This thesis contributes quantitative and qualitative evidence from Southeast Asia to the literature on earnings management and audit quality. The association between earnings management and audit quality is investigated primarily by a new measure of audit quality and a new probit model. The new probit model tests the degree to which audit firms can be tolerant of earnings management at different levels of the artificial audit materiality. The tests of the probit model cover time-series data from 1992 to 2011 for 2,148 listed companies in Indonesia, Malaysia, the Philippines, Singapore and Thailand with a total of 20,757 firm-year observations. To extend the results of the probit model and to explore the stakeholders of audits’ perspectives on earnings management and audit quality, 16 semi-structured interviews with respondents from the audit firms, the listed companies, the regulators and the academic institute in Malaysia, Thailand and Singapore are also conducted.
This thesis finds that the term earnings management is seen differently from its extant definition. In addition, this thesis provides evidence that the different roles in the process of an audit lead to different definitions and measures of audit quality; therefore, audit quality remains an imprecise measure. The evidence of this thesis also indicates that in the context of Southeast Asia, big firms have higher quality audits than non-big firms. This is likely to be because big firms are more concerned with their reputation and the serious consequences of an audit failure and big firms are perceived to have more resources.
This thesis also explores whether long audit tenure impacts audit quality. The evidence on this suggests that audit partner rotation rather than audit firm rotation is being appointed by key stakeholders. Moreover, there is evidence that in comparison to audit firms from Singapore, those from Malaysia are more tolerant of earnings management whilst those from Indonesia, the Philippines and Thailand are less so. This thesis goes on to explore the possible impact of some national level factors such as the number of registrant audit firms, the use of non-English standards, corporate governance and type of legal system. From the probit model tests, it was found that non-English accounting standards and a limited number of registrant audit firms did not restrict audit quality, as perceived by some key stakeholders of audits.
