The Impact of Japan s Stewardship Code on Shareholder Voting

The Impact of Japan s Stewardship Code on Shareholder Voting Yasutomo Tsukioka * School of Business Administration, Kwansei Gakuin University Abstract This study examines the impact of the Japanese version of the stewardship code on shareholder voting. Japan s stewardship code was published in February 2014, under which institutional shareholders are expected to discharge their stewardship responsibility through engagement and the exercise of voting. Using data of voting outcomes in shareholder meetings from 2010 to 2015, we find that ownership by trust banks (Japanese institutions that combine the functions of commercial banks, depositary institutions, and trust companies) and insurance companies has significantly positive impacts on votes against directors and CEO appointments after the implementation of Japan s stewardship code, when firms exhibit lower profitability than do their industry peers. These results suggest that Japan s stewardship code has changed the voting behavior of trust banks and insurance companies. Moreover, mutual fund, pension fund, and foreign investor ownership is positively related to voting against the appointment of directors and CEOs. On the other hand, corporate ownership is significantly negatively related to voting against the appointment of directors and CEOs. Keywords: Stewardship code; Voting; Shareholder meeting; Corporate governance; Institutional ownership; Cross-shareholding JEL codes: G32, G38 * School of Business Administration, Kwansei Gakuin University 1-155 Uegahara Ichiban-cho, Nishinomiya, Hyogo 662-8501, Japan. Phone: +81 7 9854 6205 E-mail: tsukioka@kwansei.ac.jp 1

1. Introduction The Japanese government has improved the country s corporate governance system by introducing some principles and codes. Japan s stewardship code is among the corporate governance reforms published by the Financial Service Agency on February 26, 2014. Based on the code, institutional shareholders are expected to enhance corporate value and growth through engagement and voting. Principal five of the code states, Institutional investors should have a clear policy on voting and disclosure of voting activity. The policy on voting should not be comprised only of a mechanical check list; it should be designed to contribute to the sustainable growth of investee companies. Voting is an essential tool of engagement and monitoring. This study aims to investigate whether Japan s stewardship code influences the voting activities of institutional shareholders. In Japan, financial institutions, such as banks, trust banks (Japanese institutions that combine the functions of commercial banks, depositary institutions, and trust companies), and insurance companies, have been considered passive shareholders that vote in line with management or do not exercise their votes. Until the 1980s, financial institutions in Japan were used to playing the important role of keiretsu (cross-shareholdings). Although financial institutions have been selling off their cross-shareholdings since the 1990s, their capital shares remained at about 20% on the Tokyo Stock Exchange in 2015. Japan s stewardship code attempts to encourage institutional investors to promote constructive engagement or purposeful dialogue in order to increase investment returns, improve corporate value, and undertake sustainable growth. Although the code is not mandatory with legally binding regulations, it expects institutional investors to voluntary disclose policies about how they discharge their stewardship responsibilities, voting policies, and voting outcomes. Moreover, when institutional investors accept the code, they are expected to express their acceptance of the code and policy on their web sites, and to register their web sites with the Financial Service Agency. The agency has released a list of institutional investors that have accepted the code. By December 27, 2016, 214 institutional investors had signed up to the code, including 7 trust banks, 22 insurance companies, 26 pension funds, and 152 investment managers. Those investment managers include not only Japanese asset management companies but also foreign investment managers, such as Fidelity, Black Rock, and Goldman Sachs Asset Management. Some institutional investors disclose their aggregated voting outcomes each fiscal year, although they do not disclose their voting activity for each company. For example, Mitsubishi UFJ trust bank has presented its aggregate voting outcomes for each major 2

kind of proposal each fiscal year on its web site since 2014. As we cannot observe institutional investor voting outcomes in the period before Japan s stewardship code was implemented, we cannot understand whether institutional investors changed their voting activities after the code was introduced. On the other hand, we can observe the outcomes of shareholders meetings for each listed firm. Listed firms have been required to disclose these outcomes since March 2010, owing to a revision of a Cabinet Office Ordinance on Disclosure Items Concerning Corporate Governance. Information on shareholder meeting outcomes includes the number of votes for and against a proposal or amendment as well as the number of spoilt votes. The purpose of this study is to clarify the impact of Japan s stewardship code on shareholder voting activities. We investigate the relationship between ownership structure and voting outcomes of director appointments in the period leading up to the introduction of Japan s stewardship code and in the period following the code s introduction. We focus on the proposals for the appointment of directors for the following reasons. First, director appointment proposals are usually initiated by managers on an annual basis. 1 Second, we consider director appointments are more influential on managers than are other proposals, as prior literature shows that votes against director appointments lead to improved corporate governance (Cai et al., 2009; Fischer et al., 2009; Iliev et al., 2015). Specifically, we examine the relationship between ownership structure and voting against director appointments in Japan from 2010 to 2015. We divide the sample into two time periods: before and after Japan s stewardship code was introduced (the pre-code period from 2010 to 2014 and the post-code period from January 2015 to June 2015). Our main finding is that ownership by trust banks, and insurance companies is significantly positively related to voting against proposals for director and CEO appointments of firms, with lower performance in only the post-code period. These results suggest that Japan s stewardship code has changed the voting activity of trust banks and insurance companies. Moreover, ownership by mutual funds, pension funds, and foreign shareholders is positively related to voting against proposals for director and CEO appointments in periods of both the pre-code and post-code. These results suggest that mutual funds, pension funds, and foreign shareholders play an important corporate governance role through their voting activities. On the other hand, there is a negative relationship between corporate ownership and voting against proposals for director and CEO appointments. 1 A few firms propose amendments to director appointments every 2 years or propose partial amendments to director appointments. 3

We contribute to the current debate on the voting activity of institutional investors and the impact of stewardship code on shareholder voting activity. Although the UK is the first country that introduced a stewardship code in July 2010, as far as we known, the impact of a stewardship code on the shareholder voting of investors has not been examined yet. 1 We clarify whether and how Japan s stewardship code changed the voting behavior of trust banks and insurance companies. Our results imply the possibility that nonbinding suggestions from governments change investor behavior and improve the corporate governance system. The rest of this paper is organized as follows. In Section 2, we review the related literature. Section 3 describes the data and a statistical summary, and provides a methodology. Section 4 presents our empirical results and Section 5 concludes. 2. Related literature Prior literature demonstrates that independent shareholders that do not have business relationships with investee firms vote against proposals arising from potential conflicts of interest between management and shareholders. Brickley et al. (1988, 1994) show that ownership by pension funds, mutual funds, and endowments is positively related to voting against management-proposed antitakeover amendments for US firms. De Jong et al. (2006) indicate there is a positive relationship between pension fund ownership and voting against the appointment of executive board members in the Netherlands. Moreover, prior studies indicate that foreign investor ownership is positively related to corporate governance and firm value. Aggarwal et al. (2011) demonstrate that foreign ownership is positively associated with a governance index and firm value across 23 countries. Ferreira and Matos (2008) report that foreign ownership is positively related to firm value and profitability across 27 countries. These studies suggest that mutual fund, pension fund, and foreign ownership plays an important role in corporate governance. We classify mutual funds, pension funds, and foreign shareholders as independent shareholders, which oppose firm activity diminishing shareholder value. We consider that mutual funds, pension funds, and foreign shareholders are likely to vote against proposals for director appointments when firms exhibit lower performance. In other words, ownership by mutual funds, pension funds, and foreign shareholders is positively related to dissenting votes against director appointments in firms with lower profitability. 1 Yermack (2010) reviews the comprehensive research of voting. Mallin (2012) reports the voting behavior of two of the UK s largest institutional investors over the period 2007 to 2009. 4

On the other hand, prior research suggests that banks and insurance companies vote with management on their proposals, because they have business relationships with investee firms and are concerned about voting to maintain the relationship with management. Brickley et al. (1988, 1994) find that insurance company and bank ownership is positively related to voting for management-proposed antitakeover amendments. In Japan, banks, trust banks, insurance companies, and corporate shareholders have business or corporate group relationships with investee firms. Therefore, these types of owners do not vote against management proposals, especially director appointments. However, most trust banks and insurance companies accept Japan s stewardship code and disclose their voting policies in Japan. These disclosures indicate that such institutions vote against director appointments when firms exploit shareholders interests, exhibit lower operating performance, are embroiled in scandals, and do not appoint outside directors or independent directors. The question is whether they changed their voting behavior in the post-code period. In other words, did Japan s stewardship code lead trust banks and insurance companies to vote against director appointments of firm with lower profitability? 3. Data and methodology 3.1 Data and statistics We collect the voting outcomes of shareholder meetings from NIKKEI NEEDS shareholder meeting data for the period August 2010 to June 2015. Based on a revision of the Cabinet Office Ordinance on Disclosure Items Concerning Corporate Governance, firms should disclose shareholder-voting outcomes from 2010. In addition, we obtain financial data, ownership data, major shareholder data, and executive data from the NIKKEI NEEDS Quest. We restrict our sample to all nonfinancial firms with available data and voting outcomes of director appointments in annual general meetings. In addition, we exclude sample firms that have outliers of ownership and voting outcomes, and that hold shareholder meetings more than 90 days after the fiscal-year end. 1 Furthermore, we exclude firms with negative shareholder equity. Our final sample consists of 3,584 firms and 14,371 firm-year observations. In Table 1, Panel A provides summary statistics of the voting outcomes of each annual general meetings of shareholders. The mean number of proposals is 3.482. Most firms 1 An annual general meeting has to be held within 3 months of the end of a firm s fiscal year. 5

make proposals for profit distribution, director appointments, and audit and supervisory appointments each year. We focus on director appointments. As director appointment proposals include individual amendments to director appointments, the mean percentage of votes against proposals for director appointments is calculated based on the mean number of votes against those amendments. Percentage of votes agaist number of votes for = number of votesfor + number of votes against + number ofspoilt votes (1) Proposals are almost never rejected: votes against director appointments comprise only 0.019% of total votes, and those against CEO appointments 0.029%. In this study, CEOs is identified as the top representative directors of firms. 1 Votes against CEO appointments are higher than those against directors are because shareholders tend to vote against CEO appointments if firms exhibit low performance or governance and are embroiled in scandals. In Table 1, Panel B provides summary statistics of firm characteristics and ownership structure. Information on ownership structure is collected from information about the distribution of shareholder ownership and major shareholder data. Specifically, ownership by mutual funds, pension funds, foreign shareholders, corporate shareholders, directors, and employees is from the distribution of shareholder ownership. Meanwhile, major shareholder data often include details on the top 30 shareholders. We extract information on ownership by banks, trust banks, and insurance companies from major shareholder data. We exclude trust account ownership. 2 Table 2 shows the correlation among voting outcomes, ownership structure, and firm character. There is significantly positive correlation among mutual ownership, pension ownership, and foreign ownership. 3.2 Methodology We divide the sample into two time periods: pre-code period (August 2010 through December 2014) and post-code period (January 2015 through June 2015). We examine the impact of Japan s stewardship code on shareholders voting behavior by estimating the following equation: 1 Although CEOs are the presidents of most Japanese firms, chairpersons or other representative directors have decision-making authority. 2 Banks, trust banks, and insurance companies are entrusted with managing funds from pension funds and mutual funds. Some trust banks also engage in custodian business. 6

Percentage of votes against i,t = α + β Ownership i,t Lower ROE dummy i,t + γ Ownership i,t + δ Lower ROE dummy i,t (2) + ρ Controls i,t + ε i,t In equation (2), we include interception terms between each type of shareholder ownership (Ownership) and a lower return on equity (ROE) dummy, which equals 1 if the ROE of a firm is less than the median ROE of the industry to which the firm belongs, and 0 otherwise. 1 This is because institutional investors that have accepted the code state that they vote against director appointments of firms that have lower profitability. Furthermore, Institutional Shareholder Services (ISS) recommends voting against CEO appointments in firms that have less than 5% average ROE for 5 consecutive fiscal years without an improvement trend. We divide shareholders into two types: independent shareholder and cross-shareholder. Independent shareholders include mutual funds, pension funds, and foreign shareholders. Cross-shareholders include banks, trust banks, insurance companies, and corporate shareholders. Control variables include director ownership, employee ownership, logarithm of total assets (AST), the ratio of debt to assets (LEV), a concentration dummy that equals 1 if the annual general meeting is held on the same day on which most companies hold annual general meetings, and 0 otherwise, and the outside director ratio (the number of outside directors/the number of directors). 2 Gordon and Pound (1993) find that ownership by employee stock ownership plans, insiders, and directors are negatively related to voting for shareholders proposals. 4. Empirical results We examine the impact of the stewardship code on shareholder voting on director appointments by estimating equation (2) in periods of both the pre-code and post-code. Table 3 presents the result of the determinants of the percentage of votes against director appointments. Panel A reports the results when we use ownership by independent 1 In addition, we change the lower ROE dummy criteria to other profitability criteria, such as net loss, the ISS recommended criteria (5% average ROE for 5 fiscal years and without an improvement trend), and return on assets (the ratio of earnings before interest and tax to assets), and then, we run the analysis. The results are similar. 2 For example, more than 900 companies held annual general meetings 1 business day before the end of June each year. 7

shareholders: mutual funds, pension funds, and foreign shareholders. All models show that both the interception terms and each type of shareholder ownership is significantly positively related to votes against director appointments, and that the coefficient of the interception term is higher than that of each type of shareholder ownership. This result suggests that mutual funds, pension funds, and foreign shareholders vote against director appointments more when firms exhibit lower ROE than their industry peers do. Moreover, the coefficient of the lower ROE dummy is positive and significant. By contrast, the coefficient on employee ownership is negative and significant. Panel B reports the result of the relationship between cross-shareholder ownership and votes against director appointments. Models (3) and (4) show that the interception term (trust bank ownership lower ROE dummy) is significantly positively associated with votes against director appointments in only the post-code period. Models (5) and (6) show that the interaction term (insurance ownership lower ROE dummy) are significantly positively related to voting against director appointments in only the postcode period. These results suggest that trust banks and insurance companies become opposed to director appointments of firms with lower profitability following the introduction of Japan s stewardship code. On the other hand, when firms perform with higher ROE, ownership of banks, trust banks, and insurance companies is negatively related to voting against director appointments. This result suggests that banks, trust banks, and insurance companies vote for director appointments of firm with higher profitability. Moreover, models (7) and (8) show that both the interception term (corporate ownership lower ROE dummy) and corporate ownership are significantly and negatively related to voting against director appointments, suggesting that corporate shareholders usually vote for director appointments. Since ISS has recommended voting against the appointment of top managers and some institutional investors have indicated they vote against the appointment of top managers in firms when dissatisfied with certain criteria, we substitute the percentage of votes against CEO appointments for the dependent variable in equation (2). The results of the relationship between ownership structure and votes against CEO appointments are reported in Table 4. Panel A reports the results of the relationship between independent shareholder ownership and votes against CEO appointments. All models show that the coefficient of the interception term and each shareholder ownership is positive and significant in both the pre-code and post-code periods, suggesting that mutual funds, pension funds, and foreign investors tend to vote against CEO appointments. In addition, the coefficient of the interception term in the post-code period is higher than in the precode period. Mutual funds, pension funds, and foreign investors seem more likely to vote 8

against CEO appointments in the post-code period, when firms exhibit lower ROE. Panel B reports the result of the relationship between cross-shareholder ownership and votes against CEO appointments. Models (3) and (4) show that trust bank ownership is significantly positively associated with votes against CEO appointments of firms with lower ROE in only the post-code period. This result suggests that trust banks change their voting behavior and come to vote against CEO appointments of firms with lower profitability following the introduction of Japan s stewardship code. Models (5) and (6) show that insurance companies are significantly positively associated with votes against CEO appointments of firms with lower ROE, and the coefficient of the interception term (insurance ownership lower ROE dummy) in the pre-code period is higher than that in the post-code period. 1 This result suggests that insurance companies are more likely to vote against CEO appointments of firms with lower ROE in the post-code period. On the other hand, models (7) and (8) indicate that corporate shareholders vote for CEO appointments. These results suggest that Japan s stewardship code has made trust banks and insurance companies change their voting policies. Actually, most Japanese trust banks and insurance companies have signed up to Japan s stewardship code and now disclose their policies and aggregate voting outcomes. Our evidence suggests that trust banks and insurance companies have begun to discharge their stewardship responsibilities through voting following the introduction of Japan s stewardship code. In addition, these results suggest that mutual funds, pension funds, and foreign investors have engaged in corporate governance through voting on director appointments. On the other hand, corporate shareholders still vote a line with managers and are passive shareholders. 5. Conclusion We document the relationship between ownership structure and outcomes of director appointment in annual general meetings before and after the introduction of Japan s stewardship code. The code, published in January 2014, is not legally binding. Institutional investors that are signatories to the code are expected to perform their stewardship responsibilities through their engagement with the investee firm and their voting activities. We focus on institutional shareholder voting activity, since voting is an essential tool. 1 The coefficient of the interception term (insurance ownership lower ROE dummy) is negative and not significant in the pre-code period when we estimate fixed model (3) controlled for firm and year effects. 9

We find that after the introduction of Japan s stewardship code, ownership by trust banks and insurance companies has a significantly positive impact on voting against the appointment of directors and CEOs when firms exhibit lower ROE than do their industry peers. Moreover, ownership by mutual funds, pension funds, and foreign investors is positively related to voting against the appointment of directors and CEOs, and the significance is higher in firms with lower ROE than those with higher ROE. On the other hand, corporate ownership is significantly negatively related to voting against the appointment of directors and CEOs. Our evidence suggests that Japan s stewardship code has led trust banks and insurance companies to begin to perform their stewardship responsibility through voting activity for the appointment of directors and CEOs. Consistent with a lot of prior literature, the results indicate that mutual funds, pension funds, and foreign investors play an important role in corporate governance in Japan. In contrast to these shareholders, corporate shareholders vote in line with managers, because they have business relationships with managers and are cross-shareholders. References Aggarwal, R., Erel, I., Ferreira, M., and Matos, P. (2011), Does Governance Travel around the World? Evidence from Institutional Investors, Journal of Financial Economics, Vol. 100, No. 1, pp. 154-181. Brickley, J. A., Lease, R. C., and Smith Jr., C. W. (1988), Ownership Structure and Voting on Antitakeover Amendments, Journal of Financial Economics, Vol. 20, No. 1, pp. 267-291. Brickley, J. A., Lease, R. C., and Smith, Jr. C. W. (1994), Corporate Voting: Evidence from Charter Amendment Proposals, Journal of Corporate Finance, Vol. 1, No. 1, pp.5-31. Cai, J., Garner, J. L., and Walking, R. A. (2009), Electing Directors, Journal of Finance, Vol. 64, No. 5, pp. 2389-2421. De Jong, A, Mertens, G., and Roosenboom, P. (2006), Shareholder s Voting at General Meetings: Evidence from the Netherlands, Journal of Management and Governance, Vol. 10, No. 4, pp. 353-380. Ferreira, M. A., and Matos, P. (2008), The Colors of Investors Money: The Role of Institutional Investors around the World, Journal of Financial Economics, Vol. 88, No. 3, pp. 499-533. 10

Fischer, P. E., Gramlich, J. D., Miller, B. P., and White, H. D. (2009), Investor Perceptions of Board Performance: Evidence from Uncontested Director Elections, Journal of Accounting and Economics, Vol. 48, No. 2-3, pp. 172-189. Gordon, L. A., and Pound, J. (1993), Information, Ownership Structure, and Shareholder Voting: Evidence from Shareholder-Sponsored Corporate Governance Proposals, Journal of Finance, Vol. 48, No. 2, pp. 697-718. Iliev, P., Lins, K. V., Miller, D. P., and Roth, L. (2015), Shareholder Voting and Corporate Governance around the World, Review of Financial Studies, Vol. 28, No. 8, pp. 2167-2202. Mallin, C. (2012), Institutional Investors: The Vote as a Tool of Governance, Journal of Management and Governance, Vol. 16, No. 2, pp. 177-196. Yermack, D. (2010), Shareholder Voting and Corporate Governance, Annual Review of Financial Economics, Vol. 2, pp. 103-125 11

Table 1 Summary statistics Panel A Shareholder voting outcome Mean Median SD 1 st Quantile 3 rd Quantile Number of proposals 3.482 3.000 1.376 3.000 4.000 Percentage of votes against director 0.018 0.009 0.024 0.003 0.024 appointment Percentage of votes against CEO appointment 0.029 0.011 0.042 0.003 0.037 Panel B Ownership structure and firm characteristics Mean Median SD 1 st Quantile 3 rd Quantile Mutual fund ownership 0.020 0.007 0.029 0.000 0.030 Pension fund ownership 0.009 0.000 0.014 0.000 0.013 Foreign ownership 0.090 0.038 0.115 0.005 0.138 Bank ownership 0.034 0.024 0.037 0.000 0.054 Trust bank ownership 0.005 0.000 0.010 0.000 0.006 Insurance ownership 0.024 0.009 0.034 0.000 0.036 Corporate ownership 0.284 0.254 0.194 0.123 0.410 Director ownership 0.077 0.016 0.124 0.003 0.094 Employee ownership 0.017 0.010 0.021 0.002 0.024 AST 10.459 10.308 1.704 9.283 11.478 LEV 0.489 0.493 0.204 0.331 0.648 ROE 0.063 0.060 0.143 0.025 0.108 Outside director ratio 0.146 0.125 0.150 0.000 0.231 Data on mutual fund ownership, pension fund ownership, foreign ownership, corporate ownership, director ownership, and employee ownership are obtained from the distribution of shareholder ownership. Data on bank ownership, trust bank ownership, and insurance company ownership are from major shareholder data. This financial institution ownership excludes trust account ownership 12

Table 2 Correlation between variables (1) (2) (3) (4) (5) (6) (7) (8) (10) (11) (12) (13) (14) (15) (1) Percentage of votes against director appointment 0.452 0.488 0.548-0.064-0.008 0.090-0.201-0.331-0.126 0.457-0.071-0.074 0.065 (2) Mutual fund ownership 0.282 0.832 0.575 0.047 0.059 0.213-0.181-0.228 0.207 0.528-0.101 0.165 0.053 (3) Pension fund ownership 0.335 0.632 0.599 0.079 0.060 0.241-0.181-0.292 0.173 0.610-0.057 0.121 0.034 (4) Foreign ownership 0.326 0.414 0.466-0.086-0.026 0.127-0.287-0.352-0.136 0.656-0.174 0.200 0.162 (5) Bank ownership -0.082-0.024 0.012-0.129 0.371 0.538-0.064-0.102 0.254 0.215 0.105-0.121-0.186 (6) Trust bank ownership -0.025 0.013 0.019-0.049 0.321 0.379 0.000-0.138 0.113 0.153 0.078-0.043-0.095 (7) Insurance ownership 0.025 0.140 0.183 0.081 0.428 0.320-0.126-0.300 0.184 0.366 0.087-0.079-0.075 (8) Corporate ownership -0.156-0.231-0.221-0.314-0.109-0.033-0.179-0.216-0.015-0.045 0.040-0.030 0.031 (10) Director ownership -0.164-0.106-0.173-0.200-0.224-0.169-0.273-0.323 0.117-0.529-0.092 0.077-0.179 (11) Employee ownership -0.133-0.021-0.038-0.201 0.163 0.026 0.022-0.049 0.034 0.019 0.052-0.020-0.155 (12) AST 0.317 0.398 0.514 0.591 0.142 0.118 0.334-0.086-0.406-0.103 0.133 0.078 0.052 (13) LEV -0.065-0.084-0.035-0.137 0.087 0.081 0.094 0.040-0.044 0.040 0.155 0.058-0.024 (14) ROE -0.081 0.130 0.076 0.111-0.045-0.002-0.023-0.004 0.101-0.002 0.075-0.017 0.062 (15) Outside director ratio 0.056 0.051 0.011 0.193-0.189-0.087-0.063 0.057-0.071-0.142 0.045-0.028 0.019 This table shows Pearson correlations (lower triangle) and Spearman correlations (upper triangle). 13

Table 3 Determinants of votes against director appointments Panel A: Independent shareholder ownership and percentage of votes against director appointments Dependent variable = Percentage of votes against director appointments Pre-code Post-code Pre-code Post-code Pre-code Post-code (1) (2) (3) (4) (5) (6) Mutual fund ownership 0.179 *** 0.147 *** lower ROE dummy 7.827 3.434 Pension fund ownership 0.389 *** 0.497 *** lower ROE dummy 8.263 3.918 Foreign ownership lower ROE dummy 0.044 *** 0.047 *** 6.186 5.099 Mutual fund ownership 0.103 *** 0.073 *** 7.934 3.855 Pension fund ownership 0.282 *** 0.259 *** 9.932 4.080 Foreign ownership 0.026 *** 0.022 *** 5.456 3.591 Lower ROE dummy 0.003 *** 0.004 *** 0.003 *** 0.004 *** 0.003 *** 0.003 *** 5.548 4.407 5.646 4.235 4.850 2.774 Director ownership -0.007 *** -0.003-0.007 *** -0.005-0.006 ** -0.006-3.351-0.851-3.134-1.163-2.504-1.388 Employee ownership -0.124 *** -0.116 *** -0.125 *** -0.111 *** -0.090 *** -0.080 *** -11.941-6.483-12.273-5.831-8.588-4.153 AST 0.003 *** 0.003 *** 0.003 *** 0.003 *** 0.003 *** 0.002 *** 11.775 9.339 9.192 7.529 9.615 5.600 DEBT -0.008 *** -0.007 *** -0.008 *** -0.007 *** -0.006 *** -0.005 ** -4.743-3.094-4.913-3.282-3.712-2.139 Concentration dummy 0.002 ** -0.001 0.002 ** 0.000 0.002 *** 0.000 14

2.242-0.603 2.346-0.460 2.594-0.429 Outside director ratio 0.004 0.005 0.004 * 0.004 0.000 0.002 1.487 1.162 1.676 1.039 0.098 0.400 Industry dummies Yes Yes Yes Yes Yes Yes Year dummies Yes No Yes No Yes No Adjusted R 2 0.196 0.170 0.210 0.177 0.188 0.174 N 11,848 2,523 11,848 2,523 11,848 2,523 15

Panel B: Cross-shareholder ownership and percentage of votes against director appointments Dependent variable = Percentage of votes against director appointments Pre-code Post-code Pre-code Post-code Pre-code Post-code Pre-code Post-code (1) (2) (3) (4) (5) (6) (7) (8) Bank ownership -0.010 0.000 lower ROE dummy -0.764-0.019 Trust bank ownership -0.009 0.182 ** lower ROE dummy -0.214 2.081 Insurance ownership 0.023 0.053 ** lower ROE dummy 1.518 2.042 Corporate ownership -0.023 *** -0.015 *** lower ROE dummy -8.830-3.457 Bank ownership -0.086 *** -0.062 *** -8.703-4.182 Trust bank ownership Insurance ownership Corporate ownership -0.127 *** -0.175 *** -4.079-3.129-0.074 *** -0.072 *** -7.029-4.185-0.011 *** -0.014 *** -5.163-4.660 Lower ROE dummy 0.006 *** 0.006 *** 0.006 *** 0.005 *** 0.005 *** 0.005 *** 0.012 *** 0.010 *** 7.860 5.274 8.591 5.253 7.175 4.655 11.585 6.302 Director ownership -0.008 *** -0.005-0.006 ** -0.004-0.006 *** -0.004-0.019 *** -0.018 *** -3.412-1.271-2.431-0.888-2.758-1.030-6.843-3.916 Employee ownership -0.094 *** -0.082 *** -0.117 *** -0.101 *** -0.113 *** -0.097 *** -0.130 *** -0.119 *** -8.439-4.190-10.616-5.212-10.230-5.031-11.961-6.279 AST 0.005 *** 0.004 *** 0.005 *** 0.004 *** 0.005 *** 0.005 *** 0.004 *** 0.004 *** 18.852 14.115 18.616 14.055 19.203 14.344 15.862 11.876 16

DEBT -0.010 *** -0.009 *** -0.011 *** -0.009 *** -0.011 *** -0.009 *** -0.010 *** -0.009 *** -6.020-4.098-6.178-4.005-6.194-4.140-6.099-3.817 Concentration dummy 0.002 *** 0.000 0.002 ** 0.000 0.002 *** 0.000 0.001-0.001 3.229 0.030 2.381-0.273 2.756-0.230 1.442-1.305 Outside director ratio 0.002 0.004 0.004 0.005 0.004 * 0.006 0.005 ** 0.005 0.644 0.866 1.557 1.274 1.680 1.348 2.067 1.140 Industry dummies Yes Yes Yes Yes Yes Yes Yes Yes Year dummies Yes No Yes No Yes No Yes No Adjusted R 2 0.173 0.151 0.161 0.146 0.164 0.147 0.191 0.171 N 11,848 2,523 11,848 2,523 11,848 2,523 11,848 2,523 The dependent variable is the percentage of votes against director appointments. Robust standard errors are clustered at firm level. The lower step indicates t-statistics. *, **, and *** denote statistical significance at the 10%, 5%, and 1% levels, respectively. 17

Table 4 Determinants of percentage of votes against CEO appointments Panel A: Independent shareholder ownership and percentage of votes against CEO appointments Dependent variable = Percentage of votes against CEO appointments Pre-code Post-code Pre-code Post-code Pre-code Post-code (1) (2) (3) (4) (5) (6) Mutual fund ownership 0.300 *** 0.352 *** lower ROE 7.687 4.374 Pension fund ownership 0.567 *** 1.441 *** lower ROE 7.245 5.324 Foreign ownership lower ROE 18 0.072 *** 0.142 *** 5.793 7.075 Mutual fund ownership 0.125 *** 0.062 * 5.686 1.930 Pension fund ownership 0.489 *** 0.422 *** 9.798 3.510 Foreign ownership 0.060 *** 0.041 *** 6.914 3.669 Lower ROE dummy 0.003 *** 0.013 *** 0.003 *** 0.011 *** 0.003 *** 0.007 *** 2.627 6.378 3.705 5.798 3.067 3.579 Director ownership -0.010 *** -0.005-0.009 *** -0.009-0.009 ** -0.012 * -2.824-0.690-2.743-1.197-2.375-1.655 Employee ownership -0.213 *** -0.180 *** -0.215 *** -0.184 *** -0.148 *** -0.108 *** -12.038-5.255-12.504-5.547-8.612-3.227 AST 0.008 *** 0.008 *** 0.006 *** 0.006 *** 0.006 *** 0.004 *** 15.605 11.246 12.141 7.832 11.141 4.954 DEBT -0.018 *** -0.020 *** -0.017 *** -0.019 *** -0.012 *** -0.012 *** -6.507-4.223-6.407-4.080-4.334-2.588 Concentration dummy 0.002-0.004 ** 0.002-0.004 ** 0.002 * -0.004 ** 1.217-2.276 1.319-2.197 1.645-2.049

Outside director ratio -0.048 *** -0.046 *** -0.047 *** -0.047 *** -0.054 *** -0.052 *** -13.355-5.662-13.495-5.813-14.311-6.252 Industry dummies Yes Yes Yes Yes Yes Yes Year dummies Yes No Yes No Yes No Adjusted R 2 0.245 0.215 0.263 0.243 0.253 0.249 N 10,544 2,222 10,544 2,222 10,544 2,222 19

Panel B: Cross-shareholder ownership and percentage of votes against CEO appointments Dependent variable = Percentage of votes against CEO appointments Pre-code Post-code Pre-code Post-code Pre-code Post-code Pre-code Post-code (1) (2) (3) (4) (5) (6) (7) (8) Bank ownership -0.027 0.009 lower ROE dummy Trust bank ownership lower ROE dummy Insurance ownership lower ROE dummy Corporate ownership lower ROE dummy Bank ownership Trust bank ownership Insurance ownership Corporate ownership -1.235 0.218-0.114 *** -0.104 *** -6.395-3.721-0.054 0.408 *** -0.671 2.687-0.122 ** -0.269 *** -2.199-3.162 20 0.049 ** 0.135 *** 2.013 2.589-0.112 *** -0.129 *** -6.286-3.842-0.033 *** -0.040 *** -7.770-4.568-0.016 *** -0.012 ** -4.522-2.126 Lower ROE dummy 0.008 *** 0.018 *** 0.007 *** 0.016 *** 0.006 *** 0.015 *** 0.016 *** 0.030 *** 6.294 7.622 6.570 7.938 4.899 7.060 9.338 8.854 Director ownership -0.011 *** -0.008-0.007 * -0.005-0.009 ** -0.006-0.028 *** -0.028 *** -2.891-1.060-1.930-0.624-2.337-0.809-6.133-3.280 Employee ownership -0.169 *** -0.130 *** -0.203 *** -0.161 *** -0.197 *** -0.155 *** -0.223 *** -0.190 *** -9.238-3.613-11.135-4.590-10.851-4.444-12.041-5.323 AST 0.010 *** 0.010 *** 0.010 *** 0.010 *** 0.010 *** 0.010 *** 0.009 *** 0.009 *** 22.091 14.835 21.915 14.784 22.154 14.835 19.169 13.274 DEBT -0.021 *** -0.023 *** -0.022 *** -0.023 *** -0.021 *** -0.024 *** -0.021 *** -0.022 ***

-7.448-4.927-7.679-4.868-7.623-4.983-7.532-4.732 Concentration dummy 0.003 ** -0.003 0.002-0.004 * 0.002 * -0.003 * 0.001-0.005 *** 2.059-1.603 1.379-1.908 1.734-1.781 0.542-2.743 Outside director ratio -0.050 *** -0.047 *** -0.047 *** -0.044 *** -0.047 *** -0.044 *** -0.045 *** -0.044 *** -13.448-5.654-12.696-5.261-12.582-5.292-12.276-5.360 Industry dummies Yes Yes Yes Yes Yes Yes Yes Yes Year dummies Yes No Yes No Yes No Yes No Adjusted R 2 0.226 0.198 0.217 0.195 0.220 0.196 0.240 0.214 N 10,544 2,222 10,544 2,222 10,544 2,222 10,544 2,222 The dependent variable is the percentage of votes against CEO appointments. Robust standard errors are clustered at firm level. The lower step indicates t-statistics. *, **, and *** denote statistical significance at the 10%, 5%, and 1% levels, respectively. 21