目前国外关于财务困境预测的研究基本形成了三种主流的方法或模型，分别是：基于会计信息的传统模型、基于未定权益(CCA)的模型以及风险模型(Hazard Model)。这些模型的判别能力经常通过以下3个维度来衡量：一是区分失败和非失败企业的能力；二是不同模型捕获企业失败或破产的增量信息程度；三是当失败和非失败企业的误分类成本不同时，模型的绩效表现。国际研究表明，模型的预测效果如何，在很大程度上取决于信息的可获得性，同时使用会计和市场信息，并采用风险分析(Hazard Analysis)的建模方法成为当前研究的一种趋势；预测因子、样本规模、建模技术以及验证方法都对模型的预测效果有很大的影响。%At present there have been three dominant approaches about predicting financial distress, they are traditional models predominantly based on accounting information, contingent claim analysis-based models and Hazard models. The discriminatory power of these models can be assessed along three dimensions: the first is the ability of discriminate between failures and non-failures, the second is the incremental information about bankruptcy captured by different models, and the third is the performance of the models when the cost of miscalssifying a failed firm is different from the cost of misclassifying a company that doesn't fail. The international studies show that the power of the models' prediction largely depend on the availability of information. Using both accounting and market information as well as the hazard analysis modeling have become a trend in current study. The predictors, sample size, modeling techniques and validation methods all have a great effect on the prediction results.
展开▼
