圣岩(1995).半参数回归模型,合肥：安徽教育出版社。
[22]高集体.洪圣岩,梁华.半参数回归模型研究的若干进展[J].应用概率统计,1994(1)：98-104。
[23]胡舒合.一类半参数回归模型的估计问题[J].数学物理学报,1999,S1：541-549.学,2000(1)：40-48。
[24]唐亚宁,赵选民.半参数回归模型的误差分布的估计的大样本性质[J].纯粹数学与应用数学,2000(1)：40-48。
[25]丁士俊,陶本藻.半参数模型核光滑估计与模拟分析[J],大地测量与地球动力学,2004,Vol.04：26-30。
[26]但尧,丁鹭飞.变换核估计和迭代算法.应用概率统计,1994,10(2)：113-118。
[27]孙海燕,吴云.半参数回归与模型精化,武汉大学学报信息科学版,加02,27(2,)： 172-174。
[28]潘雄,孙海燕.半参数P范极大似然回归,测绘学报,2005,34(1)：30-35。
[29]钱伟民,柴根象,半参数回归模型的估计的渐近性质,高校应用数学学报A(辑),1999,14(2)：161-165。
[30]黄四民,梁华.用半参数部分线性模型分析居民消费结构[J].数量经济技术经济研究,1994,10：33-39。
[31]梁华,熊健.再论半参数部分线性模型在居民消费结构分析中的应用[J].数理统计与管理,1995(6)：9-17。
[32]G. W. De Wit, W. M. Kestelijn. The Solvency Margin in Non-life Insurance Companies, Astin Bulletin,1980,1(11):136-144.
[33]ENGKLE R F,RICE J.Semiparametric estimates of the relation between weather and electricity sales[J].JASA,1986,81:310-320.
[34]FRAIMAN R,INBARREN P.Nonparametric regression estimation in models with weak error's structure[J].J Mult Anal,1991,37(2):180-196.
[35]Chen H. Shiau J.H.Data-driven efficient estimator for a partially linear model[J],Ann.Statist.22,211-237.
[36]HongChang Hu.Ridge estimation of a semiparametric regression model[J],Journal ofcomputational and applied mathematics,176(2005):215-222.
[37]Shibata R.An optimal selection of regression variables[J],Biometrika 68:45-54.
[38]Sheng-Yan Hong.Automatic Bandwidth Choice In A Semiparametric Regression Model[J],Statistica Sinica 9(1999):775-794.
[39]P.J.Green,B.W.Silverman.Nonparametric Regression and Generalized Liniear Models[M],Champman&Hall,London,1994.
[40]Speckman,P.E(1988),"Regression Analysis for Partially Linear Models".Journal of the Royal Statistical Society, Ser.B,50:413-436.
[41]Harris.I.R(1993),"Quasi-Likelihood for Mixed Model",Technical Report93-01, University of Texas.Center for Statistical Science.
[42]McCullagh.P(1983),Quasi-Likelihood Functions.The Annals of Statistics,11: 59-67.
[43]Green,P.J.(1987),"Penalized Likelihood For General Semi-Parametric Regression Models",International Statistical Review,55:245-259.
[44]Thomas A.S. and Joan G.S.(1994),Quasi-likelihood Estimation in Semiparametric Model Journal of the American Statistical Association, Vol.99,No.426:501-511
[2]刘云海,张琳.论保险业监管模式的现实选择[J].保险研究.2005(10)：67—69。
[3]粟芳.中国非寿险保险公司的偿付能力研究[M].上海.复旦大学出版社,2002：106-107。
[4]谢志刚、韩天雄.《风险理论与非寿险精算》[M].天津：南开大学出版社.2000.
[5]俞自白.《建立中国保险公司偿付能力监管指标体系的建议》.世界华人论坛专题.2002.9.3-9.4。
[6]周力生.欧盟偿付能力监管体系改革及意义[J].保险研究,2005(2),90—92。
[7]李建标,杜传忠.试析保险企业的偿付能力及监管对策,经济学研究,2001(3)：10-14。
[8]刘茂山.保险经济学,天津：南开大学出版社,2000。
[9]伍超标.保险精算学基础,北京：中国统计出版社,1999。
[10]毛宏.保险企业偿付能力的监控,运筹与管理,2001(2)：144-148。
[11]高虹.论保险公司偿付能力监管,市场周刊,管理探索,200,5年第7期。
[12]刘畅.因子分析法在我国寿险公司偿付能力监测中的应用,统计与决策,2005年第5期。
[13]黄琦.西方偿付能力监管对我国保险监管的启示,科技创业月刊,2005年第2期。
[14]李欣霞.浅析我国非寿险保险公司偿付能力现状。特区经济,2004年第12期。
[15]孙祁祥.保险学.北京：北京大学出版社,1999：168-197。
[16]闫春,赵明清,张彦梅.非寿险保险公司偿付能力影响因素的灰色关联分析[J].山东科技大学学报(自然科学版),2003第4期：115-118。
[17]占梦雅.中国非寿险业法定偿付能力额度标准的合理性分析与实证分析[J].财经理论与实践,2006,第2期：25-28。
[18]张伟,邱长溶.财产保险公司偿付能力实证分析[J].江西则经大学学报,2004 第4期：16-18。
[19]周晶晗.从欧美财险公司破产看我国非寿险偿付能力监管[J].金融教学与研究,2007第2期：46-48。
[20]刘栋.保险企业偿付能力的理论与使用模型.保险研究,1991第2期。
[21]柴根象,洪圣岩(1995).半参数回归模型,合肥：安徽教育出版社。
[22]高集体.洪圣岩,梁华.半参数回归模型研究的若干进展[J].应用概率统计,1994(1)：98-104。
[23]胡舒合.一类半参数回归模型的估计问题[J].数学物理学报,1999,S1：541-549.学,2000(1)：40-48。
[24]唐亚宁,赵选民.半参数回归模型的误差分布的估计的大样本性质[J].纯粹数学与应用数学,2000(1)：40-48。
[25]丁士俊,陶本藻.半参数模型核光滑估计与模拟分析[J],大地测量与地球动力学,2004,Vol.04：26-30。
[26]但尧,丁鹭飞.变换核估计和迭代算法.应用概率统计,1994,10(2)：113-118。
[27]孙海燕,吴云.半参数回归与模型精化,武汉大学学报信息科学版,加02,27(2,)： 172-174。
[28]潘雄,孙海燕.半参数P范极大似然回归,测绘学报,2005,34(1)：30-35。
[29]钱伟民,柴根象,半参数回归模型的估计的渐近性质,高校应用数学学报A(辑),1999,14(2)：161-165。
[30]黄四民,梁华.用半参数部分线性模型分析居民消费结构[J].数量经济技术经济研究,1994,10：33-39。
[31]梁华,熊健.再论半参数部分线性模型在居民消费结构分析中的应用[J].数理统计与管理,1995(6)：9-17。
[32]G. W. De Wit, W. M. Kestelijn. The Solvency Margin in Non-life Insurance Companies, Astin Bulletin,1980,1(11):136-144.
[33]ENGKLE R F,RICE J.Semiparametric estimates of the relation between weather and electricity sales[J].JASA,1986,81:310-320.
[34]FRAIMAN R,INBARREN P.Nonparametric regression estimation in models with weak error's structure[J].J Mult Anal,1991,37(2):180-196.
[35]Chen H. Shiau J.H.Data-driven efficient estimator for a partially linear model[J],Ann.Statist.22,211-237.
[36]HongChang Hu.Ridge estimation of a semiparametric regression model[J],Journal ofcomputational and applied mathematics,176(2005):215-222.
[37]Shibata R.An optimal selection of regression variables[J],Biometrika 68:45-54.
[38]Sheng-Yan Hong.Automatic Bandwidth Choice In A Semiparametric Regression Model[J],Statistica Sinica 9(1999):775-794.
[39]P.J.Green,B.W.Silverman.Nonparametric Regression and Generalized Liniear Models[M],Champman&Hall,London,1994.
[40]Speckman,P.E(1988),"Regression Analysis for Partially Linear Models".Journal of the Royal Statistical Society, Ser.B,50:413-436.
[41]Harris.I.R(1993),"Quasi-Likelihood for Mixed Model",Technical Report93-01, University of Texas.Center for Statistical Science.
[42]McCullagh.P(1983),Quasi-Likelihood Functions.The Annals of Statistics,11: 59-67.
[43]Green,P.J.(1987),"Penalized Likelihood For General Semi-Parametric Regression Models",International Statistical Review,55:245-259.
[44]Thomas A.S. and Joan G.S.(1994),Quasi-likelihood Estimation in Semiparametric Model Journal of the American Statistical Association, Vol.99,No.426:501-511