A Study of Constructing Corporation Financial Distress Prediction Models Using LOGIT and Discriminant Analyses
| Autor: | Hong, Yu-Chiun, 洪于珺 |
|---|---|
| Rok vydání: | 2014 |
| Druh dokumentu: | 學位論文 ; thesis |
| Popis: | 102 The purpose of this study is to construct corporation financial distress prediction models based upon financial ratios and corporate governance. The samples are from Taiwan listed companies, which had been announced distressed during the period of 2002 to 2012. The matching healthy samples are collected using similar year, industry and total assets. This study uses five year financial ratios and corporate governance variables prior to financial distress of both distress companies and healthy companies. A total of 96 companies with 36 financial ratio variables and 8 corporate governance variables are collected. This study compares both LOGIT and Discriminant models to check which model is suitable for predicting potential financial distressed firms. The results are summarized as follows: 1. Based on LOGIT and Discriminant models, the most important variables for prediction among financial ratios are debt ratio, EPS, inventory turnover ratio, sales growth rate, fixed assets turnover, the fund requirement, current ratio, payment days, days of working capital supply. Manager’s holding ratio is the only corporate governance variable that is significant for prediction. 2. In terms of each year prediction, both LOGIT and Discriminant models have the same 100% prediction accuracy. However, for the 5-year total sample, the LOGIT model demonstrated an 84.58% predictive accuracy and the Discriminant model only resulted in 68.54% accuracy. Thus, it might be inferred that the LOGIT model could provide better prediction. |
| Databáze: | Networked Digital Library of Theses & Dissertations |
| Externí odkaz: |
|