基于非完全信息模型和Logistic模型的中国上市公司信用风险研究
|
摘要
信用风险是金融市场参与者普遍面临的风险。目前我国的信用风险分析仍然以对财务指标的分析为主,现代信用风险分析方法在中国的运用尚处于起步阶段。因此,建立衡量信用风险的模型来构造更为精确、实用的信用风险机制以提高信用风险度量和管理水平既有迫切性,也具重要的现实意义。本文分别选择了非完全信息模型以及主成分Logistic模型对我国上市公司的信用风险进行了实证分析。作为现代信用风险评估模型的一个发展方向,非完全信息模型为金融机构进行风险评估提供了一个全新的思路。但是非完全信息模型产生于发达的资本市场的背景之下,其使用更多的受到资本市场有效性的影响,在我国资本市场的现状之下,企业资本价值的信息更多的时候脱离了企业的真实价值,非完全信息模型的应用受到了一定的限制。
Every investor faced credit risk in the capital market. With the happening of the international financial crisis, organizations from the world have begun to pay more attention to the influence of the credit risk model to asset pricing and risk management. Now countries whose financial markets is more mature have established mature credit evaluation system. But in China the credit risk of our country still based on the analysis of financial indicators, the modern credit risk analysis methods used in China still in its infancy. Therefore, the establishment of credit risk measurement models which are used to construct more precise and practical mechanisms for the improving of the credit risk credit risk measurement and management is very important.
The first chapter introduces the research background, research papers of credit risk by civil and foreign scholars on the Review of the model and briefly introduces the research methods and structure of the article. With the development of financial markets, economists have established a series of models of credit risk measurement methods. More traditional methods rely on the financial indicators. Based on statistical methods and econometric basis, the modern credit risk measurement methods more depend on the complex mathematical models as tools.
The second chapter is the core of the thesis. The chapter is about the empirical study of credit risk of listed companies which is based on incomplete information model. Incomplete information model is based on incomplete information under the assumption of credit risk measurement model. As the company's value and limits of information, the uncertainty of default, credit risk measurement of incomplete information model can be divided into the following three: (1) The full value of the company information, the information about the boundary of the breach of contract is incomplete,(2) the value of the company information is not complete, default boundary information completely,(3) Both the information about the value of the company and the default boundaries are incomplete. This chapter is to study the second case. This chapter selected the companies which have happened debt default event and public disclosure as empirical research samples. And we have studied the value of the assets, asset price volatility, the stock value and the relationship between the probabilities of default. This chapter also selected eight listed companies of the same industry and the closing market value for the study.
The empirical study in chapter three is based on logistic regression based on principal component analysis. From the asset structure, profitability, solvency, liquidity and other aspects of the situation we select financial indicators, and add trade by such factors as the method of principal component analysis representation of components were calculated and analyzed by principal component regression.
Chapter four compared principal component Logistic model and the result of incomplete information model of the same listed company for the same default probability calculation and analyzes the applicability of the model.
Logistic model represents the traditional credit risk assessment model and pay more attention to the situation for their own business risk profile of credit. However, due to information availability reasons, we cannot and receive the internal information in time; the model can only judge the likelihood of default risk using historical data according to enterprise companies. Enterprise real-time changes cannot be reflected in the model. This is also a defect model application.
As a modern credit risk assessment model, incomplete information model provide a new way of risk assessment for financial institutions. However, the assumption for the value of the assets which are Brownian motion is limited. Changing of the company value is affected by the macroeconomic factors, capital market conditions and business operating conditions and other factors and has certain trend. Incomplete information model was generated in the developed capital markets, and the use of the model affected by the validity of the capital market impact. In an efficient market, the business capital value of the information is internal situation and a comprehensive reflection of external influence. However, the status of the capital market in China under the capital value of corporate information more often break away from the real value of enterprises, the application of the model of incomplete information is limited.
Every investor faced credit risk in the capital market. With the happening of the international financial crisis, organizations from the world have begun to pay more attention to the influence of the credit risk model to asset pricing and risk management. Now countries whose financial markets is more mature have established mature credit evaluation system. But in China the credit risk of our country still based on the analysis of financial indicators, the modern credit risk analysis methods used in China still in its infancy. Therefore, the establishment of credit risk measurement models which are used to construct more precise and practical mechanisms for the improving of the credit risk credit risk measurement and management is very important.
The first chapter introduces the research background, research papers of credit risk by civil and foreign scholars on the Review of the model and briefly introduces the research methods and structure of the article. With the development of financial markets, economists have established a series of models of credit risk measurement methods. More traditional methods rely on the financial indicators. Based on statistical methods and econometric basis, the modern credit risk measurement methods more depend on the complex mathematical models as tools.
The second chapter is the core of the thesis. The chapter is about the empirical study of credit risk of listed companies which is based on incomplete information model. Incomplete information model is based on incomplete information under the assumption of credit risk measurement model. As the company's value and limits of information, the uncertainty of default, credit risk measurement of incomplete information model can be divided into the following three: (1) The full value of the company information, the information about the boundary of the breach of contract is incomplete,(2) the value of the company information is not complete, default boundary information completely,(3) Both the information about the value of the company and the default boundaries are incomplete. This chapter is to study the second case. This chapter selected the companies which have happened debt default event and public disclosure as empirical research samples. And we have studied the value of the assets, asset price volatility, the stock value and the relationship between the probabilities of default. This chapter also selected eight listed companies of the same industry and the closing market value for the study.
The empirical study in chapter three is based on logistic regression based on principal component analysis. From the asset structure, profitability, solvency, liquidity and other aspects of the situation we select financial indicators, and add trade by such factors as the method of principal component analysis representation of components were calculated and analyzed by principal component regression.
Chapter four compared principal component Logistic model and the result of incomplete information model of the same listed company for the same default probability calculation and analyzes the applicability of the model.
Logistic model represents the traditional credit risk assessment model and pay more attention to the situation for their own business risk profile of credit. However, due to information availability reasons, we cannot and receive the internal information in time; the model can only judge the likelihood of default risk using historical data according to enterprise companies. Enterprise real-time changes cannot be reflected in the model. This is also a defect model application.
As a modern credit risk assessment model, incomplete information model provide a new way of risk assessment for financial institutions. However, the assumption for the value of the assets which are Brownian motion is limited. Changing of the company value is affected by the macroeconomic factors, capital market conditions and business operating conditions and other factors and has certain trend. Incomplete information model was generated in the developed capital markets, and the use of the model affected by the validity of the capital market impact. In an efficient market, the business capital value of the information is internal situation and a comprehensive reflection of external influence. However, the status of the capital market in China under the capital value of corporate information more often break away from the real value of enterprises, the application of the model of incomplete information is limited.
引文
[1] Dither Cossin,Hugues Pirotte,殷剑峰等译,高级信用风险分析[M].北京:机械工业出版社,2005年第一版.
[2]张美灵,欧志伟.信用评估理论与实务[M].上海:复旦大学出版社,2004年7月. B.学位论文
[1]程功.基于结构化模型的信用及其应用研究[D].天津:天津大学,2007.
[2]刘秀凤.企业债券违约风险研究——基于不完全信息的分析框架[D].广州:暨南大学,2008.
[3]石伟.我国上市公司信用风险度量实证研究——基于I^2模型的分析框架[D].广州:暨南大学,2006.
[4]续云丰.不完全信息条件下信用风险模拟与数值模拟[D].武汉:武汉理工大学,2006.
[1]管七海,冯宗宪.信用违约概率测度研究:文献综述与比较,世界经济,2004(11).
[2]何运强,方兆本.债券信用评级与信用风险[J].管理科学,2003(4):45-50.
[3]李秉祥.基于期望违约率模型的上市公司财务困境预警研究[J].中国管理科学.2004(5):12-16.
[4]李建华,韩岗,韩晓普.基于Credit Portfolio View的信用风险度量模型研究[J].工业技术经济,2008(3).
[5]李萌.Logit模型在商业银行信用风险评估中的应用研究[J].管理科学,2005,18(2):33-38.
[6]梁琪.企业信用风险的主成分判别模型及其实证研究[J].财经研究,2003(5):52-57.
[7]梁世栋,李勇,方兆本.信用风险模型比较分析[J].中国管理科学,2002(2):17-22.
[8]鲁炜,赵恒晰.KMV模型在公司价值评估中的应用[J].管理科学,2003,(3):30-33.
[9]马九杰,郭宇辉,朱勇.县域中小企业贷款违约分析与信用风险实证分析[J].管理世界,2004(4):58-87.
[10]沈沛龙,任若恩.现代信用风险管理模型和方法的比较研究[J].经济科学,2002(3):32-41
[11]石晓军.Logistic违约率模型的最优样本配比与分界点研究[J].财经研究,2005(9):38-48.
[12]王春峰,万海晖,张维.商业银行信用风险评估及其实证研究[J].管理科学学报,1998(1):68-72
[13]吴世农,卢贤义.我国上市公司财务困境的预测模型研究[J].经济研究,2001(6):46-55.
[14]杨星.I^2模型在中国的实证检验,统计与决策,2007 (7):17-19.
[15]杨星.信用风险管理理论的新发展——I^2模型[J].南方金融,2007(1):29-31.
[16]于立勇,詹捷辉.基于Logistic回归分析的违约概率预测研究[J].财经研究,2004(9):15-32.
[17] Artzner Philippe and Freddy Delbaen, Default risk insurance and incomplete markets [J].Mathematical Finance, 1995, 5:187-195.
[18] Black F., M. Scholes. The pricing of options and corporate liabilities[J].Journal of Political Economy,1973,81:637-354.
[19] Duffle D., D.Lando. Term structures of credit spreads with incomplete accounting information[J].Econometrica,2001,69,3 :633-664.
[20] Duffle Darrell and Kenneth Singleton. Modeling term structures of defaultable bonds, Review of Financial Studies 1999,12:687-720.
[21] Frey Rudiger and Alexander J. McNeil. Modeling dependent defaults[J]. Journal of risk, 2003, 6:59-92.
[22] Jarrow, A. Robert and Fan Yu. Counterparty risk and the pricing of defaultable secutities[J].Journal of Finance 2001, 56:555-576.
[23] Kay Giesecke and Lisa Goldberg. A simple exponential model for dependent defaults[J]. Journal of Fixed Income 2003, 13:74-83.
[24] Kay Giesecke and Lisa Goldberg. Forecasting default in the case of uncertainty[J]. Journal of Derivatives 2004, 12(1): 1-15.
[25] Kay Giesecke and Lisa Goldberg. Sequential Defaults and Incomplete Information, Forthcoming[J]. Journal of Risk, 2004,7(1):1-26.
[26] Kay Giesecke. Default and information[J]. Journal of Economic Dynamics Control, 2006,30: 2281-2303.
[27] Kusuoka S. A remark on default risk models [J]. Advances in Mathematical Economics, 1999, 1: 69-82.
[28] Merton, C. Robert. Theory of rational option pricing [J]. Bell Journal of Economics and Management Science, 1973,4:141-183.
[29] Moraux Franck. Structural credit risk models for defaultable securities, Working Paper, University de Rennes I-IGR, 2001.
[30] Ohlson J A,Financial Ratios and the Probabilistic Prediction of Bankruptcy,Journal of Accounting Research,1980,18(1):109~131.
[2]张美灵,欧志伟.信用评估理论与实务[M].上海:复旦大学出版社,2004年7月. B.学位论文
[1]程功.基于结构化模型的信用及其应用研究[D].天津:天津大学,2007.
[2]刘秀凤.企业债券违约风险研究——基于不完全信息的分析框架[D].广州:暨南大学,2008.
[3]石伟.我国上市公司信用风险度量实证研究——基于I^2模型的分析框架[D].广州:暨南大学,2006.
[4]续云丰.不完全信息条件下信用风险模拟与数值模拟[D].武汉:武汉理工大学,2006.
[1]管七海,冯宗宪.信用违约概率测度研究:文献综述与比较,世界经济,2004(11).
[2]何运强,方兆本.债券信用评级与信用风险[J].管理科学,2003(4):45-50.
[3]李秉祥.基于期望违约率模型的上市公司财务困境预警研究[J].中国管理科学.2004(5):12-16.
[4]李建华,韩岗,韩晓普.基于Credit Portfolio View的信用风险度量模型研究[J].工业技术经济,2008(3).
[5]李萌.Logit模型在商业银行信用风险评估中的应用研究[J].管理科学,2005,18(2):33-38.
[6]梁琪.企业信用风险的主成分判别模型及其实证研究[J].财经研究,2003(5):52-57.
[7]梁世栋,李勇,方兆本.信用风险模型比较分析[J].中国管理科学,2002(2):17-22.
[8]鲁炜,赵恒晰.KMV模型在公司价值评估中的应用[J].管理科学,2003,(3):30-33.
[9]马九杰,郭宇辉,朱勇.县域中小企业贷款违约分析与信用风险实证分析[J].管理世界,2004(4):58-87.
[10]沈沛龙,任若恩.现代信用风险管理模型和方法的比较研究[J].经济科学,2002(3):32-41
[11]石晓军.Logistic违约率模型的最优样本配比与分界点研究[J].财经研究,2005(9):38-48.
[12]王春峰,万海晖,张维.商业银行信用风险评估及其实证研究[J].管理科学学报,1998(1):68-72
[13]吴世农,卢贤义.我国上市公司财务困境的预测模型研究[J].经济研究,2001(6):46-55.
[14]杨星.I^2模型在中国的实证检验,统计与决策,2007 (7):17-19.
[15]杨星.信用风险管理理论的新发展——I^2模型[J].南方金融,2007(1):29-31.
[16]于立勇,詹捷辉.基于Logistic回归分析的违约概率预测研究[J].财经研究,2004(9):15-32.
[17] Artzner Philippe and Freddy Delbaen, Default risk insurance and incomplete markets [J].Mathematical Finance, 1995, 5:187-195.
[18] Black F., M. Scholes. The pricing of options and corporate liabilities[J].Journal of Political Economy,1973,81:637-354.
[19] Duffle D., D.Lando. Term structures of credit spreads with incomplete accounting information[J].Econometrica,2001,69,3 :633-664.
[20] Duffle Darrell and Kenneth Singleton. Modeling term structures of defaultable bonds, Review of Financial Studies 1999,12:687-720.
[21] Frey Rudiger and Alexander J. McNeil. Modeling dependent defaults[J]. Journal of risk, 2003, 6:59-92.
[22] Jarrow, A. Robert and Fan Yu. Counterparty risk and the pricing of defaultable secutities[J].Journal of Finance 2001, 56:555-576.
[23] Kay Giesecke and Lisa Goldberg. A simple exponential model for dependent defaults[J]. Journal of Fixed Income 2003, 13:74-83.
[24] Kay Giesecke and Lisa Goldberg. Forecasting default in the case of uncertainty[J]. Journal of Derivatives 2004, 12(1): 1-15.
[25] Kay Giesecke and Lisa Goldberg. Sequential Defaults and Incomplete Information, Forthcoming[J]. Journal of Risk, 2004,7(1):1-26.
[26] Kay Giesecke. Default and information[J]. Journal of Economic Dynamics Control, 2006,30: 2281-2303.
[27] Kusuoka S. A remark on default risk models [J]. Advances in Mathematical Economics, 1999, 1: 69-82.
[28] Merton, C. Robert. Theory of rational option pricing [J]. Bell Journal of Economics and Management Science, 1973,4:141-183.
[29] Moraux Franck. Structural credit risk models for defaultable securities, Working Paper, University de Rennes I-IGR, 2001.
[30] Ohlson J A,Financial Ratios and the Probabilistic Prediction of Bankruptcy,Journal of Accounting Research,1980,18(1):109~131.