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

The Relationship between Shared Audit Opinions and the Audit Quality of Group Enterprises’ Financial Statements – Based on the Audit Adjustment 62 Relevant research from the past also supports this argument (Kellogg, 1984; Wilson and Grimlund, 1990; Stice, 1991; Lys and Watts, 1994; Bonner, Palmrose, and Young, 1998) The analysis above on whether or not the financial statements of all the components are audited by the same auditor will impair audit quality has no consistent conclusion in theory. Thus, this study does not predict the direction of the hypothesis and presents the following hypothesis: H1: A group company with component firms engaging with the same audit firm will have higher audit quality than group companies with component firms engaging with different audit firms. This study constructs a sample from the firms listed on both the TWSE and GTSM from 1999 to 2014. Firms in finance, insurance, and securities sectors as well as non- group companies are excluded. The full sample contains 2,191 observations after this study’s excluding observations with missing value in variables. The data is collected from the financial databases of Taiwan Economic Journal (TEJ). For the research model, this study refers to previous studies to construct the following regression model: | MISSTATE |( MISSTATE + , | MISSTATE – |) = β 0 + β 1 SHARE ( SHARE% ) + ∑ i β i Control Variables + ∑ j β j INDUSTRY DUMMY + ∑ k β k YEAR DUMMY + ε For the measurement of dependent variables, the audit adjustment ( MISSTATE ) server acts as the proxy of audit quality. MISSTATE is measured by the difference between pre-tax income before audit and after audit deflated by the total assets (Lennox, Wu, and Zhang, 2014). Prior studies point out that auditors are more susceptible to litigation and suffer reputational damage as a result of companies overstating their earnings or net assets (Kellogg, 1984; Kinney and Martin, 1994; Bonner et al., 1998; Francis and Krishnan, 1999). This has led auditors to hold different attitudes toward the increase or decrease in the manipulation of surpluses by management. Hence, this study separates the sample into two parts – positive audit adjustment ( MISSTATE + ) and absolute value of negative audit adjustment (| MISSTATE – |) – to further examine if the relationship differs between these two groups. The main experimental variable is SHARE . If the group company is issued the shared-opinion audit report by the engagement auditor (meaning that part of the components’ financial statement of a group company is audited by other component