本論文旨在研究中國大陸與台灣兩岸危機公司的存活時間與發生危機機率,並分析比較影響的因素。首先採用存活模型探討公司存活時間、存活率與退出率。將樣本分為大陸下市與特別處理股、台灣下市與全額交割股等四種樣本,探討1999年至2003年影響公司存活時間之因素與其失敗率。研究顯示,存活模型最適機率分配的選擇,大陸下市樣本以Weibull分配較佳之機率分配、大陸特別處理股與台灣下市皆以Lognormal為最佳時間機率分配、台灣全額交割股則以Log-logistic分配為最佳。 實證發現流動比率愈低、速動比率愈低、資產週轉率愈小、總負債比率愈小、固定比率愈高、銷貨品質愈小、償還新增資本比愈大皆會使大陸上市公司的存活時間縮短。另一方面,流動比率愈小、速動比率愈小、總資產週轉率愈低、總資產報酬率愈低、總負債比愈小、固定比愈大、現金利息保障倍數愈高、保留現金比率愈小、現金股利準備率愈高、股東權益與長期負債的現金報酬率愈小、股東權益現金報酬率愈大皆會使台灣上市公司的存續期間減短。 第二實證部分,採用Logit模型分析公司發生財務危機之機率。研究樣本為大陸特別處理股與台灣全額交割股兩種樣本。研究將預警模型分別以傳統財務比率指標、現金流量指標、傳統財務比率與現金流量兩種指標來建構預警模型。發現,當預警模型只有現金流量指標時,此模型的誤判率偏高,且正確區別率亦較低。實證顯示流動比率愈小、速動比率愈大、營運資金比率愈低、現金利息保障倍數愈高、保留現金比率愈小、銷貨品質愈大皆會使大陸上市公司危機發生率提高。另一方面,速動比率愈小、固定比率愈大、銷貨品質愈低則提高台灣上市公司危機發生率。
The purpose of the study is to build warning model of survival time and hazard rate for Mainland’s and Taiwan’s companies and analyze the reasons of financial crisis. First, we used survival model to analyze survival time, survival rate and hazard rate. The sample was divided into four parts which were Mainland delisted stock, Mainland ST stock, Taiwan delisted stock and Taiwan full delivery stock from 1999 to 2003. The Empirical results of the best fit for probability distribution showed that Mainland delisted stock was Weibull distribution, while Mainland ST stock was Lognormal distribution. Futher, Taiwan delisted stock was Lognormal distribution, while Taiwan full delivery stock was Lognormal. Our empirical results revealed that current ratio, quick ratio, total assets turnover ratio, debt ratio and quality of sale were positive relation to survival time of Mainland’s firms. However, fixed ratio and capital acquisitions ratio were negative relation to survival time in Mainland. We also found that current ratio, quick ratio, total assets turnover ratio, return on total assets, debt ratio, cash debt coverage and cash return on debt and equity were showed positive relation to survival time of Taiwan’s firms. Fixed ratio, cash interest coverage and cash return on stockholders’ equity were showed negative relation to survival time of Taiwan’s firms. Second empirical research employed Logit model to analyze crisis probability. The study sample was designed to two parts. They were Mainland delisted stock and Taiwan full delivery stock. We used traditional financial ratios and cash flow financial to build warning models. We found that the performance for employing cash flow to build warning model was poor. The empirical result showed that current ratio, capital to total assets and cash debt coverage are negative relation to the occurrence of crisis probability of Mainland’s delisted stock, while cash interest coverage and quality of sale were positive relation to the occurrence of crisis probability. Meanwhile, quickly ration and quality of sale were negative relation to the occurrence of crisis probability of Taiwan’s fully delivery stock. They were showed positive relation between fixed ratio and the occurrence of crisis probability.