Research On Discernment Of Audit Collusion And The Effectiveness Of Prevention Measures

Audit collusion seriously damaged the basic order and reputation of the securities market. Although severely punish those who conspire to bring psychological comfort to victims but compensation effect is little. Investors, creditors and other stakeholders more concern about how to detect early and stop it timely, and hope that the Punishment Strategies is effective to prevent similar incidents from recurring. How to identify and prevent audit collusion is still important research topic.The thesis use the basic theory such as economics, management science, law and game theory to analyses the formation mechanism of audit collusion. The thesis gets Characteristic variables of audit collusion from perspective of the listed company's financial condition, ownership structure and governance structure, the characteristics of accounting firms, auditing fees, then constructs the audit collusion discerning model by using a variety of methods.The thesis has organized a total of six chapters from the three major aspects to study the problem of identification and governance of audit collusion. First, the thesis uses the basic theory such as economic man hypothesis, asymmetric information theory, the reputation theory, game theory, to analyses the collaborators how to transform his opportunistic motives to opportunistic behavior when there are deficiencies in institutional arrangements and reputation mechanism works weakly. By constructing a multi-game model, the thesis analyses the characteristics of collusive behavior and the role in the prevention of the audit collusion. These studies provide a solid foundation for the following theoretical analysis and empirical testing. Second, using empirical methods to study the relationship between anomalies listed company's financial situation, special ownership structure and governance structure, accounting firm characteristics, abnormal audit fees and collusion, and then obtain the characteristic variable of audit collusion by multiple regression analysis. On this basis, the thesis use the methods of single-variable model, principal component analysis, LPM, LOGIT and PROBIT regression methods to construct the model of discernment of audit collusion respectively. As audit collusion is highly concealed, it is not ideal to discern by single-variable model, but multi-variable model is better than single-variable model greatly. The discerning accuracy rate of PROBIT model is higher than LPM model, but Logistic model is higher than PROBIT model. Third, expanding the model constructed by Tirole (1986), Kofman and Lawarree (1993), Fahad (2003), the thesis carries on reasoning and demonstrating to analyze the validity of preventing collusion by rewarding or punishing the auditor. Then uses empirical methods to test the effectiveness of punishment mechanism in the fight against collusion from the aspect of abnormal stock market reaction, the value of listed companies, the reputation of auditors, audit quality and audit independence. The results show that the market value of listed companies and the reputation of accounting firms had not been impacted substantially, but punishment mechanism enhance the independence of the audit quality and the audit independence.The main innovation of the thesis lies in: First, the thesis analyses the collaborators how to transform his opportunistic motives to opportunistic behavior from the aspect of institutional arrangements, governance structure, legal liability and reputation stress, which provides theoretical foundation of the research. Second, based on the research about identification of financial reporting fraud, the thesis fully considers the characteristics of the audit conspiracy and constructs the identification models of audit collusion, which provide a better approach to identify the audit collusion case. Third, the thesis analyses the effects of preventive measures and tests it by empirical study, which provide theoretical basis and evidence to improve the existing system for preventing audit collusion.