財務危機預警指標與會計師查核意見關聯之研究

Abstract

本研究以2000年至2012年台灣上市公司為研究對象，結合Z-score財務預警模型中的財務變數與會計師查核意見，透過Logistic回歸分析檢視其財務危機預測能力。本研究亦探討Z-score財務預警模型之結果與會計師查核意見是否相關以及進一步檢視Z-score值對財務正常公司預測之準確性分析。另外驗證文獻上認為會計師面臨企業可能發生財務危機時，傾向簽發修正式無保留意見之假說。

實證結果顯示，透過Logistic回歸分析Z-score財務預警模型中的財務變數若結合會計師查核意見後，能有效提升其模型的判斷性。其中在危機前一年預測正確率最高可達91.8%，而加入會計師意見後型一錯誤與型二錯誤也有所下降。另外，本研究進一步將修正式無保留意見依照其在TEJ台灣經濟新報資料庫中之內容分類，結果顯示會計師出具修正式無保留意見(類型二)與保留意見的情況下，Z-score財務預警模型之結果與實際結果較吻合。最後，本研究亦證實即便企業可能發生財務危機，會計師仍舊傾向簽發修正式無保留意見，而不會直接簽發保留意見，且修正式無保留意見之比率大多在90%以上，可看出台灣會計師在簽發保留意見時仍相當保守。

This study’s subjects are the listed companies in Taiwan from year 2000 to 2012. Combined with Altman’s (1968) Z-score model variables and CPA opinions, using logistic regression analysis to establish the financial crisis warning model and view the ability to predict the financial crisis. Besides, the study examined the relevant between the results of Z-score model and CPA opinions. Finally, we also want to verify the literature suggestions that auditors prefer to issue modified unqualified opinion than qualified opinion as their clients have financial crisis. The empirical results show that if the Z-score model variables combine with the CPA opinions variable can effectively improve the judgment and enhance accuracy of the models. The prediction accuracy of the logistic model is 91.8%. Besides, after adding CPA opinions variable, the type I and II errors also declined. In accordance with the contents of the opinion in TEJ database, we further divided modified unqualified opinion into two categories. The results show that when auditors issued the modified unqualified opinion (category II) and the qualified opinion, the results of Z-score model were more equivalent to the actual results. Finally, The study also confirm that even though financial crisis may occur, auditors still tend to issue the modified unqualified opinion, but not directly issue the qualified opinion. Therefore, we can see auditors still quite conservative in Taiwan.