Top-executives Compensation: The Role of Corporate Ownership Structure in Japan

Abstract

This paper explores the impact of corporate control, measured by ownership structure, on top-executives' compensation in Japan. According to agency theory, the pay-performance link is expected to be affected by the firm's ownership structure. Using a sample of 4,411 firm-year observations (401 firms for the 11-years period from 2001 to 2011) for Japanese non-financial firms publicly traded on the first section and second section of the Tokyo Stock Exchange (TSE), this study demonstrates that institutional ownership (both financial and corporate) is negatively related to the level of executives' compensation. Such finding is in line with efficient monitoring hypothesis which claims that the presence of institutional shareholders provides direct monitoring over managers, limits managerial self-dealing and curves the increase in top-executives pay. On the other hand, the results also show that managerial ownership is positively related to their compensation which supports managerial power theory hypothesis, i.e. management-controlled firms are more likely to extract more compensation from the business than other firms. Overall, this study confirms that corporate control has significant impact on cash compensation paid to Japanese top-executives after controlling the conventional pay-performance relationship.

1. Introduction

Executive compensation is a popular topic in both the popular press and the scholarly literature (Boyd, Santos & Shen, 2012). Nevertheless, the same has been less of an issue of concern in Japan, where the number of empirical research on executives’ compensation is very few. One of the reasons might be the widely held perception that Japanese top-executives get very lower amount in comparison to their western counterparts. But, Jensen and Murphy (2010) argue that it is less important to focus on how much top-executives are being paid rather the real problem is how they are being paid. In addition, Japan is a country of particular interest not only because it has considerably lower levels of compensation than the US, but also because it has very different corporate governance structure, characterized by insider dominated board and high institutional ownership. Therefore, further empirical research that contributes to the managerial compensation literature in Japanese context is expected to be highly appreciated.

Existing theoretical framework and empirical research have identified many hypotheses and determinants that explain the amount and structure of executive compensation in the corporate world. The use of Japanese data allows testing of the universality of those hypotheses and determinants in a country where different governance mechanism persists. Most of the previous studies in Japanese context have focused on relatively narrow aspects of the executive compensation process, i.e., testing the relative significance of accounting profitability or stock market performance (e.g., Abe, Gaston, & Kubo, 2005; Ang & Constand, 1997; Basu, Hwang, Mitsudome, & Weintrop, 2007; Joh, 1999; Kaplan, 1994; Kato, 1997; Kato, Lemmon, Luo, & Schallheim, 2005; Kato & Kubo, 2006; Mitsudome, Weintrop, & Hwang, 2008; Shuto, 2007; Xu, 1999). But, according to agency theory, the compensation-performance link is also affected by the level of firm’s corporate control. The reason is that top-executives in firms with weak corporate control may persuade to ignore real corporate performance and set compensation according to their wish. Any study on managerial compensation which does not address the aspect of corporate control along with corporate performance is incomplete and dubious in findings. Sakawa, Moriyama and Watanabel (2012) claim that corporate control in Japanese companies is important for controlling executive compensation. Therefore, this current study tries to contribute to the existing literature by examining how the corporate control (measured by ownership structure) influences the level of top-executives' compensation in Japanese companies.

The paper has very interesting results. Corporate control is found to have a significant influence on managerial compensation of Japanese firms after controlling corporate performance. In particular, institutional ownerships measured by financial institutional shareholdings and business corporations’ shareholdings have significant negative impact on top-executives' compensation which is consistent with ‘Efficient Monitoring Hypothesis’. On the other hand, in line with ‘Managerial Power Theory Hypothesis’, directors (executive)’ shareholdings have the positive impact on top-executives' compensation. The findings clearly justify the role of corporate control as a determinant of on top-executives' compensation.

The remainder of this paper is organized as follows. Section 2 presents theory, prior studies, and hypotheses to be empirically tested in this study. Section 3 describes research design and sample characteristics. Section 4 presents empirical results, and Section 5 concludes the paper.

2. Theoretical Background and Hypotheses Development

Numerous studies have tried to identify the most important factors impacting executive compensation in the listed companies. Lazear (1986) argues that the appropriateness (optimality) of contingent performance based managerial compensation contracts depends critically on the degree of direct monitoring of the agent’s action. The demand for incentives to align managerial interest will be relatively weak if control and regulation ensures that (Agrawal, Makhija & Mandelker, 1991). Therefore, it is also important to consider the level of corporate control, along with economic determinants, to explore how it influences the managerial compensation.

Ownership structure of business firms is often used in the empirical research as the proxy of corporate control. The types of owner and their holdings have significant influence on corporate major decisions. Haid and Yurtoglu (2006) argue that the classical conflict between managers and dispersed shareholders is not pronounced in the firms, where large shareholders have substantial incentives and the ability to monitor the managers. Diamond (1984), Del Guercio and Hawkins (1999), and Almazan, Hartzell and Starks (2005) argue that institutional investors can provide direct monitoring and disciplining over managers, which is more difficult for other investors who are typically smaller, more passive and less-informed. Institutional investors can use various formal and informal mechanisms to influence management (Cubbin & Leech, 1983). This aspect is often denoted as ‘Efficient Monitoring Hypothesis’. In the absence of any control either internal or external, there exists an opportunity for senior executives to award themselves by pay raises that are not congruent with shareholders’ interests. Bertrand and Mullainathan (2000) posit that the presence of large shareholders moderates the increase in CEO pay and limits managerial self-dealing.

There is a large body of literature in the international context examining the role of institutional and/or large owners in the determination of director pay (e.g., Conyon & He, 2011; Cosh & Hughes, 1997; Dong & Ozkan, 2008; Elston & Goldberg, 2003; Hartzell & Starks, 2003; Mehran, 1995; Ozkan, 2007). For US companies, Mehran (1995) finds that the use of incentive-based compensation also declines with the percentage of stock held by outsider blockholders. Hartzell and Starks (2003) provide evidence that the level of executive pay decreases with institutional ownership. In addition, Khan, Dharwadkar and Brandes (2005) also find that higher level of ownership concentration is associated with lower level of pay. For UK companies, Ozkan (2007) provides evidence that institutional ownership has a significant and negative impact on CEO compensation.

Contrary to the evidence provided by Hartzell and Starks (2003) and others, Gallagher, Smith and Swan (2005) find that institutional influence increases the level of compensation, and thus, conclude that institutions are not effective monitors of executive compensation in the US companies. Cosh and Hughes (1997) show that the presence of major financial institutions among owners make no appreciable difference to either the level of pay or the sensitivity of pay to performance in the UK companies. Based on Chinese evidence, Firth, Fung and Rui (2007) and Conyon and He (2011) find that executive pay and CEO incentives are lower in state controlled firms and firms with concentrated ownership structure. In German context, Schmid (1997) indicates that bank ownership insulates managers from effective disciplining, and thus, has a positive impact on the level of pay. In contrast to evidence provided by Schmid (1997), Elston and Goldberg (2003), and Haid and Yurtoglu (2006) find that bank influence in German firms has negative impact on compensation. They also confirm that block ownership restrains compensation. Based on 75 largest Dutch firms for the period 2006-2008, Knop and Mertens (2010) find that the presence of external block-holders significantly lowers the levels of total CEO compensation. In a review paper, Goergen and Renneboog (2011) claim that most of the existing literature supports that the executives are able to extract rents in the presence of weak corporate governance. They also argue that the increase in CEO pay is highest in firms without large shareholders as the latter are instrumental in preventing excessive rent extraction.

In Japanese context, limited research has investigated at how the volume of institutional ownership affects executive compensation. Abe et al. (2005) argue that when outside monitors can observe the behavior of directors, the relative importance of financial performance in executive compensation contract will be smaller. Bases on the early 1980s Japanese firms’ data, Kato (1997) find the evidence that the compensation paid to directors is significantly smaller in the companies with keiretsu or cross-corporate shareholdings. Basu et al. (2007) and Sakawa and Watanabel (2008) report that the monitoring role of banks facilitates the reduction of top executive compensation for Japanese nonfinancial firms in the 1990s. These earlier findings in the literature suggest that Japanese banks served a substitute role for incentive compensation packages in the 1990s through substantial stockholdings. In contrast, using very recent but small number of firm observations, Sakawa et al. (2012) find that the bank ownership do not substitute for incentive compensation, unlike in the 1990s in Japan. So, it is really commendable to investigate the relationship between institutional ownership and managerial compensation level of Japanese firms based on recent and large samples. Recent corporate ownership survey (2011) by Tokyo stock exchange (TSE) shows that, on average, the largest shareholders in a firm owns more than 50% of the issued shares. Furthermore, the financial institutions and the business corporations are the largest shareholder in many listed firms. Thus, it is expected that financial institutions and the business corporations usually have effective control on the listed Japanese firms. Based on the above discussion, we can form hypotheses like:

H-1: Ceteris paribus, top-executives’ compensation (cash) is significantly associated with financial institutional ownership.

H-2: Ceteris paribus, top-executives’ compensation (cash) is significantly associated with business corporate ownership.

Bebchuk, Fried and Walker (2002) and Bebchuk and Fried (2004) consider another dimension of corporate control i.e. managerial control which influence executive compensation. Ferris, Kim, Kitsabunnarat and Nishikawa (2007) conclude that managerial power effects are present in the design of executive compensation. Managementcontrolled firms are more likely than owner-controlled firms to extract more compensation from the business. Williamson (1964) and Monsen and Downs (1965) argue that the managers of large corporations with diffuse ownership have considerable discretion in guiding the affairs of their firms and that discretionary power is used to divert some resources from corporate shareholders. This is often denoted as ‘Managerial Power Theory Hypothesis’.

Japanese commercial law requires that the top-executives’ compensation in the listed firms should be designed by the board of directors, and the ceiling of maximum gross compensation should be approved by the shareholders. As most of the board members in Japanese companies are also top-executives, it is expected that top-executives would have significance dominance over pay-settings. Sakawa et al. (2012) mention that lack of independence of the board from the company’s top-executives might be a cause of greater managerial power in Japanese companies. Moreover, when the top-executives own shares in the companies, the ability to control the amount of compensation raise further as they can substantially control the shareholders’ approval on any amount of managerial compensation. Grabke-Rundell and Gomez-Mejia (2002) point out that executive power, such as their stock ownership, is an important determinant that comes into play when designing executive compensation. However, this aspect of managerial control has not been explored yet in the Japanese context. Therefore, it is also important to investigate the effect of managerial control on the compensation level of Japanese top-executives. From this point of view the next hypothesis to examine empirically is:

H-3: Ceteris paribus, top-executives’ compensation (cash) is significantly associated with their share ownership.

3. Research Design

3.1. Basic Equation

Any statistical model of the relation between compensation and its determinants must begin with the equation describing pay for the managers. A general form for the compensation equation in this paper to test the before-mentioned hypotheses is:

Compensation

=f(Corporate Performance, Corporate control, Firms - specific Characteristics)

3.2. Empirical Models

Based on the basic equation in sub-section 3.1, in this section, this study constructs empirical as following:

Model 1 :

COMP=

\(\begin{array}{l} \alpha_{0}+\beta_{1} E A R N+\theta_{1} \text { Dir_ } O w n+\theta_{2} \text { Fin own }+ \\ \theta_{3} \text { Corp } \text { Own }+ \\ \sum \delta_{j} \text { control variables }+\sum \gamma_{j} \text { Industry dummies }+ \\ \sum \varphi_{j} \text { Year dummies }+\varepsilon \end{array}\)

3.3. Variable Descriptions

3.3.1. Dependent Variable: Executive Compensation (COMP)

In Japanese context, there is a wider application of cash incentives which represent the major portion of managerial compensation. In contrast, the use of stock based compensation is very limited and insignificant in the total compensation package.² Moreover, there is no publicly available data on CEO and individual compensation in Japan. Only cash compensation data, including all directors (top-executives) are available in the Nikkei NEEDS database. That’s why, this study focuses on cash compensation paid to top-executives of Japanese companies. This study measures top-executives’ cash compensation (COMP) as the sum of salary and bonus earned by all directors, deflated by total assets.

3.3.2. Research Variable: Corporate Performance (EARN)

Net income deflated by total assets (denoted in this study as “Earnings”) is commonly used proxy of accounting performance. This study applies the same proxy to measure corporate performance.

3.3.3. Research Variable: Corporate Control

In this study, corporate control can be addressed from two perspectives: managerial control and shareholder control. Greater managerial control allows managerial dominance over the pay setting process, and thus, could reasonably increase the overall payment to top-executives. Japanese Commercial Act provisions that top-executive compensation will be determined by the board of directors and the ceiling of the gross payments will be approved by the shareholders. Japanese boards are historically dominated by insiders who are also the top executives. As most of the board members are internally promoted top-executives and there is no mandatory provision of independent remuneration committee, it is pretty much possible that top-executives will have to tendency to extract more compensation in the absence of any control. Therefore, this study considers ownership by directors (denoted as “Dir_Own”) as a proxy of managerial power over pay settings. According to managerial power theory hypothesis, this study expects to find a positive significant relationship between managerial compensation and directors’ ownership.

On the contrary, according to the current ownership structures of Japanese companies, almost fifty percent of their outstanding shares are owned by financial institutions and business corporations. Several empirical studies support that institutional owners are very dominating in Japanese companies to control the overall operating and financial decisions. According to efficient monitoring hypothesis, controlling shareholders have the incentive, ability and expertise to control the behavior of the managers to extract more compensation from the business. This study expects that institutional ownership which has been defined by financial institutional ownership (denoted as “Fin_Own”) and business corporation ownership (denoted as “Corp_Own”) will have significant impact on top-executives’ compensation.

3.3.4. Control Variables

This study uses several firm-specific control variables in the empirical model to control the micro-economic or firmspecific conditions. Uchida (2006) argues that Japanese companies are more reliant on debt financing and that the agency costs of debt is a central issue in corporate governance. As a firm’s agency cost of debt increases, it reduces the tendency of the managers to get more compensation from the firm. Therefore, this study adopts the proportion of total liabilities to total assets (denoted as “Leverage”) as a proxy of agency costs of debts or debtholder monitoring.

The relationship between firm size and managerial compensation is confounding. While some researchers argue that larger firms are more inclined to give more compensation to their executives, others argue that larger firms are more subject to greater monitoring and institution control which reduce the tendency of managers to extract more compensation. It is also argued that in small and medium companies, top executives are all in all to set their compensation as large as possible, but the same is not possible in large firms. In Japanese context, Sakawa et al. (2012) finds that firm’s size negatively linked with managerial compensation. This study applies natural logarithm of total sales, as a proxy of firm size (denoted as “Size”).

This study also includes another control variable which is defined as the proportion of total operating cash flow to total assets (Cho, 1998; Davies et al., 2005), as a proxy of firm liquidity (denoted as “Liquidity”). Liquidity measure whether the level of firm’s operating cash flow has any significant impact on managerial compensation or not. In addition, industry dummies and year dummies are included to control for the industry-specific and time-specific macro-economic conditions that might have some influence on empirical findings.

3.4. Sample Selection and Source of Data

The firm-wise data is extracted from the Nikkei NEEDS database. The Nikkei NEEDS data are sourced from annual company reports. The selected non-financial firms are subdivided into industry sectors according to Tokyo Stock Exchange (TSE) industry classifications. Initial sample is adjusted to exclude the firms with following criteria: having fiscal year end other than March and having insufficient or missing data for the whole period under this study. Adjustments are also made in the sample for some variables with top and bottom 1 percent outliers. The selection process yields 4,411 firm-year observations (401 firms for the 11-years period from 2001 to 2011) for Japanese non-financial firms publicly traded on the first section and second section of the TSE. Table 1 details the breakdown of observations across different industry-sectors.

Table 1: Industry wise Sample Classifications

4. Empirical R1esults and Discussion

4.1. Descriptive Statistics and Collinearity Results

Table 2 presents descriptive statistics of sample firms. The average of total cash compensation (scaled by total assets) across entire sample is 0.4 percent. Regarding corporate performance variables, the average of net income (scaled by total assets) across entire sample is about 2.3 percent. Regarding stock ownership variables, the average of directors’ ownership is about 4 percent. Ownership by financial institutions is about 26 percent and ownership by business corporations is about 27 percent, indicating the strong control of institutional shareholders on Japanese companies. Regarding firm characteristic variables, the average of firm size (measured by firm’s sales) is 229,820 million yen. The average of leverage is about 58 percent, which is almost an identical value to that reported by Basu et al. (2007). This implies that the Japanese firms are still largely dependent on debt financing. The average of liquidity ratio, measure by operating cash flows to total asset ratio, is around 7 percent.

Table 2: Descriptive Statistics

One important consideration in the empirical research is whether the variables are subject to collinearity effect. When multicollinearity exists between two independent variables, the predictive power of a particular independent variable is reduced by the extent to which it is associated with the other independent variable. To assess for the presence of multicollinearity, tolerance value and the variance inflation factor (VIF) are examined. For brevity, this study does not report the results of multi-collinearity check. The study finds that the maximum VIF value is 2.90 for EARN. According to Hair, Anderson, Tatham, and Black (1998), VIF values greater than 10 indicate a multicollinearity problem. As VIF values for all the variables do not exceed the generally accepted threshold, this suggests that there is no multicollinearity problem among the independent variables.

4.2. Regression Results

Table 3 represents the results of regression analysis in this study. As expected, the coefficient on ‘EARN’ is significantly positive at less than the 0.01 level. The result suggests that firms with higher levels of accounting performance or net income appear to pay greater managerial compensation than firms with lower levels of net income. The results show that the coefficients of ‘Fin_Own’ as well as ‘Corp_Own’ are significant and negatively associated with the compensation level (COMP). These findings are consistent with the ‘efficient monitoring hypothesis’ which argues that institutional owners play controlling role in the firms and prevent managers from enjoying more compensation. On the other hand, the study also finds that the coefficient of ‘Dir_Own’ is significant and positively associated with the compensation level (COMP). This result is consistent with managerial power theory hypothesis which argues that the greater ownership by executive directors entails greater managerial control over pay settings, and thus, allows the managers to extract more compensation.

Table 3: Regression of the sensitivity of the level of total compensation on firm performance and corporate control

Note: (1) The t-statistics given are based on the heteroskedasticityconsistent standard errors of White(1980); (2) The symbols *, **, *** denote significance at the 10%, 5% and 1% level, respectively.​​​​​​​

With respect to the control variables, the coefficient on “Size” is significantly negative in all models, which indicates that larger firms pay less to managers than smaller firm, and vice-versa. These findings are consistent with the argument that large firms are subject to more outside monitoring and less managerial discretion that lead to lower managerial compensation. This study fails to justify any impact of costs of debt on top-executives’ compensation level as the coefficient of on “Leverage” is not significant. Moreover, the evidence shows that the “Liquidity” has no significant effect on the level of compensation.

4.3. Robustness and Further Analysis

Because the regression analysis are based on 11 years of pooled, cross-sectional data in which the same firm can appear multiple times in the sample, observations may not be independent both cross-sectionally and over different years. This procedure may involve cross-sectional (spatial dependence) and auto correlation (serial dependence) problems (Shuto & Takada, 2010). Papers by Thompson (2011) and by Cameron, Gelbach and Miller (2011) suggest a way to account correlations among different firms in the same year and different years in the same firm. To control for time-series as well as cross-sectional correlation simultaneously, this study further run pooled regressions with t-statistics based on standard errors clustered at firm and year levels (two-way clustering). The results reported in Table 4 suggest that the findings are qualitatively similar.

Table 4: Regression of the sensitivity of the level of total compensation on firm performance and corporate control (Two-way Clustering)

Note: (1) The t-statistics given are based on standard errors with firm-level clustering for arbitrary serial-correlation and yearwise clustering for contemporaneous-correlation(two-ways clustering); (2) The symbols *, **, *** denote significance at the 10%, 5% and 1% level, respectively​​​​​​​

5. Summary and Conclusion

The dearth of research exploring the relationship between top-executives' compensation and corporate control encourages the author to do research in this field based on Japanese data. Empirical findings examining the compensation-corporate control link suggest that greater director ownership signifies managerial dominance scenario which is associated with greater compensation in Japanese firms. However, controlling shareholders (financial and corporate shareholders in Japanese context) restrain the behavior of top executives to extract more compensation, and thus, has significant negative effects on compensation in Japanese companies. While prior research on topexecutives’ compensation in Japan is narrowly focused and mixed in findings, current evidence provide broader and contemporaneous insights in this aspect with more rational findings. This study definitely contributes to the existing literature of managerial compensation by addressing the roles of corporate control along with corporate performance in determining top-executives’ compensation in Japanese firms.