预期财务困境成本在债务违约预测模型中的信息含量

Information Content of Expected Costs of Financial Distress in Default Predicting Model

本文首先对预期财务困境成本、财务政策与违约概率之关系进行理论分析，模型显示，预期财务困境成本愈高，债务违约概率愈小。在此基础上，作者选择2002-2004年间首次出现债务违约的97家A股上市公司为违约样本，并予以配对，应用二元Logistic回归模型实证考察预期财务困境成本与违约概率之关系。实证结果表明，预期财务困境成本在债务违约预测模型中既具有相对信息含量，也具有增量信息含量，即在基准模型的基础上引入预期财务困境成本的代理变量能够显著降低模型的误判率。

This paper theoretically analyzes the relation among expected cost of financial distress （ECFD）, financial policy and default probability. Theoretical model reveals that an increase in the ECFD reduces the probability of default. Based on theoretical analysis, the paper constructs a default sample of 97 listed A-share firms firstly defaulted during the year of 2002-2004, matching a non-default sample, and empirically examines the relation between ECFD and default probability. Empirical results are concluded that ECFD has not only relative, but also incremental information content in predicting default.