none

碩士 === 東吳大學 === 經濟學系 === 98 === In this study, we construct an early warning model to predict financial crisis of electronic companies in Taiwan. The models we use are panel ordered logit and panel logit models. There are 663 listed electronic companies in sample. We try to find important financial...

Full description

Bibliographic Details
Main Authors: Yi-Hua Tseng, 曾怡華
Other Authors: none
Format: Others
Language:zh-TW
Published: 2010
Online Access:http://ndltd.ncl.edu.tw/handle/36385281590545362217
Description
Summary:碩士 === 東吳大學 === 經濟學系 === 98 === In this study, we construct an early warning model to predict financial crisis of electronic companies in Taiwan. The models we use are panel ordered logit and panel logit models. There are 663 listed electronic companies in sample. We try to find important financial ratios which are capable of distinguishing abnormal companies from normal companies. In addition, we try to test the hypothesis of Pompe and Bilderbeek (2005), which assumes that the deteriorated process of a company starts from financial ratios of efficiency, then financial ratios of profitability, safety, and liquidity in order. We also considered macroeconomic variables- average benchmark interest rate and exchange rate- to see their impacts on predictive power of model for company’s financial crisis. There are five conclusions in the study. First, whether the model is panel ordered logit model or panel logit model, the goodness of fit is enhanced after considering macroeconomic variables. Second, the financial ratios of liquidity and the average benchmark interest rate affect a company’s default probability. The impact of the average benchmark interest rate on the probability of a company as core default or other default is much more than the impact of the financial ratios of liquidity. Third, the hypothesis of Pompe and Bilderbeek (2005) is not confirmed for Taiwan’s electronic industry. Fourth, the impact of financial ratios of liquidity and the average benchmark interest rate is more for companies with core default than companies with other defaults. Fifth, for prediction power, panel logit model with macroeconomic variables and financial ratios has more predictive power than which only with financial ratios.