我国上市公司财务困境预测模型的参数问题研究

发布时间：2018-03-29 16:17

  本文选题：财务困境　切入点：预测模型　出处：《首都经济贸易大学》2012年硕士论文

【摘要】：企业陷入财务困境，不仅会给企业投资者、债权人以及其他企业相关利益者带来经济损失，而且会影响社会稳定。找到上市公司陷入财务困境的原因，构造适合中国上市公司财务困境的预测模型，及时获得上市公司财务状况出现严重恶化的预警信号，不论对投资者、债权人、经营者还是监管者，，都具有重大意义。 在以往对于财务困境预测的研究中，Logistic回归是主流的统计模型之一。然而，Logistic回归中隐含的模型系数固定不变假设可能与事实相悖，有待于统计检验。与此同时，企业所在行业属性在财务困境预测模型中经常被忽略或回避。因此，有必要将行业属性纳入预测模型，并验证企业行业属性对回归系数变动的解释力。因此，本文选择分层Logistic模型进行预测模型的建立，该模型既能适用于离散型响应变量，又能处理分层数据结构的统计模型，能够将宏观变量即行业信息引入模型，并对其效应检验。 本文首先对上市公司财务困境预测问题进行理论研究，在此基础上给出了基于流动比率的财务困境定义方法；之后进行样本选择和指标选取，选取2010年我国沪、深股市的A股上市公司作为研究总样本，得到样本公司1249家，其中财务困境公司272家，选择财务指标和非财务指标作为预测变量；而后建立多层Logistic模型，并对回归系数固定效应和随机效应检验。本研究的数据来自于国泰安CSMAR数据库。 基于分层Logistic模型的财务困境预测模型显示，微观层次回归系数的变动情况在组间（跨行业）存在，且该变动可分解为两个方面，一部分可由宏观层次解释变量解释，一部分由微观层次变量回归系数的随机斜率解释。同时，在财务管理意义上，模型在宏观层次支持了行业前景水平（由行业平均营业收入增长率表示）对企业财务困境风险的影响，在微观层次支持了流动比率、流动负债比率、资产报酬率、流动资产周转率和审计意见对企业财务困境风险的影响。除资产报酬率为随机效应外，其余均为固定效应。基于模型的拟合效果，多层Logistic模型对于数据拟合情况良好，适用于上市公司财务困境预测问题。
[Abstract]:Financial distress will not only bring economic losses to investors, creditors and other stakeholders of the enterprise, but also affect social stability. Find out why listed companies are in financial distress. It is of great significance for investors, creditors, managers and regulators to construct a forecasting model suitable for the financial distress of listed companies in China and to obtain early warning signals of serious deterioration of the financial situation of listed companies in time. Logistic regression is one of the mainstream statistical models in previous researches on financial distress prediction. However, the assumption of fixed coefficient of the model implied in logistic regression may be contrary to the facts and need to be tested by statistics. The industry attribute of the enterprise is often ignored or evaded in the forecasting model of financial distress. Therefore, it is necessary to incorporate the industry attribute into the forecasting model and verify the explanatory power of the enterprise industry attribute to the change of regression coefficient. In this paper, the hierarchical Logistic model is chosen to build the prediction model. The model can not only be applied to discrete response variables, but also can deal with the statistical model of hierarchical data structure. It can introduce macro variables (i.e. industry information) into the model and test its effects. This paper firstly studies the financial distress prediction of listed companies, and then gives the definition method of financial distress based on current ratio, then selects samples and indicators, and selects 2010 Shanghai, China. The A-share listed companies in Shenzhen stock market are taken as the total sample of 1249 companies, including 272companies with financial distress. The financial index and non-financial index are selected as the predictive variables, and then the multi-layer Logistic model is established. The data of this study are obtained from the Cathay Pacific CSMAR database. The prediction model of financial distress based on hierarchical Logistic model shows that the change of micro-level regression coefficient exists in inter-group (cross-industry), and the change can be decomposed into two aspects, some of which can be explained by macro-level explanatory variables. Part of it is explained by the random slope of regression coefficient of microcosmic variables. At the same time, in the sense of financial management, The model supports the influence of industry prospect level (expressed by the average growth rate of industry income) on the risk of financial distress at the macro level, and supports the current ratio, current debt ratio, asset return rate at the micro level. The influence of current assets turnover rate and audit opinion on the risk of financial distress of enterprises. Except for the stochastic effect of return on assets, the rest are fixed effects. Based on the fitting effect of the model, the multi-layer Logistic model can fit the data well. It is suitable for forecasting financial distress of listed companies.
【学位授予单位】：首都经济贸易大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：F275;F832.51;F224

【参考文献】

中国期刊全文数据库 前10条

1 吴德胜,梁j,殷尹;不同模型在财务预警实证中的比较研究[J];管理工程学报;2004年02期

2 李义超;蔡晶晶;;基于财务与非财务指标的财务困境预警研究——以制造业上市公司为例[J];中国管理信息化;2010年21期

3 鲜文铎;金延华;;基于Probit模型的上市公司财务失败实证研究[J];贵州财经学院学报;2007年06期

4 刘迎春;;基于Logistic回归的中国上市公司信用风险度量研究[J];黑龙江对外经贸;2010年11期

5 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年06期

6 吕长江,徐丽莉,周琳;上市公司财务困境与财务破产的比较分析[J];经济研究;2004年08期

7 姜国华,王汉生;上市公司连续两年亏损就应该被“ST”吗?[J];经济研究;2005年03期

8 郭斌;戴小敏;曾勇;方洪全;;我国企业危机预警模型研究—以财务与非财务因素构建[J];金融研究;2006年02期

9 刘京军;秦宛顺;;上市公司陷入财务困境可能性研究[J];金融研究;2006年11期

10 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年02期

中国博士学位论文全文数据库 前1条

1 徐鹿;上市公司财务困境预测实证研究[D];东北林业大学;2002年

本文编号：1681891