基于Logistic模型的上市公司财务危机预警研究
|
摘要
这些年我国证券市场发展比较快速,对国民经济的增长起着重要的作用,但在资本市场规模不断扩大、上市公司数量不断增加的同时,我国上市公司中因发生财务危机而被特别处理的公司也在不断增加。因此,不管是企业自身还是债权人都渐渐意识到有必要建立一套财务危机预警系统来规避风险。
上市公司财务危机的发生是有过程的,并不会一下子爆发,这就使得我们完全可以建立一套预警系统来预测上市公司是否在未来会发生财务危机。本文以上市公司是否因特别处理而戴上ST帽子作为其发生财务危机的标志,利用样本公司的指标数据建立逻辑回归模型来找出具有预警作用的指标变量。
行业不同,预警模型所要求的指标变量也应该有所不同,因为不同的行业如果采用相同的模型来研究的话,模型的预测准确度会降低。理论上讲,每个行业应该都建立一个适合于自己的预警模型,但这样有的行业的样本量就会由于过少的原因使模型的预测准确度也降低。因此,本文在样本期2005-2007年间,选取了制造业小类别下因“财务状况异常”而首次被ST的45家上市公司及45家相应的非ST公司进行实证研究及检验。检验结果表明加入非财务指标的模型与只包含财务指标的模型相比,模型预测的正确率有较大的提高,如果仅仅使用财务指标,其准确率仅为73.35%,但是当引入非财务指标后,其准确率上升到83.3%。这两个模型的预测精确度还都比较高,说明两个模型在实际工作中能有效地来预测财务危机的发生。
With the quick development of stock market in our country,the number of listed company has increased,market scale also has been found enlarged.This has become major force which promotes growth in national economy and has monumental contributions to national economy.In spite of this.it is obvious that there remains a trend in the polarization of listed company's performance and it shows a trend of further expansion.More and more companies have got into financial trouble,as a result,more and more investors and managers are aware that it is emergent to build a quick financial trouble forecast system to avoid investment risk.Listed company's financial trouble forecast is a key concern to the investors,managers and supervision institutions.
The financial trouble of listed company doesn't occur suddenly,but a gradual process.So it is possible for us to take effective measures to avoid that.This test is based on the overall study of present situation of domestic and foreign listed company' finantional trouble and existing achievements in this field.It marked by the symbol "out going warning" which means whether it or not has got into financial trouble.It uses financial trouble company's variables and non-financial one's variables and sets logistic regression analysis.Through such way,we will seek for accurate variable which can predict listed company's financial trouble and thus set financia trouble forecast model to adapt to the situation of our country.
Due to the nature of different industries, financial models, including the financial ratios and its parameters will also be different, if the neglect of the financial characteristics of different industries differences, the accuracy of the model will be affected, that is, different models should be used to study the different industries. Therefore, this article tests 45 financial trouble company and 45 non-financial ones under the category of manufacturing (Succompanies has been dealt with specially first time from 2005 to 2007). Test results show that compared with the model which contains only financial indicators.the model which contains non-financial indicators can do better in demonstration.and can be practical in practical application. the model can forecast the possible financial trouble in following three years. the accuracy rate is only 73.35 percent with the use of financial indicators only, but when the introduction of non-financial indicators, their accuracy rate rose to 83.3%. The test accuracy rate is more higher and the use of financial indicators has some practical value the use of financial indicators.
上市公司财务危机的发生是有过程的,并不会一下子爆发,这就使得我们完全可以建立一套预警系统来预测上市公司是否在未来会发生财务危机。本文以上市公司是否因特别处理而戴上ST帽子作为其发生财务危机的标志,利用样本公司的指标数据建立逻辑回归模型来找出具有预警作用的指标变量。
行业不同,预警模型所要求的指标变量也应该有所不同,因为不同的行业如果采用相同的模型来研究的话,模型的预测准确度会降低。理论上讲,每个行业应该都建立一个适合于自己的预警模型,但这样有的行业的样本量就会由于过少的原因使模型的预测准确度也降低。因此,本文在样本期2005-2007年间,选取了制造业小类别下因“财务状况异常”而首次被ST的45家上市公司及45家相应的非ST公司进行实证研究及检验。检验结果表明加入非财务指标的模型与只包含财务指标的模型相比,模型预测的正确率有较大的提高,如果仅仅使用财务指标,其准确率仅为73.35%,但是当引入非财务指标后,其准确率上升到83.3%。这两个模型的预测精确度还都比较高,说明两个模型在实际工作中能有效地来预测财务危机的发生。
With the quick development of stock market in our country,the number of listed company has increased,market scale also has been found enlarged.This has become major force which promotes growth in national economy and has monumental contributions to national economy.In spite of this.it is obvious that there remains a trend in the polarization of listed company's performance and it shows a trend of further expansion.More and more companies have got into financial trouble,as a result,more and more investors and managers are aware that it is emergent to build a quick financial trouble forecast system to avoid investment risk.Listed company's financial trouble forecast is a key concern to the investors,managers and supervision institutions.
The financial trouble of listed company doesn't occur suddenly,but a gradual process.So it is possible for us to take effective measures to avoid that.This test is based on the overall study of present situation of domestic and foreign listed company' finantional trouble and existing achievements in this field.It marked by the symbol "out going warning" which means whether it or not has got into financial trouble.It uses financial trouble company's variables and non-financial one's variables and sets logistic regression analysis.Through such way,we will seek for accurate variable which can predict listed company's financial trouble and thus set financia trouble forecast model to adapt to the situation of our country.
Due to the nature of different industries, financial models, including the financial ratios and its parameters will also be different, if the neglect of the financial characteristics of different industries differences, the accuracy of the model will be affected, that is, different models should be used to study the different industries. Therefore, this article tests 45 financial trouble company and 45 non-financial ones under the category of manufacturing (Succompanies has been dealt with specially first time from 2005 to 2007). Test results show that compared with the model which contains only financial indicators.the model which contains non-financial indicators can do better in demonstration.and can be practical in practical application. the model can forecast the possible financial trouble in following three years. the accuracy rate is only 73.35 percent with the use of financial indicators only, but when the introduction of non-financial indicators, their accuracy rate rose to 83.3%. The test accuracy rate is more higher and the use of financial indicators has some practical value the use of financial indicators.
引文
[1]胡援成,田满文.上市公司财务困境预测模型的再比较[J].经济学(季刊),2005(10)
[2]高培业,张道奎.企业失败判别模型实证研究[J].统计研究,2000(10)
[3]陈静.上市公司财务恶化预测的实证分析[J].会计研究,1999(4)
[4]赵国忠.ST公司财务特征分析[J].金融与财会,2008(3)
[5]桂梦吟,伍硕频.非财务视角下财务困境预警模型研究[J].金融经济,2007(22)
[6]吴世农,卢贤义.我国上市公司财务困境预测[J].经济研究,2001(6)
[7]张玲.财务危机预警分析判别模型及其应用[J].预测,2000(6)
[8]刘洪,何光军.基于人工神经网络方法的上市公司经营失败预警研究[J].会计研究,2004(2)
[9]陈井勇,王克敏.股权结构、投资者保护与公司绩效[J].管理世界,2004(7)
[10]陈晓,江东.股权多元化、公司业绩与行业竞争性[J].经济研究,2000(8)
[11]郝其友,傅冠男.财务预警模型实证研究综述[J].山东经济,2006(3)
[12]吴春雷,马林梅.上市公司财务预警模型构建与变量选取方法[J].中国高新技术企业,2007(6)
[13]邓晓岚,王宗军.非财务视角下的财务困境预警[J].管理科学,2006(3)
[14]姜秀华,任强,孙铮.治理结构与公司业绩的相关性研究[J].财经研究,2001(27)
[15]陈晓,陈治鸿.中国上市公司的财务困境预测[J].中国会计与财务研究,2000(9)
[16]白重恩等.中国上市公司治理结构的实证研究[J].经济研究,2005(2)
[17]熊莉.基于公司治理结构的财务危机预警模型研究[D].硕士学位论文.同济大学,2007(2)
[18]吕长江,王力斌.上市公司董事会设置与公司业绩的实证研究一以东北地区为例[J].财经问题研究,2004(10)
[19]王满玲,杨德礼.国外公司财务危机预测研究进展评述[J].预测,2004(6)
[20]姜秀华.治理内核与综合业绩的相关性研究[D].博士学位论文.上海财经大学,2001
[21]端木正.基于人工神经网络的财务预警模型[J].统计与决策,2004(10)
[22]张鸣,程涛.上市公司财务预警实证研究的动态视角[J].财经研究,2005(1)
[23]李增泉.激励机制与企业绩效——一项基于上市公司的实证研究[J].会计研究,2000(1)
[24]王济川,郭志刚.Logistic回归模型一方法与应用[M].高等教育出版社,2001年第1版
[25]余建英,何旭宏.数理统计分析与SPSS应用[M].人民邮电出版社,2003
[26]马庆国.管理统计一数据获取、统计原理、SPSS工具应用研究[M].科学出版社,2002
[27]吴喜之.统计学：从数据到结论[M].中国统计出版社,2004
[28]张文彤,闫洁.SPSS统计分析基础教程[M].高等教育出版社,2004
[29]王子林,陆庆平.《中国上市公司业绩评价报告》[M].北京：经济科学出版社,2002
[30]黄岩,李元旭.《上市公司财务失败预测实证研究》[M].系统工程理论方法应用,2001
[31]周首华.《企业财务预警方法与机制》[M].现代财务理论前沿专题,2001(4)
[32]Aharony J., Jones C. P. and Swary I., "An Analysis of Risk and Return Characteristics of Corporate BankruptcyUsing Capital MarkData", Journal ofFinance,1980(9):1001-1016.
[33]Altman, E., "Financial Ratios, Discriminant Analysis and the Prediction of Corporate Bankruptcy",Journal ofFinance,1968(9):589-609.
[34]Aziz, A., D. Emanuel and G. Lawson, "Bank Prediction:An Investigation of Cash Flow Based Models",Journal ofManagemeStudies,1988(25):419-437.
[35]Barth, M.E., Beaver, W.H., andLandsman, W.R.,"Relative Valuation Roles of Book Equity Value and Net Income as a Functioof Financial Health",Journal ofAccounting and Economics,1998(25):1-33.
[36]Beaver, W.,"Financial Ratios as Predictors of Failure",Supplement to Journal ofAccounting Research,,1966:71-111.
[37]Best, R.W.,"The Role of Default Risk in Determining the Market Reaction to Debt Announcement", The Financial Review,1997 (2):87-105.
[38]Carmichael,D.R, The Auditor's Reporting Obligation.Auditing Research Monograph No.1 (NewYork:AICPA),1972:94.
[39]Charles E. Mossman, Geoffrey G. Bell, L. Mick Swartz, Harry Turtle,1998,"An Empirical Comparison of Bankruptcy Models",The Fnancial Review,1998(33):35-54.
[40]Deakin, E. B,"ADiscriminantAnalysisofPrediction of BusinessFailure",Journal of Accounting Research.,1972 (Spring):67-169
[41]Gregor, A., and Kaplan, S.N.,How Costly Is Financial Distress?Journal of Finance,1998 (10):1442-1493.
[42]Hamer, M.,"Failure Prediction:Sensitivity of Classification Accuracy toAlternative Statistical Methods and Variable Sets", Journal Accounting and Public Policy.1983(2):289-307.
[43]Meyer, P. A. and H. W. Pifer,"Prediction of Bank Failures",Journal ofFinance,1970 (9)
[44]Secrist, H.,"National Bank Failure andNon-failure. The Principia Press, Bloomington,1938
[45]Stephen A. Ross, Randolph W. Westerfield, Jeffrey F. Jaff,1999,吴世农、沈艺峰等译：《公司理财》,机械工业出版社,2000年。
[46]Agrawal,Anup and Knoeber,Charles R Firm performance and mechanisms to control agency problems between managers and shareholders,Journal of finance and quantitative analysis, 1996(31):377-97,
[47]Donaldson,Lex;Davis James H.Stewardship theory or agency theory:CEO governance and shareholder returns.Australian Journal of Management.l991(1):49-66.
[48]FitzPartriek, P.J.A comparison of Ratios of Sueeessful Industrial Enierprise with those of Failed Firms.Certified Public Accountant,1932(2):589—b
[2]高培业,张道奎.企业失败判别模型实证研究[J].统计研究,2000(10)
[3]陈静.上市公司财务恶化预测的实证分析[J].会计研究,1999(4)
[4]赵国忠.ST公司财务特征分析[J].金融与财会,2008(3)
[5]桂梦吟,伍硕频.非财务视角下财务困境预警模型研究[J].金融经济,2007(22)
[6]吴世农,卢贤义.我国上市公司财务困境预测[J].经济研究,2001(6)
[7]张玲.财务危机预警分析判别模型及其应用[J].预测,2000(6)
[8]刘洪,何光军.基于人工神经网络方法的上市公司经营失败预警研究[J].会计研究,2004(2)
[9]陈井勇,王克敏.股权结构、投资者保护与公司绩效[J].管理世界,2004(7)
[10]陈晓,江东.股权多元化、公司业绩与行业竞争性[J].经济研究,2000(8)
[11]郝其友,傅冠男.财务预警模型实证研究综述[J].山东经济,2006(3)
[12]吴春雷,马林梅.上市公司财务预警模型构建与变量选取方法[J].中国高新技术企业,2007(6)
[13]邓晓岚,王宗军.非财务视角下的财务困境预警[J].管理科学,2006(3)
[14]姜秀华,任强,孙铮.治理结构与公司业绩的相关性研究[J].财经研究,2001(27)
[15]陈晓,陈治鸿.中国上市公司的财务困境预测[J].中国会计与财务研究,2000(9)
[16]白重恩等.中国上市公司治理结构的实证研究[J].经济研究,2005(2)
[17]熊莉.基于公司治理结构的财务危机预警模型研究[D].硕士学位论文.同济大学,2007(2)
[18]吕长江,王力斌.上市公司董事会设置与公司业绩的实证研究一以东北地区为例[J].财经问题研究,2004(10)
[19]王满玲,杨德礼.国外公司财务危机预测研究进展评述[J].预测,2004(6)
[20]姜秀华.治理内核与综合业绩的相关性研究[D].博士学位论文.上海财经大学,2001
[21]端木正.基于人工神经网络的财务预警模型[J].统计与决策,2004(10)
[22]张鸣,程涛.上市公司财务预警实证研究的动态视角[J].财经研究,2005(1)
[23]李增泉.激励机制与企业绩效——一项基于上市公司的实证研究[J].会计研究,2000(1)
[24]王济川,郭志刚.Logistic回归模型一方法与应用[M].高等教育出版社,2001年第1版
[25]余建英,何旭宏.数理统计分析与SPSS应用[M].人民邮电出版社,2003
[26]马庆国.管理统计一数据获取、统计原理、SPSS工具应用研究[M].科学出版社,2002
[27]吴喜之.统计学：从数据到结论[M].中国统计出版社,2004
[28]张文彤,闫洁.SPSS统计分析基础教程[M].高等教育出版社,2004
[29]王子林,陆庆平.《中国上市公司业绩评价报告》[M].北京：经济科学出版社,2002
[30]黄岩,李元旭.《上市公司财务失败预测实证研究》[M].系统工程理论方法应用,2001
[31]周首华.《企业财务预警方法与机制》[M].现代财务理论前沿专题,2001(4)
[32]Aharony J., Jones C. P. and Swary I., "An Analysis of Risk and Return Characteristics of Corporate BankruptcyUsing Capital MarkData", Journal ofFinance,1980(9):1001-1016.
[33]Altman, E., "Financial Ratios, Discriminant Analysis and the Prediction of Corporate Bankruptcy",Journal ofFinance,1968(9):589-609.
[34]Aziz, A., D. Emanuel and G. Lawson, "Bank Prediction:An Investigation of Cash Flow Based Models",Journal ofManagemeStudies,1988(25):419-437.
[35]Barth, M.E., Beaver, W.H., andLandsman, W.R.,"Relative Valuation Roles of Book Equity Value and Net Income as a Functioof Financial Health",Journal ofAccounting and Economics,1998(25):1-33.
[36]Beaver, W.,"Financial Ratios as Predictors of Failure",Supplement to Journal ofAccounting Research,,1966:71-111.
[37]Best, R.W.,"The Role of Default Risk in Determining the Market Reaction to Debt Announcement", The Financial Review,1997 (2):87-105.
[38]Carmichael,D.R, The Auditor's Reporting Obligation.Auditing Research Monograph No.1 (NewYork:AICPA),1972:94.
[39]Charles E. Mossman, Geoffrey G. Bell, L. Mick Swartz, Harry Turtle,1998,"An Empirical Comparison of Bankruptcy Models",The Fnancial Review,1998(33):35-54.
[40]Deakin, E. B,"ADiscriminantAnalysisofPrediction of BusinessFailure",Journal of Accounting Research.,1972 (Spring):67-169
[41]Gregor, A., and Kaplan, S.N.,How Costly Is Financial Distress?Journal of Finance,1998 (10):1442-1493.
[42]Hamer, M.,"Failure Prediction:Sensitivity of Classification Accuracy toAlternative Statistical Methods and Variable Sets", Journal Accounting and Public Policy.1983(2):289-307.
[43]Meyer, P. A. and H. W. Pifer,"Prediction of Bank Failures",Journal ofFinance,1970 (9)
[44]Secrist, H.,"National Bank Failure andNon-failure. The Principia Press, Bloomington,1938
[45]Stephen A. Ross, Randolph W. Westerfield, Jeffrey F. Jaff,1999,吴世农、沈艺峰等译：《公司理财》,机械工业出版社,2000年。
[46]Agrawal,Anup and Knoeber,Charles R Firm performance and mechanisms to control agency problems between managers and shareholders,Journal of finance and quantitative analysis, 1996(31):377-97,
[47]Donaldson,Lex;Davis James H.Stewardship theory or agency theory:CEO governance and shareholder returns.Australian Journal of Management.l991(1):49-66.
[48]FitzPartriek, P.J.A comparison of Ratios of Sueeessful Industrial Enierprise with those of Failed Firms.Certified Public Accountant,1932(2):589—b