Related Party Transactions

Related Party Transactions By Haihao (Ross) Lu A thesis submitted in conformity with the requirements for the degree of Doctor of Philosophy Rotman School of Management University of Toronto Ó Copyright by Haihao Lu 2017

Related Party Transactions Haihao (Ross) Lu Doctor of Philosophy Rotman School of Management University of Toronto 2017 Abstract Related party transactions (hereafter RPTs) are common business practices. If misused, however, they can result in significant losses for investors. This study attempts to examine empirical issues pertaining to RPTs. Specifically, the paper has two goals. The main goal is related to a 2006 Securities and Exchange Commission (SEC) regulation mandating public firms to disclose their governance policies on RPTs. I intend to investigate whether firms change their RPT behaviors in response to this mandated RPT governance disclosure and whether investors update the implied cost of capital on these RPT behaviors accordingly. Employing hand-collected RPT data for S&P 1500 firms, I find the initiation of RPT governance disclosure significantly reduces the occurrence of RPTs, suggesting an ex-post improvement in RPT governance. This reduction is more pronounced on firms with higher agency costs. I also find the types of RPTs and the identities of related parties play a role in the detected disclosure effect, suggesting certain types of RPTS could be more likely to be used as expropriation. Furthermore, I document that the disclosure of RPT governance reduces the implied costs of capital premium (ICC) associated with RPTs. Taken together, this analysis provides initial evidence about the effects of the SEC s 2006 regulation on RPT governance disclosures, and shows how the disclosures on RPT governance can enhance firms ii

governance on RPT activities and whether investors could recognize the firm behavior change by reducing firms implied cost of capital. The second goal is to assess whether independent directors compensation is associated with the occurrence of RPTs. Studying independent directors attracts my interest because they are responsible for reviewing and approving RPTs in most companies. By decomposing independent directors compensation into the market-level component and the over-compensated component, I show that firms with over-compensated directors or with a lower portion of equity-based compensation in contrast to cash-based compensation incur more RPTs. These results suggest that independent directors overcompensation reduce directors independence and equity-based compensation aligns directors interests with those of shareholders, consistent with the private benefit theory. This study provides novel evidence that appropriate directors compensation design can better control RPT behaviors. iii

Acknowledgements First and foremost, I would like to thank my supervisor, Ole-Kristian Hope, for his valuable advices and unconditional support during the term of my Ph.D. study. His passion to research, high standards to work, and excellent professionalism left a profound impact on me that will keep motivating me for the rest of my life. He never hesitates to offer help and always pushes me to strive for the next level. I couldn t achieve what I have done without him. I am also very much indebted to my mentor, Gus De Franco for his generous help in my research. Since the very first year, he has encouraged me to discuss any topics with him, no matter how immature it initially sounds like. I benefit a lot from his patience, kindness and professional judgment. I would also like to thank the other two committee members, Ping Zhang and Baohua Xin, for their invaluable suggestions and encouragements in developing the dissertation. I would especially thank Mark Kohlbeck for donating his time to serve as my external committee member and providing me with constructive and beneficial comments. I also want to thank all Rotman professors who helped me during my Ph.D. study, including Partha Mohanram, Jeff Callen, Gordon Richardson, Hai Lu, Franco Wong, Scott Liao, Dushyant Vyas, Francesco Bova, Alex Edwards, Daehyun Kim, Aida Wahid and Feng Chen. I also enjoyed my time with my fellow Ph.D. students Sasan Saiy, Yu Hou, Leila Peyravan, Wuyang Zhao, Danqi Hu, Barbara Su, Stephanie Cheng, Mahfuz Chy, Muhammad Azim and Mingyue Zhang. iv

Finally, I am very grateful to my beautiful wife and best friend, Stella Peng, and my lovely kids, Dennis Lu and Deanna Lu, who supported me wholeheartedly throughout my journey toward Ph.D. I would also like to acknowledge with great gratitude, the support and love from my parents, sister, and parents-in-law. Their help and love were essential to my achievement. v

Table of Content Abstract... ii Acknowledgement.iv Table of Contents... v Introduction... xi Chapter 1: Economic Consequences of Corporate Governance Disclosure: Evidence from the 2006 SEC Regulation on Related-Party Transactions... 1 1. Introduction... 1 2. Hypotheses Development... 7 2.1 Background RPT regulations in the U.S.... 7 2.2 Literature review and hypotheses development... 8 3. Data and Research Design... 15 3.1 Related-party transaction data... 15 3.2 Regression model... 16 3.3 Control variables... 17 4. Empirical Results... 19 4.1 Descriptive statistics... 19 4.2 Primary Analyses... 20 4.2.1 RPT governance disclosure and occurrence of RPTs... 20 4.2.2 RPT governance disclosure effects conditional on potential agency costs... 22 4.2.3 RPT governance disclosure and the implied cost of capital on RPT... 23 4. 3 Robustness Tests... 24 4.3.1 Alternative measure of RPTs... 24 vi

4.3.2 Regulation effects on Already-Disclosed firms... 25 4.3.3 Two-stage selection model... 26 4.3.4 News coverage effects on RPTs... 28 4.3.5 Measurement of CGQ on regulation effects... 28 4.3.6 Effect of increased disclosure threshold of RPT... 28 4.4 Additional Analyses... 29 4.4.1 Policy characteristics and RPTs... 29 4.4.2 Heterogeneity in RPTs... 30 5. Conclusion... 31 References for Chapter 1... 33 Appendix A: Variable Definitions... 35 Appendix B: Categories of Related-Party Transactions... 37 Table 1 for Chapter 1: Descriptive Statistics... 38 Table 2 for Chapter 1: Pearson Correlation Matrix... 39 Table 3 for Chapter 1: RPTs Control-Procedures Disclosure Effects on Occurrence of RPTs... 40 Table 4 for Chapter 1: Disclosure Effects Conditional on External Monitoring... 41 Table 5 for Chapter 1: ICC Premium on RPTs Conditional on Governance Disclosure... 42 Table 6 for Chapter 1: Information Channel vs. Governance Channel... 44 Table 7 for Chapter 1: Alternative RPT Measurement using Transaction Dollar Amounts... 46 Table 8 for Chapter 1: Alternative Measurement of RPT Governance Disclosure... 48 Table 9 for Chapter 1: Two-Stage Selection Models... 49 Table 10 for Chapter 1: RPT Governance Disclosure Effects on RPTs Conditional on Control- Procedures Characteristics... 51 Table 11 for Chapter 1: Heterogeneity of Related-Party Transactions... 53 vii

Chapter 2: Director Compensation and Related Party Transactions... 56 1. Introduction... 56 2. Hypothesis Development... 60 3. Research Design... 63 3.1 Sample collection... 63 3.2 Research Design... 63 4 Empirical Results... 67 4.1 Descriptive statistics... 67 4.2 Main Analysis... 68 4.3 Additional analysis and robustness tests... 69 4.3.1. Audit committee members vs. Non Audit Committee members... 69 4.3.2. External monitoring and growth potentials... 70 4.3.3. Firm level analysis... 71 4.3.4. Alternative RPT measurement... 71 5. Conclusion... 72 References for Chapter 2... 73 Table 1 for Chapter 2: Descriptive Summary... 75 Table 2 for Chapter 2: Correlation Matrix... 76 Table 3 for Chapter 2: Director Compensation effect on the occurrence of RPTs OLS 77 Table 4 for Chapter 2: Determinants of Director Compensation... 78 Table 5 for Chapter 2: Excessive Director Compensation Effects on The occurrence of RPTs... 79 viii

Table 6 for Chapter 2: Effect of Director Compensation on RPT Conditional on Approval Authority... 81 Table 7 for Chapter 2: Moderator of Director Compensation Effect on RPTs... 82 Table 8 for Chapter 2: Director compensation effect on RPTs at firm level... 83 Table 9 for Chapter 2: Alternative RPT measurement... 84 ix

Introduction Related party transactions are common business practices. If properly used, they can be beneficial for firms. However, misused related party transactions can cause significant losses for investors, as in the case of Enron and Adelphia. Even though the transaction amount could be small, misused RPTs could signal serious governance problem. Given that RPTs represent potential means for insiders to expropriate wealth from other investors via self-dealing, both the Financial Accounting Standards Board (FASB) and the Securities and Exchange Commission (SEC) require detailed disclosure of material RPTs in annual reports and proxy statements. In 2006, the SEC amended its rule and regulation for RPT disclosure, which includes an entirely new requirement to disclose a description of a company s policies and procedures for the review, approval, and ratification of RPTs. Chapter 1 shows how a 2006 SEC regulation mandating RPT governance disclosure affects the firm s RPT behavior and the investors risk perception on firm s RPTs. Before 2006, only a few firms voluntarily disclosed their RPT governance procedures. After 2006, the RPT governance disclosure becomes mandatory. I expect this mandatory disclosure of RPT governance reduce the occurrence of RPTs because it enhances the RPT governance. For firms that did not have a formal RPT governance policy before the regulation, this mandatory disclosure requirement would encourage them to establish formal RPT policies to avoid bad signals to investors. For firms that had an RPT policy but did not disclose before the regulation, public RPT governance disclosure would enhance the monitoring effects as the disclosure increases the litigation risk of responsible directors, mitigates the opportunistic behaviors of insiders, and hence reduces agency problems arising from RPTs. Consistent with my expectation, x

I find when firms initiate their RPT governance disclosures following the mandatory requirement, they significantly reduce their RPT behaviors. Chapter 1 also examines whether investors alter their risk perceptions on those firm s RPT behaviors following the 2006 regulation, using implied cost of capital as risk perception proxy. I expect this mandatory disclosure will reduce implied cost of capital premium on firms RPTs as this mandatory disclosure reduces the information asymmetry between investors and firms and brings better controls on RPTs. I do find affirmative results. Chapter 2 investigates how independent directors compensation can affect firm s RPT behaviors. As independent directors are responsible for RPT governance in most U.S. companies, the way in which they are compensated is likely to affect their incentives to effectively govern firms behaviors. On the one hand, their compensation has to be high enough to attract capable candidates; on the other hand, over-compensation may impair the independence of directors, making them more subject to managers influences. By decomposing independent directors compensation into the market level component and the over-compensated component, I find that firms with over-compensated directors or with a lower portion of equity-based compensation in contrast to cash-based compensation incur more RPTs. These results suggest that independent directors overcompensation reduces directors independence and equity-based compensation aligns director interests with those of shareholders. xi

Chapter 1 Economic Consequences of Corporate Governance Disclosure: Evidence from the 2006 SEC Regulation on Related-Party Transactions 1. Introduction A related-party transaction (hereafter RPT) is a transfer of resources, service, or obligations between a reporting entity and a related party, which usually refers to executives, the board of directors, and primary shareholders. 1 RPTs can be beneficial to firms daily operations because appropriate RPTs can reduce transaction costs, improve operating efficiency, and share financial and intangible resources. However, if misused, they can result in significant losses for investors, as in the case of Enron and Adelphia. 2 Even though the transaction amounts could be small, misused RPTs could signal serious governance problems. Given that RPTs represent potential means for insiders to expropriate wealth from other investors via self-dealing, both the Financial Accounting Standards Board (FASB) and the Securities and Exchange Commission (SEC) require detailed disclosure of material RPTs in annual reports and proxy statements. To facilitate investors assessment of the potential conflict of interests arising from RPTs, the SEC amended its rules and regulations for RPT disclosure in 2006. 3 The revised SEC regulation includes an entirely new requirement to disclose a description of a company s policies and procedures for the review, approval, and ratification of RPTs. In this paper, I investigate economic consequences of this mandatory RPT governance disclosure. Specifically, I ask two questions: (1) does the disclosure of firms RPT governance 1 Related parties also include their immediate family members. 2 For example, the top management team of Enron used special purpose entities to manipulate profit. Adelphia Communications Corp guaranteed related party debts and provided extensive loans to its top executives. 3 https://www.sec.gov/rules/final/2006/33-8732a.pdf, retrieved on July 25, 2017 1

lead to a lower level of RPT activities? (2) does the disclosure of firms RPT governance lead to a lower implied cost of capital? Before 2006, only a few firms voluntarily disclosed how they govern their RPTs. 4 The 2006 SEC regulation requires that all firms to disclose their RPT review policy, representing an exogenous increase in RPT governance disclosure. For the convenience of presentation, I define firms that voluntarily disclosed their RPT governance before the 2006 regulation as Already- Disclosed firms and firms that initiated RPT governance disclosure after the 2006 regulation as Newly-Disclosed firms. Since the Already-Disclosed firms have voluntarily disclosed their RPT governance prior to 2006, the impact of 2006 regulation on the Already-Disclosed firms, on average, is expected to be significantly less than that on the Newly-Disclosed firms. As discussed in Section 2, this study predicts that the mandatory disclosure of RPT governance leads to a lower level of RPTs and lower cost of capital, both involving real impacts on firms operating activities. This prediction is consistent with the conflict-of-interest view (Kohlbeck and Mayhew 2010, 2017), which considers RPTs as a potentially harmful form of expropriating wealth from shareholders. For firms that did not have a formal RPT governance policy before the regulation, this mandatory disclosure requirement encourages them to establish a formal policy to avoid being scrutinized by investors and regulators. For firms that had RPT policies but had not disclosed their existing RPT policies before the regulation, public RPT governance disclosure potentially enhances the monitoring as the disclosure increases the litigation risk of responsible directors, mitigates the opportunistic behaviors of insiders, and hence reduces agency problems arising from RPTs. To test these hypotheses, I hand-collect 4 Approximately 20% of S&P 1500 firms in my sample voluntarily disclosed RPT governance in the fiscal year 2004. 2

information regarding RPTs and RPT governance for all S&P 1500 non-financial firms from annual proxy statements for fiscal years 2004, 2007, and 2010. 5 For the first research question, I find that after the 2006 regulation, Newly-Disclosed firms significantly reduce their RPT activities relative to Already-Disclosed firms. This result is robust to controlling for several determinants of RPT occurrence, suggesting that (1) the disclosure of RPT governance enhances firms governance on RPTs and (2) the 2006 SEC regulation affects firms RPT activities. The conclusions also hold for a propensity-score matched sample between Already-Disclosed firms and Newly-Disclosed firms. Next, I examine whether such RPT governance disclosure effects on RPTs are conditional on mechanisms that can reduce agency problems. Previous literature shows that external monitoring helps reduce the opportunistic behaviors of insiders and hence reduce agency costs (Jensen and Meckling 1976; Huddart 1993; Giroud and Mueller 2011). When external monitoring is relatively weak, potential agency costs are high, and thus managers are more likely to be involved in self-dealing. I expect that the RPT governance disclosure effects on RPTs should be more pronounced when the external monitoring is less effective, because these RPT activities are more likely to be harmful ex ante. Consistent with my expectation, I find that the disclosure effects of RPT governance are stronger when there is lower institutional ownership, lower analyst following, and less severe industry competition. In the second research question, I investigate whether this mandatory disclosure alters the implied cost of capital. Prior literature finds that RPT firms are associated with poor corporate governance and lower accounting quality (Kohlbeck and Mayhew 2017; Gordon and Henry 2005). While not all RPTs are misused, there is a view that RPTs represent a high-risk factor for 5 Following Kohlbeck and Mayhew (2017), I do not collect RPTs data in consecutive years because many RPTs are sticky transactions that either have multiple-year terms or appear every year with similar contract terms. Collecting data every three years allows for changes in RPTs and is more efficient. 3

investors to consider before investing (Cheung, Rau, and Stouraitis 2006). There is evidence that insiders may use RPTs to directly expropriate wealth from other investors (Cheung, Qi, Rau and Stouraitis 2009; Kohle and Shastri 2004). High information asymmetry and high expropriation risk suggest that RPTs firms could be associated with high implied costs of capital. I argue that this mandatory disclosure can reduce the implied costs of capital associated with RPTs. Consistent with this expectation, my empirical analyses show that when Newly-Disclosed firms disclose their RPT governance, investors reduce the implied cost of capital associated with RPTs of these firms. The implied cost of capital associated with RPTs of Already-Disclosed firms, on the other hand, does not change significantly from the pre-regulation to the post-regulation periods. Previous literature shows that the effects of disclosures on firms implied cost of capital could be through two possible channels: an information channel or a governance channel (Cheynel 2013; Verrecchia 2007; La Porta et al. 2000). The information channel implies that the disclosure provides more information regarding how firms control their RPTs, reducing the information asymmetry between firms and investors. Through the governance channel, the disclosure improves firms governance on RPTs and reduces harmful RPTs, which leads to fewer self-dealing activities and lower litigation risks. My analysis shows that, between these two channels, the governance channel is the main contributing factor to the reduction of implied cost of capital associated with RPTs. The empirical results show that for firms that only improved their disclosure but that did not change their RPT behaviors in the post period, their implied cost of capital associated with RPTs did not decrease. By contrast, for firms that reduce their RPTs in the post period, their implied cost of capital associated with RPTs significantly reduced. 4

The SEC requires, among other disclosures, a disclosure of (1) whether the policies and procedures are in writing, and (2) the persons or groups of persons who are responsible for the review, approval, and ratification of RPTs. Hence I further test if these particular disclosure requirements affect the frequency of RPTs and the implied cost of capital associated with RPTs. My results show that whether firms have a written policy does not affect the occurrence of RPTs nor the implied cost of capital. The group responsible for approving RPTs, however, is associated with outcomes of RPTs. Specifically, when an Audit Committee is responsible for reviewing and approving RPTs, firms have fewer RPTs and lower implied cost of capital associated with RPTs than firms that do not assign a specific committee to review RPTs. This research contributes to the existing literature in several aspects. My study shows that disclosure of RPT governance effectively increases firms monitoring of RPTs, reducing both RPT activities and the implied cost of capital, suggesting that the disclosure of governance enhances firms real governance on RPTs. Such outcomes add to literature studying real disclosure effects (Leuz and Wysocki 2016), showing that mandatory disclosure requirements have economic consequences on firms operating activities besides financial reporting. I also find that the effects of RPT governance disclosure on RPTs are more pronounced when industry competition is weak, when institution holding is low, and when analyst following is low. These findings provide evidence that the effects of disclosure on governance are contingent on potential agency costs. In addition, I find that the disclosure effects depend on the transaction types and identities of related parties, suggesting that some types of RPTs are more likely to be used as expropriation and that it is advisable to recognize the heterogeneity and treat them differently. 5

This is one of the few papers investigating the economic consequences of the SEC s 2006 regulation related to RPT governance disclosure. This mandatory requirement reflects the continuous efforts of the SEC to encourage firms to provide more transparent disclosure about RPTs. Studying how firms react to this new regulation and whether investors benefit from such mandatory requirements can provide evidence on the effectiveness of the disclosure regulation and should be relevant to future regulations. Balsam, Gifford, and Puthenpurackal (2017) also find a change of RPTs around this 2006 SEC RPT disclosure change. My paper differs from theirs in several ways. First, this article investigates a governance effect of a mandatory disclosure using a difference-in-differences model around the 2006 SEC disclosure change, while their focus is on the association between RPTs and CEO compensation. Recognizing their findings, I also control for CEO compensation in my research design. Second, my sample includes a broader scope of companies. Specifically, my sample consists of S&P 1500 composite firms while Balsam et al. (2017) randomly select 500 firms. Third, my paper examines more detailed RPT categories based on the identity of related parties and transaction types. Such analysis helps to better understand the nature of RPTs and how various types of RPTs can reflect conflict of interests differently. Furthermore, among the several RPT disclosures required by the SEC in 2006, I provide evidence on which particular disclosures affect investors' implied cost of capital. Such findings can provide further guidance to regulators, and firms as well, to adopt the best practices in RPT monitoring. Finally, my research contributes to the limited empirical literature on RPTs in the U.S. setting. Existing RPTs research focuses mainly on developing countries because firms are more likely to have self-dealing problems when legal protection for investors is weak. In contrast, 6

there are few studies in more developed countries such as the U.S. My study shows that even in the U.S., where strong minority shareholder protection is in place, RPTs can still imply risks to investors, adding to the findings of Kohlbeck and Mayhew (2010, 2017) and Ryngaert and Thomas (2012). 2. Hypotheses Development 2.1 Background RPT Regulations in the U.S. RPTs represent potential self-dealing between the company and its executives, directors, major shareholders, and other related parties. Such transactions usually require additional monitoring from shareholders. In the U.S., regulators do not require shareholders approval of RPTs and instead rely on disclosure regulation and ex-post litigation to protect minority shareholders. 6 The FASB and the SEC set the main RPT disclosure requirements. The FASB ASC Topic 850 requires disclosure of material RPTs in the financial statements. The SEC regulation S-K 404(a) requires disclosure of RPTs in both 10-Ks and proxy statements. Required disclosures by SEC include (1) the nature of the relationship, (2) a description of the transaction, (3) the dollar amount of the transaction, and (4) the amount due to or from related parties at the balance sheet date. In August 2006, the SEC amended its regulations for RPT disclosure. The revised SEC regulation includes an entirely new requirement to describe a company s policies and procedures for the review, approval, and ratification of RPTs. The regulation states that the description should include: (a) the types of transactions that are covered by such policies and procedures; (b) the standards to be applied pursuant to such policies and procedures; (c) the persons or groups of 6 Before 2002, loans to executives were prevalent RPTs in public companies. The Sarbanes-Oxley Act of 2002, however, prohibits any public company from providing loans to its executive officers and directors. 7

persons on the board of directors or otherwise who are responsible; and, (d) a statement of whether such policies and procedures are in writing and, if not, how such policies and procedures are evidenced. 2.2 Literature Review and Hypotheses Development Prior literature has established two prevailing theories regarding RPTs: the conflict of interests theory and the efficient transaction theory. The former views RPTs as a potentially harmful form of expropriating wealth from shareholders, while the latter describes RPTs as an efficient economic exchange. Existing RPTs studies are more consistent with the conflict of interests theory. For example, research finds that RPTs are associated with poorer financial reporting quality and lower operating performance (Berkman, Cole, and Fu 2009; Ryngaert and Thomas 2012). Research also shows that RPTs relate to weak corporate governance. Denis and Sarin (1999) and Klein (2002) document that RPTs may undermine non-executive directors functions, turning them into affiliated or gray directors who are not independent anymore, and hence are associated with weaker corporate governance. Similarly, Kohlbeck and Mayhew (2010) and Gordon, Henry, and Palia (2004) provide evidence that weaker corporate governance makes RPTs more likely to occur. 7 If RPTs represent the conflict of interests between investors and insiders, then effective corporate governance should mitigate this conflict. When studying the corporate governance effects on RPTs, prior literature uses traditional governance proxies such as the percentage of independent directors on the board. In my study, I 7 Nekhili and Cherif (2011) and Lo, Wong, and Firth (2010) find similar associations in France and China, respectively. 8

focus on a more directly relevant measure of RPTs governance, namely, the presence of an RPT governance policy. Before the 2006 SEC RPT regulation, approximately one fifth of firms voluntarily disclosed their RPT governance policy. Following the 2006 RPT regulation, all public firms in the U.S. are required by the SEC to disclose their RPT governance policies. I posit that this mandatory disclosure has real effects on firms behavior and reduces the occurrence of RPTs. First, under this new regulation, firms are likely to establish formal RPT governance policies if they did not have one yet, given that the absence of formal RPT governance policies could lead to scrutiny from investors and regulators. Another benefit for such firms to establish a new RPT governance policy is the signaling effects. Firms with newly-created RPT policies could signal to the market of their RPT governance, as compared to those firms that do not have ex-ante RPT policies and choose not to establish one ex-post. Under the mandatory disclosure requirement, these latter firms will have to disclose the fact that they do not have a formal RPT governance policy, making them less desirable in the eyes of investors. When firms do not have formal RPT governance policy, theoretically they could conduct any RPTs as long as these transactions are disclosed. If the related parties are able to selfapprove their own transactions, it is likely that they benefit themselves at the cost of other stakeholders. When a firm has an effective RPT governance policy, the policy potentially aligns management interests with shareholders interests by minimizing harmful RPTs. Hence a more transparent disclosure of governance policy may constrain the occurrence of harmful RPTs and mitigate investor concerns regarding the self-dealing effects. If firms had RPT governance policies but choose not to disclose them before the adoption of the 2006 SEC regulation, this public disclosure requirement can still enhance their RPT 9

control as it decreases the monitoring cost of investors, increases the litigation risk of responsible authorities, and therefore encourages the firm to improve the effectiveness of their RPT control, also leading to a reduction of RPTs. In sum, a formal RPT governance policy, no matter whether it is newly established or already in place, could help ex-ante no-disclosure firms (i.e. Newly-Disclosed firms) to reduce harmful RPTs when the SEC s new mandatory disclosure requirement is in place, since it reduces information asymmetry and aligns more with investors interests. On the other hand, it is also possible that RPT governance policies are just window dressing and do not have any practical impact on the occurrence of RPTs. For example, Enron had set up a formal procedure to examine the fairness of transactions, yet this procedure did not prevent Enron from its RPT wrongdoings. In such case, the disclosure of RPT governance policy would not help firms reduce harmful RPT activities. Taken together, I expect that, on average, the disclosure of RPT governance policy reduces the occurrence of RPTs given the potential conflict of interests associated with RPTs. Therefore, I construct my first hypothesis as follows (all hypotheses are stated in the alternative form): H1: The disclosure of RPT governance policy is associated with a lower level of RPTs. If RPTs represent opportunistic behaviors of insiders, then RPTs are more likely to be harmful when agency problems are more severe. As agency problems are not only constrained by corporate governance but also affected by other factors such as industry competition, ownership structure, and analyst following, I expect that the pressures faced by firms from these 10

external factors can also affect the nature of RPTs. Consequently, I expect the disclosure effects to be conditional on the environment in which firms operate. Prior research shows that ownership structure is a factor affecting firms behavior. With a diffused ownership structure, there is a reduced incentive for any individual owner to monitor corporate management, because the individual owner would bear the entire monitoring cost, yet all shareholders benefit. As a result, only large non-controlling shareholders, such as institutional investors, have sufficient incentives and means to monitor firms and limit agency problems (Shleifer and Vishny 1986; Huddart 1993). Accordingly, firms with low institutional ownership face lower monitoring pressures and are more likely to involve in harmful RPTs ex-ante. Consequently, I expect that the ex-post reduction in RPTs is more pronounced for these firms. H2a: The reduction of RPT occurrence following the initial disclosure of RPT governance is more pronounced when the proportion of institutional ownership is low. Equity analysts provide another monitoring mechanism to align insiders and shareholders interests (Jensen and Meckling 1976). To increase their market influence and reputation through timely and reasonable recommendations, analysts exert significant efforts in examining firms strategy, governance, and activities. With scrutiny and attention from more analysts, insiders face greater risks of negative market reactions or personal reputation damage when involved in self-dealing transactions. Consequently, I expect that the reduction of RPTs following disclosure of RPT governance be more pronounced when analyst following is low. 11

H2b: The reduction of RPT occurrence following the initial disclosure of RPT governance is more pronounced when analyst following is low. Research also shows that industry competition can put direct pressure on firms and reduce agency costs (Giroud and Mueller 2011; Baggs and De Bettignies 2007; Jagannathan and Srinivasan 1999). This is because firms in highly competitive industries are more likely to go bankruptcy than other firms. This high bankruptcy risk gives managers strong incentives to work hard to remain competitive, aligning their interests more with shareholders. With respect to RPTs, these firms are more likely to forgo harmful RPTs and take beneficial RPTs to reduce transaction costs and improve efficiency, mitigating expected bankruptcy costs. As a result, I expect that reduction in RPTs following disclosure of RPT governance be more pronounced when industry competition is weak. H2c: The reduction of RPT occurrence following the initial disclosure of RPT governance is more pronounced when industry competition is weak. The conflict of interests theory suggests that RPTs firms could have higher implied costs of equity capital for at least two reasons. First, RPTs are associated with poor accounting quality, increasing information risks. Jian and Wong (2010) find that Chinese listed firms use abnormal related sales to their controlling owners to prop up earnings. Chen, Cheng, and Xiao (2010) find that controlling shareholders in China structure RPTs in the pre-ipo period to affect IPO performance. Using a sample of 360 companies in the U.S., Gordon and Henry (2005) find that some types of RPTs are associated with higher abnormal accruals. Kohlbeck and Mayhew (2017) 12

show that RPT firms are more likely to restate their financial reports. Similarly, Cullinan, Du, and Wright (2006) document a significant association between executive loans and financial misstatements. The literature shows that information risks are associated with higher cost of equity both theoretically (Easley and O Hara 2004; Lambert et al. 2007) and empirically (e.g., Francis et al. 2005). Second, many RPTs are associated with poor corporate governance (Denis and Sarin 1999; Klein 2002; Gordon, Henry and Palia 2004). There is also evidence that insiders can use RPTs to directly expropriate wealth from other investors. For example, Cheung, Qi, Rau, and Stouraitis (2009) examine 254 related-party acquisitions and sales of assets in Hong Kong and find that firms pay higher prices when purchasing from related parties and receive lower prices when selling to related parties. Kahle and Shastri (2004) show that the loans made to executives are usually issued at below-market interest rates. Consistent with these studies, the stock market reacts negatively to RPTs (Cheung, Rau, and Stouraitis 2006; Ryngaert and Thomas 2012; Kohlbeck and Mayhew 2010). This concern of expropriation could also increase the investigation risk from SEC and litigation risk through conflict of interests lawsuits. 8,9 Disclosure of RPT governance policy could mitigate the relation between RPTs and the implied cost of capital. When firms withhold information, rational investors tend to assume the worst scenario and increase the risk premium (Grossman 1981). Cheynel s (2013) model predicates that firms that voluntarily disclose their information have lower cost of equity capital than firms that do not disclose. Lambert, Leuz, and Verrecchia s (2007) model demonstrates that 8 For example, in October of 2015, the SEC charged Home Loan Servicing Solutions Ltd. (HLSS) with making material misstatements about its handling of related party transactions and for having inadequate internal accounting controls. HLSS agreed to pay a $1.5 million penalty to settle the SEC s charges. 9 For example, in June 2016, Tesla Motors Inc. proposed a $2.6 billion merger with a related party, SolarCity Corp. Two individuals and two institutional shareholders of Tesla filled four lawsuits alleging board members breached their fiduciary duty. 13

disclosure can affect the cost of capital because higher quality disclosures affect the firm s assessed covariance with other firms cash flows, which is non-diversifiable. In addition, they show that disclosure can affect a firm s real decisions, which likely changes the firm s ratio of the expected future cash flows to the covariance of these cash flows with the sum of all the cash flows in the market. Numerous empirical studies support such predictions (Botosan 1997; Lang, Lins, and Maffett 2012). When firms disclose their RPT governance policies, I expect that this behavior affects the implied cost of capital through two channels: an information channel and a governance channel. Through the information channel, a disclosure of RPT policy can provide investors with more information regarding how firms control their RPTs, reducing the information asymmetry premium between firms and investors. In such a case, investors are able to assess the consequence of RPTs with less uncertainty. Through the governance channel, a disclosure of RPT governance could improve firms governance and reduce harmful RPTs, mitigating the conflict of interests between insiders and outsiders. Less risk of both expropriation and consequent litigation could also reduce the cost of capital associated with RPTs. Therefore, my third hypothesis is: H3: The initial disclosure of RPT governance policy is associated with lower implied costs of capital associated with RPTs. 14

3. Data and Research Design 3.1 Related-party transaction data I hand-collect RPT data and RPT governance data from S&P 1500 firms proxy statements on the SEC s website for 2004, 2007, and 2010. Following Kohlbeck and Mayhew (2017), I choose every three years rather than consecutive years because this design allows for changes in RPTs to occur, as many RPTs are sticky transactions that involve multiple years. I start from 2004 because post-sox period data can mitigate the SOX impact on corporate governance and RPT disclosures. I choose 2010 as the last year of data collection, believing that the three-year data are sufficient to study the effects of RPT policy disclosures. I focus on S&P 1500 because these firms cover approximately 90% of the U.S. market capitalization. Out of the 1500 S&P composite firms, I exclude 430 financial firms, resulting in 3,210 firm-year observations for which proxy statements are available to identify whether or not the firm reported RPTs. After removing firms with missing data, my final sample includes 2,850 firm-year observations. RPTs can be measured by either the number of transactions or the amount of transactions. In my main analyses, I adopt the number of RPTs as I consider that several small RPT transactions likely reflect more severe agency problems than a single larger transaction if the aggregate transaction amounts are similar. However, in additional analyses I use the transaction amount of RPTs as an alternative measurement. 3.2 Regression model The objective of this paper is to examine whether firms change their RPT behavior in response to the mandated RPT governance disclosure and whether investors update the implied 15

cost of capital associated with these RPT behaviors accordingly. To test the relation between the RPT governance disclosure and RPT occurrence, I use a difference-in-differences research design as follows. Log#RPTs = β 1 Newly-Disclosed Post + β 2 Newly-Disclosed + β 3 Post + β 4 Size + β 5 LEV + β 6 MB + β 7 ROA + β 8 Dividend Yield + β 9 CGQ +β 10 Ind. Director %+ β 11 Inst. Holdings % + β 12 Industry Competition + β 13 CEO Pay + β 14 Internal Control Weakness + β 15 Analyst Following + Industry Fixed Effects + ε (1) Where Log#RPTs measures the logarithm of the total number of RPTs that firms disclose in their proxy statements. The Newly-Disclosed variable equals one when firms did not disclose their RPT governance in the pre-regulation period, and zero otherwise. Post equals one when the sample year is after 2006. I expect β 1 <0, meaning that these Newly-Disclosed firms reduce more RPTs following the regulation than Already-Disclosed firms do. To test the relation between RPTs occurrence and the implied cost of capital, I estimate the following model for Newly-Disclosed firms and for Already-Disclosed firms, respectively. Implied cost of capital = β 1 Log#RPTs Post + β 2 Log#RPTs + β 3 Post + β 4 Beta + β 5 Size + β 6 Leverage + β 7 MB + β 8 ROA + β 9 Dividend Yield + β 10 CGQ + β 11 Ind. Director % + β 12 Inst. Holdings % + β 13 Industry Competition + β 14 CEO Pay + β 15 Internal Control Weakness (2) 16

+ β 16 Analyst Following +Industry Fixed Effects + ε To measure the implied cost of capital, I follow Hail and Leuz (2006) and estimate the cost of equity using four different models: the Claus and Thomas (2001) model, the Gebhardt et al. (2001) model, the Ohlson and Juettner-Nauroth (2005) model, and the Easton (2004) model. Because there is little consensus in the literature regarding which model works best, I use an average value of the four models to reduce the idiosyncratic measurement error across various models following prior literature (Ghoul et al 2011; Hou 2015). I run this regression separately for Newly-Disclosed firms and Already-Disclosed firms to avoid inconvenient interpretation of triple interaction terms. I expect β 1 to be significantly more negative for Newly-Disclosed firms than for Already-Disclosed firms. 3.3 Control variables In the multivariate analyses, my control variables include firm size (Size), measured as the natural logarithm of total revenue; market-to-book ratio (MB), measured as the ratio of market value of equity to the book value of equity; return on assets (ROA), measured as net income before extraordinary items scaled by total asset; leverage (Leverage), measured as the ratio of total debt to the book value of equity; and dividend yield (Dividend Yield), measured as the ratio of total dividends to the book value of equity. As prior research shows that overall corporate governance affects the occurrence of RPTs, I also control for general governance, including the percentage of independent director on the Board (Ind. Director %), presence of internal control weakness (Internal Control Weakness), and the Corporate Governance Quotient (CGQ) index, a rating developed by Institutional Shareholder Services (ISS) that rates publicly 17

traded companies in terms of the quality of their corporate governance. 10 A higher CGQ implies stronger corporate governance. To control for external monitoring factors affecting opportunistic behaviors, I include institutional ownership (Inst. Holdings %), measured as percentage of shares owned by institutional investors, industry competition (Industry Competition), measured as the Herfindahl-Hirschman Index of industry market share, and analyst following (Analyst following), measured as the number of analysts following the firm. I also control for CEO compensation (CEO PAY), measured as the natural logarithm of total CEO annual compensation scaled by firm size. In addition, I control for Beta in the analysis of risk associated with RPT. Finally, all regressions include Fama-French 48 industry fixed effects to control for heterogeneity at the industry level and error clustering at the firm level. 10 This CGQ index measures the overall corporate governance relative to a firm s industry group. Ratings are calculated on the basis of 61 data points from eight core categories: (1) Board of directors, (2) audit, (3) charter and bylaw provision, (4) laws of the state of incorporation, (5) executive and director compensation, (6) qualitative factors, (7) ownership, and (8) director education. 18

4. Empirical Results 4.1 Descriptive statistics Table 1 provides descriptive statistics. On average each firm-year observation reports 1.37 related-party transactions. The mean implied cost of capital is 9.8% and the average beta is 1.17, slightly above the market average. The average revenue is $7.5 billion and the average leverage is 0.17. The mean of the governance score (CGQ) is 73 compared to an industry average score of 50, reflecting that my sample firms are better governed than their industry peers. The percentage of independent directors is 81% and the percentage of average institutional holdings is 80%, showing that both internal and external monitoring in my sample firms are relatively strong. Table 2 reports the Pearson correlations and shows that firms that are larger, have higher leverage, have lower market-to-book ratios, and are poorly governed tend to report more RPTs. In addition, I do not find high correlations between the explanatory variables, suggesting multicollinearity is not a serious concern in my analyses. 4.2 Primary Analyses 4.2.1 RPT governance disclosure and occurrence of RPTs I begin my empirical analysis by examining whether disclosures of RPT governance affect the occurrence of RPTs in Table 3. In Column 1, I regress the number of RPTs on Post while controlling for other firm characteristics and fixed effects. The coefficient on Post is -0.04 (significant at the 10% level using a two-sided test), providing preliminary support for the 2006 SEC regulation effectively reducing the occurrence of RPTs. 19

In Column 2, I examine whether RPT governance disclosure is associated with fewer RPTs in the pre-regulation period. The coefficient on Newly-Disclosed firms is 0.22 and is significant at the 1% level, showing that Newly-Disclosed firms have more RPTs than Already- Disclosed firms in the pre-regulation period. When RPT governance disclosure is voluntary, the Newly-Disclosed firms report significantly more RPTs than Already-Disclosed firms. In Column 3, I conduct a difference-in-differences analysis around the 2006 SEC regulation. The interaction term of Newly-Disclosed and Post is negative and significant at the 5% level, implying that compared with Already-Disclosed firms, the reduction of RPTs is more salient for Newly-Disclosed firms. The above analysis suggests that the RPT governance disclosure affects the occurrences of RPTs, consistent with H1. One possible concern is that Already-Disclosed firms may be fundamentally different from Newly-Disclosed firms. Including numerous firm characteristics and fixed effects in the regression may not be sufficient to control for such pre-existing differences. To better separate out the treatment effects, I use propensity-score matching to match samples by firm characteristics and redo the analysis in Column 4. 11 Using this balanced panel, I continue to find that Newly-Disclosed firms reduce more RPTs than Already-Disclosed firms in the post-regulation period, consistent with prior findings. The signs of control variables are consistent with expectations. Larger firms tend to have more RPTs than small firms, because larger firms tend to have more related parties and their business is more complicated. The coefficient on MB is negative, consistent with the perception that related-party firms are valued lower by the market. Finally, the corporate governance measurements are negatively related to RPTs. 11 In this PSM, I first run a logistic regression of RPT governance disclosure on all control variables to obtain the propensity score. Then I match the treatment group sample (i.e. Newly-Disclosed Firms) with the control group sample (i.e. Already-Disclosed Firms) with a caliper matching set at 0.1. 20

In sum, the analyses show that the disclosure of RPT governance significantly reduces the occurrence of RPTs even after controlling for traditional internal and external monitoring mechanisms. 4.2.2 RPT governance disclosure effects conditional on potential agency costs In this section, I examine under which conditions the regulation effects are more pronounced. Because RPTs can reflect a conflict of interests, and the disclosure of RPT governance potentially mitigates the potential risks associated with RPTs, I expect that the effects of RPT governance disclosure are conditional on agency costs. I choose three environmental factors reflecting potential agency problems: institutional ownership, analyst following, and industry competition. In Columns 1 and 2 of Table 4, I partition sample firms into high and low institutional ownership by the median of the sample and rerun model (1) for each group. The estimated coefficients on Newly-Disclosed and Post and the interaction of Newly-Disclosed and Post are not significant in Column 1, suggesting that the disclosure of RPT governance has little effect on the occurrence of RPTs when institutional holdings are high. In Column 2 where the outcomes for low institutional ownership group is presented, I find that the coefficients on the interaction of Newly-Disclosed and Post are significant and negative, suggesting the effects of RPT disclosure governance concentrate on firms with low institutional ownership. The coefficient difference between high and low institutional holding firms (using a fully-interacted model) is significant at the 5% level. In Columns 3 and 4 of Table 4, I partition the sample by the median of the number of analysts following the firm. Columns 3 and 4 represent the analysis for firms with high and low analyst following, respectively. The results show that the coefficients on the interaction of 21