A Study on Prediction Model of Financial Distress Firms--Taking Listed and OTC Firms in Taiwan for Instance

• Main Authors: Huang-Feng Chang, 張晃峰
• Other Authors: Chien-Hui Lee
• Format: Others
• Language: zh-TW
• Published: 2010
• Online Access: http://ndltd.ncl.edu.tw/handle/j5j8ce

碩士 === 國立高雄應用科技大學 === 國際企業系 === 98 === This paper aims to study the relative issues about the firms with financial distress, hoping that the use of statistical methods could construct a corporate financial distress alert system to predict the risk in advance and to reduce the risk of payment default. The studying period is from 2001 to 2007, sampling from domestic listed and OTC firms. The study takes 30 firms with financial distress in 2006, as well as 30 matching sample companies as empirical data. In addition, a total of 10 companies with financial crisis in 2007 are taken as an out-of-sample. Logit regression model has been applied to construct the predictive models. Twenty financial indicators and fourteen non-financial indicators are considered simultaneously as explanatory variables in an attempt to construct financial ratios, non-financial ratios and combined ratios as three different distress predictive models. The main conclusions of this empirical research are as follows: (1) The forecast result from the out-of-sample research shows that when either the financial ratios or corporate governance variables are used alone to construct early warning system of the firm distress, the predictive ability is slightly poor. (2) If appropriate corporate governance variables are further added to the early warning models constructed by the use of financial ratios, the predictive ability and the stability of predictability of the model can be increased when it is applied to the case of predicting risks one to three years prior to the possible distress of the firm. (3) The empirical results show that, as to the financial, non-financial or combined ratios, no matter which ratios are constituted as predictive variables, the predictive ability is not good in the case of predicting risks four to six years prior to the possible distress of the firm. (4) Finally, it is suggested that the non-financial indicators be appropriately added to the early warning system to greatly improve the predictive ability in the case of predicting risks one to three years prior to the happening of the distress.