博達事件對會計師事務所客戶股價影響之實證研究

Abstract

本研究以代理/監督假說及保險假說的角度，探討博達事件對會計師事務所客戶股價之影響，並分別以博達無預警聲請重整及博達前後任簽證會計師遭金管會停業兩年處分為第一個事件日及第二個事件日。 客戶股價單變量分析的研究結果指出，博達前後任簽證會計師事務所之客戶於兩個事件期間皆無顯著的負累積異常報酬率。博達前任簽證會計師之客戶於兩個事件期間皆有顯著的負累積異常報酬率，而博達後任簽證會計師之客戶於第二個事件期有顯著的負累積異常報酬率，此結果顯示博達事件的負面影響僅限於博達前後任簽證會計師之客戶，而未及於其他會計師之客戶。此外，博達後任會計師事務所之財務不健全客戶於第一個事件期股價下跌幅度較大，支持保險假說。迴歸分析的研究結果亦發現累積異常報酬與客戶的代理成本及財務狀況有關，符合代理/監督假說及保險假說之推論。

Based on agency hypothesis and insurance hypothesis, this study examines the effect of two Procomp related events on stock prices of audit clients. The first event date is June 15, 2004 when Procomp Informatics unexpectedly filed to the court for restructuring. The second event date is July 15, 2004 when the Financial Supervisory Commission decided to punish the predecessor and successor of Procomp’s CPAs by suspending their rights to audit listed companies’ financial reports for a period of two years. Regarding the stock market reaction, the results indicate that the stock prices of clients of the Procomp’s both predecessor and successor audit firms showed no significant reaction during the 1st and 2nd event periods. The stock prices of clients of the Procomp’s predecessor CPAs showed negative cumulative abnormal returns during the 1st and 2nd event periods, and the stock prices of clients of the Procomp’s successor CPAs showed negative significant reaction during the 2nd event period. The results suggest that the Procomp scandal only had an adverse effect on market prices of clients of the predecessor and successor CPAs, but not on market prices of clients of other CPAs. Moreover, results are consistent with the inference of the insurance hypothesis that the financially distressed clients of the Procomp’s successor audit firm experienced more negative abnormal returns than financially healthy clients during the 1st event period. The multivariate analysis indicates that cumulative abnormal returns were related to the agency cost and financial conditions of the audit clients. These findings generally support agency hypothesis and insurance hypothesis.