| Summary: | 碩士 === 輔仁大學 === 會計學系碩士班 === 93 === Once the credit risk and/or financial distress occur in the enterprise, it does not only hurt the rights and interests of the stakeholders but also pay for society cost. It will be a great help to all societies if we can predict the signals of credit risks that might possibly occur. That’s why different credit risk models are came out. They are all for the avoidance of the damage and even get the warnings that enterprise might fail in the earlier.
This study uses financial rate to measure the possibility of the credit risk. The listed and OTC companies in Taiwan ranging from 1997 to 2004 are sampled. First of all, we utilize factor analysis method to extract components of three sample groups, i.e. electronic industry, traditional industry and whole industry then rename each of them. Second, utilize Logit Model to build up each sample group’s related credit risk model for the further analysis.
The result of experiments shows that through the factor analysis method we can get 10, 8 and 9 components of each sample group respectively. Eight components are the same, i.e. short-term liquidity, operational efficiency, assets and equity efficiency, profitability, cash flow, operation growth, debt paying ability and profitability growth. The other components of each sample group are quite different. It shows that different industry comes out different components and also indicates that there exist differentiations between electronic industry and traditional industry.
Then we build up the credit risk model base on these extracted components and find out the significant components are different between electronic and traditional industry. The component of assets and equity efficiency is quite significant in electronic industry, but not in traditional industry. It apparently shows the electronic industry might easily get into credit risk due to lack of the efficient utilization of the assets and equity. Below data are the accuracies of each model, 95.58% for the all sample, 97.50% for the electronic industry and the traditional industry is 94.64%. The explanatory power of each model is 30.23%, 17.14% and 33.49% respectively. Although the explanatory power of the electronic industry model is the lowest, but it does not affect the accuracy of the classification in our study.
|
|---|
