2.1 Literature on Dual Audit System
2.1.1 Definition of Dual Audit
Wu (2005) [226] provides a comprehensive study on the dual audit system in China. He holds that a dual audit refers to an audit system where the two engaged audit firms audit a company at the same time and issue separate audit reports in order to produce more reliable audit outcomes. (1)According to Wu (2005) [226] , there are mainly four types of listed companies that should have dual audits: ①A-share listed companies with initial public offering or seasoned equity offering no less than 300 million shares; ②To-be-listed or listed financial companies, including commercial banks, insurance companies and securities companies; ③B-share companies; ④AH companies.
The reasons why this study focuses on the dual audit system for AH companies are as follows. First, A-share listed companies with issuance no less than 300 million shares have a quite small number, and they lack representativeness and universality. Second, owing to particular requirements imposed by the Chinese financial policies on the financial industry and issues left over from history, financial companies are too special to be an object in research on listed companies. Third, although the B-share market thrived in the past, it now has weakened financing function and is marginalized, tending to be merged into the A-share market. As a result, studying the audit systems in that market seems to have little instructive significance to reality. Finally, the financing function of the H-share market becomes increasingly stronger. The H-share market plays an important role in achieving the mainland Chinese companies' internalization; more and more mainland companies have the willingness to be listed in Hong Kong. After the deregulation of the mandatory dual audit system, investors are very concerned about AH companies' audit system change and their audit quality (Sun, 2014) [1] . Therefore, it is of researchers and practitioners' interest to study how the deregulation of the mandatory dual audit system for AH companies affects the cost of equity and audit quality.
Overall, this study focuses on the dual audit system for AH companies. On September 17 1992, the MOF released the "Interim Provisions on Certified Public Accountants (CPA) Providing Services for Pilot Joint-Stock Enterprises" . In the same year, the HKEx released the "Listing Rules of the Stock Exchange of Hong Kong Limited" . According to these rules and regulations, a dual audit refers to an audit system where an AH company hires a mainland-based audit firm to audit (as per the mainland auditing standards) its A-share financial statements prepared in accordance with the CAS and a Hong Kong-based audit firm to audit[using the international (or Hong Kong) auditing standards]its H-share financial statements prepared according to the IFRS (or HKFRS). Each of the two engaged audit firms releases a separate annual audit report for the same company. This audit system aims at meeting the different information demands of Chinese and international investors.
2.1.2 Economic Consequences of Dual Audit System
2.1.2.1 Dual Audit System for AH companies
Given the uniqueness of the dual audit system for the AH-share market, studies on this regime are mainly conducted by Chinese scholars. Some researchers conduct normative analyses. For example, Zheng et al.(2011) [10] analyze the basic pattern and competition status in the H-share audit market. They argue that the deregulation of the mandatory dual audit system for AH companies provides an important opportunity for mainland local audit firms' development and improvement. They propose some strategies of entering the H-share audit market for mainland local auditors. Weng and Huang (2012) [11] list the weaknesses of mainland local audit firms. They also put up with some suggestions to improve mainland local auditors' competitiveness. They point out that although the deregulation of the mandatory dual audit system creates advantageous conditions for local audit firms when competing with others, they still need to become bigger and stronger through their own efforts.
In the empirical aspect, some studies compare audit quality across different groups in the period when the mandatory dual audit system was still in place. The common practice is to compare cross-listed firms with non-cross-listed firms. For example, Lin (2014) [229] and Che (2014) [230] show that AH companies have higher audit quality than non-cross-listings, providing indirect evidence on the role of the dual audit system in maintaining audit quality. Lin et al.(2014) [12] find that China's A-share listed companies that are required to adopt the dual audit system have greater auditor conservatism than those that are not bound by this requirement. In addition, auditor conservatism is highest when the engaged mainland-and Hong Kong-based audit firms belong to different brands. Wang and Xin (2011) [15] compare Big 4 and non-Big 4 auditors under the dual audit system. They find that, in most cases, when an AH firm hires a mainland-based member firm belonging to a Big 4 brand for its A-share audits, it also hires a Hong Kong-based member firm from the same brand for its H-share audits. Thus, any misconduct by one of the audit firms would impair the Big 4's international reputation and the other audit firm would also be adversely affected. Non-Big 4 auditors, however, have no such concerns. Their empirical results show that such a difference in motivation increases the audit quality of AH companies audited by Big 4 auditors compared to that of non-Big 4 auditors' AH clients.
Other empirical studies also involve observations in the period after the deregulation of the mandatory dual audit system. Wang (2014) [13] studies the effect of the dual audit system on audit quality among non-Big 4's audit clients using data from 2001 to 2012. She finds that the audit quality of AH clients is higher than that of clients listed only on the A-share market, suggesting that the dual audit system contributes toward improving audit quality. Focusing on A-share audits conducted by Big 4 auditors in 1995-2012, Ke et al.(2015) [5] find that Big 4 assign more-experienced partners and guarantee higher audit quality for their AH-share clients using the dual audit system than for other A-share clients. They attribute their findings to a potential positive spillover effect from Hong Kong-based auditors through the dual audit system. Using data on A-share firms from 2007 to 2014, Tian et al.(2017) [6] find that mainland-based auditors (including Big 4 and non-Big 4 auditors in mainland China) exert more audit effort and provide higher audit quality for AH firms using the dual audit system than for other A-share firms. Sun (2013) [14] focuses on changes in audit market structure and audit fees after the deregulation of the mandatory dual audit system. Using data on AH firms from 2006 to 2011, she finds that Big 4 auditors still occupy large market share in the mainland audit market. She also finds that after the policy change, audit fees even increase in some firms.
2.1.2.2 Dual Audit System in Other Sectors
Mainland Chinese companies issuing B shares were also bound by the mandatory dual audit system from 2001, the year when the China Securities Regulatory Commission (CSRC) required A-share listed companies who had issued B shares to have overseas audits. (2)Before that, they voluntarily adopted the dual audit system. Early studies on the dual audit system for B-share companies mainly focus on the differences between the CAS adopted by the A-share market and the international accounting standards adopted by the B-share market. Based on the A-share and B-share firms, Li (1997) [231] concludes that the main reason why accounting standards affect corporate earnings is the differences in accounting system and professional judgement. Chen et al.(2001) [232] use data of B-share companies and find that earnings under the CAS differ from those under the international accounting standards. They suggest that these differences in earnings mainly result from differences in accounting standards. Focusing on the differences between the net income adjusted and audited according to the CAS and the net income adjusted and audited based on the international accounting standards, Jiang (2002) [233] finds that the former is significantly higher than the latter even in 1999 when the CAS is closer to the international accounting standards. This finding implies that mainland auditors have lower independence than overseas auditors.
In 2007, the CSRC officially abolished the mandatory dual audit system for B-share companies. After that, studies mainly focus on testing changes in audit quality and audit fees before and after the abolition of the mandatory dual audit system. Li and Wu (2009) [234] analyze the influence of the abolition of the mandatory dual audit system for B-share companies from two aspects: auditor selection and audit fees. They find that most B-share companies no longer use the dual audit system after the abolition. B-share companies that previously hired a mainland audit firm and an overseas audit firm from different brands have stronger willingness to retain the original mainland auditor. In addition, the audit fee reduction driven by opting out of the dual audit system varies with firm characteristics. Fang (2010) [235] tests market reaction to the abolition of the mandatory dual audit system for B-share companies, and finds that investors do not react to this policy change positively. However, market reaction to firms previously hiring a mainland auditor and an overseas auditor from the same brand is significantly better than that to firms previously hiring two auditors from different brands. He also finds that the audit quality and information environment of firms previously hiring two auditors from different brands become worse after they opt out of the dual audit system. Zhang (2011, 2012) [236,237] test the effect of this policy change on corporate governance efficiency (represented by earnings management) using sample in 2001-2009. The empirical results show that compared with the A-share market, corporate governance efficiency in the B-share market decreases significantly after the abolition of the mandatory dual audit system, suggesting that the mandatory dual audit system has significant corporate governance benefits. Wang and Liu (2012) [238] , however, find that the abolition of the mandatory dual audit system has no significant impact on the audit quality of companies issuing both A shares and B shares (AB companies). After controlling for the impact of new accounting standards, AB companies' audit quality remains higher than that of pure A-share companies. Lin et al.(2014) [12] also find that the abolition of the mandatory dual audit system does not necessarily lead to a decrease in auditor conservatism for AB companies that no longer use this audit system.
As listed in Section 2.1.1, A-share listed companies with initial public offering or seasoned equity offering no less than 300 million shares and to-be-listed or listed financial companies also should have dual audits. However, as these two kinds of companies lack representativeness and universality, research on the dual audit systems for them is rather scarce.