2013年6月20日学术报告通知.docx

Does Public Enforcement in Weak Investor Protection Countries Matter? Evidence from a Natural Experiment Bin Ke1, and Xiaojun Zhang2 ABSTRACT Views differ among investors, policymakers, and researchers on whether public enforcement matters in protecting investors. The objective of this study is to use a natural experiment of public enforcement from China to test the causal effect of public enforcement on shareholder value. Specifically, the China Securities Regulatory Commission (CSRC) conducted a Special Public Enforcement Activity in March 2007 regarding listed firms’ compliance with five major corporate governance regulations issued in 2002-2006. The public enforcement activity appears to be a success because the listed firms were forced to disclose over 10,000 corporate governance noncompliance problems, more than 90% of were claimed to be corrected by the end of the Special Public Enforcement Activity in 2008. We find some evidence that the Special Public Enforcement Activity helps reduce controlling shareholders’ tunneling, but we find little evidence that correcting the identified corporate governance noncompliance problems increases overall shareholder value. Preliminary draft Key words: public enforcement; weak investor protection countries; China; firm value JEL classifications: xxxx June 11, 2013 We wish to thank Huai Zhang, Terence Ng and workshop participants at the Nanyang Business School for helpful comments. 1Division of Accounting, Nanyang Business School, Nanyang Technological University, S3-01b-39, 50 Nanyang Avenue, Singapore 639798.Tel: +65 6790 4832. Fax: +65 67913697. Email: kebin@ntu.edu.sg. 2Division of Accounting, Nanyang Business School, Nanyang Technological University, S3-01b-73, 50 Nanyang Avenue, Singapore 639798.Tel: +65 9811 6160. Fax: +65 67913697. Email: C090040@e.ntu.edu.sg. 1. Introduction The strength of a country’s investor protection depends on the quality of the laws that protect investors’ rights and the strength of legal institutions that facilitate law enforcement. A widely held view in the finance literature, which we share, is that investor protection in general and law enforcement in particular are vital for corporate financing, financial market development and economic growth (La Porta et al. 1997; Beck et al. 2000). The objective of this study is to use a natural experiment from China to examine whether public enforcement of securities laws, a specific investor protection mechanism, can help increase shareholder value in weak investor protection countries. Our research question is motivated by an ongoing debate on the relative efficacy of two common approaches to law enforcement: public enforcement via a public regulator versus private enforcement via private parties (mainly using private litigation). In an influential study on the effect of law enforcement on financial market development across 49 countries, La Porta et al. (2006) find little evidence that public enforcement benefits stock markets, but they find that laws mandating disclosure and facilitating private enforcement through liability rules benefit stock markets. La Porta et al.’s conclusions have gained wide acceptance in the academic circle (e.g., Wurgler 2000; Shleifer and Wolfenzon 2002). Influential policy makers such as the World Bank, International Monetary Fund, and the European Central Bank also share a similar view (World Bank 2006; Bruno and Claessens 2008; Hartmann, Heider, Papaioannou, and Duca 2007). For example, the World Bank (2006, p.1) asserts that “[i]n banking and securities markets, characteristics related to private monitoring and enforcement drive development more than public enforcement measures.” However, Jackson and Roe (2009) have directly challenged La Porta et al.’s (2006) conclusions. Jackson and Roe (2009) argue that there is no a priori reason to believe that private enforcement dominates public enforcement in protecting investors. The reason is that 1 both approaches have serious defects and strong advantages and therefore it is an empirical question which enforcement mechanism is more effective in protecting investors (see Jackson and Roe 2009 for a detailed discussion). Jackson and Roe (2009) also question the reliability of La Porta et al.’s (2006) public enforcement proxy, which is based on the financial supervisor’s formal qualities (i.e., independence from the executive, its investigative powers, its capacity to issue remedial orders, and the range of criminal sanctions available). Using securities regulators’ resources as a more direct proxy for the intensity of public enforcement, Jackson and Roe (2009) find that financial market development significantly correlates with stronger public enforcement. In horse races between the resource-based measures of public enforcement intensity and the most common measures of private enforcement, Jackson and Roe (2009) find that public enforcement is overall as important as disclosure in explaining financial market outcomes around the world and more important than private liability rules. Overall, the extant literature is mixed on whether public enforcement is effective in protecting investors. Whether public enforcement can help protect investors carries additional significance in weak investor protection countries because private enforcement mechanisms usually do not work well due to the lack of an independent judiciary. In addition, it is extremely difficult to develop credible private enforcement institutions in many developing economies and the only viable option readily available to protect investors is public enforcement mechanisms (Layton 2008). Therefore, it is of utmost importance to understand whether public enforcement works in weak investor protection countries and if so, how. In this study we use a natural experiment from China to study whether public enforcement helps increase investor protection and therefore shareholder value. In March 2007 the China Securities Regulatory Commission (CSRC) issued a notice regarding a onetime Special Public Enforcement Activity on listed firms’ compliance with several important 2 corporate governance regulations issued over the period 2002-2006. The Special Public Enforcement Activity covers a comprehensive list of corporate governance issues, including matters related to controlling shareholders, shareholders’ meeting, board of directors, management’s responsibilities, internal control, executive compensation and accountability, and corporate disclosure. A unique aspect of the Special Public Enforcement Activity is that the public enforcement is carried out in two sequential steps. First, all listed firms are required to self-report identified corporate governance noncompliance problems and suggest remedial solutions and timetable in a self-assessment report. Next, the CSRC conducts its own independent investigation of listed firms’ corporate governance compliance status. To the extent that the CSRC identifies additional noncompliance problems, it will recommend further remedial solutions in a separate remediation report. The Special Public Enforcement Activity is required to be finished by the end of October 2007. In addition, in a separate notice issued in late June 2008, the CSRC further requires listed firms to provide a follow-up report on the status of the Special Public Enforcement Activity no later than July 20, 2008. Judging by the number of identified and corrected corporate governance noncompliance problems, the Special Public Enforcement Activity is a clear success. For our sample of 1,094 unique firms, the Special Public Enforcement Activity identified a total of 5,320 self-reported governance problems and 5,402 CSRC-identified governance problems. For both the listed firms’ self-assessment reports and the CSRC’s remediation reports, more than half of the identified problems are related to the board of directors and internal control. By the time of the follow-up report in 2008, the mean (median) firm claimed to have corrected 91% (100%) of the self-reported problems and 93.5% (100%) of the CSRC identified problems. We next examine whether correcting the identified corporate governance problems helps improve shareholder value. The Special Public Enforcement Activity covers many 3 corporate governance areas and therefore there are many possible channels through which the Special Public Enforcement Activity could affect shareholder value. The extant corporate governance literature indicates that controlling shareholders’ tunneling and low-quality financial reporting are two major challenges facing investors in emerging markets, including China. Hence, in this study we focus on the effect of the Special Public Enforcement Activity on the following two specific channels: controlling shareholders’ tunneling via intercorporate loans (Jiang et al. 2010) and earnings quality. To take into consideration the multiple effects of the Special Public Enforcement Activity on shareholder value, many of which are unobservable, we also consider the effect of the Special Public Enforcement Activity on net shareholder value, measured using operating accounting performance and Tobin’s Q. Using a difference-in-differences firm fixed effects regression research design, we find strong evidence that correcting self-reported governance problems helps reduce insiders’ tunneling. However, we find no evidence that correcting CSRC-identified governance problems helps reduce insiders’ tunneling. We find that the effect of correcting self-reported governance problems on tunneling is driven by the correction of governance problems related to controlling shareholders and internal control. However, we find little evidence that correcting either self-reported governance problems or CSRC-identified governance problems helps increase earnings quality and net shareholder value measured by operating accounting performance and Tobin’s Q. Overall, our results suggest that public enforcement of securities laws in weak investor protection countries may have some visible impact on listed firms’ behavior but it results in little significant net improvement in shareholder value. Our study makes several important contributions to the existing literature. First, to our knowledge, we are the first empirical study that directly examines the causal effect of public enforcement on shareholder value. Both La Porta et al. (2006) and Jackson and Roe (2009) 4 examine the association between public enforcement and financial market development because their public enforcement proxies are endogenously determined. Jackson and Roe (2009, p.233) readily acknowledge this problem and discuss the difficulty of identifying valid exogenous instruments for public enforcement. Our study does not suffer from the same problem because the identification and correction of corporate governance noncompliance is mandated by the CSRC and thus can be regarded as exogenous from a firm’s perspective. Second, to our knowledge, we are the first study to cleanly separate the effect of public enforcement from the effect of laws. This separation is made possible in our setting because the Special Public Enforcement Activity occurred long after the relevant corporate governance regulations took effect. In contrast, the effects of public enforcement and laws usually coexist in the settings of prior research (e.g., La Porta et al. 2006; Jackson and Roe 2009). Because both public enforcement and laws are imperfectly measured, it is typically difficult to isolate the effect of public enforcement from the effect of laws. Third, our study provides useful insights on the specific mechanisms through which public enforcement works. As noted in Jackson and Roe (2009, p.235), public enforcement could affect shareholder value through several channels: (i) public enforcers could spend resources to write more sophisticated regulations; (ii) conditional on the existing regulations, public enforcers could also conduct ex ante surveillance by preempting potential wrongdoing and governance problems; and (iii) once a wrongdoing is detected, public enforcers can also bring enforcement actions ex post against the perpetrators. The 2007 Special Public Enforcement Activity falls in channel (ii). Since the costs and benefits of the above three public enforcement channels are likely different, understanding how each channel affects shareholder value is important. Several studies have examined the economic consequences associated with channel (iii) in both the U.S. and emerging markets (e.g., Karpoff et al. 2008a; 2008b; Chen et al. 2005), but this line of research usually does not distinguish the roles of 5 private enforcement vs. public enforcement. To our knowledge, prior research has not isolated the effects of channels (i) and (ii). Our contribution is to provide direct empirical evidence on the efficacy of one example of channel (ii): ex ante surveillance of firms’ noncompliance with corporate governance regulations. Finally, our study provides some preliminary evidence on how public enforcement could be made more effective. A common criticism of public enforcement is that public enforcers are not as well informed as private enforcers. Hence, public enforcers may not be able to accurately identify the most severe governance problems. Consistent with this hypothesis, we find that correcting the governance problems self-reported by the firms themselves results in a reduction in insiders’ tunneling but we find no evidence that correcting the governance problems identified by the CSRC reduces insiders’ tunneling. This preliminary evidence suggests that public enforcement could become more effective if government regulators could find ways to utilize the private information possessed by other parties (insiders in our case). The rest of the paper is organized as follows. Section 2 provides the institutional background about the Special Public Enforcement Activity. Section 3 discusses the sample selection procedures and descriptive statistics. Section 4 examines the economic consequences of the Special Public Enforcement Activity. Section 5 analyzes the effects of correcting the individual categories of corporate governance noncompliance problems on firm behavior. Section 6 concludes. 2. Institutional background China is known for its poor investor protection (Allen et al. 2005). Over the years since China’s reestablishment of the domestic stock markets in the early 1990s, the Chinese 6 Government has introduced a series of investor protection regulations. Unfortunately, most of these regulations provide little benefit to investors because they are not strictly enforced. In recent years the CSRC starts to invest significant resources in the enforcement of investor protection regulations. To date, the 2007 Special Public Enforcement Activity, which applies to all publicly listed Chinese firms, is probably the most comprehensive and rigorous public enforcement activity undertaken by the CSRC. 1 The enforcement activity focuses on listed firms’ compliance with the following five investor protection regulations that went effect in 2006 or earlier: (i) the Corporate Governance Code of Listed Companies issued in 2002; (ii) The Company Law issued in 2005; (iii) The Opinions of China Securities Regulatory Commission on Improving the Quality of Listed Companies issued in 2005; (iv) Guidelines for Articles of Association of Listed Companies issued in 2006; and (v) Notice of the China Securities Regulatory Commission on Promulgating the Rules for the General Assemblies of Shareholders of Listed Companies issued in 2006. The major corporate governance areas targeted by the Special Public Enforcement Activity include the following: (1) Matters related to the controlling shareholder, such as the independence of the listed firm’s management from the controlling shareholder, related party transactions between the listed firm and the controlling shareholder, and direct product market competition between the listed firm and the controlling shareholder; 1 Recently the CSRC took a similar public enforcement approach targeting IPO firms and their underwriters and public accounting firms (Zhen 2013), suggesting this approach is one favored enforcement approach by the CSRC. 7 (2) Matters related to the shareholders’ meeting, such as shareholders’ participation rate, the availability of the online voting and cumulative voting, and the shareholders’ meeting’s procedures and records; (3) Matters related to the board of directors (including the supervisory board), such as the establishment and responsibilities of board committees, board meeting procedures, board meeting attendance, board meeting records, and director training; (4) Matters related to management’s responsibilities, such as working protocols, training, and insider trading policy; (5) Matters related to the listed firm’s internal control, such as policy on the use of proceeds from external financing, staff training, internal control policy, internal audit, risk management, and financial reporting internal control; (6) Matters related to management’s compensation and accountability, such as policy on managerial evaluation and incentive compensation, and policy on the board’s supervision of managerial compensation; (7) Matters related to the listed firm’s disclosure, such as investor relations, and disclosure policy; and (8) Matters related to all other miscellaneous governance issues. According to the CSRC’s Notice to listed firms on March 9, 2007 (CSRC 2007), the Special Public Enforcement Activity is implemented in three sequential stages and must be finished no later than October 31, 2007. The first stage is self-reporting. During this period all listed firms are required to perform a self-assessment of the firms’ compliance with the aforementioned regulations, propose remedial solutions to the self-identified noncompliance problems, including the remediation timetable, and then publicly disclose the relevant information in a board-approved self-assessment report submitted to the CSRC. The self- 8 assessment report must contain a detailed description of the identified governance problems and suggested remedial solutions, including whether the identified problems have been corrected or not by the time of the self-assessment report. The second stage is public comments. During this period listed firms are required to establish dedicated phone lines and internet-based communication channels to allow investors and the general public to make comments and suggestions on the listed firms’ investor protection. The public comment period should be no less than 15 calendar days. During this period the CSRC’s regional offices will also conduct an independent assessment of the listed firms’ compliance with the aforementioned regulations. After gathering the information from the listed firms’ self-assessment reports, public comments, and the CSRC’s independent assessments, the CSRC’s regional offices may propose remedial solutions to the identified problems beyond the listed firms’ self-assessment reports. No public disclosures by either the listed firms or the CSRC are made during this period. The third stage of the Special Public Enforcement Activity is implementation. During this period the listed firms are required to implement the suggested remedial solutions to all identified problems. In addition, the listed firms are required to publicly disclose a boardapproved remediation report submitted to the CSRC. The remediation report must disclose the corporate governance noncompliance problems identified by the firm, the public, the stock exchange, and the CSRC separately. For each identified problem, the remediation report needs to provide a summary of the identified governance problem and discuss the suggested remedial solutions, including whether the identified problem has been corrected or not by the time of the remediation report. To encourage listed firms to cooperate with the Special Public Enforcement Activity, the CSRC requires the listed firms to complete all three stages of the Special Public Enforcement Activity before the firms are allowed to propose and implement managerial 9 stock incentive schemes. For firms that have serious governance problems and refuse to correct them, the CSRC will not accept the firms’ applications for stock incentive schemes. This is a significant binding constraint because mainland Chinese listed firms were allowed to use equity-based executive compensation only after 2005 and therefore many firms were interested in proposing equity incentive schemes around the Special Public Enforcement Activity period. For firms that have corporate governance noncompliance problems, the CSRC’s regional offices may also request to meet the firms’ top executives, issue attention letters or criticism letters internally circulated among listed firms. In addition, the CSRC may deny a firm’s applications for seasonal equity offerings, share transfers, and mergers and acquisitions if the firm has serious uncorrected governance problems, such as lack of independence from the controlling shareholder and tunneling by the controlling shareholder. To make sure listed firms had fully complied with the requirements of the Special Public Enforcement Activity, the CSRC issued another Notice (CSRC 2008) on June 20, 2008 that further required listed firms to publicly issue a board-approved follow-up report about the status of the previously issued remediation report no later than July 20, 2008.2 This follow-up report should disclose whether the remedial solutions enclosed in the remediation report are fully implemented before the end of the previously proposed timetable. If certain remedial solutions fail to be implemented on time, the firms need to explain the reasons for the failure and discuss any follow-up plans, including the punishment for the responsible parties. The 2008 follow-up report singles out the following three major corporate governance problem areas in particular: (a) Strengthening the independence of the listed firm from the controlling shareholder; 2 To the extent that certain proposed remedial solutions are not implemented by the time of the follow-up report, listed firms are required to continuously update the progress of the unfinished remedial solutions proposed in the remediation report in the 2008, 2009, and 2010 annual reports. 10 (b) Preventing the controlling shareholder from tunneling the listed firm’s resources via inter-corporate loans; and (c) Strengthening the listed firm’s timely disclosure of price sensitive information. In summary, the 2007 Special Public Enforcement Activity is a comprehensive and rigorous public enforcement activity that applies to all publicly listed Chinese firms. To our knowledge, this is the most sweeping public enforcement activity the CSRC has ever taken with an aim to improve the listed firms’ compliance with existing corporate governance laws. 3. Sample selection procedures and descriptive statistics We limit our sample to the Chinese firms that are listed on the main boards of the two domestic stock exchanges, Shanghai and Shenzhen. Since the Special Public Enforcement Activity was launched in 2007 and finished by July of 2008, we test the impacts of the Special Public Enforcement Activity using the pre-period 2004-2006 and post-period 20082010. We exclude year 2007 because it is a transition year. We require each firm to have data for at least one year in both the pre-period and post-period in order to avoid the possibility that our inferences are due to changing mix of sample firms over time. To avoid IPO-related complications, we further require our sample firms’ IPO dates to be no later than December 31, 2002. Due to their unique business and special government regulation, we also delete firms in the financial industry. These sample restrictions result in a sample of 1,104 unique firms. We further exclude ten unique firms because, for unknown reasons, they were not subject to the Special Public Enforcement Activity. Our final sample contains 1,094 unique firms, representing approximately 78% of the stocks listed on the main boards of the two stock exchanges. Due to missing values on control variables, the actual number of unique firms included in the following regression results could be slightly smaller than 1,094. 11 We first provide some descriptive statistics on our sample firms’ degree of compliance with the Special Public Enforcement Activity. Figure 1 shows the frequency distribution of the announcement dates for each of the three key reports associated with the Special Public Enforcement Activity (i.e., self-assessment report, the remediation report, and the follow-up report). Recall that both the self-assessment report and the remediation report are required to be completed no later than October 31, 2007 while the follow-up report is required to be completed no later than July 20, 2008. It is clear from Figure 1 that most firms finished the required tasks mandated by the Special Public Enforcement Activity by the end of 2007. Table 1 reports the frequency distribution of our sample firms’ corporate governance noncompliance problems identified by the firms in the self-assessment report (Panel A) and the CSRC in the remediation report (Panel B), respectively, across the eight corporate governance categories mentioned in the previous section. While both the public and the two stock exchanges could identify additional corporate governance problems, in reality this rarely occurs.3 Hence, to avoid unnecessary complications, we only focus on the problems identified by either the firms themselves or the CSRC. Inferences are qualitatively the same if the problems identified by the public and the stock exchanges are controlled for in our regression analyses. The listed firms themselves voluntarily identified a total of 5,320 corporate governance problems. The CSRC identified an additional 5,402 corporate governance problems. These results suggest that the CSRC played a significant role in exposing the listed firms’ degree of noncompliance with existing investor protection regulations. Among the 5,320 self-reported problems, 5.5% are related to controlling shareholders, 3.95% related to shareholders’ meeting, 26.37% related to the board of directors, 7.01% 3 In fact, the 90th percentile of the number of problems identified by either the public or by the two stock exchanges as a percentage of the total combined number of problems identified by the firms and the CSRC is zero. 12 related to management’s responsibilities, 30.85% related to internal control, 5.34% related to executive compensation and accountability, 19.06% related to corporate disclosure, and 1.88% related to the miscellaneous category. Among the 5,402 CSRC-identified problems, 10.63% are related to controlling shareholders, 10.98% related to shareholders’ meeting, 31.17% related to the board of directors, 5,96% related to management’s responsibilities, 29.27% related to internal control, 1.46% related to executive compensation and accountability, 9.74% related to corporate disclosure, and 0.80% related to the miscellaneous category. For both the self-reported problems and the CSRC-identified problems, more than half of the identified problems are related to either the board of directors or the listed firm’s internal control. The Special Public Enforcement Activity is a huge success based on the high correction rates of the identified problems by the time of the follow-up report. For example, as shown in Table 1, the listed firms claimed that 91.3% of the self-reported problems and 94.0% of the CSRC-identified problems had been corrected by the time of the follow-up report. The comparable percentages are also very high (usually above the 70% threshold) for each of the eight categories of corporate governance problems. The only exception is that only 36.3% of the self-reported problems related to management’s compensation and accountability were corrected by the time of the follow-up report. However, the total number of self-reported problems in this category is only 284, representing 5.34% of the total number of self-reported problems. To provide another perspective on the correction rates of identified corporate governance problems, Table 2 shows the distribution of the correction rates, at the individual firm level, for the self-reported problems (Panel A) and the CSRC-identified problems (Panel B) separately, by the time of each of the three key report dates. For our sample of 1,090 firms that disclosed at least one self-reported problem, the mean firm claimed to have corrected 91.04% of the identified problems by the time of the follow-up report. For our sample of 13 1,014 firms that disclosed at least one CSRC-identified problem, the mean firm claimed to have corrected 93.51% of the identified problems by the time of the follow-up report. Overall, the statistics shown in Tables 1 and 2 suggest that the Special Public Enforcement Activity is a success. We next examine whether the Special Public Enforcement Activity helps increase shareholder value. 4. Consequences of the Special Public Enforcement Activity 4.1. Predictions As explained in the Introduction, we examine the effect of the Special Public Enforcement Activity on the following four corporate outcomes: two specific investor protection targets (inter-corporate loans and earnings quality), and two proxies for net shareholder value (operating accounting performance and Tobin’s Q). While the Special Public Enforcement Activity may have a positive effect on a specific corporate governance area, the effect on net shareholder value may not be positive for several reasons. First, the Special Public Enforcement Activity could be just ineffective. For example, corporate insiders may simply check the box declaring they have complied with the relevant corporate governance regulations without fundamentally changing their incentives to increase shareholder value. Second, even if the Special Public Enforcement Activity is effective in improving investor protection in the targeted areas, net shareholder value may not increase because controlling shareholders may resort to alternative tunneling channels that could be more costly to other shareholders (i.e., unintended consequences). Third, China’s investor protection laws have been enacted hastily over a fairly short period of time and therefore the quality of China’s investor protection laws may be low (i.e., bad laws from shareholders’ perspective). As a result, strict enforcement of such laws could hurt 14 shareholder value. Therefore, it is an empirical question whether the Special Public Enforcement Activity brings a net benefit to shareholders. 4.2. Research design We use the following difference-in-differences firm fixed effects regression model to examine the effect of the Special Public Enforcement Activity on inter-corporate loans (OREC), earnings quality (ABS_AQ), and operating accounting performance (OROA): DVit = β1SOLVED_SELFi × AFTER + β2SOLVED_CSRCi × AFTER + β3CONTROLSit - 1 + β4CONTROLSit - 1 × AFTER + µt + µi + εit (1) where i and t are firm and year indicators, respectively. See the Appendix for all variable definitions. The sample includes the firm years 2004-2006 and 2008-2010. DV represents one of the first three dependent variables: OREC, ABS_AQ, or OROA and CONTROLS represents a list of control variables. We will discuss both DV and CONTROLS later. We allow the coefficients on CONTROLS to vary with AFTER because China adopted a set of new accounting standards in 2007 that is substantially converged with the IFRS and therefore the relation between DV and CONTROLS could have changed post the Special Public Enforcement Activity. We include year fixed effects to control for time trends and firm fixed effects to control for time-invariate determinants of DV (e.g., whether a firm is statecontrolled or not). We use SOLVED_SELF and SOLVED_CSRC to test the effects of the Special Public Enforcement Activity. Both SOLVED_SELF and SOLVED_CSRC are firm fixed effects, denoting the number of self-reported corporate governance problems that were corrected by the time of the follow-up report and the number of CSRC-identified corporate governance problems that were corrected by the time of the follow-up report, respectively. Note that the coefficients on AFTER, SOLVED_SELF and SOLVED_CSRC are subsumed by the year and firm fixed effects. To the extent that the Special Public Enforcement Activity results in a 15 genuine increase (decrease) in investor protection resulting from the correction of identified corporate governance problems, we should expect the coefficients on AFTER×SOLVED_SELF and AFTER×SOLVED_CSRC to be negative (positive) when DV is OREC and ABS_AQ and to be positive (negative) when DV is OROA. We use the following firm fixed effects regression model to test the effect of the Special Public Enforcement Activity on Tobin’s Q: Tobin's Qit = β1SOLVED_SELFi × AFTER + β2UNSOLVED_SELFi × AFTER + + β3 SOLVED_CSRCi × AFTER + β4UNSOLVED_CSRCi × AFTER + β5CONTROLSit + β6CONTROLSit × AFTER + µt + µi + εit (2) where i and t are firm and year indicators, respectively. See the Appendix for all variable definitions. The control variables included in CONTROLS follow Gompers et al. (2003) and Bebchuk and Cohen (2005). Since Chinese listed firms don’t disclose R&D expenditure, we use INTANGIBLE as a rough substitute. Because stock prices are forward looking and reflect new information quickly, we estimate regression model (2) using only the listed firms’ Tobin’s Q at the calendar year end of 2006, which predates the CSRC’s announcement of the Special Public Enforcement Activity, and the listed firms’ Tobin’s Q at the calendar year end of 2008, by which time all listed firms’ follow-up reports should have been made public. In addition to using only two years (2006 and 2008), another major difference of model (2) relative to model (1) is that model (2) includes UNSOLVED_SELF and UNSOLVED_CSRC. We exclude these two terms from model (1) because we don’t expect unresolved corporate governance problems to have a significant impact on tunneling, earnings quality, and operating accounting performance.4 However, it is necessary to include them in model (2) because investors may not be fully aware of the scope of the listed firms’ 4 Inferences for AFTER×SOLVED_SELF and AFTER×SOLVED_CSRC are similar if UNSOLVED_SELF and UNSOLVED_CSRC and their interactions with AFTER are included in model (1). The coefficients on AFTER×UNSOLVED_SELF and AFTER×UNSOLVED_CSRC are never significant. 16 corporate governance problems prior to the Special Public Enforcement Activity and therefore the public disclosure of UNSOLVED_SELF and UNSOLVED_CSRC could provide new information to investors. The coefficients on AFTER×UNSOLVED_SELF and AFTER×UNSOLVED_CSRC will be negative if investors were not fully aware of the corporate governance problems prior to the CSRC’s announcement of the Special Public Enforcement Activity. To the extent that correcting identified corporate governance problems results in a genuine increase (decrease) in corporate governance, we expect the coefficients on AFTER×SOLVED_SELF and AFTER×SOLVED_CSRC to be positive (negative). The reason is that corporate governance failures are frequently discussed topic among Chinese investors and hence we expect the stock prices prior to the CSRC’s announcement of the Special Public Enforcement Activity to partially reflect the negative consequences of the corporate governance problems. Therefore, if the Special Public Enforcement Activity is value increasing (value decreasing) on a net basis, the forced correction of such problems should result in an increase (decrease) in Tobin’s Q. 4.3. The effect of public enforcement on tunneling We first examine whether the Special Public Enforcement Activity helps reduce the controlling shareholder’s tunneling via inter-corporate loans. Following Jiang et al. (2010), we use OREC as a proxy for inter-corporate loans by the listed firm to the controlling shareholder. Jiang et al. (2010) show that OREC is a reliable measure of inter-corporate loans. The selection of CONTROLS follows Jiang et al. (2010). Panel A of Table 3 shows the descriptive statistics of the regression variables. Both the median SOLVED_SELF and median SOLVED_CSRC are 4 but the values of both variables vary significantly across our sample firms, suggesting that the impact of the Special Public Enforcement Activity varies across our sample firms. 17 Panel B of Table 3 shows the regression results of OREC. The coefficient on AFTER×SOLVED_SELF is significantly negative but the coefficient on AFTER×SOLVED_CSRC is insignificant. These results suggest that correcting self-reported corporate governance problems results in a significant decline in inter-corporate loans, but there is no evidence that correcting CSRC-identified corporate governance problems results in any positive or negative effect. Therefore, the overall evidence is mixed on the efficacy of the Special Public Enforcement Activity on tunneling by controlling shareholders. If there is anything, it appears that requiring listed firms to self-report corporate governance noncompliance problems works better than asking the CSRC to identify listed firms’ corporate governance noncompliance problems in curbing tunneling by controlling shareholders. The coefficients on CONTROLS are consistent with the findings in Jiang et al. (2010). The only exceptions are the coefficients on SIZE and MARKETIZATION. However, this difference in results is due to the inclusion of firm fixed effects. We obtain similar inferences as Jiang et al. (2010) for SIZE and MARKETIZATION when we replace firm fixed effects with industry fixed effects. 4.4. The effect of public enforcement on earnings quality We next examine whether the Special Public Enforcement Activity improves the listed firms’ earnings quality (ABS_AQ). Following Kothari et al (2005), we use the absolute performance-matched cross-sectional modified Jones (1991) model abnormal accruals to proxy for earnings quality. Specifically, we first obtain the abnormal accruals from the following modified Jones (1991) model estimated cross-sectionally each year using all firms in the same industry: TAit = β0 + β1(1/ASSETSit - 1) + β2(∆SALESit - ∆ARit) + β3PPEit + εit (3) 18 where TA is total accruals scaled by lagged total assets (ASSETS), ∆SALES is change in sales scaled by lagged total assets, ∆AR is change in accounts receivable scaled by lagged total assets, and PPE is net property, plant and equipment scaled by lagged total assets. Total accruals are directly computed as the difference between net income and operating cash flows. The abnormal accruals for firm i in year t are the residual from the above model. The absolute performance-matched abnormal accruals for firm i in year t (ABS_AQ) are the absolute difference in abnormal accruals between firm year it and a matched firm with the closest ROA (defined as net income divided by total assets) in the same industry in year t. For the purpose of performance matching, we require the absolute difference in ROA between firm i and its matching firm to be no greater than 20% of firm i’s ROA. Higher values of ABS_AQ mean lower earnings quality. The control variables (CONTROLS) included in the regression of ABS_AQ follow Francis and Yu (2009). However, we exclude two variables used in Francis and Yu because they are not appropriate in our setting. First, we exclude BANKRUPTCY, a summary measure of financial distress based on the Altman bankruptcy model because the risk of bankruptcy is almost close to zero for listed Chinese firms. Second, we exclude WEAKNESS, the number of SOX-mandated material internal control weaknesses because such information is not available in China. More importantly, internal control quality is part of the Special Public Enforcement Activity and therefore should not be included as a control variable. Table 4 shows the descriptive statistics and regression results for the dependent variable ABS_AQ. The coefficient on AFTER×SOLVED_SELF is insignificant while the coefficient on AFTER×SOLVED_CSRC is significantly positive (two-tailed p=0.085). Hence, there is no evidence that correcting either self-reported or CSRC-identified corporate governance problems leads to higher earnings quality. To the contrary, the coefficient on 19 AFTER×SOLVED_CSRC suggests that correcting CSRC-identified corporate governance problems is associated with reduced earnings quality. 4.5. The effect of public enforcement on operating accounting performance Because the Special Public Enforcement Activity could affect the listed firms through multiple channels, many of which are unobservable, we next examine the effect of the Special Public Enforcement Activity on the listed firms’ operating accounting performance (OROA), a proxy for the net effect on shareholder value. Inferences are similar if we use ROA (defined as net income divided by total assets) instead. The list of control variables (CONTROLS) follows Fan et al. (2007) and Core et al. (1999). Table 5 shows the descriptive statistics and regression results for the dependent variable OROA. The coefficients on AFTER×SOLVED_SELF and AFTER×SOLVED_CSRC are both insignificant at the ten-percent two-tailed significance level. Hence, there is no evidence that correcting either self-reported or CSRC-identified corporate governance problems leads to higher operating accounting performance. 4.6. The effect of public enforcement on Tobin’s Q Table 6 shows the descriptive statistics and regression results for the dependent variable Tobin’s Q. As shown in Panel A, the mean values of UNSOLVED_SELF and UNSOLVED_CSRC are less than 0.50. Hence, the distributions of these two variables may lack power in detecting any effects even if the effects exist. The coefficients on AFTER×UNSOLVED_SELF and AFTER×UNSOLVED_CSRC are both negative but insignificant. Hence, there is no evidence that the Special Public Enforcement Activity provides incremental new information to stock investors. This finding is not too surprising because corporate governance problems are a very common topic of discussion among 20 investors. In addition, the coefficients on AFTER×SOLVED_SELF and AFTER×SOLVED_CSRC are both insignificant. This evidence is consistent with the OROA result, suggesting that correcting the identified corporate governance problems does not result in any (positive or negative) change in shareholder value. 5. A detailed look at the disclosed corporate governance problems The results presented so far suggest no evidence that the Special Public Enforcement Activity as a whole results in a significant increase in shareholder value. Since the Special Public Enforcement Activity covers eight different types of corporate governance noncompliance problems (see Section 2), it is possible that improvement in certain aspects of the listed firms’ corporate governance may result in a net increase in shareholder value. To test this idea, we decompose each of our key variables of interest in models (1) and (2) into the eight specific types discussed in Section 2. Table 7 shows the regression results. For brevity, we only report the interaction coefficients between AFTER and the eight specific types of corporate governance problems. A general conclusion from Table 7 is that we find little consistent evidence that correcting a particular type of corporate governance problem results in a significant positive or negative effect on net shareholder value. With regard to OREC, the coefficients on AFTER×SOLVED_SELF1 and AFTER×SOLVED_SELF5 are significantly negative but the coefficients on AFTER×SOLVED_SELFi (i=2, 3, 4, 6, 7, and 8) are insignificant. These results suggest that the negative coefficient on AFTER×SOLVED_SELF in Table 3 is largely attributed to correcting self-reported corporate governance problems related to the controlling shareholder and the listed firm’s internal control. None of the coefficients on AFTER×SOLVED_CSRCi (i=1 to 8) are significantly different from zero. Hence, there is no evidence that the CSRC has 21 the ability to identify any individual categories of corporate governance weaknesses that are critical to shareholder value. With regard to ABS_AQ, none of the interaction coefficients are significantly negative, suggesting no evidence that correcting any individual categories of corporate governance problems improves shareholder value. The significantly positive coefficient on AFTER×SOLVED_CSRC in Table 4 is driven by AFTER×SOLVED_CSRC2 (shareholders’ meeting), AFTER×SOLVED_CSRC4 (management’s responsibilities), and AFTER×SOLVED_CSRC8 (miscellaneous governance problems). Because we use both OROA and Tobin’s Q to proxy for net shareholder value, we examine the regression results of both OROA and Tobin’s Q to draw inferences. We find no consistent evidence across the two regressions that correcting any individual categories of corporate governance noncompliance problems improves shareholder value. Specifically, the coefficient on AFTER×SOLVED_SELF8 (the miscellaneous category) is significantly positive for the OROA regression but insignificant for the Tobin’s Q regression. Likewise, the coefficient on AFTER×SOLVED_CSRC2 (shareholders’ meeting) is not significant for the OROA regression but significantly positive for the Tobin’s Q regression. We also find no consistent evidence suggesting that correcting any individual categories of corporate governance problems reduces shareholder value. Specifically, the coefficients on AFTER×SOLVED_SELF6 (management’s compensation and accountability), AFTER×SOLVED_CSRC7 (corporate disclosure), and AFTER×SOLVED_SELF8 (the miscellaneous category) are significant in either the OROA regression or the Tobin’s Q regression but not both. Interestingly, the coefficients on AFTER×UNSOLVED_SELFi (i=6 for management’s compensation and accountability, 7 for corporate disclosure, and 8 for the miscellaneous category) are significantly negative in the Tobin’s Q regression. Taken at face value, these 22 coefficients suggest that the listed firms’ self-reported corporate governance problems related to these three categories are new to the stock market and thus result in a significant reduction in Tobin’s Q after the revelation of these uncorrected problems. However, we caution the reader to interpret these coefficients with caution because the majority values of UNSOLVED_SELF are zero. 6. Conclusion A widely held view in the finance literature is that investor protection in general and law enforcement in particular are vital for corporate financing, financial market development and economic growth. However, the extant literature disagrees on the optimal approaches to law enforcement. La Porta et al. (2006) concludes public enforcement does not work and advocate the adoption of private enforcement while Jackson and Roe (2009) support public enforcement. The objective of this study is to use a natural experiment from China to directly test whether public enforcement matters in improving investor protection for firms domiciled in weak investor protection countries. Specifically, in March 2007 the China Securities Regulatory Commission (CSRC) conducted a one-time Special Public Enforcement Activity on listed firms’ compliance with several important corporate governance regulations issued over the period 2002-2006. We find that the Special Public Enforcement Activity forced our sample firms of 1,094 unique listed firms to disclose more than 10,000 corporate governance noncompliance problems. The most common problems relate to the board of directors and internal control. More importantly, we find that the affected firms claimed to have corrected more than 90% of the disclosed problems by the end of the Special Public Enforcement Activity in late 2008. Based on this evidence, the 2007 Special Public Enforcement Activity is a success. 23 However, we find little evidence that correcting the identified corporate governance problems resulted in a significant increase in shareholder value measured by both operating accounting performance and Tobin’s Q. We also study the effect of the 2007 public enforcement on two important specific outcomes frequently considered in prior literature: controlling shareholders’ tunneling via inter-corporate loans (Jiang et al. 2010) and earnings quality. We find little evidence that the 2007 public enforcement helps reduce tunneling or improve earnings quality. The only exception is that correcting the corporate governance problems voluntarily reported by the firm insiders resulted in a significant reduction in tunneling. Taken as a whole, the results from our study suggests no evidence that public enforcement matters in helping improve investor protection and shareholder value in weak investor protection countries. Our results are significant because we are the first study to directly test the causal effects of public enforcement on shareholder value. However, we also acknowledge that our study is limited to one specific example of public enforcement only and we don’t rule out the possibility that other types of public enforcement could be more effective in protecting investors. We call for more research to better understand the economic consequences of public enforcement on shareholder value. 24 REFERENCES Allen F, Qian J, and Qian M. 2005. Law, finance, and economic growth in China. Journal of Financial Economics 77(1):57-116. Bebchuk LA, Cohen A. 2005. The costs of entrenched boards. Journal of Financial Economics 78(2):409-433. Beck T, Levine R, Loayza N. 2000.Finance and the sources of growth. Journal of Financial Economics 58(1–2):261-300. Bruno V, Claessens S. 2010. Corporate governance and regulation: Can there be too much of a good thing? Journal of Financial Intermediation 19(4):461-482. Chen, G., M. Firth, D. N. Gao, and O.M. Rui. 2005. Is China’s securities agency a toothless tiger? Evidence from enforcement actions. Journal of Accounting and Public Policy 24, 451-488. Chen Z, Yang Z and Ke B. 2013. Minority shareholders’ control rights and the quality of corporate decisions: A natural experiment from China, The Accounting Review, forthcoming. Core JE, Holthausen RW, and Larcker DF. 1999. Corporate governance, chief executive officer compensation, and firm performance. Journal of Financial Economics 51(3):371-406. CSRC (China Securities Regulatory Commission). 2007. Notice regarding special check activity for strengthening the corporate governance in publicly listed firms. March 9, 2007. In Chinese. CSRC. 2008. CSRC Notice NO. [2008]27. June 20, 2008. In Chinese. Fan G, Wang X and Zhu H. 2011. The Report on the Relative Process of Marketization of Regions in China. The Economic Science Press, Beijing (in Chinese) Fan JP, Wong T, and Zhang T. 2007. Politically connected CEOs, corporate governance, and Post-IPO performance of China's newly partially privatized firms. Journal of Financial Economics 84(2):330-357. Francis JR, Yu MD. 2009. Big 4 Office Size and Audit Quality. Accounting Review 84(5):1521-1552. Gompers P, Ishii J, and Metrick A. 2003. Corporate Governance and Equity Prices. The Quarterly Journal of Economics 118,(1,):107-156. Hartmann, P., Heider, F., Papaioannou E., and Duca, M.. 2007. The role of financial markets and innovation in productivity and growth in Europe. European Central Bank Occasional Paper no. 72. Jackson HE, Roe MJ. 2009. Public and private enforcement of securities laws: Resourcebased evidence. Journal of Financial Economics 93(2):207-238. Jiang G, Lee CM, Yue H. 2010. Tunneling through intercorporate loans: The China experience. Journal of Financial Economics 98(1):1-20. Jones, J. 1991. Earnings management during import relief investigations. Journal of Accounting Research (Autumn): 193–228. Karpoff, J.M. D. S. Lee, G. S. Martin. 2008a. The consequences to managers for financial misrepresentation. Journal of Financial Economics 88, 193–215. Karpoff, J.M. D. S. Lee, G. S. Martin. 2008b. The Cost to Firms of Cooking the Books. Journal of Financial and Quantitative Analysis 43 (3), 581–612. Kothari S, Leone AJ, and Wasley CE. 2005. Performance matched discretionary accrual measures. Journal of Accounting and Economics 39(1):163-197. La Porta, R., F. Lopez-De-Dilanes, A. Shleifer, and R. Vishny. 1997. Legal determinants of external finance. Journal of Finance 52(3): 1131-1150. 25 La Porta, R., F. Lopez-De-Dilanes, and A. Shleifer. 2006. What works in securities laws? Journal of Finance 61 (1): 1-32. Layton, Marlon A. 2008. Is Private Securities Litigation Essential for the Development of China's Stock Markets? New York University Law Review, 83(6): 1948-1978. Shleifer A, Wolfenzon D. 2002. Investor protection and equity markets. Journal of Financial Economics 66(1):3-27. World Bank. 2006. Institutional foundations for financial markets. Available at http://siteresources.worldbank.org/INTTOPACCFINSER/Resources/Institutional.pdf. Wurgler J. 2000. Financial markets and the allocation of capital. Journal of Financial Economics 58(1–2):187-214. Zhen W. 2013. The China Securities Regulatory Commission starts a special enforcement activity on the financial reports of firms applying for IPO (in Chinese). Caixin.com, January 9, 2013. http://finance.caixin.com/2013-0109/100481145.html?utm_source=mail.caixin.com&. 26 Appendix. Variable definitions Variable Definition OREC ABS_AQ = = Net other receivables deflated by total assets Absolute performance-matched cross-sectional modified Jones model abnormal accruals (see Section 4.4 for the detail) Total Operating income divided by total assets OROA = Tobin’s Q = The ratio of the market value of assets and the book value of assets, where the market value of assets is computed as book value of assets plus the market value of common stock less the sum of book value of common stock and balance sheetdeferred taxes liabilities SOLVED_SELF = The number of corporate governance noncompliance problems identified and corrected by the firm itself as of the filling date of the follow-up report SOLVED_CSRC = The number of corporate governance noncompliance problems identified and corrected by the CSRC as of the filling date of the follow-up report UNSOLVED_SELF = The number of unsolved corporate governance noncompliance problems identified by the firm itself as of the filling date of the follow-up report UNSOLVED_CSRC = The number of unsolved corporate governance noncompliance problems identified by the CSRC as of the filling date of the follow-up report SOLVED_SELFi = The number of corporate governance noncompliance problems in category i (i=1 to 8) identified and solved by the firm itself as of the filing date of the follow-up report SOLVED_CSRCi = The number of corporate governance noncompliance problems in category i (i=1 to 8) identified and solved by the CSRC as of the filing date of the follow-up report UNSOLVED_SELFi = The number of unsolved corporate governance noncompliance problems in category i (i=1 to 8) identified by the firm itself as of the filing date of the follow-up report UNSOLVED_CSRCi = The number of unsolved corporate governance noncompliance problems in category i (i=1 to 8) identified by the CSRC as of the filing date of the follow-up report AFTER = A dummy variable that equals one for the years 2008-2010, and zero for years 20042006 SIZE LAYER = = Natural log of total assets The number of intermediate layers between a listed firm and its controlling owner through the longest pyramidal chain ROA LARGEHLD = = Net income divided by total assets Percentage of shares held by the largest shareholder MARKETIZATION = A comprehensive index measuring the development of the regional market in which the firm is registered (see Fan et al. 2011), where higher values indicate greater regional market development; SALES_GROWTH STDSALES = = One-year growth rate of a firm’s sales revenue Standard deviation of SALES_GROWTH in the prior three years (t-3, t-1). OCF CFOVOLATILITY = = Operating cash flows deflated by lagged total assets Standard deviation of CFO in the prior three years (t-3, t-1). LEV LOSS VOLATILITY = = = Total liabilities divided by total assets Dummy variable that takes the value of 1 if net income is negative, and 0 otherwise; Standard deviation of 12 monthly stock returns; BM STDOROA = = The ratio of a company's book value of equity to its market value of equity Standard deviation of OROA in the prior three years (t-3, t-1). 27 CAPEX = HUSHEN300 = The ratio of capital expenditures to total assets; Capital expenditures equal Cash Paid to Acquire and Construct Fixed Assets, Intangible Assets and Other Long-term Assets LESS Net Cash Received from Disposals of Fixed Assets, Intangible Assets and Other Long-term Assets An indicator variable for firms in HUSHEN 300 index INTANGIBLE = Intangible asset deflated by total assets AGE = Natural logarithm of company age; Age is the number of years since the firm’s initial public offering. 28 Figure 1. The frequency distribution of the announcement dates of the self-assessment report (Panel A), the remediation report (Panel B), and the follow-up report (Panel C) Panel A. Announcement dates of the self-assessment report Distribution of the announcement dates of the self-assessment report N=429 39.21% 45.00% 40.00% N=36 1 33.00% 35.00% 30.00% N=232 21.21% 25.00% 20.00% 15.00% N=60 5.48% 10.00% 5.00% N= 0.18% N= 3 0.27% N= 1 0.09% Mar-07 Apr-07 May-07 N=5 0.46% N= 1 0.09% Oct-07 Nov-07 0.00% Jun-07 Jul-07 Aug-07 Sep-07 Panel B. Announcement dates of the remediation report Distribution of the announcement dates of the remediation report N=509 46.53% 50.00% 45.00% N=387 35.37% 40.00% 35.00% 30.00% 25.00% 20.00% 15.00% 29 8 Ju l-0 n08 Ju 8 -0 ay M r-0 8 0.09% N=1 Ap 8 ar -0 M 08 bFe n08 Ja -0 7 ec D ov N ct -0 -0 7 7 1.28% 0.27% 0.09% N=1 N=3 N=1 4 O 07 Se p- 07 Au g- Ju n07 0.00% 3.93% N=4 0.27% 0.82% 3 0.09% N=1 N=3 N=9 7 5.00% 9.78% N=107 Ju l-0 10.00% Panel C. Announcement dates of the follow-up report Distribution of the announcement dates of the follow-up report N=1030 94.15% 100.00% 90.00% 80.00% 70.00% 60.00% 50.00% 40.00% 30.00% 20.00% 10.00% N=48 4.39% N=1 0.09% N=2 0.18% N=1 0.09% Sep-08 Oct-08 0.00% May-08 Jun-08 Jul-08 30 Aug-08 Table 1. The frequency distribution of the corporate governance noncompliance problems identified by the firm itself in the self-assessment report and by the CSRC in the remediation report, respectively, across the eight corporate governance categories. Panel A. Noncompliance problems identified by the firm itself Categories IDENTIFY_SELF %IDENTIFY_SELF SOLVED_SELF %SOLVED_SELF %CORRECTION 1. Controlling shareholders 295 5.55% 232 4.78% 78.6% 2. Shareholders' Meeting 210 3.95% 205 4.22% 97.6% 3. Board of Directors 1403 26.37% 1348 27.76% 96.1% 4. Management's Responsibilities 373 7.01% 355 7.31% 95.2% 5. Internal Control 1641 30.85% 1544 31.80% 94.1% 6. Executive Compensation and Accountability 284 5.34% 103 2.12% 36.3% 7. Corporate Disclosure 1014 19.06% 996 20.51% 98.2% 8. Other Miscellaneous Governance Issues 100 1.88% 73 1.50% 73.0% Total 5320 100.00% 4856 100.00% 91.3% Panel B. Noncompliance problems identified by the CSRC Categories IDENTIFY_CSRC %IDENTIFY_CSRC SOLVED_CSRC %SOLVED_CSRC %CORRECTION 1. Controlling shareholders 574 10.63% 475 9.35% 82.8% 2. Shareholders' Meeting 593 10.98% 590 11.62% 99.5% 3. Board of Directors 1684 31.17% 1636 32.22% 97.1% 4. Management's Responsibilities 322 5.96% 297 5.85% 92.2% 5. Internal Control 1581 29.27% 1477 29.09% 93.4% 6. Executive Compensation and Accountability 79 1.46% 54 1.06% 68.4% 7. Corporate Disclosure 526 9.74% 517 10.18% 98.3% 8. Other Miscellaneous Governance Issues 43 0.80% 32 0.63% 74.4% Total 5402 100.00% 5078 100.00% 94.0% 31 IDENTIFY_SELF is the number of corporate governance noncompliance problems identified by the firm itself. %IDENTIFY_SELF is IDENTIFY_SELF for a particular category divided by the total number of noncompliance problems identified by the firm itself. SOLVED_SELF is the number of noncompliance problems identified and solved by the firm itself as of the filling date of the follow-up report. % SOLVED_SELF is SOLVED_SELF for a particular category divided by the total number of noncompliance problems identified and solved by the firm itself as of the filling date of the follow-up report. IDENTIFY_CSRC is the number of corporate governance noncompliance problems identified by the CSRC. %IDENTIFY_CSRC is IDENTIFY_CSRC for a particular category divided by the total number of noncompliance problems identified by the CSRC. SOLVED_CSRC is the number of corporate governance noncompliance problems identified and solved by the CSRC as of the filling date of the follow-up report. % SOLVED_CSRC is SOLVED_CSRC for a particular category divided by the total number of noncompliance problems identified and solved by the CSRC as of the filling date of the follow-up report. %CORRECTION is SOLVED_SELF divided by IDENTIFY_SELF or SOLVED_CSRC divided by IDENTIFY_CSRC. 32 Table 2. The distribution of the correction rates, at the individual firm level, by the time of the self-assessment report, the remediation report, and the follow-up report, respectively, for the corporate governance noncompliance problems identified by the firm itself in the self-assessment report and by the CSRC in the remediation report Panel A. Correction rates for noncompliance problems identified by the firm itself By the time of Self-assessment report Remediation report Follow-up report N 1090 1090 1090 mean 39.93% 78.87% 91.04% Median 40.00% 83.33% 100.00% Std. Dev. 29.05% 25.75% 17.08% Min 0.00% 0.00% 0.00% 25% 16.67% 66.67% 83.33% 75% 62.50% 100.00% 100.00% Max 100.00% 100.00% 100.00% Min 0.00% 0.00% 25% 60.00% 100.00% 75% 100.00% 100.00% Max 100.00% 100.00% Panel B. Correction rates for noncompliance problems identified by the CSRC By the time of Remediation report Follow-up report Observations 1014 1014 Mean 74.76% 93.51% Median 83.33% 100.00% Std. Dev. 29.81% 16.93% The number of unique firms in Table 2 is smaller than 1,094 because a few firms have either no self-reported problems or CSRC-identified problems. 33 Table 3. Regression results of OREC Panel A. Descriptive statistics of regression variables Descriptive statistics Variable OREC SIZE LAYER ROA LARGEHLD MARKETIZATION N 6254 6254 6254 6254 6254 6254 Mean Median Std. Dev. 0.049 0.017 0.091 21.407 21.332 1.158 3.484 3.000 0.911 0.019 0.026 0.084 0.390 0.369 0.162 8.036 7.970 2.082 Min 0.000 14.158 2.000 -0.413 0.036 0.380 25% 0.006 20.698 3.000 0.008 0.262 6.440 75% 0.048 22.058 4.000 0.053 0.516 9.770 Max 0.567 27.488 10.000 0.202 0.852 11.800 Min 0 0 25% 3 2 75% 6 6 Max 18 28 Descriptive statistics at the individual firm level Variable SOLVED_SELF SOLVED_CSRC N 1051 1051 Mean 4.3892 4.5585 Median Std. Dev. 4 2.009 4 3.569 Panel B. Regression results of OREC Coefficient -0.004*** -0.001 -0.000 0.018*** -0.002 -0.004 -0.284*** 0.237*** -0.074*** -0.009 0.009*** 0.001 AFTER×SOLVED_SELF AFTER×SOLVED_CSRC SIZE AFTER×SIZE LAYER AFTER×LAYER ROA AFTER×ROA LARGEHLD AFTER×LARGEHLD MARKETIZATION AFTER×MARKETIZATION Year fixed effects Firm fixed effects N Adj. R2 Two-tailed p value 0.002 0.355 0.988 0.000 0.407 0.143 0.000 0.000 0.001 0.573 0.006 0.292 YES YES 6254 0.244 See the Appendix for variable definitions. ***, **, * denote statistical significance at the 1%, 5%, 10% levels (two-tailed), respectively. Two-tailed robust p values are clustered at the firm level. To reduce the influence of outliers, we winsorize all the continuous ratio variables at the top and bottom one percentiles. 34 Table 4. Regression results of earnings quality Panel A. Descriptive statistics of regression variables Descriptive statistics Variable ABS_AQ SIZE SALES_GROWTH STDSALES OCF CFOVOLATILITY LEV LOSS VOLATILITY BM N Mean Median Std. Dev. 5626 0.098 0.068 0.111 5626 21.478 21.405 1.101 5626 0.253 0.131 1.400 5626 0.376 0.189 0.812 5626 0.065 0.057 0.093 5626 0.088 0.074 0.064 5626 0.525 0.518 0.230 5626 0.139 0.000 0.346 5626 0.145 0.133 0.066 5626 0.472 0.413 0.316 Min 0.001 16.962 -1.046 0.017 -0.257 0.006 0.073 0.000 0.038 -0.240 25% 0.030 20.767 -0.016 0.106 0.014 0.044 0.383 0.000 0.094 0.235 75% 0.128 22.096 0.305 0.344 0.110 0.115 0.644 0.000 0.185 0.657 Max 0.913 27.488 45.518 8.127 0.389 0.408 1.835 1.000 0.477 1.342 Min 0 0 25% 3 2 75% 6 6 Max 18 28 Descriptive statistics at the individual firm level Variable SOLVED_SELF SOLVED_CSRC N 1079 1079 Mean 4.437 4.623 Median Std. Dev. 4 2.040 4 3.630 Panel B. Regression results of ABS_AQ Coefficient 0.000 0.001* -0.027*** -0.002 0.001 0.006 -0.002 0.006 0.033 -0.030 -0.052 -0.006 0.017 0.006 -0.005 -0.005 -0.076 -0.016 -0.003 -0.007 AFTER×SOLVED_SELF AFTER×SOLVED_CSRC SIZE AFTER×SIZE SALES_GROWTH AFTER×SALES_GROWTH STDSALES AFTER×STDSALES OCF AFTER×OCF CFOVOLATILITY AFTER×CFOVOLATILITY LEV AFTER×LEV LOSS AFTER×LOSS VOLATILITY AFTER×VOLATILITY BM AFTER×BM Year fixed effects Two-tailed p value 0.874 0.085 0.002 0.652 0.872 0.302 0.553 0.341 0.336 0.567 0.449 0.952 0.374 0.760 0.362 0.554 0.255 0.838 0.767 0.640 YES 35 Firm fixed effects N Adj. R2 YES 5626 0.043 See the Appendix for variable definitions. ***, **, * denote statistical significance at the 1%, 5%, 10% levels (two-tailed), respectively. Two-tailed robust p values are clustered at the firm level. To reduce the influence of outliers, we winsorize all the continuous ratio variables at the top and bottom one percentiles. 36 Table 5. Regression results of operating accounting performance Panel A. Descriptive statistics of regression variables Descriptive statistics Variable OROA SIZE BM LEV LARGEHLD STDOROA N 6199 6199 6199 6199 6199 6199 Mean 0.022 21.440 0.470 0.531 0.388 0.065 Median Std. Dev. 0.027 0.089 21.363 1.131 0.415 0.318 0.519 0.245 0.367 0.162 0.047 0.072 Min -0.413 11.348 -0.240 0.073 0.036 0.005 25% 0.004 20.734 0.235 0.382 0.261 0.027 75% 0.059 22.074 0.655 0.647 0.514 0.079 Max 0.303 27.488 1.342 1.835 0.852 0.625 Min 0 0 25% 3 2 75% 6 6 Max 18 28 Descriptive statistics at the individual firm level Variable SOLVED_SELF SOLVED_CSRC N 1094 1094 Mean 4.439 4.642 Median Std. Dev. 4 2.038 4 3.663 Panel B. Regression results of OROA AFTER×SOLVED_SELF AFTER×SOLVED_CSRC SIZE AFTER×SIZE BM AFTER×BM LEV AFTER×LEV LARGEHLD AFTER×LARGEHLD STDOROA AFTER×STDOROA Year fixed effects Firm fixed effects N Adj. R2 Coefficient 0.001 -0.000 0.004 -0.008*** -0.052*** -0.018 -0.088*** 0.003 0.081*** 0.019 0.081 0.029 Two-tailed p value 0.161 0.503 0.557 0.006 0.000 0.124 0.000 0.838 0.000 0.249 0.147 0.697 YES YES 6199 0.074 See the Appendix for variable definitions. ***, **, * denote statistical significance at the 1%, 5%, 10% levels (two-tailed), respectively. Two-tailed robust p values are clustered at the firm level. To reduce the influence of outliers, we winsorize all the continuous ratio variables at the top and bottom one percentiles. 37 Table 6. Regression results of Tobin’s Q Panel A. Descriptive statistics of regression variables Descriptive statistics Variable Tobin’s Q CAPEX ROA SIZE LEV HUSHEN300 INTANGIBLE AGE N 2092 2092 2092 2092 2092 2092 2092 2092 Mean 1.734 0.053 0.019 21.508 0.542 0.224 0.048 2.174 Median Std. Dev. 1.399 1.110 0.036 0.059 0.024 0.076 21.448 1.151 0.538 0.241 0.000 0.417 0.028 0.060 2.197 0.388 Min 0.878 -0.049 -0.360 16.704 0.079 0.000 0.000 1.099 25% 1.143 0.012 0.006 20.777 0.398 0.000 0.009 1.946 75% 1.899 0.079 0.050 22.155 0.653 0.000 0.061 2.485 Max 10.136 0.279 0.194 27.346 1.818 1.000 0.288 2.890 25% 3 2 0 0 75% 6 6 1 0 Max 18 28 5 8 Descriptive statistics at the individual firm level Variable SOLVED_ZC SOLVED_CSRC UNSOLVED_ZC UNSOLVED_CSRC N 1046 1046 1046 1046 Mean 4.425 4.605 0.402 0.286 Median Std. Dev. 4 2.030 4 3.589 0 0.777 0 0.743 Min 0 0 0 0 Panel B. Regression results of Tobin’s Q Coefficient 0.007 -0.045 0.003 -0.048 0.467 1.116* 1.549 -1.353 -0.897*** 0.016 0.568 0.595** -0.126 0.285 1.389 0.395 AFTER×SOLVED_SELF AFTER×UNSOLVED_SELF AFTER×SOLVED_CSRC AFTER×UNSOLVED_CSRC CAPEX AFTER×CAPEX ROA AFTER×ROA SIZE_RAW AFTER×SIZE_RAW LEV AFTER×LEV INTAN AFTER×INTAN AGE AFTER×AGE Year fixed effects Firm fixed effects N Adj. R2 38 Two-tailed p value 0.583 0.265 0.611 0.139 0.464 0.094 0.118 0.195 0.000 0.710 0.172 0.017 0.887 0.577 0.152 0.187 YES YES 2092 0.300 See the Appendix for variable definitions. ***, **, * denote statistical significance at the 1%, 5%, 10% levels (two-tailed), respectively. Two-tailed robust p values are clustered at the firm level. To reduce the influence of outliers, we winsorize all the continuous ratio variables at the top and bottom one percentiles. 39 Table 7. The effects of correcting the eight individual types of corporate governance noncompliance problems AFTER×SOLVED_SELF1 AFTER×SOLVED_SELF2 AFTER×SOLVED_SELF3 AFTER×SOLVED_SELF4 AFTER×SOLVED_SELF5 AFTER×SOLVED_SELF6 AFTER×SOLVED_SELF7 AFTER×SOLVED_SELF8 AFTER×SOLVED_CSRC1 AFTER×SOLVED_CSRC2 AFTER×SOLVED_CSRC3 AFTER×SOLVED_CSRC4 AFTER×SOLVED_CSRC5 AFTER×SOLVED_CSRC6 AFTER×SOLVED_CSRC7 AFTER×SOLVED_CSRC8 (1) DV=OREC -0.010* (0.060) 0.004 (0.402) -0.004 (0.197) -0.005 (0.376) -0.009*** (0.001) -0.010 (0.237) 0.000 (0.909) 0.009 (0.517) -0.005 (0.154) 0.002 (0.634) -0.001 (0.606) -0.008 (0.150) 0.002 (0.319) -0.008 (0.459) -0.001 (0.755) -0.003 (0.890) AFTER×UNSOLVED_SELF1 AFTER×UNSOLVED_SELF2 AFTER×UNSOLVED_SELF3 AFTER×UNSOLVED_SELF4 AFTER×UNSOLVED_SELF5 40 (2) DV=ABS_AQ -0.005 (0.409) 0.003 (0.718) -0.002 (0.524) 0.007 (0.267) 0.004 (0.194) -0.000 (0.990) -0.002 (0.671) 0.017 (0.184) -0.002 (0.701) 0.007* (0.084) -0.002 (0.405) 0.018*** (0.005) -0.001 (0.539) -0.004 (0.739) 0.003 (0.424) 0.068*** (0.010) (3) DV=OROA 0.005 (0.215) 0.001 (0.883) 0.003 (0.206) -0.002 (0.680) 0.000 (0.996) 0.012 (0.103) 0.000 (0.878) 0.015** (0.026) 0.001 (0.604) -0.001 (0.772) -0.001 (0.442) 0.004 (0.321) 0.001 (0.349) -0.007 (0.245) -0.004 (0.280) -0.036** (0.041) (4) DV=Tobin’s Q 0.018 (0.719) 0.023 (0.735) 0.010 (0.703) -0.019 (0.716) 0.014 (0.595) -0.157** (0.044) -0.011 (0.715) 0.039 (0.627) 0.002 (0.931) 0.062* (0.084) -0.015 (0.465) -0.002 (0.963) 0.014 (0.427) 0.036 (0.721) -0.066** (0.048) 0.078 (0.631) -0.104 (0.405) 0.567 (0.155) 0.129 (0.350) -0.156 (0.490) 0.058 AFTER×UNSOLVED_SELF6 AFTER×UNSOLVED_SELF7 AFTER×UNSOLVED_SELF8 AFTER×UNSOLVED_CSRC1 AFTER×UNSOLVED_CSRC2 AFTER×UNSOLVED_CSRC3 AFTER×UNSOLVED_CSRC4 AFTER×UNSOLVED_CSRC5 AFTER×UNSOLVED_CSRC6 AFTER×UNSOLVED_CSRC7 AFTER×UNSOLVED_CSRC8 Relevant control variables Year fixed effects Firm fixed effects N Adj. R2 YES YES YES YES YES 6254 0.249 YES YES 5626 0.046 YES YES 6199 0.078 (0.425) -0.165*** (0.008) -0.228* (0.067) -0.474** (0.034) 0.012 (0.811) 0.234 (0.558) -0.017 (0.881) -0.134 (0.635) -0.107 (0.304) -0.083 (0.574) -0.155 (0.357) -0.220 (0.180) YES YES YES 2092 0.309 See the Appendix for variable definitions. ***, **, * denote statistical significance at the 1%, 5%, 10% levels (two-tailed), respectively. Two-tailed p values clustered at the firm level are reported in parentheses. To reduce the influence of outliers, we winsorize all the continuous ratio variables at the top and bottom one percentiles. 41