本文主要在探討用財務比率、公司治理變數、盈餘管理與投資決策對企業財務危機發生機率之預測能力,分別建構了七種財務危機預警模型,並運用羅吉斯迴歸分析方法來探究各變數所組成的模型,對企業財務危機發生機率之預測能力。 研究期間以2001年至2009年之80家財務危機公司配對160家財務正常公司為本研究之樣本,從研究的結果顯示,在財務比率變數中,以速動比率及現金流量比率對財務危機機率有最顯著負向影響,表示速動率及現金流量比率越高,發生財務危機的機率愈低。公司治理變數中,以董監事持股質押比率有顯著的正向影響,董監事持股質押比率越高,發生財務危機的機率亦愈高。而在盈餘管理與投資決策變數中,發現企業聘請四大事務所的比率愈高,發生財務危機的機率愈低,投資損失的金額越高,發生財務危機的機率亦愈高。而由上述相關決策指標所建構的財務危機預警模型顯示最佳的預測正確率,在財務危機發生前一年有達94.2%之正確率。因此本研究對財務危機預警模型的建構具有重要啟示與貢獻。
This paper aimed to discuss the predictability of financial distress using variables in relation to financial ratios, corporate governance, earnings management, and investment decisions. Seven financial distress alert models were created and analyzed using Logistic Regression to determine the predictability of financial distress using models based on different variables. The research sample consisted of 80 companies in financial distress between 2001 and 2009, compared with 160 companies who remained financially sound over the same period. The research showed that, among all financial ratios, the Quick Ratio and the Cash Flow Ratio had the most significant negative correlations on the probability of financial distress. This implied that the higher the Quick and Cash Flow Ratios are, the less likely it is for a company to undergo financial distress. With respect to corporate governance variables, the percentage of shares collateralized by directors/supervisors had the most significant positive correlations. This indicated that the higher percentage a company’s directors and/ or supervisors place their shares as collateral, the more likely the company in question is to undergo financial distress. With respect to earnings management and investment decisions, the retaining of the Top Four accounting firms and investment losses were the variables with the most significant correlations; those research samples also suggested that companies who retained the Top Four accounting firms for services were less likely to undergo financial distress, whereas companies who incurred higher investment losses were more likely to undergo financial distress. The financial distress alert model that incorporated financial ratios, corporate governance, earnings management, and investment decision variables delivered the best predictability, up to 94.2% accurate in the year before their respective financial crisis. This research, therefore, provided an important reference and contribution to the construction of a financial distress alert model.