| Summary: | 碩士 === 國立雲林科技大學 === 財務金融系碩士班 === 95 === Based on the traditional Altman Z-Score model, the main purpose of this thesis is to investigate a rather complete set of financial ratios linked to SMEs in Taiwan and find out which are the most predictive ratios for the entities’ credit worthiness in order to construct a default prediction model.Using a panel data of 69 defaulted firms(with sales less than NTD2500 million) and 1380 non-defaulted firms over the period 1998~2004, we pick up the most predictive five ratios for default prediction by employing accuracy ratio and stepwise regression technique, by comparing regression results from multivariate discriminant analysis(MDA) model, logisitc model, and the traditional Altman Z-Score model and by testing from 2005-2006 out of sample. In addition, we consider the effect of macroeconomic variable and industrial variable on the corporate default by adding those variables to the original model and run a robust test under considering different sample ratio of the defaulted firms and non-defaulted firms.
Empirical results show that the five corporate financial ratios selected through two-step process of this thesis are better than ones of Z-Score model by comparing regression results from MDA model, logisitc model, and the Z-Score model and by testing from out of sample. This result appears robust under the different sample ratios of the defaulted firms and non-defaulted firms. Howver, we do not find out the existence of the significant effect of macroeconomic variable and industrial variable on the corporate default.
|
|---|
