基于会计信息的KMV模型实证研究

An Empirical Analysis of the KMV Model Based On Fiscal Data

本文借助KMV模型框架,运用统计方法对大量的公司财务报表数据进行模型参数估计,计算得到了非上市公司的违约距离和经验EDF函数,实现了违约概率的模型估计。实证表明,我国公司在违约距离或违约数量上的真实概率分布均呈现显著的T分布和肥尾特性;违约距离具有较高的风险区分能力;由会计信息进行参数估计的模型导出的EDF具有较高的风险标识精度;进而表明基于会计报表数据的违约风险模型和基于资本市场数据的模型在实证上的有效性非常近似。

Based on the principle of KMV model and statistical method, applied with much fiscal data to forecast the parameter of KMV model, the paper simulates the distance to default and the EDF for non-listed companies and calculate the probability of default. The research demonstrated that the distribution of samples have the characteristic of T distribution and ＂fat tail＂, the distance to default has a more precise ability to distinguish the credit risk, the EDF which based on fiscal data can reveals the degree of credit risk precisely, the model based on bond market data and fiscal data have the same precision for the default risk estimation.