臺大管理論叢 NTU Management Review VOL.30 NO.2

237 NTU Management Review Vol. 30 No. 2 Aug. 2020 Note: REASON_DISCLOSURE is coded 1 if a firm discloses the reason for its auditor change and 0 otherwise. STATE is coded 1 if the observation is obtained from a state-owned enterprise and 0 otherwise. POST_2007 is coded 1 if the observation is obtained on or after 2007 and 0 otherwise. LNTA is the natural logarithm of total assets. CFOTA is the amount of cash flow from operations divided by the amount of total assets. NEWFINANCING is the sum of new debt and equity divided by the amount of total assets. TIE is the amount of operating income divided by the amount of interest expenses (times interest earned). DA is the amount of total liabilities divided by the amount of total assets. LEGALFORMLLP is coded 1 if the audit firm is formed as a limited liability partnership and 0 otherwise. LEGALFORMLLC is coded 1 if the audit firm is formed as a limited liability corporation and 0 otherwise. TOP10 is coded 1 if the auditor is one of China’s local Top 10 audit firms, measured by the market share of total audit fees and 0 otherwise. LOCAL is coded 1 if a firm engages an audit firm that is neither an international Big-4 nor a local Top 10 audit firm and 0 otherwise. CMI is the credit market index, derived by Wang et al. (2008). GDI is the government decentralization index, derived by Wang et al. (2008). LEI is legal environment index, derived by Wang et al. (2008). ISSUEB is coded 1 if the client issues B shares and 0 otherwise. ISSUEH is coded 1 if the client issues H shares and 0 otherwise. EXCHANGE is coded 1 if the firm is listed in the Shanghai Stock Exchange and 0 otherwise. The variable labelled with “ D_ ” reflects that annual report data are calculated between years. As shown in Table 5, the model examined for this inquiry is significant at the 1% level with a pseudo R 2 of 0.1488 ( p -value < 0.001). Moreover, the coefficient of STATE is positive and significant at the 1% level (coefficient = 0.430, p- value = 0.009). Overall, these statistics suggest that SOEs are more likely to disclose the reasons for auditor-client realignment than NSOEs. Furthermore, POST_2007 significantly and positively affects REASON_DISCLOSURE at the 1% level (coefficient = 1.001, p -value < 0.001). This statistic demonstrates that firms are more likely to disclose the reasons for auditor-client realignment after 2007. Comparing the results presented in Table 5 to the findings shown in Panel D of Table 1 and Table 3, we find it consistent that in terms of making disclosure of the realignment reasons, the percentage of SOEs is higher than that of NSOEs, which supports the notion that there is a significantly positive coefficient of STATE , as illustrated in Table 5. Since the percentage increase in firms disclosing realignment reasons from the pre-2007 to post- 2007 periods among NSOEs is higher than that of SOEs, we also find that the interaction term, STATE × POST_2007 , is significantly negative (coefficient = -0.645, p = 0.001). Taking these results together, we conclude that state ownership and the government initiative jointly affect firm management decisions to disclose realignment reasons.