学海网

毕业论文-引入非报表指标的财务危机预警研究—以制造业公司为例，共83页，41277字
中文摘要
随着经济全球化进程步伐的加快，我国社会主义市场经济体制也在不断完
善，我国企业在面临竞争与机遇的同时也面临许多潜在的风险和危机。尤其在
2008 年的全球经济危机过后，人们对于在复杂的金融环境下，如何及早地预知
风险、提前预判危机的需求日益增加。因此，建立一个稳定并且准确度高的企
业危机预警模型成为研究学者们关注的焦点。
本文的第 1 章是引言部分，主要阐述了在全球经济不景气的大背景下，我
们进行财务危机预警对于政府、企业经营者及投资人的重大意义。随后的文献
综述部分，对国内外研究学者的研究成果，按照研究方法的不同和变量选取的
差异进行进一步的比较、分类和总结。
本文的第 2 章至第 3 章是研究方法及研究成果部分，也是本文的重点。首
先，本文选取了我国 A 股市场上制造业 2007-2010 年被证监会实施特别处理的
上市公司作为研究样本。随后，根据上市公司公布的数据选择了 20 个合适的报
表指标，并创新性的引入了 4 个非报表因素的指标。接下来的第 3 章，在经过
对变量进行检验后，使用 Logistic 回归方法，构建了仅包含财务报表指标信息
的单一模型和引入了非报表信息的混合模型。单一模型的回归结果说明，对于
制造业上市公司财务危机具有显著判别作用的报表指标有资产负债率、资产报
酬率和总资产现金回收率；混合模型的回归结果说明，对于财务危机具有显著
判别作用的非报表指标有审计意见类别和国有股比例；对于处于不同生命周期
的企业，判定财务危机的指标也会有所差异。单一模型的模拟效果不如混合模
型，但是非报表因素的判别稳定性不如报表因素。
第 4 章是本文的总结和展望部分，主要对本文的四个模型进行理论性的分
析和概述，并陈述了本文存在的不足和未来类似研究的发展趋势。通过研究发
现，加入非报表因素指标后的模型模拟效果更好。但是本文从样本量、变量选
择以及模型的适用性方面，都可以做进一步深入的挖掘和研究。更加全面的反
应企业的生产经营状况，更加准确的预测未来企业的发展状况是类似研究的发
展趋势。
关键词：财务危机预警 非报表因素指标 Logistic 回归 分割点

ABSTRACT
With the acceleration of the economic globalization, the establishment and
improvement of socialist market economic system is developed constantly in our
country. The enterprises in our capital market up against important competition and
opportunities which are also confronted with potential problems and risks.
Especially in the 2008 global economic crisis, people who are living in the complex
financial environment are really concerned about how to predict the risk and how to
anticipate the crisis. Therefore, it is necessary to establish a stable and accurate
model of the prediction of corporation bankruptcy which can forecast effectively of
management situation and financial condition in the future.
Chapter 1 of this article is the background section, focusing on the significance
and purpose of this topic to the Government, business operators and investors in this
desolate circumstance. The following section is literature review. This part is about
the comparison, classification and summarization of the paper home and abroad,
classifying them with two categories in different models and diversified ratios.
The second and third sections are the focus of this article, dealing with the
research methodology and results. First, based on the previous studies, the listed
manufacturing industries which were specially treated during 2007-2010 in A-share
market are selected. Next, according to the data released by the listed companies, 20
financial combined ratios and 4 innovative factors are introduced in the model. After
a series of layered tests such as Kolmogorov-Smirnov test, T test, Mann-Whitney
test and Multi-collinearity test, utilizing the logistic regression, two different models
are constructed, one of including financial statement information only, the other
containing other information comprehensively. The single model shows that debt to
asset ratio, ROA EBIT ratio and operation cash into asset are significant to the
discrimination function. The mixed model illuminates that the two indicators
opinion type and state shares percentage have ability to distinguish groups. For
different corporations in different stage of industry life cycle, the model may have
heterogonous indicators of bankruptcy forecasting. The single model has a lower
II
ability in discrimination relative to the hybrid one, but a higher precision as well.
The forth section is about the summary and expectation of this thesis, including
the analysis and overview of the four main models and explain the shortcomings in
this study objectively. The tangible result of research shows that the bankruptcy of
enterprise can be predicted accurately more than 90%. The mixed model with
comprehensive variables is more significant. However, as for the aspects of sample
size, variable selection and modification of model, the study subsequent can be
further excavation and research. The trend of similar article should be more
preciseness and reflect more comprehensive situation of the enterprise in the future.
Keywords: financial early warning; non-reporting financial ratio; logistic
regression; cut-off point

目 录
第 1 章
引言 .... 1
选题背景 ........ 1
研究目的和意义 ..... 2
本文的研究思路及框架 .. 2
文献综述 ........ 3
理论研究方法的发展 ........ 5
模型指标选取的演变 ........ 8
第 2 章 研究设计与方法 10
2.1 主要应用的概念及模型 ... 10
国内外学者对于财务危机的定义 .... 10
我国上市公司ST标识...... 10
行业选择 ........ 12
2.1.4 LOGIT模型 .... 12
2.2 样本公司的选取.... 13
第 3 章
指标选取 ...... 16
模型的构建 ........ 20
基于单一财务指标的预警模型 ...... 20
财务指标的正态性检验 .. 20
指标描述性统计 ..... 22
均值差异检验 25
多重共线性检验 ..... 26
单一模型的logistic回归... 27
加入企业成长性的Logit模型. 31
引入非报表指标的Logit模型构建.. 33
模型中引入非报表指标的意义 ........ 33
混合模型的Logistic回归结果及检验 ........ 34
确定判别阈值 ....... 36
第 4 章 模型预测效果 ....... 38
结论 .. 40
结论分析 ...... 40
本文存在的不足及对未来相关研究的展望 .... 41
插图索引 . 43
表格索引 . 44
参考文献 . 46
致谢 48
声明 49
附录A 外文资料的调研阅读报告或书面翻译.. 50
原文索引 . 60
附录B 2007-2010 年制造业上市公司样本数据 ........ 61