Description
In contemporary business corporations, the main external stakeholder groups are shareholders, debtholders, trade creditors, suppliers, customers and communities affected by the corporation's activities. Internal stakeholders are the board of directors, executives, and other employees.
Stockholm School of Economics Institute of International Business (IIB) Master’s Thesis
Internationalization of Corporate Governance in Japan
Mikael Johansson [email protected] Kristina Sawaguchi [email protected]
Abstract This thesis investigates the impact of internationalization on corporate governance in Japan. As a consequence of increasingly global competition, different systems of corporate governance come into direct contact with each other. The Japanese economy has become more international over the last two decades, and it is therefore interesting to see whether internationalization has an impact on corporate governance of Japanese companies. Using OLS regressions on a sample of 236 companies listed on the Tokyo Stock Exchange First Section, we test whether internationalization in the form of foreign presence and foreign ownership are related to board size and composition in Japanese companies. We find that foreign presence and foreign ownership are positively related to the adoption of Anglo-American board practices, i.e. smaller boards with a higher ratio of outside directors.
Keywords: Japanese corporate governance, internationalization, convergence, board structure, stakeholder, shareholder
Tutor: Dr. Jesper Edman Presentation: June 12, 2009, 10:15-12:00, Room C606 Discussant: Pascal Lindqvist (20083)
Acknowledgements: We would like to express our thankfulness to our tutor Dr. Jesper Edman for his continued support and valuable comments throughout the project. We would also like to thank Per-Olov Edlund and Håkan Lyckeborg for their help with econometrics.
Contents
1. Introduction ........................................................................................................................................... 2 1.1. Delimitation .............................................................................................................................. 3 1.2. Outline....................................................................................................................................... 3 2. Background ........................................................................................................................................... 4 2.1. The Traditional Japanese Corporate Governance System......................................................... 4 2.2. The Anglo-American Corporate Governance System .............................................................. 5 2.3. Weaknesses in the Japanese Model........................................................................................... 7 2.4. International Convergence of Corporate Governance............................................................... 7 2.5. Changes in Japanese Corporate Governance ............................................................................ 8 2.6. Internationalization of the Japanese Economy .......................................................................... 9 3. Theory and Hypothesis Development ................................................................................................. 11 3.1. Foreign Presence ..................................................................................................................... 11 3.2. Foreign Ownership.................................................................................................................. 13 4. Methodology and Data ........................................................................................................................ 17 4.1. Data ......................................................................................................................................... 17 4.2. Dependent Variables ............................................................................................................... 17 4.3. Independent Variables............................................................................................................. 18 4.4. Control Variables .................................................................................................................... 18 4.5. Quality of Research Design .................................................................................................... 19 4.6. Descriptive Statistics ............................................................................................................... 21 4.7. Recent Developments ............................................................................................................. 22 5. Model Development and Results ........................................................................................................ 24 5.1. Board Size ............................................................................................................................... 24 5.2. Percentage of Outside Directors.............................................................................................. 26 5.3. Outside Directors .................................................................................................................... 27 5.4. Quality of Models ................................................................................................................... 29 6. Analysis............................................................................................................................................... 30 7. Conclusion and Discussion ................................................................................................................. 33 8. References ........................................................................................................................................... 36 9. Appendix A ......................................................................................................................................... 40 10. Appendix B ......................................................................................................................................... 42 10.1. Heteroscedasticity ................................................................................................................... 42 10.2. Normality ................................................................................................................................ 43 10.3. Multicollinearity...................................................................................................................... 45
1
1.
Introduction
Capitalism is the dominating economic system in most of the world today. However, there are many varieties of capitalism across nations (Dicken 2007) and at the heart of the distinction between them are differences in corporate governance. Corporate governance can be split into the marked-based, shareholder-oriented system applied in Anglo-American countries and the stakeholder-oriented system used in Germany and Japan (Ahmadjian and Robbins 2005). These two systems have each had their own heyday. In the 1980s, when the Japanese economy boomed and Japanese companies captured market shares all over the world (Jacoby 2004), the Japanese system was considered to be an efficient model of corporate governance, but because of the outstanding performance of the U.S. economy since the 1990s, the Anglo-American corporate governance has become regarded as the best-practice model. The American economic upswing coincided with an economic downturn in Japan. At the end of the 1980s, the Japanese asset price bubble burst, and many Japanese banks were found insolvent due to tremendous amounts of non-performing loans. Much of the troubles that the Japanese economy has undergone for the last two decades have been blamed on corporate governance problems (Thomsen 2007). Japanese companies have been faced with demands to adopt Anglo-American shareholder-oriented corporate governance practices (Jacoby 2004) and corporate governance in Japan is now starting to change. The domestic economic crisis is one factor behind the corporate governance reforms now taking place in Japan. However, since competition in product and capital markets is becoming increasingly global, corporate governance needs to be analyzed in an international context. As companies internationalize they are inevitably forced to compete with actors from other corporate governance systems (Coffee 1999). Internationalization is, therefore, a crucial factor to take into consideration when studying the development of corporate governance. Japanese companies have become increasingly dependent on overseas markets for sales and operations, and the shareholdings by foreign investors in Japanese listed companies have grown substantially since the 1990s (Jackson and Miyajima 2007). Few previous studies have, however, investigated the impact of internationalization on corporate governance in Japan. The purpose of our essay is therefore to fill this gap in existing research. The question we aim to answer is: Does internationalization impact on corporate governance in Japan?
2
The research question is interesting to look at for several reasons. Japan, being the second largest economy of the world, makes the future of Japanese companies of general interest. Good management, which corporate governance aims to assure, is essential for firm performance and economic efficiency, and therefore for the economic welfare of society as a whole. To the best of our understanding, this thesis is the first attempt to look at how the Japanese corporate governance is influenced by internationalization. Therefore, we hope that it will contribute to the understanding of corporate governance reform in Japanese companies and at the same time on a larger scale be a case study of how globalization is an influencing force in the modern world.
1.1.
Delimitation
This study will look at the impact of internationalization on corporate governance in Japanese companies. We have limited the scope of the thesis to look at one particular governance mechanism, the size and composition of the board of directors. The reason for this is that Japanese corporate boards have traditionally been very large, with up to 50 members, and dominated by insiders (Jacoby 2004). This makes it possible to use the board of directors as a proxy to study if and how internationalization is having an impact on corporate governance in Japan. Our dataset consists of 236 of the 250 largest companies listed at the Tokyo Stock Exchange First Section, as measured by annual turnover in JPY. Due to the limited availability of data on board structure for previous years we will look at one point in time, the fiscal year 2007.
1.2.
Outline
The remainder of the thesis is organized as follows. First, the background is introduced in section 2, describing the traditional Japanese corporate governance system and comparing it with the AngloAmerican system. We then look at the internationalization of Japanese companies and the likelihood of convergence of corporate governance systems. In section 3, we review the literature and theory concerning the influence of internationalization on corporate governance practices leading us up to our hypotheses. We will then go on to describing the methodology and data in section 4. Our models and results are presented in section 5 and analyzed in section 6. A concluding discussion with some limitations of the study and suggestions for further research is found in section 7.
3
2.
Background
The basic problem of corporate governance is the “agency problem” associated with the separation between ownership and management. Managers (the agents) act on behalf of the shareholders (principals) and the question to address is how to ensure that managers do not act in their own self-interest rather than in the interests of shareholders. Several corporate governance mechanisms are available to mitigate the agency problem. These mechanisms include laws and regulation, shareholder pressure, creditor monitoring, boards, incentive systems, the market for corporate control, trust, reputation, etc. (Thomsen 2007). In this study of corporate governance in Japan, we have chosen to focus on the board of directors because of the following three reasons. Firstly, this is one of the most widely debated institutions of corporate governance (Thomsen 2008). Secondly, boards in Japan have traditionally been extremely large and dominated by insiders but are now starting to gradually change (Charkham 2005). Thirdly, board size and composition is a relatively visible and measurable governance mechanism. Corporate governance varies a lot across countries, but researchers mainly split corporate governance practices into two broad groups, the Anglo-American shareholder-oriented system and the German and Japanese systems of stakeholder-oriented corporate governance (Gilson 2001). The different governance mechanisms are used to varying degree in these systems. Especially the board of directors has had very different forms and function in the Japanese and the Anglo-American models. In the following sections, we highlight the differences between the two systems and briefly explain the role of the board in the Japanese and the Anglo-American systems, respectively.
2.1.
The Traditional Japanese Corporate Governance System
The Japanese corporate governance system that developed in the post-war period is described as a stakeholder-oriented system (Thomsen 2008). The role of the company in the Japanese system has been to balance the interests of several stakeholders, such as employees, suppliers and customers, instead of only prioritizing shareholder value (Kester 1996). Investors, typically financial institution and other corporations, used their shareholdings to support long term relationships rather than for purely financial gains (Sheard 1994). Thus, contrary to the liquid portfolio investors in the United States, owners of Japanese companies were mostly stable shareholders (Yafeh 2000). Japanese corporate shareholding also
4
involved cross ownership, or interlocking shareholdings, meaning that two firms own each other’s shares. The extensive share-interlocks practically insulated the management of Japanese firms from external takeover threats (Sheard 1994) and, thus, limited the market for corporate control (Jackson 2003). The Japanese system is also known as a bank-based system, as Japanese companies have traditionally relied on bank finance rather than equity finance for capital provision (Thomsen 2008). The Japanese financial system was characterized by close relationships between companies and banks. A company usually had a “main bank”, which was its largest lender as well as its largest shareholder (Hoshi 1994). In return for supplying finance, banks were given the right to monitor management’s use of the provided credit from inside the firm, e.g. by appointing directors to their boards. Since the market for corporate control was almost non-existing, banks played an important monitoring role in Japanese corporate governance (Jackson 2003). Complementary to the stable shareholding and close bank relationships, was the system of “lifetime” employment (Sheard 1994). Stable financing made it easier for firms facing financial distress to renegotiate its obligations and avoid corporate failure, which supported long-term relationships with stakeholders, such as employees (Abe and Hoshi 2007). In turn, employment practices like senioritybased wage and promotion systems, and investment in firm-specific skills encouraged loyal labor relations (Jackson 2003). The size and composition of traditional Japanese board of directors is, partly, an extension of this internal promotion system (Jackson 2007). Boards in Japanese companies have traditionally been very large, often with 30-40 members (Thomsen 2008), and composed primarily of senior managers promoted from within the company (Aoki 1988). This has provided an incentive for employees to aspire for a position on the board as a final goal of a long career within the same company (Jackson and Miyajima 2007). Listing rules do not require companies to have outside directors (Jackson 2003).
2.2.
The Anglo-American Corporate Governance System
In contrast to the stakeholder-oriented Japanese corporate governance, the Anglo-American system, as implemented in the US and the UK, is a market-based model, in which shareholder value creation is assumed to be the primary goal of the company. Ownership is dispersed among diversifying shareholders, mainly institutional investors and individuals (Thomsen 2008). Contrary to the Japanese owners, these investors do not hold the shares for business strategic reasons, but are mainly interested in the financial
5
return on their investment. Shares are easily traded and shareholders who are dissatisfied with management can seek to replace it or sell their shares. The latter option includes the possibility to sell the share to another group of investors that may chose to replace management, i.e. there exists an open market for corporate control (Jacoby 2001). Since shareholders are clearly separated from other stakeholder groups such as creditors, suppliers or employees, governance mechanisms that enhance monitoring and align the interest of shareholders and mangers have become important in the Anglo-American system (Kester 1996). One of these mechanisms is the election of independent outside directors. In a study by Li (1994), for the year 1987, American boards had an outsider percentage of 74 percent, compared to 9 percent for Japanese companies. The rules of both the New York Stock Exchange and NASDAQ now require listed companies to have board of directors with a majority of independent directors (NYSE 2009; NASDAQ 2009). The size of boards in American companies has also decreased over time. For large companies, the median number of directors has decreased from 14 to about 10 directors, while the median for medium-cap companies have decreased from 12 to 9 directors since 1972 (Charkham 2005). The major differences between the two systems of corporate governance are summarized in Table 1 below. Table 1 Characteristics of the Anglo-American and Japanese corporate governance systems The Anglo-American System Shareholder oriented Dispersed ownership Equity financed Many takeovers Flexible labor market Small boards of directors Majority of outsiders on the boards The Japanese system Stakeholder oriented Stable owners, cross-ownership Bank financed Few takeovers Lifetime employment Large boards of directors Insider-dominated boards
6
2.3.
Weaknesses in the Japanese Model
During the 1980s, the Japanese system of corporate governance caught researchers’ interest due to its strength in supporting economic performance and social organization (O’Sullivan 2003). The ability of Japanese corporations to build long-term relations between shareholders, banks, employees and suppliers was seen as a competitive advantage (Jackson 2003) that resulted in company growth and increased global market shares (Yafeh 2000). Proponents of the Japanese system characterized it as a superior substitute for the market for corporate control, since Japanese managers were less subject to short-term pressure from the market (Denis and McConnell 2003). However, as a result of the prolonged economic crisis, corporate governance has been much debated in Japan. The asset price bubble at the end of the 1980s and the following period of economic stagnation have been widely blamed on corporate governance problems (Thomsen 2007). A chief shortcoming of Japanese corporate governance is its low ability to control the agency problem associated with the separation of ownership and control (Kester 1996). Tight stakeholder relationships, internal promotion, as well as the insider-dominated board structure assure continuity of business strategies and long-term perspectives, but can also result in business conservatism and delay in restructuring (Jackson and Miyajima 2007). Senior managers and directors promoted from within the company are often resistant to make decisions that would cost jobs even when company profits are declining (Ahmadjian and Robbins 2005). In practice, board control of managers has been almost absent in the Japanese system. Furthermore, the large size of the boards complicates decision making (Thomson 2008).
2.4.
International Convergence of Corporate Governance
The opinions as to the scope of the convergence vary from one extreme to the other. Some researchers foresee a total convergence between systems and others think that there will be almost no convergence at all. To start at one end of the spectrum, Hansmann and Kraakman (2000) argue that the Anglo-American system is now so dominant that other systems of corporate governance can do nothing than just follow suit and that we are already seeing a de facto convergence on and “victory” of the Anglo-American system. There are different arguments as to why this convergence is taking place. Some argue that this system generates higher levels of economic efficiency, since it features liquid labor markets, an external market for skills and an emphasis on profitability over growth (Ahmadjian and Robbins 2005). Others
7
think that the convergence is a product of the political hegemony of U.S. financial interests (O’Sullivan 2003). Gilson (2001) proposes that a functional rather than a formal convergence will take place. A functional convergence can take place without changing the formal and legal frameworks of the corporate governance system. The reason for this can be that a formal convergence can be deemed impossible due to high costs and or path dependence. Functional convergence can take many different forms such as listing the company on a foreign stock exchange and thus show compliance with stricter governance principles, or setting up voluntary contractual regimes that promotes a good governance structure. It is common to these practices that they are not necessarily regulated in the national law of the country where the company is listed, but are instead undertaken on a more voluntary basis. There are also researchers who think that almost no convergence will take place. Babchuk and Roe (1999) argue that there is strong path dependence due to initial conditions in the economy. There can also be rent-seeking behavior by the parties that are being favored by the existing system, which makes change hard. They conclude that notwithstanding the powerful forces of internationalization, many important differences between different corporate governance systems will continue to exist.
2.5.
Changes in Japanese Corporate Governance
Regardless of the question whether the Japanese system will completely converge to the Anglo-American or any other system, corporate governance practices have in fact started to change in Japan. In 1997, Sony initiated corporate board reform in Japan by reducing the size of its board from 38 to 10 directors and at the same time increasing the number of outside directors. Since then, many leading companies have reduced their board sizes to speed up decision making. However, many Japanese mangers are still reluctant to introduce outside directors since they doubt that they have enough company knowledge to make valuable contributions (Jackson 2003). One force for convergence of Japanese corporate governance practices toward the Anglo-American model is, as mentioned above, the poor economic performance of the Japanese economy (Jacoby 2005). Moreover, the committed capital that used to characterize the Japanese governance model has been weakened since the 1990s (Jackson 2003). As a result of financial deregulations dependence on equity finance has increased, and corporate ownership by stable investors such as financial institutions has declined. Japanese banks were also severely hit by the crises, which greatly affected their monitoring role.
8
Furthermore, as a response to changing economic conditions and changes in the technological environment, the decision-making in Japanese firms is becoming increasingly decentralized. These new conditions have induced corporate governance reforms such as the introduction of outside directors, and separation of monitoring and management functions (Jackson and Miyajima 2007). Another important driver for international convergence of corporate governance is that competition, on both product and capital markets, has become increasingly globalized. As a consequence, the different systems of corporate governance have come into direct contact and been forced to compete (Coffee 1999). The question of what impact internationalization has on a company is one of the most discussed topics in the international business literature. However, few studies have looked into the impact of internationalization on a company’s corporate governance, especially in a Japanese context. The aim of our thesis is to fill this gap in existing research. The research question we aim to answer in this thesis is: Does internationalization impact on corporate governance in Japan?
2.6.
Internationalization of the Japanese Economy
A firm’s degree of internationalization refers to the extent to which it depends on foreign markets for customers, factors of production, and the capacity to create value, as well as the geographical dispersion of such dependence (Sullivan 1994). In the last two decades, Japan has become increasingly internationalized in a number of ways. Firstly, Japan has begun to attract inward foreign direct investment (FDI). Secondly, Japanese companies are selling more and more on foreign markets, and thirdly, foreign ownership of Japanese shares is going up. Inward FDI has become a key policy in Japan aimed at stimulating the Japanese economy (Jackson and Miyajima 2007). The Japan External Trade Organization (JETRO), that only used to promote exports, is now also working to attract inward FDI to Japan in order to raise the level of the inward FDI stock to 5 percent of GDP (JETRO 2009). This is an ambitious goal for a country with a current level of about 3 percent, but the figure is still low compared to the United States with 15.1 percent and the United Kingdom with 48.6 percent (UNCTAD 2008). The reason for Japan to attract more inward FDI is to spur reform and innovation in Japanese companies. Foreign companies coming to Japan are often seen as catalysts for change that bring ideas to Japan, ideas that then are adopted and get adapted to the Japanese system (Economist 2007). Mergers and acquisitions are also increasing steadily and Japanese companies
9
are realizing that tying up with foreign companies offers a way of becoming part of a global strategy that they might not be able to achieve on their own (Muramatsu 2001). Alongside inward FDI, Japanese companies have become increasingly internationalized, with overseas production facilities and operations (Jackson and Miyajima 2007). During the decades following the World War II, Japan enforced quite severe trade restrictions, but in the middle of the 1960s these were relaxed and Japanese exports started to grow substantially (Flath 2005). During the 70s and 80s Japanese companies were very successful on the international export markets, to the degree that trade conflicts arose, especially with the U.S. In the late 1980s, the focus shifted somewhat as Japanese overseas investments exploded and Japan rapidly became a major source of FDI worldwide (Beamish and Inkpen 2001). One of the reasons for this shift is the appreciated Japanese yen that made the production costs in Japan high (Itoh 2000). Japanese subsidiaries opened up all over the globe, but they were criticized for mainly using Japanese expatriates for manager positions. This is starting to change though, indicating that Japanese companies are taking a more global approach by recognizing the value of empowering local management (Beamish and Inkpen 2001). The share held by foreign investors in publicly listed companies in Japan has increased steadily over the last two decades. Measured by market value, the share of foreign ownership in all publicly listed companies in Japan increased from 6.0 percent in 1992 to 13.4 percent in 1998, and to 27.6 percent in 2008 (Tokyo Stock Exchange 2005; Tokyo Stock Exchange 2009). A majority of the foreign investors are institutional investors, mainly American and European funds. Individuals represent only a negligible fraction of the foreign investors, and strategic investments by foreign corporations are also small compared to portfolio investments. Inward FDI influences the economy directly through capital formation, tax revenue etc., but it can also indirectly influence the host economy and the conduct of locally owned firms. Historically, inward FDI has played an important role in the restructuring process in Japan (Blomström, Konan and Lipsey 2001), and, thus, has probably had an indirect influence on the corporate governance structure in Japanese companies. However, this form of indirect influence is extremely difficult to measure and in this thesis we have therefore chosen to focus on the two other forms of internationalization, i.e. Japanese presence on overseas markets and foreign ownership of Japanese shares. The combination of these two measures of internationalization has not been applied to Japanese corporate governance by any previous study.
10
3.
Theory and Hypothesis Development
In this section we will review the theories of how foreign presence and foreign ownership may influence a company’s corporate governance, leading to our hypotheses.
3.1.
Foreign Presence
Foreign presence of a business can be measured in a number of ways, but we have chosen to concentrate on two measures of foreign presence, namely foreign sales and subsidiaries abroad. There are many reasons for companies to internationalize their business. Since competition today is increasingly global, firms are no longer competing only with their domestic rivals (Dicken 2007). Hennart and Park (1994) describe the international expansion decision as a two-step process. First, the firm has to decide whether to produce at home or abroad. If the firm decides to produce abroad, it then has to make a decision of how to organize the foreign production, either by licensing or by integrating into a production subsidiary. They mention location factors such as transportation costs, economies of scale, production costs and presence of customers and governance factors such as knowledge transfer and trademark reputation as some of the main variables that are important when a firm decides to internationalize. The decision to internationalize the business may also be a response to the actions or expected actions by rivals (Knickerbocker 1973). Entering a foreign market will almost inevitably mean that the complexity of the firm increases (Melin 1992). According to Bartlett and Ghoshal (1998) a firm needs certain organizational capabilities in order to handle the complex and often contradicting demands that international expansion implies. The requirements to learn about and develop new capabilities in environments, where language, culture, buyers, suppliers and political and legal systems may be different, increase uncertainty significantly. A company’s corporate governance structure is an important factor for successfully handling the increased complexity and uncertainty that internationalization brings about (Sanders and Carpenter 1998). Many studies have investigated the effect of increased complexity and uncertainty on board structure, and most studies find a positive relationship between the increased uncertainty and board size and board composition. Adding members to the board might be a way of increasing the viewpoints of the board, but can also slow down the decision making process (Mitchell 1997). Birnbaum (1984) finds a positive relationship between increased environmental uncertainty and both board size and the number of outside directors. He argues that adding board members, some of which might be outsiders will make boards
11
more able to make forecasts about the organization’s task environment, and that outsiders increase the representation of important external constituent groups. Pearce and Zahra (1992) argue along the same line, saying that increased environmental uncertainty will lead companies to adopt larger boards with more outsiders. In their sample of 119 Fortune 500 companies, they find positive correlations between external uncertainty and board size and proportion of outsiders on the boards. A study by Sanders and Carpenter (1998) is one of the few that have investigated the effect of internationalization on different aspects of a firm’s corporate governance structure. Using a sample of American companies, they assume that increased internationalization will lead to bigger boards with fewer outsiders. More insiders will lead to better board control, they argue. They find support for their hypothesis that the degree of internationalization is positively correlated to board size. However, they do not find support for the hypothesis that the internationalization is negatively related to the ratio of outside directors on the board. They conclude that this result is due to the fact that the increased complexity that internationalization implies is strong and that outside directors can contribute to a successful managing of this complexity. Another factor that might influence a company to change its corporate governance structure as it becomes more dependent on international markets is the need for convergence and adaptation of international “best-practices”. In order to compete successfully in the global economy, actors in different business systems try to interpret each other’s “best-practices” and strategically incorporate these practices in their own routines and organizational processes (Wei-Chung Yeung 2000). According to Coffee (1999), those firms seeking to grow in size to a global scale are likely to adopt “higher” governance practices already observed in the United States. For Japanese companies this would imply increased internal transparency and implementation of the kind of global standards that are familiar to stakeholders abroad (Jackson and Miyajima 2007). When it comes to outside directors on Japanese boards and the impact of internationalization, we assume that increased internationalization will also in Japan be positively related to the number of outsiders on the boards due to fact that increased complexity of international operations will put pressure on the firms to make use of the external competence that outsiders on the board bring to the company. This is in line with what the literature has found so far. Looking at the board size in Japanese companies, we think that internationalization might have a different effect than previous studies have found, mainly because the Japanese boards have played a rather
12
different role than their American counterparts, and because they have traditionally been larger. Even though increased internationalization will mean increased complexity for Japanese companies as well, we think that a more streamlined board of directors is more likely to be able to make the fast decisions that this complexity requires. The Japanese case is therefore different compared to the American case as investigated by Sanders and Carpenter (1998). This implies that the marginal effect of adding an additional board member to an already large board is miniscule at the best, but most likely will have a negative effect on the functionality of the board. For Japanese companies, decreasing the size of the board implies moving towards best-practice. This leads us up to our hypotheses on how internationalization in the form of foreign presence affects the boards of directors of Japanese companies. Hypothesis 1a: Foreign presence is negatively correlated with board size in Japanese companies. Hypothesis 1b: Foreign presence is positively correlated with the percentage of outsider directors in Japanese companies.
3.2.
Foreign Ownership
A recent trend among many major companies worldwide is that they are becoming more international not only in their scope of business activities but also in their capital structure. The incentive for companies to internationalize their ownership structure is to lower the cost of capital (Eun and Resnick 2007). However, the increase in international equity investments has also been driven by the rise of institutional investors, especially pension funds, seeking to diversify their portfolios by overseas investments (Ahmadjian 2007). These institutional investors have become dominant players in the financial markets and an increasingly important factor affecting corporate governance worldwide (Gillan and Starks 2003).The trend started in the United States during the 1980s and 1990s, where American institutional investors revolutionized corporate governance and contributed to the development of the shareholder value model. Encouraged by the success they then turned to other markets (Jacoby 2007). In Japan, the shareholdings by foreign institutional investors have, as mentioned above, increased dramatically since the 1990s. The entry of American and European institutional investors in Japan has inevitable implied a “clash” between the shareholder-oriented Anglo-American system and the Japanese system of relational governance. Standing apart from the Japanese stakeholder system and without any ties to Japanese companies other than their equity investment, these investors have been pushing for
13
corporate governance reform in the direction of the Anglo-American shareholder-oriented model. The introduction of outside directors on corporate boards is one of the reforms that have been particularly emphasized by foreign institutional investors in Japan (Ahmadjian 2007). The American Chamber of Commerce in Japan (ACCJ) has, for example, tried to persuade the Japanese government to introduce listing rules that require listed companies to increase the proportion of independent outside directors to at least one third of their boards. According to the ACCJ, many foreign fund managers have come to regard the reform of insider-dominated boards as an important indicator of how serious Japan actually is in its efforts to reform corporate governance (ACCJ 2009). Shareholders have the possibility to affect companies in several ways. If shareholders are dissatisfied with the performance of the board of directors they can choose between selling their shares, “exit”, or hold their shares and express their dissatisfaction, using “voice”. Exit can affect companies since heavy institutional selling can put downward pressure on the stock price, and thus make the company more vulnerable for takeovers. Institutional selling can also be interpreted as bad news, which trigger sales by other investors and further depresses the stock price (Gillan and Starks 2003). In Japan, foreign investors have been much more active buyers and sellers of stocks than domestic investors, and have therefore had a larger influence through the exit mechanism than the shareholding percentages would suggest. Even though Japanese managers have traditionally been more concerned with market share and growth rather than the share price, they have recently started to pay more attention to share price (Ahmadjian 2007). Voice can be exercised by shareholders either formally through the use of voting rights, or by expressing opinions through more informal channels. Foreign investors are increasingly using their voting rights since the early 2000s as Japanese voting rules have been made easier. However, foreign investors in Japan seem to have been more influential through less formal channels such as analyst and shareholder meetings as well as private meetings with CEOs (Ahmadjian 2007). CalPERS, a giant California-based pension fund, was one of the first foreign institutional investors to make major investments in Japan. As a wellknown advocate of shareholder value in the United States, CalPERS has also been a major promoter of corporate governance reform in Japan. In a case study of CalPERS, Jacoby (2007) concluds that the fund did have an effect on governance practices related to transparency, but that it was less successful in changing board structure, takeover norms and executive compensation. Foreign investors can also affect a firm’s corporate governance through the indirect supply-demand effect, i.e. firms may be motivated to improve their corporate governance in order to attract foreign capital (Gillan and Starks 2003). In an empirical study of outside directors in Korean companies, Rhee
14
and Lee (2008) find that the growth of foreign ownership is positively affected if a higher proportion of outside directors hold advanced foreign degrees, have former or current affiliations with governmental organizations, or have job experience in the same industry. In other words, board directors can also have a signalling role when stakeholders face significant information asymmetries, as for example in the case of foreign investors. The size and composition of the board of directors then provide an observable secondary source of information for foreign investors with limited information about domestic firms (Rhee and Lee 2008). This view is supported by other studies concluding that governance is a significant factor in the investment decision of institutional investors. In a McKinsey survey of more than 200 institutional investors worldwide, 80 percent of these investors stated that they would pay more for the shares of a well governed company than a poorly governed company with similar financial performance. A well governed company was defined as one that, among other things, has a majority of outside directors with no management ties on its board (Coombes and Watson 2000). There are few quantitative studies on the relationship between board structure and foreign shareholdings in Japanese companies. In an analysis based on a corporate governance index measuring how close a firm adheres to Anglo-American corporate governance standards, Ahmadjian (2007) finds that firms scoring high on board structure and function had higher levels of foreign ownership than firms with low scores. The corporate governance index used in the analysis is generated by the Japan Corporate Governance Research Institute on the basis of email surveys to all Tokyo Stock Exchange First Section companies in 2002 and 2003. However, the response rate was rather low (201 of 1,523 companies in 2003). In 2006, the Tokyo Stock Exchange (TSE) decided to require all listed companies to prepare a standardized corporate governance report, which means that every company is now obligated to disclose information on several corporate governance issues, including the number of outside directors. According to an analysis of the first reports, conducted by TSE, the percentage of companies with outside directors and the number of outside directors increases as the foreign shareholding ratios increase. The analysis also shows that as the foreign shareholding ratio increases the total number of directors increases, implicating that the percentage of outside directors does not necessarily increase with the foreign shareholding ratio. As pointed out in the report the reason that the number of outside directors increase as the foreign shareholding ratio increases may be that larger companies, which have a larger number of directors, have relatively larger foreign shareholding ratios (Tokyo Stock Exchange 2007). Based on the above studies we want to test whether internationalization in the form of foreign ownership have an impact on board size and the ratio of outside directors in Japanese companies. As pointed out
15
before, Japanese boards have traditionally been very large and insider dominated. Since foreign investors advocate more Anglo-American corporate governance practices, we expect board size to decrease and the ratio of outside directors to increase as the foreign shareholding ratio increases if foreign investors have an impact on board size and composition. Hypothesis 2a: Foreign ownership is negatively correlated with board size in Japanese companies. Hypothesis 2b: Foreign ownership is positively correlated with the percentage of outside board directors in Japanese companies.
16
4.
Methodology and Data
In order to test the hypotheses, we have chosen a quantitative approach with a large sample of companies. The advantage of quantitative research is that it gives rise to statistically meaningful results. We have run OLS regressions with board size and the percentage of outside directors as dependent variables in two different models. The OLS regression analysis allows us to test the impact of several explanatory variables simultaneously. We have used SPSS statistical software for the regressions. In the sections below, we explain the data that we are using, and how it was collected. We also present descriptive statistics of the data. Finally we have included a smaller study of the historical developments in a limited sample of companies.
4.1.
Data
Our initial data set consisted of the 250 largest companies, based on revenue in the fiscal year 2007, listed on the First Section of the Tokyo Stock Exchange. In total 2,370 companies are listed at the Tokyo Stock Exchange and of these 1,719 companies are listed at the First Section, which is the section for large companies with a market capitalization of more than JPY 50 billion. We collected this data from Bureau van Dijk’s ORBIS database. The motivation for choosing the largest companies is that most of the corporate governance literature also tends to focus on the largest companies (Thomsen 2008). We excluded banking and insurance companies, since these companies are subject to stricter regulations. After the exclusions we ended up with a data set of 236 companies.
4.2.
Dependent Variables
As dependent variables we have board size measured as the number of board directors (BSIZE), the number of outside directors (OUTDIR) and the ratio of outside directors of total directors (ODTD). Data for these variables have been collected from the Tokyo Stock Exchange (TSE) homepage. Since 2006, all companies listed on the TSE are required to prepare a corporate governance report, in which they disclose the size of their boards as well as the number of outside directors (Tokyo Stock Exchange 2007). In our data set, we have included all directors that according to TSE are classified as outside directors regardless of their relationship with the company in question.
17
4.3.
Independent Variables
Foreign Ownership (FOROWN) This variable captures the percentage of outstanding shares that are held by non-Japanese investors. The data were collected from the companies’ annual reports downloaded from their homepages. We primarily used the English versions, but collected data from the Japanese language versions when English annual reports were not available. For most of the observations, there was no split between the holdings of foreign institutions and foreign individuals, but, as mentioned above, the vast majority of Japanese shares held by foreigners are held by foreign institutions. Our screening of the annual report also strongly supports this.
Foreign Presence (FS_OS) In order to measure foreign presence we constructed a combined variable by adding foreign sales as a ratio of total sales (FSTS) and the number of overseas subsidiaries as a ratio of total number of subsidiaries (OSTS). This is done in order to test the total effect of these two aspects of foreign presence. The variable can theoretically take values between 0 and 2, where 0 indicates that the company has no overseas subsidiaries and no foreign sales. A composite variable is more likely to correctly measure a firm’s degree of internationalization and also enables better error control (Sullivan 1994). The method of adding measures of internationalization has also been used by for example Sanders and Carpenter (1994). The data for total sales of each company were collected from ORBIS database, while the data for foreign sales were collected from the annual reports of the companies. Foreign sales are defined the amount of sales generated by the companies’ foreign operations and, which is not the same as the export sales. Export sales means that the factors of production are located in Japan. For total number of subsidiaries as well as the number of overseas subsidiaries we used the data available in ORBIS. The definition of a subsidiary is a company in which the parent company holds more than 25 percent of the shares.
4.4.
Control Variables
Dispersion (DISP) Dispersion measures the total number of countries in which the company has a subsidiary. We used the same data source as was used for overseas subsidiaries. Given that the ratio of overseas subsidiaries of
18
total subsidiaries can be very high for companies with few total subsidiaries and that some companies might have most of their overseas subsidiaries in one or a few countries, a dispersion variable complements the FS_OS measure of foreign presence. We assume that this variable will be positively related to ODTD and negatively related to board size.
Assets (LNASSETS) Previous studies have found positive relations between firm size and board size as well as the percentage of outside directors (Pearce and Zahra 1992; Li 1994). We therefore want to control for firm size in our regressions. The total assets of the companies in 2007 were collected from ORBIS. The natural logarithmic form of the variable was used in order to make interpretations more understandable.
Establishment Dummy (EST_DUM) We included an establishment dummy, which takes the value 1 if the company was incorporated after 1997 and 0 otherwise. The reason for including this variable is that corporate governance reform in Japan is generally considered to have started in 1997 (Shishido 2007). This was also the year that Sony initiated board reforms in Japan by drastically reducing the size of their board and simultaneously increasing the number of outsiders. We assume that companies founded after this event are more likely to adopt a Western style corporate governance system. Data on the year of incorporation was collected from ORBIS.
Solvency Ratio (SOLV) Solvency ratio, defined as shareholders funds divided by total assets, was used as a control variable, the underlying argument being that companies with a high solvency ratio are more dependent on equity market for financing and thus more exposed to pressure from investors to change corporate governance (Pearce and Zahra 1992). The data come from ORBIS.
4.5.
Quality of Research Design
According to Yin (1994), four criteria are commonly used in order to analyze the quality of the research, namely construct validity, internal validity, external validity and reliability. We will discuss all four of these and relate them to our study. The explanations of the different criteria below are all based on the definitions used by Yin (1994).
19
Construct Validity Construct validity refers to the selection of a correct set of measures for the research question. Boards of directors are a major part of any corporate governance system, and we therefore believe that they can well be used as a proxy in order to study the corporate governance system as a whole. Since the size and composition of Japanese boards have traditionally displayed some unique characteristics, we think that this is a good way of seeing whether Japanese corporate governance is converging on Anglo-American practices. Foreign presence and foreign ownership are commonly used measures of internationalization. Therefore we conclude that the construct validity is high. Internal Validity Internal validity concerns the establishment of causal direction between variables. Since we have done a cross-sectional study, we have not been able to test causal relationships between our variables. This could have been remedied by looking at panel-data. We have used a small selection of companies in order to see some trends that can give us some clue as to in what direction the causality is going. When it comes to other influencing factors on our dependent variable, we have tried to control for as many additional variables as our data allows for. Therefore we regard the internal validity to be satisfactory. External Validity External validity concerns whether the findings are generalizable outside the study sample. Since Japan is somewhat of a rara avis when it comes to corporate governance, the generalizability of our results to other countries might not be very high. However, we think that our sample is representative for other large, listed companies in Japan. Due to this we deem the external validity to be fair. Reliability Reliability concerns the possibility to reiterate the study and obtain the same results. The goal is to minimize errors and biases in a study. We have clearly documented the data collection and the regression analysis procedures. The data we have used are straight-forward and readily available in databases and annual reports. We therefore conclude that the reliability of our study is high.
20
4.6.
Descriptive Statistics
Table 2 Descriptive statistics of variables
Standard deviation 4.54
Variable BSIZE
Unit Discrete number Discrete number Ratio Ratio Discrete number Ratio MJPY Ratio Dummy
N 236
Minimum 4
Maximum 30
Mean 12.22
OUTDIR ODTD FS_OS DISP FOROWN ASSETS SOLV EST_DUM
236 236 236 236 236 236 236 236
0 0 .00 .00 .03 187,555 -.03 -
12 .80 1.57 55 .69 32,458,320 .84 -
1.48 .13 .45 8.69 .25 2,085,900 .36 -
1.63 .15 .41 10.43 .12 3,201,410 .16 -
Table 2 shows a summary of descriptive statistics for all variables. The asset variable is presented in million JPY instead of in its logarithmic form in order to give the reader a better indicator of firm size. We use the natural logarithm of the asset variable, LNASSETS, when running the regressions. In our sample of 236 companies, the average board size is 12.2 directors and the average ratio of outside directors of total directors is 13 percent for this particular point in time, fiscal year 2007. An average board size of 12.2 directors is not to be regarded as extremely large. For the sample of 192 listed Japanese companies during the period 1990-1999 used in a study by Abe and Jung (2004) the average board size was 18 directors, which is considerably higher. The average board size in the sample of 258 listed U.S. companies in 1992 used by Sanders and Carpenter (2004) was 12.3 directors, which is approximately the same as for our sample. However, the proportion of outside directors in Sanders and Carpenter’s sample was 76 percent, i.e. on average a majority of the directors were outsiders. The average
21
proportion of 13 percent outside directors in our sample indicates that board composition in Japanese companies is still far from Anglo-American standards.
4.7.
Recent Developments
In order to assess the sequence of events, we randomly chose 20 companies from our sample that had at least one outside director and then searched in the annual reports available on the company websites for data on foreign ownership, total number of board directors and outside directors. The measure for number of outside directors used is not the same as used in our full sample since TSE does only disclose the most recent information. Therefore, we used the information regarding outsiders as given in the annual reports. Companies were dropped if they did not have annual reports dating back to at least 2002 or if the relevant information was not available. The remaining sample contained 13 companies. See Appendix A for graphs. On average the ratio of foreign shareholdings increased from 17.3 percent to 31.2 percent for these 13 companies over the period 2002-2007. At the same time the number of directors decreased from 17.9 to 13.9, the number of outside directors increased from 0.85 to 3.31 and the ratio of outside directors increased from 6.2 percent to 25.3 percent. As can be seen in Figure 1 below, the average trends in foreign ownership, number of board directors, outside directors and the ratio of outside directors show a simultaneous increase in the ratio of foreign ownership and the ratio of outside board directors. The increase in the ratio of outside directors is a result of both a decreasing trend in the total number of directors and an increase in the number of outside directors.
22
Figure 1 Historical trend average, 13 random companies (percentage and number on y-axis)
23
5.
Model Development and Results
Table 3 Correlations OUT DIR FOR OWN LNAS SETS EST_ DUM
BSIZE BSIZE OUTDIR ODTD FS_OS DISP FOROWN LNASSETS EST_DUM SOLV -.004 -.249 -.080 .073 -.069 .299 -.148 -.046
ODTD
FS_OS
DISP
SOLV
.910 .155 .130 .241 .278 .063 .053 .205 .098 .256 .141 .097 .135 .595 .363 .118 -.045 .358 .306 .438 -.065 .058 .313 -.060 .304 -.012 -.105 -.010
Table 3 presents a correlation matrix of the variables in order to show that multicollinearity is not a big problem in our sample.
5.1.
Board Size
Based on the discussion in our theory and data sections, we end up with the following model for board size. Board Size Full Model
BSIZE
7 1 2
FS _ OS
3
FOROWN
4
DISP
5
LNASSETS
6
EST _ DUM
SOLV
u
We started by running the full regression model, including all independent variables. The results are summarized in column “Full Model” of Table 4 below. The full model yielded an R-square of 0.152 and
24
an adjusted R-square of 0.130. Three of the variables were found significant at the 5 percent level, namely FOROWN, LNASSETS and EST_DUM. We continued by removing the insignificant variables, one at the time, and then compared the adjusted R-square for the different models. The adjusted R-square is a better measure for the explanatory power than R-square when comparing models with the same dependent variable, since it takes the degrees of freedom into account. Removing the variable DISP, which was highly insignificant, proved to render us with the model with the best adjusted R-square. This final model is presented below. Board Size Final Model
BSIZE
6 1 2
FS _ OS
3
FOROWN
4
LNASSETS
5
EST _ DUM
SOLV
u
Table 4 Board size
Dependent Variable: BSIZE Constant
Full Model -10.405** (.026)
Final Model -10.675** (.013) -1.030 (.172) -6.884*** (.009)
FS_OS
-1.115 (.237)
FOROWN
-6.893*** (.009)
DISP
.006 (.880)
LNASSETS
1.721*** (.000)
1.742*** (.000) -2.692** (.010) .024 (.228) 236 .147 .134
EST_DUM
-2.685** (.011)
SOLV
.024 (.226)
N R-square Adjusted R-square
236 .152 .130
p-values: ***<1%, **<5%, *<10%. Significance levels in brackets.
25
The R-square of the final model, 0.147, is slightly lower than in the full model, but as described above this is compensated for by a higher adjusted R-square, which indicates that the significance of the remaining variables are higher. The R-square is rather low, but in line with what is found in similar studies that we have covered in the theory section. FS_OS has the predicted negative sign, but is not significant at the 10 percent level. Still, a p-value of 0.172 is not extremely high. FOROWN has a negative effect on board size as predicted and is significant on the 1 percent level. The control variables LNASSETS and EST_DUM also have the predicted signs and are significant on the 1 percent level.
5.2.
Percentage of Outside Directors
We use the same full model when testing the influence of internationalization on the percentage of outside directors in relation to total directors. The full model is presented below. ODTD Full Model
ODTD
7 1 2
FS _ OS
3
FOROWN
4
DISP
5
LNASSETS
6
EST _ DUM
SOLV
u
Coefficients and significance levels for the variables are presented in Table 5. Two variables, FS_OS and FOROWN, were significant at the 5 percent level. Using the same methodology as with the board size regression above, we removed the insignificant variables one by one and compared the adjusted Rsquares for the different models. With ODTD as the dependent variable, it proved that the removal of SOLV gave us the highest adjusted R-square. Even though DISP is found insignificant, we keep it for the sake of improved adjusted R-square. The final model is presented below. ODTD Final Model
ODTD FS _ OS FOROWN DISP LNASSETS EST _ DUM u
1
2
3
4
5
6
26
Table 5 Outside directors of total directors
Dependent Variable: ODTD Constant
Full Model -.197 (.206)
Final Model -.179 (.234) .071** (.017) .231*** (.007) -.002 (.169) .017 (.132) .062* (.077)
FS_OS
.067** (.035)
FOROWN
.221** (.012)
DISP
-.002 (.198)
LNASSETS
.018 (.121)
EST_DUM
.062* (.079)
SOLV
.000 (.651)
N R-square Adjusted R-square
236 .105 .082
236 .105 .085
p-values: ***<1%, **<5%, *<10%. Significance levels in brackets. Also in this model, the R-square is fairly low at 0.105, but as mentioned above, similar studies get Rsquares in the same range. Both explanatory variables, FS_OS and FOROWN, have the predicted positive signs and are significant at the 5 percent level and the 1 percent level, respectively. EST_DUM is positively related to ODTD as expected and significant at the 10 percent level. The coefficient of LNASSETS has a positive sign as predicted, but is not significant.
5.3.
Outside Directors
We have not hypothesized about the relation between internationalization and the total number of outside directors. Nevertheless, since the ratio of outside directors on the board (ODTD) is related to board size, we want to test whether internationalization has an effect on the absolute number of outside directors. A
27
high ODTD could be the result of a small board size and does not necessarily imply that there are many outsiders on the board. We run the same regression for OUTDIR as we did with ODTD. OUTDIR Final Model
OUTDIR FS _ OS FOROWN DISP LNASSETS EST _ DUM u
1
2
3
4
5
6
Table 6 Outside directors
Dependent Variable: OUTDIR Constant
Full Model -5.255*** (.002)
Final Model -5.304*** (.001) .549* (.093) 1.918** (.038) -.017 (.217) .441*** (.000) .451 (.237)
FS_OS
.562 (.104)
FOROWN
1.947** (.042)
DISP
-.017 (.217)
LNASSETS
.439*** (.000)
EST_DUM
.452 (.237)
SOLV
.000 (.908)
N R-square Adjusted R-square
236 .121 .098
236 .121 .101
p-values: ***<1%, **<5%, *<10%. Significance levels in brackets. As for the model with ODTD as the dependent variable, the removal of SOLV gave the model with best adjusted R-square. The final model has an R-square of 0.121. FS_OS is positive and significant at the 10 percent level, while FOROWN is positive and significant on the 5 percent level. Both internationalization variables thus have the expected signs.
28
5.4.
Quality of Models
We have controlled the quality of the regression models with a number of tests. White’s estimated standard errors do not change much from the OLS standard errors indicating low heteroscedasticity. Multicollinearity does not seem to be a problem, since we have low VIF-values in combination with low pair-wise correlations and low R-square values. Finally our histograms and Jarque-Bera statistics indicate that the residuals might not be normally distributed, but this is not perceived as a major issue since our sample size is quite large. For further explanation of the quality tests, see Appendix B.
29
6.
Analysis
The regression results do not support Hypothesis 1a, i.e. that foreign presence is negatively related to board size as we had predicted. Nevertheless, the result is still interesting since we get the predicted negative sign, which indicates that internationalization in the form of foreign presence is at least not positively related to board size as was the case in U.S. companies according to previous studies (Sanders and Carpenter 1998). A possible interpretation of this result is that foreign presence has dual effects on board size in Japanese companies, since Japanese boards tend to be quite large from the outset. On the one hand, the increased complexity that foreign presence implies requires companies to add more directors to their boards in order to bring more knowledge to the boards and effectively monitor management as argued by Sanders and Carpenter (1998). On the other hand, higher foreign presence means that the company is more exposed to global competition and thus needs to conform to global practices regarding for example corporate governance, i.e. smaller boards. From our results we conclude that the latter effect is the dominating one. We predicted that foreign presence would be positively related to the ratio of outsiders on Japanese boards (Hypothesis 1b), which is supported by our results. This is in line with the literature that has investigated the effects of foreign presence on outside directors. We think that Japanese companies are adding outsiders to their board for two reasons as foreign presence increases. Firstly, the argument of outsiders being more able to deal with the complexity that foreign presence brings probably holds for Japanese companies as well. In addition to this, Japanese companies with a high degree of foreign presence are more likely to adhere to Anglo-American practices of more outsider-dominated boards. We therefore see a double effect of foreign presence working to push up the ratio of outsiders. In order to make sure that a high ratio of outside directors is not only explained by a small board size, we also ran a regression with the absolute number of outside directors as dependent variable. The result shows that the total number of outside directors is also positively correlated to foreign presence, and therefore corroborates our findings that the ratio of outside directors being high is not merely due to a small board size, but is in fact a result of a larger number of outside directors as well. One of our strongest and most interesting results is the effect that foreign ownership has on the size of Japanese boards. The variable is negative and significant on the 1 percent level, which supports Hypothesis 2a. This contradicts the TSE Whitepaper (Tokyo Stock Exchange 2007), which finds a positive relationship between foreign shareholdings and board size. It concludes that this might be
30
explained by the fact that foreign investors tend to invest in larger companies that have relatively large boards. Our correlation analysis confirms that large companies generally have large boards, and that foreign ownership is positively related to firm size. However, when controlling for firm size, we find a negative correlation between foreign ownership and board size. By this we can conclude that companies with a high degree of foreign ownership do in fact adopt more Anglo-American corporate governance practices in terms of board size. We also find strong support for Hypothesis 2b, that the foreign shareholding ratio is positively related to the percentage of outsiders on the board. Foreign ownership is also positively correlated to the absolute number of outside directors. From our cross-sectional analysis we cannot prove the direction of causality between foreign ownership and the adoption of Anglo-American board practices. Foreign investor might have been able to push companies to change their practices, but it is also possible that foreign investors tend to target companies that already have reformed their board structure. However, despite the endogenous nature of the relations that make it difficult to establish direct causality, foreign investor can be said to have influence over corporate governance either way. On the one hand, companies may be motivated to improve their corporate governance in order to attract foreign capital, and on the other hand, increased investment by foreign institutional investors may provide the companies with the power to enforce governance change (Gillan and Starks 2003). In our small sample of companies with data for the period 2002-2007, we can see that the ratio of foreign ownership and the ratio of outside directors increase almost simultaneously on average (Figure 1). Graphs of the trends for individual companies show somewhat disperse outcomes (see Appendix A). In some cases, however, the increase in foreign ownership clearly precedes the increase in the ratio of outside directors, which indicates that foreign investors may have influenced the companies to reform their boards after becoming shareholders. This was the case in, for example, KDDI Corporation, Honda Motor Co Ltd and NEC Corporation. Internationalization in the form of foreign ownership seems to have a greater impact on board size and board composition than foreign presence. One explanation for this might be the contradicting effects of foreign presence on board structure discussed above. A more plausible explanation, though, is that foreign investors are a much stronger force pushing the companies to adopt Anglo-American governance standards. The influence of foreign investors either through voice or through the threat of exit is more tangible for companies, than the need to conform to these standards due to a presence on foreign markets. We think that the influence through foreign presence might be strong in the longer run, but nevertheless it is a more indirect pressure compared to the power that shareholders wield. An alternative explanation is
31
that Japanese companies’ tend to have a low degree of local adaption (Bartlett and Ghoshal 1998). Japanese subsidiaries have traditionally not been very good at utilizing local competence, as they were mainly using Japanese expatriates for manager position. This could indicate that Japanese companies are, to some extent, immune to the impact that overseas subsidiaries have. Our dummy variable for companies established after 1997 is negatively related to both board size and positively related to the ratio of outside directors. In other words companies established after 1997 tend to have smaller boards with a higher ratio of outside directors. One reason for the negative relationship with board size is that younger companies are generally smaller and therefore not as prone to have large boards. However, another explanation is that companies founded after 1997, when corporate governance reform had already begun, are more likely to adopt Anglo-American board practices from the outset. Our dispersion measure is found insignificant and it does not have the predicted sign. It can be the case that the companies with a high dispersion only have sales offices in many countries without being particularly internationalized. Although not included in our results, we tested to split the two variables foreign sales of total sales and overseas subsidiaries of total subsidiaries, but these two measures separately gave us inconclusive results and we therefore chose to use the composite measure as suggested by the literature. This implies that we can not conclude what form of foreign presence influence what, but that the total effect is as expected.
32
7.
Conclusion and Discussion
This study has added some insights to the understanding of corporate governance in Japan and the impact of internationalization on corporate governance. Our results show that foreign presence and foreign ownership are correlated to board structure in Japanese companies. Based on this we draw the conclusion that internationalization does have an impact on corporate governance in Japan. No previous study on Japanese corporate governance has used the foreign presence measures that we use in this thesis. For our sample of Japanese companies, we find that foreign presence is negatively correlated to board size (even though the results are not statistically significant), which contradicts earlier findings for American companies (Sanders and Carpenter 1998), but is in line with our hypothesis. We also find a negative relationship between foreign ownership and board size, when controlling for firm size, which contradicts the study by Tokyo Stock Exchange (2007). Although we cannot make general statements about convergence of the Japanese system as a whole, the fact that we have observed changes in the board structure is an indication that the Anglo-American system is gaining ground in Japan. Coffee (1999) argues that firms aiming to compete on a global scale need to adopt global best-practices. In the Japanese annual reports that we used when collecting data for the econometric analysis, we have found some illustrating quotes on this line: “Ideally, companies must adapt their management supervision and business execution functions to global standards, while retaining sound corporate philosophies and traditions. If it is deemed best to hire external directors and other U.S.- style systems then such steps should be taken.” (Minolta 2002) “IHI conducted a study on the best management practices to establish a corporate governance structure befitting a global corporation and a system able to flexibly respond to sudden changes in the operating environment.” (IHI 2003)
The quotes above point to the kind of functional rather than formal convergence that, for example, Gilson (2001) is referring to. This shows that even if the corporate laws are not changing very quickly in Japan, companies find it essential to adjust their systems to more efficient forms of governance.
33
In the future, as sufficiently many Japanese companies reform their corporate governance practices, we could expect a spill-over effect on domestic companies although they are not directly exposed to the pressure of internationalization. Smaller boards with higher ratios of outside directors would then become the norm, or “best-practice”, even in Japan. A question that remains is why corporate governance reforms in Japan started when it did. We think that a possible explanation is that the financial crisis and following economic stagnation revealed some inherent faults in the Japanese system, faults that made the system inapt of handling an ailing economy. Because of their embeddedness in the existing institutional environment, it has been proved hard for Japanese companies to change their behavior even when their practices are obviously inefficient. As we have seen in the theory of path dependence as described by for example Bebchuk and Roe (1999), there might be rent-seeking behavior by parties that are favored by the existing system. Therefore, we think that the financial crisis was a prerequisite for change, but in itself it was not enough to push the Japanese corporate governance reform in the right direction. An external impetus was needed in order for reform to gain momentum and the increased internationalization of Japanese companies worked as part of this pushing force. There are of course several shortcomings in our thesis. One limitation is that we, due to time and data constraints, have conducted a cross-sectional study that focuses on only one point in time. Thus we have not been able to statistically establish causality. For future research it would therefore be interesting to look at panel data for a large sample of companies to be able to establish the influence of internationalization on corporate governance over time. The study could also have been corroborated by adding a qualitative approach such as interviews in order to investigate how internationalization influences. We have also concentrated our study to one specific governance mechanism. In order to get a better understanding of the influence of internationalization on the corporate governance system as a whole, studies including more governance mechanisms are needed. Furthermore, the study is limited to Japanese companies. Therefore, it would be interesting for future studies to look at other countries and see if there exists a relationship between internationalization and corporate governance structure in these countries as well. As this thesis is being written, we are in the midst of a financial crisis that many categorize as the worst one since the great depression of the 1930s. The reasons for the current crisis are plenty, but the chasing of short-term shareholder value creation is by many considered as one of the causes of the current economic downturn. Although completely outside the scope of our thesis, the authors find it interesting to
34
see how history will judge the current Anglo-American system. Clearly, a full-fledged stakeholder system of corporate governance, as the one that prevailed in Japan, with its slow maneuvering, cozy interlocking shareholdings and inability to change, has proven not to be the optimal solution. On the other hand, the Anglo-American system, focusing primarily on short-term gains has recently revealed its deficiencies and come under criticism. Like the rest of the economy, corporate governance systems are in a constant flux, and what is considered good today, might be deemed detrimental tomorrow. Maybe, as the world economy starts to recover and gets back on track, a new and modified system of corporate governance will emerge as best-practice. After all, the ability of capitalism to reinvent itself and adapt to environmental changes is one of the reasons why it is dominating economic system of the world today.
35
8.
References
Abe, M. & Hoshi, T. 2007, "Corporate Finance and Human Resource Management in Japan" in Corporate Governance in Japan: Institutional Change and Organizational Diversity [Electronic resource], eds. M. Aoki, G. Jackson & H. Miyajima, Oxford University Press, Oxford. Abe, N. & Jung, T. August, 2004, Cross-Shareholdings, Outside Directors, and Managerial Turnover: The Case of Japan, Institute of Economic Research, Hitotsubashi University. ACCJ 2009, ACCJ Viewpoint Reform Corporate Governance by Requiring Listed Companies to Have Independent Outside Directors. Available: http://www.accj.or.jp/user/211/Advocacy/#Viewpoints [2009, March 10]. Ahmadjian, C. 2007, "Foreign Investors and Corporate Governance in Japan" in Corporate Governance in Japan: Institutional Change and Organizational Diversity [Electronic resource], eds. M. Aoki, G. Jackson & H. Miyajima, Oxford University Press, Oxford. Ahmadjian, C.L. & Robbins, G.E. 2005, "A Clash of Capitalisms: Foreign Shareholders and Corporate Restructuring in 1990s Japan", American Sociological Review, vol. 70, no. 3, pp. 451-471. Aoki, M. 1988, Information, incentives, and bargaining in the Japanese economy, Cambridge Univ. Press, Cambridge. Bartlett, C.A. & Ghoshal, S. 1998, "Managing Across Borders: The Transnational Solution, 2nd Edition (Hardcover)", Harvard Business School Press Books, pp. 1. Bathala, C.T. & Rao, R.P. 1995, "The Determinants of Board Composition: An Agency Theory Perspective", Managerial & Decision Economics, vol. 16, no. 1, pp. 59-69. Beamish, P.W. & Inkpen, A.C. 2001, "Japanese firms and the decline of the Japanese expatriate" in Japanese subsidiaries in the new global economy, eds. P.W. Beamish, A. Delios & S. Makino, Elgar, Cheltenham. Bebchuk, L.A. & Roe, M.J. "A Theory of Path Dependence in Corporate Ownership and Governance", Stanford Law Review, vol. 52, November, 1999, pp. 127. Blomström, M., Konan, D. & Lipsey, R.E. 2001, "FDI in the Reconstructuring of the Japanese Economy" in Japan's New Economy: Continuity and Change in the Twenty-First Century, eds. M. Blomström, B. Gangnes & S.J. La Croix, Oxford University Press, New York. Charkham, J.P. 2005, Keeping better company: corporate governance ten years on, Oxford University Press, Oxford. Coffee, J.C., Jr. 1999, The Future as History: The Prospects for Global Convergence in Corporate Governance and Its Implications, SSRN.
36
Coombes, P. & Watson, M. 2000, "Three surveys on corporate governance", McKinsey Quarterly, no. 4, pp. 74-77. de Andres, P., Azofra, V. & Lopez, F. 2005, "Corporate Boards in OECD Countries: size, composition, functioning and effectiveness", Corporate Governance: An International Review, vol. 13, no. 2, pp. 197-210. Denis, D.K. & McConnell, J.J. 2003, "International Corporate Governance", The Journal of Financial and Quantitative Analysis, vol. 38, no. 1, pp. 1-36. Dicken, P. 2007, Global shift: mapping the changing contours of the world economy, 5. ed., SAGE, London. Economist 2007, "JapAnglo-Saxon capitalism", Economist, vol. 385, no. 8557, pp. 17-18. Edlund, P. 1997, SPSS för Windows 95: multipel regressionsanalys: [version 7.5], Studentlitteratur, Lund. Flath, D. 2005, The Japanese Economy, 2:a uppl. edn, Oxford University Press, New York. Gillan, S.L. & Starks, L.T. 2003, "Corporate Governance, Corporate Ownership, and the Role of Institutional Investors: A Global Perspective", Journal of Applied Finance, vol. 13, no. 2, pp. 4-22. Gilson, R.J. 2001, "Globalizing Corporate Governance: Convergence of form and function", American Journal of Comparative Law, vol. 49, pp. 329. Gujarati, D.N. 2002, Basic econometrics, Boston: McGrawHill. Hansmann, H. & Kraakman, R. 2001, "The End of History for Corporate Law", Georgetown Law Journal, vol. 89, no. 2, pp. 439. Hennart, J. & Park, Y. 1994, "Location, Governance, and Strategic Determinants of Japanese Manufacturing Investment in the United States", Strategic Management Journal, vol. 15, no. 6, pp. 419-436. IHI 2003, IHI Annual Report 2003. Available: http://www.ihi.co.jp/ihi/ir/annual/annual20030331.pdf [2009, May 22]. Itoh, M. 2000, The Japanese economy reconsidered, Palgrave, Basingstoke. Jackson, G. 2003, "Corporate Governance in Germany and Japan: Liberalization Pressures and Responses during the 1990s" in The end of diversity?: prospects for German and Japanese capitalism, eds. K. Yamamura & W. Streeck, Cornell University Press, Ithaca. Jackson, G. & Miyajima, H. 2007, "Introduction: The Diversity and Change of Corporate Governance in Japan" in Corporate Governance in Japan: Institutional Change and Organizational Diversity [Electronic resource], eds. M. Aoki, G. Jackson & H. Miyajima, Oxford University Press, Oxford.
37
Jacoby, S.M. 2007, "Principles and Agents: CalPERS and corporate governance in Japan", Corporate Governance: An International Review, vol. 15, no. 1, pp. 5-15. Jacoby, S.M. 2005, The embedded corporation: corporate governance and employment relations in Japan and the United States, Princeton University, Princeton, N. J. Jacoby, S.M. 2001, Corporate Governance in Comparative Perspective: Prospects for Convergence, SSRN. JETRO 2009, Foreign Investment - JETRO. Available: http://www.jetro.go.jp/en/jetro/activities/fdi/ [2009, May 22]. Kester, C.W. 1996, "American and Japanese Corporate Governance: Convergence to Best Practice?" in National diversity and global capitalism, eds. S. Berger & R.P. Dore, Cornell University Press, Ithaca, N.Y. Knickerbocker, F.T. 1973, Oligopolistic reaction and multinational enterprise, Boston. Konica Minolta 2002, Annual Report 2002. Available: http://www.konicaminolta.com/about/investors/pdf/ar/minolta/ar2002/ar2002.pdf [2009, May 22]. Li, J. 1994, "Ownership Structure and Board Composition: A Multi-Country Test of Agency Theory Predictions", Managerial and Decision Economics, vol. 15, no. 4, Special Issue: Aspects of Corporate Governance, pp. 359-368. Melin, L. 1992, "Internationalization as a Strategy Process", Strategic Management Journal, vol. 13, Special Issue: Fundamental Themes in Strategy Process Research, pp. 99-118. METI 2009, Foreign Direct Investment into Japan - Policy Information - Ministry of Economy, Trade and Industry. Available: http://www.meti.go.jp/english/policy/index_FDI_into_Japan.html [2009, May 22]. Mitchell, J. 1997, "Representation in Government Boards and Commissions", Public administration review, vol. 57, no. 2, pp. 160-167. Muramatsu, S. 2001, "Foreign Capital and the Recent M&A Environment in Japan" in Can Japan globalize?: studies on Japan's changing political economy and the process of globalization: in honour of Sung-Jo Park, ed. A. Holzhausen, Physica-Verl., Heidelberg. NASDAQ 2009, Listing Standards & Fees, April 2009. Available: http://www.nasdaq.com/about/nasdaq_listing_req_fees.pdf [2009, May 22]. NYSE 2009, NYSE, New York Stock Exchange > Regulation > NYSE > Listed Companies. Available: http://www.nyse.com/regulation/nyse/1101074746736.html [2009, May 22]. Orbis Database [Electronic resource], Bureau van Dijk Electronic.
38
O'Sullivan, M. 2003, "The Political Economy of Comparative Corporate Governance", Review of International Political Economy, vol. 10, no. 1, pp. 23-72. Pearce II, J.A. & Zahra, S.A. 1992, "Board Composition from a Strategic Contingency Perspective", Journal of Management Studies, vol. 29, no. 4, pp. 411-438. Rhee, M. & Ji-Hwan Lee 2008, "The Signals Outside Directors Send to Foreign Investors: Evidence from Korea", Corporate Governance: An International Review, vol. 16, no. 1, pp. 41-51. Sanders, W.M.G. & Carpenter, M.A. 1998, "Internationalization and Firm Governance: the Roles of CEO Compensation, Top Team Composition, and Board Structure", Academy of Management Journal, vol. 41, no. 2, pp. 158-178. Sheard, P. 1996, "Interlocking Shareholdings and Corporate Governance in Japan" in The Japanese firm: sources of competitive strength, eds. R. Dore & M. Aoki, Clarendon, Oxford. Sullivan, D. 1994, "Measuring the Degree of Internationalization of a Firm", Journal of International Business Studies, vol. 25, no. 2, pp. 325-342. Thomsen, S. 2008, An introduction to corporate governance: mechanisms and systems, 1. ed., DJØF Publishing, Copenhagen. Tokyo Stock Exchange 2005, Fact Book 2005. Available: http://www.tse.or.jp/english/market/data/factbook/fact_book_2005.pdf [2009, May 22]. Tokyo Stock Exchange 2007, TSE-Listed Companies White Paper of Corporate Governance 2007. Available: http://www.tse.or.jp/english/rules/cg/white_paper.pdf [2009, May 22]. Tokyo Stock Exchange 2009, Fact Book 2009. Available: http://www.tse.or.jp/english/market/data/factbook/fact_book_2009.pdf [2009, May 22]. UNCTAD 2008, World Investment Report 2008, Japan. Available: http://www.unctad.org/sections/dite_dir/docs/wir08_fs_jp_en.pdf [2009, May 22]. Yafeh, Y. 2000, "Corporate governance in Japan: past performance and future prospects", Oxford Review of Economic Policy, vol. 16, no. 2. Yin, R.K. 1994, Case study research: design and methods, Sage, Thousand Oaks, CA.
39
9.
Appendix A
The graphs below show the development of foreign ownership, board size, number of outside directors and percentage of outside directors (y-axis) over time for a sample of 13 companies.
KDDI CORPORATION
FUJITSU LIMITED
HONDA MOTOR CO LTD
NEC CORPORATION
KANSAI ELECTRIC POWER CO INC
MARUBENI CORPORATION
40
MITSUI CHEMICALS INC
OLYMPUS CORP.
MITSUBISHI CORPORATION
COSMO OIL CO LTD
MITSUBISHI HEAVY INDUSTRIES LTD
SOFTBANK CORPORATION
YAMAHA MOTOR CO LTD
41
10.
Appendix B
10.1. Heteroscedasticity
We have controlled for heteroscedasticity by using White’s procedure, through which heteroscedasticitycorrected standard errors are estimated. For some variables White’s heteroscedasticity-corrected standard errors are slightly higher than the OLS standard errors and for others they are slightly lower. Overall the heteroscedasticity-corrected standard errors do not change much from the OLS standard errors, and therefore the p-values do not change much either. Most importantly, no variable becomes insignificant when we control for heteroscedasticity. Our conclusion from these results is that heteroscedasticity is not a problem in our models. BOARD SIZE
------------------- White's estimated standard errors ------------------b se(b) wse(b) wt wp Constant -10,67477 4,27071 5,17476 -2,06285 ,04025 FS_OS -1,03016 ,75182 ,73800 -1,39587 ,16410 ForOwn -6,88439 2,59997 2,53208 -2,71887 ,00705 LnAssets 1,74213 ,30499 ,36690 4,74823 ,00000 AgeDummy -2,69244 1,03804 ,67545 -3,98615 ,00009 Solvency ,02371 ,01963 ,02006 1,18195 ,23845 b = estimated coefficient, se(b) = OLS standard error wse(b) = White's standard error, wt = White's t value, wp = White's p value -------------------------------------------------------------------------
ODTD
------------------- White's b Constant -,17930 FS_OS ,07141 Disp -,00171 ForOwn ,23084 LnAssets ,01680 AgeDummy ,06178 estimated standard errors ------------------se(b) wse(b) wt wp ,15014 ,15106 -1,18689 ,23650 ,02978 ,03482 2,05089 ,04141 ,00124 ,00148 -1,15957 ,24743 ,08405 ,08260 2,79468 ,00563 ,01112 ,01079 1,55751 ,12072 ,03480 ,03000 2,05927 ,04059
b = estimated coefficient, se(b) = OLS standard error wse(b) = White's standard error, wt = White's t value, wp = White's p value -------------------------------------------------------------------------
42
OUTDIR
------------------- White's estimated standard errors ------------------b se(b) wse(b) wt wp Constant -5,30376 1,64111 1,85824 -2,85419 ,00471 FS_OS ,54939 ,32550 ,37840 1,45187 ,14790 Disp -,01679 ,01357 ,01671 -1,00484 ,31603 ForOwn 1,91795 ,91873 ,90757 2,11328 ,03565 LnAssets ,44102 ,12154 ,13223 3,33527 ,00099 AgeDummy ,45118 ,38039 ,33280 1,35572 ,17652 b = estimated coefficient, se(b) = OLS standard error wse(b) = White's standard error, wt = White's t value, wp = White's p value -------------------------------------------------------------------------
10.2. Normality
The histograms below show that the residuals of our models follow a normal distribution fairly well, but not perfectly. We also tested the normality assumption using Jarque-Bera tests, see results below. The Jarque-Bera statistics we obtain exceeds the critical value, and therefore we have to reject the nullhypothesis of a normally distributed error term. However, for large samples it may be possible to relax the normality assumption.
43
Jarque-Bera test of normality We test the null-hypothesis of a normally distributed error term.
The Jarque-Bera statistic (JB) asymptotically follows the chi-square distribution with 2 degrees of freedom. Thus, the critical value is 5.9915.
N Statistic Res_BSize Valid N (listwise) 236 236
Mean Statistic ,0000000
Std. Deviation Statistic 4,17784326
Skewness Statistic 1,193 Std. Error ,158
Kurtosis Statistic 1,917 Std. Error ,316
44
N Statistic Res_ODTD Valid N (listwise) 236 236
Mean Statistic ,0000000
Std. Deviation Statistic ,13989755
Skewness Statistic ,956 Std. Error ,158
Kurtosis Statistic 1,130 Std. Error ,316
N Statistic Res_OutDir Valid N (listwise) 236 236
Mean Statistic ,0000000
Std. Deviation Statistic 1,52918005
Skewness Statistic 1,267 Std. Error ,158
Kurtosis Statistic 4,392 Std. Error ,316
The observed JB-values exceed the critical value, and the null-hypothesis of a normally distributed error term is therefore rejected for all models.
10.3. Multicollinearity
High R-square together with few significant variables is the classic symptom of multicollinearity. The Rsquare for our models is rather low and most variables are significant, indicating that multicollinearity is not a problem in our sample. Another sign of multicollinearity is high pair-wise correlations among independent variables. As shown in the correlation matrix in Table 3 we have no correlation coefficients exceeding 0.8 and only one correlation coefficient exceeding 0.5, which supports that there is no multicollinearity (Edlund 1997). We also estimated the variance-inflating factor (VIF) for our variables to test for multicollinearity. As rule of thumb, if the VIF of a variable exceeds 10, the variable is said to be highly collinear (Gujarati 2003). The VIF for all our variables is below 2, and thus far below the critical value of 10. We conclude that our sample does not suffer from multicollinearity.
45
Coefficients
a
Collinearity Statistics Model 1 FS_OS ForOwn LnAssets AgeDummy Solvency Tolerance ,796 ,738 ,853 ,995 ,791 VIF 1,257 1,355 1,172 1,005 1,263
a. Dependent Variable: BSize
a
Coefficients
Collinearity Statistics Model 1 FS_OS Disp ForOwn LnAssets AgeDummy Tolerance ,569 ,507 ,792 ,720 ,993 VIF 1,758 1,971 1,263 1,389 1,007
a. Dependent Variable: ODTD
Coefficients
a
Collinearity Statistics Model 1 FS_OS Disp ForOwn LnAssets AgeDummy Tolerance ,569 ,507 ,792 ,720 ,993 VIF 1,758 1,971 1,263 1,389 1,007
46
Coefficients
a
Collinearity Statistics Model 1 FS_OS Disp ForOwn LnAssets AgeDummy Tolerance ,569 ,507 ,792 ,720 ,993 VIF 1,758 1,971 1,263 1,389 1,007
a. Dependent Variable: OutDir
47
doc_965551924.pdf
In contemporary business corporations, the main external stakeholder groups are shareholders, debtholders, trade creditors, suppliers, customers and communities affected by the corporation's activities. Internal stakeholders are the board of directors, executives, and other employees.
Stockholm School of Economics Institute of International Business (IIB) Master’s Thesis
Internationalization of Corporate Governance in Japan
Mikael Johansson [email protected] Kristina Sawaguchi [email protected]
Abstract This thesis investigates the impact of internationalization on corporate governance in Japan. As a consequence of increasingly global competition, different systems of corporate governance come into direct contact with each other. The Japanese economy has become more international over the last two decades, and it is therefore interesting to see whether internationalization has an impact on corporate governance of Japanese companies. Using OLS regressions on a sample of 236 companies listed on the Tokyo Stock Exchange First Section, we test whether internationalization in the form of foreign presence and foreign ownership are related to board size and composition in Japanese companies. We find that foreign presence and foreign ownership are positively related to the adoption of Anglo-American board practices, i.e. smaller boards with a higher ratio of outside directors.
Keywords: Japanese corporate governance, internationalization, convergence, board structure, stakeholder, shareholder
Tutor: Dr. Jesper Edman Presentation: June 12, 2009, 10:15-12:00, Room C606 Discussant: Pascal Lindqvist (20083)
Acknowledgements: We would like to express our thankfulness to our tutor Dr. Jesper Edman for his continued support and valuable comments throughout the project. We would also like to thank Per-Olov Edlund and Håkan Lyckeborg for their help with econometrics.
Contents
1. Introduction ........................................................................................................................................... 2 1.1. Delimitation .............................................................................................................................. 3 1.2. Outline....................................................................................................................................... 3 2. Background ........................................................................................................................................... 4 2.1. The Traditional Japanese Corporate Governance System......................................................... 4 2.2. The Anglo-American Corporate Governance System .............................................................. 5 2.3. Weaknesses in the Japanese Model........................................................................................... 7 2.4. International Convergence of Corporate Governance............................................................... 7 2.5. Changes in Japanese Corporate Governance ............................................................................ 8 2.6. Internationalization of the Japanese Economy .......................................................................... 9 3. Theory and Hypothesis Development ................................................................................................. 11 3.1. Foreign Presence ..................................................................................................................... 11 3.2. Foreign Ownership.................................................................................................................. 13 4. Methodology and Data ........................................................................................................................ 17 4.1. Data ......................................................................................................................................... 17 4.2. Dependent Variables ............................................................................................................... 17 4.3. Independent Variables............................................................................................................. 18 4.4. Control Variables .................................................................................................................... 18 4.5. Quality of Research Design .................................................................................................... 19 4.6. Descriptive Statistics ............................................................................................................... 21 4.7. Recent Developments ............................................................................................................. 22 5. Model Development and Results ........................................................................................................ 24 5.1. Board Size ............................................................................................................................... 24 5.2. Percentage of Outside Directors.............................................................................................. 26 5.3. Outside Directors .................................................................................................................... 27 5.4. Quality of Models ................................................................................................................... 29 6. Analysis............................................................................................................................................... 30 7. Conclusion and Discussion ................................................................................................................. 33 8. References ........................................................................................................................................... 36 9. Appendix A ......................................................................................................................................... 40 10. Appendix B ......................................................................................................................................... 42 10.1. Heteroscedasticity ................................................................................................................... 42 10.2. Normality ................................................................................................................................ 43 10.3. Multicollinearity...................................................................................................................... 45
1
1.
Introduction
Capitalism is the dominating economic system in most of the world today. However, there are many varieties of capitalism across nations (Dicken 2007) and at the heart of the distinction between them are differences in corporate governance. Corporate governance can be split into the marked-based, shareholder-oriented system applied in Anglo-American countries and the stakeholder-oriented system used in Germany and Japan (Ahmadjian and Robbins 2005). These two systems have each had their own heyday. In the 1980s, when the Japanese economy boomed and Japanese companies captured market shares all over the world (Jacoby 2004), the Japanese system was considered to be an efficient model of corporate governance, but because of the outstanding performance of the U.S. economy since the 1990s, the Anglo-American corporate governance has become regarded as the best-practice model. The American economic upswing coincided with an economic downturn in Japan. At the end of the 1980s, the Japanese asset price bubble burst, and many Japanese banks were found insolvent due to tremendous amounts of non-performing loans. Much of the troubles that the Japanese economy has undergone for the last two decades have been blamed on corporate governance problems (Thomsen 2007). Japanese companies have been faced with demands to adopt Anglo-American shareholder-oriented corporate governance practices (Jacoby 2004) and corporate governance in Japan is now starting to change. The domestic economic crisis is one factor behind the corporate governance reforms now taking place in Japan. However, since competition in product and capital markets is becoming increasingly global, corporate governance needs to be analyzed in an international context. As companies internationalize they are inevitably forced to compete with actors from other corporate governance systems (Coffee 1999). Internationalization is, therefore, a crucial factor to take into consideration when studying the development of corporate governance. Japanese companies have become increasingly dependent on overseas markets for sales and operations, and the shareholdings by foreign investors in Japanese listed companies have grown substantially since the 1990s (Jackson and Miyajima 2007). Few previous studies have, however, investigated the impact of internationalization on corporate governance in Japan. The purpose of our essay is therefore to fill this gap in existing research. The question we aim to answer is: Does internationalization impact on corporate governance in Japan?
2
The research question is interesting to look at for several reasons. Japan, being the second largest economy of the world, makes the future of Japanese companies of general interest. Good management, which corporate governance aims to assure, is essential for firm performance and economic efficiency, and therefore for the economic welfare of society as a whole. To the best of our understanding, this thesis is the first attempt to look at how the Japanese corporate governance is influenced by internationalization. Therefore, we hope that it will contribute to the understanding of corporate governance reform in Japanese companies and at the same time on a larger scale be a case study of how globalization is an influencing force in the modern world.
1.1.
Delimitation
This study will look at the impact of internationalization on corporate governance in Japanese companies. We have limited the scope of the thesis to look at one particular governance mechanism, the size and composition of the board of directors. The reason for this is that Japanese corporate boards have traditionally been very large, with up to 50 members, and dominated by insiders (Jacoby 2004). This makes it possible to use the board of directors as a proxy to study if and how internationalization is having an impact on corporate governance in Japan. Our dataset consists of 236 of the 250 largest companies listed at the Tokyo Stock Exchange First Section, as measured by annual turnover in JPY. Due to the limited availability of data on board structure for previous years we will look at one point in time, the fiscal year 2007.
1.2.
Outline
The remainder of the thesis is organized as follows. First, the background is introduced in section 2, describing the traditional Japanese corporate governance system and comparing it with the AngloAmerican system. We then look at the internationalization of Japanese companies and the likelihood of convergence of corporate governance systems. In section 3, we review the literature and theory concerning the influence of internationalization on corporate governance practices leading us up to our hypotheses. We will then go on to describing the methodology and data in section 4. Our models and results are presented in section 5 and analyzed in section 6. A concluding discussion with some limitations of the study and suggestions for further research is found in section 7.
3
2.
Background
The basic problem of corporate governance is the “agency problem” associated with the separation between ownership and management. Managers (the agents) act on behalf of the shareholders (principals) and the question to address is how to ensure that managers do not act in their own self-interest rather than in the interests of shareholders. Several corporate governance mechanisms are available to mitigate the agency problem. These mechanisms include laws and regulation, shareholder pressure, creditor monitoring, boards, incentive systems, the market for corporate control, trust, reputation, etc. (Thomsen 2007). In this study of corporate governance in Japan, we have chosen to focus on the board of directors because of the following three reasons. Firstly, this is one of the most widely debated institutions of corporate governance (Thomsen 2008). Secondly, boards in Japan have traditionally been extremely large and dominated by insiders but are now starting to gradually change (Charkham 2005). Thirdly, board size and composition is a relatively visible and measurable governance mechanism. Corporate governance varies a lot across countries, but researchers mainly split corporate governance practices into two broad groups, the Anglo-American shareholder-oriented system and the German and Japanese systems of stakeholder-oriented corporate governance (Gilson 2001). The different governance mechanisms are used to varying degree in these systems. Especially the board of directors has had very different forms and function in the Japanese and the Anglo-American models. In the following sections, we highlight the differences between the two systems and briefly explain the role of the board in the Japanese and the Anglo-American systems, respectively.
2.1.
The Traditional Japanese Corporate Governance System
The Japanese corporate governance system that developed in the post-war period is described as a stakeholder-oriented system (Thomsen 2008). The role of the company in the Japanese system has been to balance the interests of several stakeholders, such as employees, suppliers and customers, instead of only prioritizing shareholder value (Kester 1996). Investors, typically financial institution and other corporations, used their shareholdings to support long term relationships rather than for purely financial gains (Sheard 1994). Thus, contrary to the liquid portfolio investors in the United States, owners of Japanese companies were mostly stable shareholders (Yafeh 2000). Japanese corporate shareholding also
4
involved cross ownership, or interlocking shareholdings, meaning that two firms own each other’s shares. The extensive share-interlocks practically insulated the management of Japanese firms from external takeover threats (Sheard 1994) and, thus, limited the market for corporate control (Jackson 2003). The Japanese system is also known as a bank-based system, as Japanese companies have traditionally relied on bank finance rather than equity finance for capital provision (Thomsen 2008). The Japanese financial system was characterized by close relationships between companies and banks. A company usually had a “main bank”, which was its largest lender as well as its largest shareholder (Hoshi 1994). In return for supplying finance, banks were given the right to monitor management’s use of the provided credit from inside the firm, e.g. by appointing directors to their boards. Since the market for corporate control was almost non-existing, banks played an important monitoring role in Japanese corporate governance (Jackson 2003). Complementary to the stable shareholding and close bank relationships, was the system of “lifetime” employment (Sheard 1994). Stable financing made it easier for firms facing financial distress to renegotiate its obligations and avoid corporate failure, which supported long-term relationships with stakeholders, such as employees (Abe and Hoshi 2007). In turn, employment practices like senioritybased wage and promotion systems, and investment in firm-specific skills encouraged loyal labor relations (Jackson 2003). The size and composition of traditional Japanese board of directors is, partly, an extension of this internal promotion system (Jackson 2007). Boards in Japanese companies have traditionally been very large, often with 30-40 members (Thomsen 2008), and composed primarily of senior managers promoted from within the company (Aoki 1988). This has provided an incentive for employees to aspire for a position on the board as a final goal of a long career within the same company (Jackson and Miyajima 2007). Listing rules do not require companies to have outside directors (Jackson 2003).
2.2.
The Anglo-American Corporate Governance System
In contrast to the stakeholder-oriented Japanese corporate governance, the Anglo-American system, as implemented in the US and the UK, is a market-based model, in which shareholder value creation is assumed to be the primary goal of the company. Ownership is dispersed among diversifying shareholders, mainly institutional investors and individuals (Thomsen 2008). Contrary to the Japanese owners, these investors do not hold the shares for business strategic reasons, but are mainly interested in the financial
5
return on their investment. Shares are easily traded and shareholders who are dissatisfied with management can seek to replace it or sell their shares. The latter option includes the possibility to sell the share to another group of investors that may chose to replace management, i.e. there exists an open market for corporate control (Jacoby 2001). Since shareholders are clearly separated from other stakeholder groups such as creditors, suppliers or employees, governance mechanisms that enhance monitoring and align the interest of shareholders and mangers have become important in the Anglo-American system (Kester 1996). One of these mechanisms is the election of independent outside directors. In a study by Li (1994), for the year 1987, American boards had an outsider percentage of 74 percent, compared to 9 percent for Japanese companies. The rules of both the New York Stock Exchange and NASDAQ now require listed companies to have board of directors with a majority of independent directors (NYSE 2009; NASDAQ 2009). The size of boards in American companies has also decreased over time. For large companies, the median number of directors has decreased from 14 to about 10 directors, while the median for medium-cap companies have decreased from 12 to 9 directors since 1972 (Charkham 2005). The major differences between the two systems of corporate governance are summarized in Table 1 below. Table 1 Characteristics of the Anglo-American and Japanese corporate governance systems The Anglo-American System Shareholder oriented Dispersed ownership Equity financed Many takeovers Flexible labor market Small boards of directors Majority of outsiders on the boards The Japanese system Stakeholder oriented Stable owners, cross-ownership Bank financed Few takeovers Lifetime employment Large boards of directors Insider-dominated boards
6
2.3.
Weaknesses in the Japanese Model
During the 1980s, the Japanese system of corporate governance caught researchers’ interest due to its strength in supporting economic performance and social organization (O’Sullivan 2003). The ability of Japanese corporations to build long-term relations between shareholders, banks, employees and suppliers was seen as a competitive advantage (Jackson 2003) that resulted in company growth and increased global market shares (Yafeh 2000). Proponents of the Japanese system characterized it as a superior substitute for the market for corporate control, since Japanese managers were less subject to short-term pressure from the market (Denis and McConnell 2003). However, as a result of the prolonged economic crisis, corporate governance has been much debated in Japan. The asset price bubble at the end of the 1980s and the following period of economic stagnation have been widely blamed on corporate governance problems (Thomsen 2007). A chief shortcoming of Japanese corporate governance is its low ability to control the agency problem associated with the separation of ownership and control (Kester 1996). Tight stakeholder relationships, internal promotion, as well as the insider-dominated board structure assure continuity of business strategies and long-term perspectives, but can also result in business conservatism and delay in restructuring (Jackson and Miyajima 2007). Senior managers and directors promoted from within the company are often resistant to make decisions that would cost jobs even when company profits are declining (Ahmadjian and Robbins 2005). In practice, board control of managers has been almost absent in the Japanese system. Furthermore, the large size of the boards complicates decision making (Thomson 2008).
2.4.
International Convergence of Corporate Governance
The opinions as to the scope of the convergence vary from one extreme to the other. Some researchers foresee a total convergence between systems and others think that there will be almost no convergence at all. To start at one end of the spectrum, Hansmann and Kraakman (2000) argue that the Anglo-American system is now so dominant that other systems of corporate governance can do nothing than just follow suit and that we are already seeing a de facto convergence on and “victory” of the Anglo-American system. There are different arguments as to why this convergence is taking place. Some argue that this system generates higher levels of economic efficiency, since it features liquid labor markets, an external market for skills and an emphasis on profitability over growth (Ahmadjian and Robbins 2005). Others
7
think that the convergence is a product of the political hegemony of U.S. financial interests (O’Sullivan 2003). Gilson (2001) proposes that a functional rather than a formal convergence will take place. A functional convergence can take place without changing the formal and legal frameworks of the corporate governance system. The reason for this can be that a formal convergence can be deemed impossible due to high costs and or path dependence. Functional convergence can take many different forms such as listing the company on a foreign stock exchange and thus show compliance with stricter governance principles, or setting up voluntary contractual regimes that promotes a good governance structure. It is common to these practices that they are not necessarily regulated in the national law of the country where the company is listed, but are instead undertaken on a more voluntary basis. There are also researchers who think that almost no convergence will take place. Babchuk and Roe (1999) argue that there is strong path dependence due to initial conditions in the economy. There can also be rent-seeking behavior by the parties that are being favored by the existing system, which makes change hard. They conclude that notwithstanding the powerful forces of internationalization, many important differences between different corporate governance systems will continue to exist.
2.5.
Changes in Japanese Corporate Governance
Regardless of the question whether the Japanese system will completely converge to the Anglo-American or any other system, corporate governance practices have in fact started to change in Japan. In 1997, Sony initiated corporate board reform in Japan by reducing the size of its board from 38 to 10 directors and at the same time increasing the number of outside directors. Since then, many leading companies have reduced their board sizes to speed up decision making. However, many Japanese mangers are still reluctant to introduce outside directors since they doubt that they have enough company knowledge to make valuable contributions (Jackson 2003). One force for convergence of Japanese corporate governance practices toward the Anglo-American model is, as mentioned above, the poor economic performance of the Japanese economy (Jacoby 2005). Moreover, the committed capital that used to characterize the Japanese governance model has been weakened since the 1990s (Jackson 2003). As a result of financial deregulations dependence on equity finance has increased, and corporate ownership by stable investors such as financial institutions has declined. Japanese banks were also severely hit by the crises, which greatly affected their monitoring role.
8
Furthermore, as a response to changing economic conditions and changes in the technological environment, the decision-making in Japanese firms is becoming increasingly decentralized. These new conditions have induced corporate governance reforms such as the introduction of outside directors, and separation of monitoring and management functions (Jackson and Miyajima 2007). Another important driver for international convergence of corporate governance is that competition, on both product and capital markets, has become increasingly globalized. As a consequence, the different systems of corporate governance have come into direct contact and been forced to compete (Coffee 1999). The question of what impact internationalization has on a company is one of the most discussed topics in the international business literature. However, few studies have looked into the impact of internationalization on a company’s corporate governance, especially in a Japanese context. The aim of our thesis is to fill this gap in existing research. The research question we aim to answer in this thesis is: Does internationalization impact on corporate governance in Japan?
2.6.
Internationalization of the Japanese Economy
A firm’s degree of internationalization refers to the extent to which it depends on foreign markets for customers, factors of production, and the capacity to create value, as well as the geographical dispersion of such dependence (Sullivan 1994). In the last two decades, Japan has become increasingly internationalized in a number of ways. Firstly, Japan has begun to attract inward foreign direct investment (FDI). Secondly, Japanese companies are selling more and more on foreign markets, and thirdly, foreign ownership of Japanese shares is going up. Inward FDI has become a key policy in Japan aimed at stimulating the Japanese economy (Jackson and Miyajima 2007). The Japan External Trade Organization (JETRO), that only used to promote exports, is now also working to attract inward FDI to Japan in order to raise the level of the inward FDI stock to 5 percent of GDP (JETRO 2009). This is an ambitious goal for a country with a current level of about 3 percent, but the figure is still low compared to the United States with 15.1 percent and the United Kingdom with 48.6 percent (UNCTAD 2008). The reason for Japan to attract more inward FDI is to spur reform and innovation in Japanese companies. Foreign companies coming to Japan are often seen as catalysts for change that bring ideas to Japan, ideas that then are adopted and get adapted to the Japanese system (Economist 2007). Mergers and acquisitions are also increasing steadily and Japanese companies
9
are realizing that tying up with foreign companies offers a way of becoming part of a global strategy that they might not be able to achieve on their own (Muramatsu 2001). Alongside inward FDI, Japanese companies have become increasingly internationalized, with overseas production facilities and operations (Jackson and Miyajima 2007). During the decades following the World War II, Japan enforced quite severe trade restrictions, but in the middle of the 1960s these were relaxed and Japanese exports started to grow substantially (Flath 2005). During the 70s and 80s Japanese companies were very successful on the international export markets, to the degree that trade conflicts arose, especially with the U.S. In the late 1980s, the focus shifted somewhat as Japanese overseas investments exploded and Japan rapidly became a major source of FDI worldwide (Beamish and Inkpen 2001). One of the reasons for this shift is the appreciated Japanese yen that made the production costs in Japan high (Itoh 2000). Japanese subsidiaries opened up all over the globe, but they were criticized for mainly using Japanese expatriates for manager positions. This is starting to change though, indicating that Japanese companies are taking a more global approach by recognizing the value of empowering local management (Beamish and Inkpen 2001). The share held by foreign investors in publicly listed companies in Japan has increased steadily over the last two decades. Measured by market value, the share of foreign ownership in all publicly listed companies in Japan increased from 6.0 percent in 1992 to 13.4 percent in 1998, and to 27.6 percent in 2008 (Tokyo Stock Exchange 2005; Tokyo Stock Exchange 2009). A majority of the foreign investors are institutional investors, mainly American and European funds. Individuals represent only a negligible fraction of the foreign investors, and strategic investments by foreign corporations are also small compared to portfolio investments. Inward FDI influences the economy directly through capital formation, tax revenue etc., but it can also indirectly influence the host economy and the conduct of locally owned firms. Historically, inward FDI has played an important role in the restructuring process in Japan (Blomström, Konan and Lipsey 2001), and, thus, has probably had an indirect influence on the corporate governance structure in Japanese companies. However, this form of indirect influence is extremely difficult to measure and in this thesis we have therefore chosen to focus on the two other forms of internationalization, i.e. Japanese presence on overseas markets and foreign ownership of Japanese shares. The combination of these two measures of internationalization has not been applied to Japanese corporate governance by any previous study.
10
3.
Theory and Hypothesis Development
In this section we will review the theories of how foreign presence and foreign ownership may influence a company’s corporate governance, leading to our hypotheses.
3.1.
Foreign Presence
Foreign presence of a business can be measured in a number of ways, but we have chosen to concentrate on two measures of foreign presence, namely foreign sales and subsidiaries abroad. There are many reasons for companies to internationalize their business. Since competition today is increasingly global, firms are no longer competing only with their domestic rivals (Dicken 2007). Hennart and Park (1994) describe the international expansion decision as a two-step process. First, the firm has to decide whether to produce at home or abroad. If the firm decides to produce abroad, it then has to make a decision of how to organize the foreign production, either by licensing or by integrating into a production subsidiary. They mention location factors such as transportation costs, economies of scale, production costs and presence of customers and governance factors such as knowledge transfer and trademark reputation as some of the main variables that are important when a firm decides to internationalize. The decision to internationalize the business may also be a response to the actions or expected actions by rivals (Knickerbocker 1973). Entering a foreign market will almost inevitably mean that the complexity of the firm increases (Melin 1992). According to Bartlett and Ghoshal (1998) a firm needs certain organizational capabilities in order to handle the complex and often contradicting demands that international expansion implies. The requirements to learn about and develop new capabilities in environments, where language, culture, buyers, suppliers and political and legal systems may be different, increase uncertainty significantly. A company’s corporate governance structure is an important factor for successfully handling the increased complexity and uncertainty that internationalization brings about (Sanders and Carpenter 1998). Many studies have investigated the effect of increased complexity and uncertainty on board structure, and most studies find a positive relationship between the increased uncertainty and board size and board composition. Adding members to the board might be a way of increasing the viewpoints of the board, but can also slow down the decision making process (Mitchell 1997). Birnbaum (1984) finds a positive relationship between increased environmental uncertainty and both board size and the number of outside directors. He argues that adding board members, some of which might be outsiders will make boards
11
more able to make forecasts about the organization’s task environment, and that outsiders increase the representation of important external constituent groups. Pearce and Zahra (1992) argue along the same line, saying that increased environmental uncertainty will lead companies to adopt larger boards with more outsiders. In their sample of 119 Fortune 500 companies, they find positive correlations between external uncertainty and board size and proportion of outsiders on the boards. A study by Sanders and Carpenter (1998) is one of the few that have investigated the effect of internationalization on different aspects of a firm’s corporate governance structure. Using a sample of American companies, they assume that increased internationalization will lead to bigger boards with fewer outsiders. More insiders will lead to better board control, they argue. They find support for their hypothesis that the degree of internationalization is positively correlated to board size. However, they do not find support for the hypothesis that the internationalization is negatively related to the ratio of outside directors on the board. They conclude that this result is due to the fact that the increased complexity that internationalization implies is strong and that outside directors can contribute to a successful managing of this complexity. Another factor that might influence a company to change its corporate governance structure as it becomes more dependent on international markets is the need for convergence and adaptation of international “best-practices”. In order to compete successfully in the global economy, actors in different business systems try to interpret each other’s “best-practices” and strategically incorporate these practices in their own routines and organizational processes (Wei-Chung Yeung 2000). According to Coffee (1999), those firms seeking to grow in size to a global scale are likely to adopt “higher” governance practices already observed in the United States. For Japanese companies this would imply increased internal transparency and implementation of the kind of global standards that are familiar to stakeholders abroad (Jackson and Miyajima 2007). When it comes to outside directors on Japanese boards and the impact of internationalization, we assume that increased internationalization will also in Japan be positively related to the number of outsiders on the boards due to fact that increased complexity of international operations will put pressure on the firms to make use of the external competence that outsiders on the board bring to the company. This is in line with what the literature has found so far. Looking at the board size in Japanese companies, we think that internationalization might have a different effect than previous studies have found, mainly because the Japanese boards have played a rather
12
different role than their American counterparts, and because they have traditionally been larger. Even though increased internationalization will mean increased complexity for Japanese companies as well, we think that a more streamlined board of directors is more likely to be able to make the fast decisions that this complexity requires. The Japanese case is therefore different compared to the American case as investigated by Sanders and Carpenter (1998). This implies that the marginal effect of adding an additional board member to an already large board is miniscule at the best, but most likely will have a negative effect on the functionality of the board. For Japanese companies, decreasing the size of the board implies moving towards best-practice. This leads us up to our hypotheses on how internationalization in the form of foreign presence affects the boards of directors of Japanese companies. Hypothesis 1a: Foreign presence is negatively correlated with board size in Japanese companies. Hypothesis 1b: Foreign presence is positively correlated with the percentage of outsider directors in Japanese companies.
3.2.
Foreign Ownership
A recent trend among many major companies worldwide is that they are becoming more international not only in their scope of business activities but also in their capital structure. The incentive for companies to internationalize their ownership structure is to lower the cost of capital (Eun and Resnick 2007). However, the increase in international equity investments has also been driven by the rise of institutional investors, especially pension funds, seeking to diversify their portfolios by overseas investments (Ahmadjian 2007). These institutional investors have become dominant players in the financial markets and an increasingly important factor affecting corporate governance worldwide (Gillan and Starks 2003).The trend started in the United States during the 1980s and 1990s, where American institutional investors revolutionized corporate governance and contributed to the development of the shareholder value model. Encouraged by the success they then turned to other markets (Jacoby 2007). In Japan, the shareholdings by foreign institutional investors have, as mentioned above, increased dramatically since the 1990s. The entry of American and European institutional investors in Japan has inevitable implied a “clash” between the shareholder-oriented Anglo-American system and the Japanese system of relational governance. Standing apart from the Japanese stakeholder system and without any ties to Japanese companies other than their equity investment, these investors have been pushing for
13
corporate governance reform in the direction of the Anglo-American shareholder-oriented model. The introduction of outside directors on corporate boards is one of the reforms that have been particularly emphasized by foreign institutional investors in Japan (Ahmadjian 2007). The American Chamber of Commerce in Japan (ACCJ) has, for example, tried to persuade the Japanese government to introduce listing rules that require listed companies to increase the proportion of independent outside directors to at least one third of their boards. According to the ACCJ, many foreign fund managers have come to regard the reform of insider-dominated boards as an important indicator of how serious Japan actually is in its efforts to reform corporate governance (ACCJ 2009). Shareholders have the possibility to affect companies in several ways. If shareholders are dissatisfied with the performance of the board of directors they can choose between selling their shares, “exit”, or hold their shares and express their dissatisfaction, using “voice”. Exit can affect companies since heavy institutional selling can put downward pressure on the stock price, and thus make the company more vulnerable for takeovers. Institutional selling can also be interpreted as bad news, which trigger sales by other investors and further depresses the stock price (Gillan and Starks 2003). In Japan, foreign investors have been much more active buyers and sellers of stocks than domestic investors, and have therefore had a larger influence through the exit mechanism than the shareholding percentages would suggest. Even though Japanese managers have traditionally been more concerned with market share and growth rather than the share price, they have recently started to pay more attention to share price (Ahmadjian 2007). Voice can be exercised by shareholders either formally through the use of voting rights, or by expressing opinions through more informal channels. Foreign investors are increasingly using their voting rights since the early 2000s as Japanese voting rules have been made easier. However, foreign investors in Japan seem to have been more influential through less formal channels such as analyst and shareholder meetings as well as private meetings with CEOs (Ahmadjian 2007). CalPERS, a giant California-based pension fund, was one of the first foreign institutional investors to make major investments in Japan. As a wellknown advocate of shareholder value in the United States, CalPERS has also been a major promoter of corporate governance reform in Japan. In a case study of CalPERS, Jacoby (2007) concluds that the fund did have an effect on governance practices related to transparency, but that it was less successful in changing board structure, takeover norms and executive compensation. Foreign investors can also affect a firm’s corporate governance through the indirect supply-demand effect, i.e. firms may be motivated to improve their corporate governance in order to attract foreign capital (Gillan and Starks 2003). In an empirical study of outside directors in Korean companies, Rhee
14
and Lee (2008) find that the growth of foreign ownership is positively affected if a higher proportion of outside directors hold advanced foreign degrees, have former or current affiliations with governmental organizations, or have job experience in the same industry. In other words, board directors can also have a signalling role when stakeholders face significant information asymmetries, as for example in the case of foreign investors. The size and composition of the board of directors then provide an observable secondary source of information for foreign investors with limited information about domestic firms (Rhee and Lee 2008). This view is supported by other studies concluding that governance is a significant factor in the investment decision of institutional investors. In a McKinsey survey of more than 200 institutional investors worldwide, 80 percent of these investors stated that they would pay more for the shares of a well governed company than a poorly governed company with similar financial performance. A well governed company was defined as one that, among other things, has a majority of outside directors with no management ties on its board (Coombes and Watson 2000). There are few quantitative studies on the relationship between board structure and foreign shareholdings in Japanese companies. In an analysis based on a corporate governance index measuring how close a firm adheres to Anglo-American corporate governance standards, Ahmadjian (2007) finds that firms scoring high on board structure and function had higher levels of foreign ownership than firms with low scores. The corporate governance index used in the analysis is generated by the Japan Corporate Governance Research Institute on the basis of email surveys to all Tokyo Stock Exchange First Section companies in 2002 and 2003. However, the response rate was rather low (201 of 1,523 companies in 2003). In 2006, the Tokyo Stock Exchange (TSE) decided to require all listed companies to prepare a standardized corporate governance report, which means that every company is now obligated to disclose information on several corporate governance issues, including the number of outside directors. According to an analysis of the first reports, conducted by TSE, the percentage of companies with outside directors and the number of outside directors increases as the foreign shareholding ratios increase. The analysis also shows that as the foreign shareholding ratio increases the total number of directors increases, implicating that the percentage of outside directors does not necessarily increase with the foreign shareholding ratio. As pointed out in the report the reason that the number of outside directors increase as the foreign shareholding ratio increases may be that larger companies, which have a larger number of directors, have relatively larger foreign shareholding ratios (Tokyo Stock Exchange 2007). Based on the above studies we want to test whether internationalization in the form of foreign ownership have an impact on board size and the ratio of outside directors in Japanese companies. As pointed out
15
before, Japanese boards have traditionally been very large and insider dominated. Since foreign investors advocate more Anglo-American corporate governance practices, we expect board size to decrease and the ratio of outside directors to increase as the foreign shareholding ratio increases if foreign investors have an impact on board size and composition. Hypothesis 2a: Foreign ownership is negatively correlated with board size in Japanese companies. Hypothesis 2b: Foreign ownership is positively correlated with the percentage of outside board directors in Japanese companies.
16
4.
Methodology and Data
In order to test the hypotheses, we have chosen a quantitative approach with a large sample of companies. The advantage of quantitative research is that it gives rise to statistically meaningful results. We have run OLS regressions with board size and the percentage of outside directors as dependent variables in two different models. The OLS regression analysis allows us to test the impact of several explanatory variables simultaneously. We have used SPSS statistical software for the regressions. In the sections below, we explain the data that we are using, and how it was collected. We also present descriptive statistics of the data. Finally we have included a smaller study of the historical developments in a limited sample of companies.
4.1.
Data
Our initial data set consisted of the 250 largest companies, based on revenue in the fiscal year 2007, listed on the First Section of the Tokyo Stock Exchange. In total 2,370 companies are listed at the Tokyo Stock Exchange and of these 1,719 companies are listed at the First Section, which is the section for large companies with a market capitalization of more than JPY 50 billion. We collected this data from Bureau van Dijk’s ORBIS database. The motivation for choosing the largest companies is that most of the corporate governance literature also tends to focus on the largest companies (Thomsen 2008). We excluded banking and insurance companies, since these companies are subject to stricter regulations. After the exclusions we ended up with a data set of 236 companies.
4.2.
Dependent Variables
As dependent variables we have board size measured as the number of board directors (BSIZE), the number of outside directors (OUTDIR) and the ratio of outside directors of total directors (ODTD). Data for these variables have been collected from the Tokyo Stock Exchange (TSE) homepage. Since 2006, all companies listed on the TSE are required to prepare a corporate governance report, in which they disclose the size of their boards as well as the number of outside directors (Tokyo Stock Exchange 2007). In our data set, we have included all directors that according to TSE are classified as outside directors regardless of their relationship with the company in question.
17
4.3.
Independent Variables
Foreign Ownership (FOROWN) This variable captures the percentage of outstanding shares that are held by non-Japanese investors. The data were collected from the companies’ annual reports downloaded from their homepages. We primarily used the English versions, but collected data from the Japanese language versions when English annual reports were not available. For most of the observations, there was no split between the holdings of foreign institutions and foreign individuals, but, as mentioned above, the vast majority of Japanese shares held by foreigners are held by foreign institutions. Our screening of the annual report also strongly supports this.
Foreign Presence (FS_OS) In order to measure foreign presence we constructed a combined variable by adding foreign sales as a ratio of total sales (FSTS) and the number of overseas subsidiaries as a ratio of total number of subsidiaries (OSTS). This is done in order to test the total effect of these two aspects of foreign presence. The variable can theoretically take values between 0 and 2, where 0 indicates that the company has no overseas subsidiaries and no foreign sales. A composite variable is more likely to correctly measure a firm’s degree of internationalization and also enables better error control (Sullivan 1994). The method of adding measures of internationalization has also been used by for example Sanders and Carpenter (1994). The data for total sales of each company were collected from ORBIS database, while the data for foreign sales were collected from the annual reports of the companies. Foreign sales are defined the amount of sales generated by the companies’ foreign operations and, which is not the same as the export sales. Export sales means that the factors of production are located in Japan. For total number of subsidiaries as well as the number of overseas subsidiaries we used the data available in ORBIS. The definition of a subsidiary is a company in which the parent company holds more than 25 percent of the shares.
4.4.
Control Variables
Dispersion (DISP) Dispersion measures the total number of countries in which the company has a subsidiary. We used the same data source as was used for overseas subsidiaries. Given that the ratio of overseas subsidiaries of
18
total subsidiaries can be very high for companies with few total subsidiaries and that some companies might have most of their overseas subsidiaries in one or a few countries, a dispersion variable complements the FS_OS measure of foreign presence. We assume that this variable will be positively related to ODTD and negatively related to board size.
Assets (LNASSETS) Previous studies have found positive relations between firm size and board size as well as the percentage of outside directors (Pearce and Zahra 1992; Li 1994). We therefore want to control for firm size in our regressions. The total assets of the companies in 2007 were collected from ORBIS. The natural logarithmic form of the variable was used in order to make interpretations more understandable.
Establishment Dummy (EST_DUM) We included an establishment dummy, which takes the value 1 if the company was incorporated after 1997 and 0 otherwise. The reason for including this variable is that corporate governance reform in Japan is generally considered to have started in 1997 (Shishido 2007). This was also the year that Sony initiated board reforms in Japan by drastically reducing the size of their board and simultaneously increasing the number of outsiders. We assume that companies founded after this event are more likely to adopt a Western style corporate governance system. Data on the year of incorporation was collected from ORBIS.
Solvency Ratio (SOLV) Solvency ratio, defined as shareholders funds divided by total assets, was used as a control variable, the underlying argument being that companies with a high solvency ratio are more dependent on equity market for financing and thus more exposed to pressure from investors to change corporate governance (Pearce and Zahra 1992). The data come from ORBIS.
4.5.
Quality of Research Design
According to Yin (1994), four criteria are commonly used in order to analyze the quality of the research, namely construct validity, internal validity, external validity and reliability. We will discuss all four of these and relate them to our study. The explanations of the different criteria below are all based on the definitions used by Yin (1994).
19
Construct Validity Construct validity refers to the selection of a correct set of measures for the research question. Boards of directors are a major part of any corporate governance system, and we therefore believe that they can well be used as a proxy in order to study the corporate governance system as a whole. Since the size and composition of Japanese boards have traditionally displayed some unique characteristics, we think that this is a good way of seeing whether Japanese corporate governance is converging on Anglo-American practices. Foreign presence and foreign ownership are commonly used measures of internationalization. Therefore we conclude that the construct validity is high. Internal Validity Internal validity concerns the establishment of causal direction between variables. Since we have done a cross-sectional study, we have not been able to test causal relationships between our variables. This could have been remedied by looking at panel-data. We have used a small selection of companies in order to see some trends that can give us some clue as to in what direction the causality is going. When it comes to other influencing factors on our dependent variable, we have tried to control for as many additional variables as our data allows for. Therefore we regard the internal validity to be satisfactory. External Validity External validity concerns whether the findings are generalizable outside the study sample. Since Japan is somewhat of a rara avis when it comes to corporate governance, the generalizability of our results to other countries might not be very high. However, we think that our sample is representative for other large, listed companies in Japan. Due to this we deem the external validity to be fair. Reliability Reliability concerns the possibility to reiterate the study and obtain the same results. The goal is to minimize errors and biases in a study. We have clearly documented the data collection and the regression analysis procedures. The data we have used are straight-forward and readily available in databases and annual reports. We therefore conclude that the reliability of our study is high.
20
4.6.
Descriptive Statistics
Table 2 Descriptive statistics of variables
Standard deviation 4.54
Variable BSIZE
Unit Discrete number Discrete number Ratio Ratio Discrete number Ratio MJPY Ratio Dummy
N 236
Minimum 4
Maximum 30
Mean 12.22
OUTDIR ODTD FS_OS DISP FOROWN ASSETS SOLV EST_DUM
236 236 236 236 236 236 236 236
0 0 .00 .00 .03 187,555 -.03 -
12 .80 1.57 55 .69 32,458,320 .84 -
1.48 .13 .45 8.69 .25 2,085,900 .36 -
1.63 .15 .41 10.43 .12 3,201,410 .16 -
Table 2 shows a summary of descriptive statistics for all variables. The asset variable is presented in million JPY instead of in its logarithmic form in order to give the reader a better indicator of firm size. We use the natural logarithm of the asset variable, LNASSETS, when running the regressions. In our sample of 236 companies, the average board size is 12.2 directors and the average ratio of outside directors of total directors is 13 percent for this particular point in time, fiscal year 2007. An average board size of 12.2 directors is not to be regarded as extremely large. For the sample of 192 listed Japanese companies during the period 1990-1999 used in a study by Abe and Jung (2004) the average board size was 18 directors, which is considerably higher. The average board size in the sample of 258 listed U.S. companies in 1992 used by Sanders and Carpenter (2004) was 12.3 directors, which is approximately the same as for our sample. However, the proportion of outside directors in Sanders and Carpenter’s sample was 76 percent, i.e. on average a majority of the directors were outsiders. The average
21
proportion of 13 percent outside directors in our sample indicates that board composition in Japanese companies is still far from Anglo-American standards.
4.7.
Recent Developments
In order to assess the sequence of events, we randomly chose 20 companies from our sample that had at least one outside director and then searched in the annual reports available on the company websites for data on foreign ownership, total number of board directors and outside directors. The measure for number of outside directors used is not the same as used in our full sample since TSE does only disclose the most recent information. Therefore, we used the information regarding outsiders as given in the annual reports. Companies were dropped if they did not have annual reports dating back to at least 2002 or if the relevant information was not available. The remaining sample contained 13 companies. See Appendix A for graphs. On average the ratio of foreign shareholdings increased from 17.3 percent to 31.2 percent for these 13 companies over the period 2002-2007. At the same time the number of directors decreased from 17.9 to 13.9, the number of outside directors increased from 0.85 to 3.31 and the ratio of outside directors increased from 6.2 percent to 25.3 percent. As can be seen in Figure 1 below, the average trends in foreign ownership, number of board directors, outside directors and the ratio of outside directors show a simultaneous increase in the ratio of foreign ownership and the ratio of outside board directors. The increase in the ratio of outside directors is a result of both a decreasing trend in the total number of directors and an increase in the number of outside directors.
22
Figure 1 Historical trend average, 13 random companies (percentage and number on y-axis)
23
5.
Model Development and Results
Table 3 Correlations OUT DIR FOR OWN LNAS SETS EST_ DUM
BSIZE BSIZE OUTDIR ODTD FS_OS DISP FOROWN LNASSETS EST_DUM SOLV -.004 -.249 -.080 .073 -.069 .299 -.148 -.046
ODTD
FS_OS
DISP
SOLV
.910 .155 .130 .241 .278 .063 .053 .205 .098 .256 .141 .097 .135 .595 .363 .118 -.045 .358 .306 .438 -.065 .058 .313 -.060 .304 -.012 -.105 -.010
Table 3 presents a correlation matrix of the variables in order to show that multicollinearity is not a big problem in our sample.
5.1.
Board Size
Based on the discussion in our theory and data sections, we end up with the following model for board size. Board Size Full Model
BSIZE
7 1 2
FS _ OS
3
FOROWN
4
DISP
5
LNASSETS
6
EST _ DUM
SOLV
u
We started by running the full regression model, including all independent variables. The results are summarized in column “Full Model” of Table 4 below. The full model yielded an R-square of 0.152 and
24
an adjusted R-square of 0.130. Three of the variables were found significant at the 5 percent level, namely FOROWN, LNASSETS and EST_DUM. We continued by removing the insignificant variables, one at the time, and then compared the adjusted R-square for the different models. The adjusted R-square is a better measure for the explanatory power than R-square when comparing models with the same dependent variable, since it takes the degrees of freedom into account. Removing the variable DISP, which was highly insignificant, proved to render us with the model with the best adjusted R-square. This final model is presented below. Board Size Final Model
BSIZE
6 1 2
FS _ OS
3
FOROWN
4
LNASSETS
5
EST _ DUM
SOLV
u
Table 4 Board size
Dependent Variable: BSIZE Constant
Full Model -10.405** (.026)
Final Model -10.675** (.013) -1.030 (.172) -6.884*** (.009)
FS_OS
-1.115 (.237)
FOROWN
-6.893*** (.009)
DISP
.006 (.880)
LNASSETS
1.721*** (.000)
1.742*** (.000) -2.692** (.010) .024 (.228) 236 .147 .134
EST_DUM
-2.685** (.011)
SOLV
.024 (.226)
N R-square Adjusted R-square
236 .152 .130
p-values: ***<1%, **<5%, *<10%. Significance levels in brackets.
25
The R-square of the final model, 0.147, is slightly lower than in the full model, but as described above this is compensated for by a higher adjusted R-square, which indicates that the significance of the remaining variables are higher. The R-square is rather low, but in line with what is found in similar studies that we have covered in the theory section. FS_OS has the predicted negative sign, but is not significant at the 10 percent level. Still, a p-value of 0.172 is not extremely high. FOROWN has a negative effect on board size as predicted and is significant on the 1 percent level. The control variables LNASSETS and EST_DUM also have the predicted signs and are significant on the 1 percent level.
5.2.
Percentage of Outside Directors
We use the same full model when testing the influence of internationalization on the percentage of outside directors in relation to total directors. The full model is presented below. ODTD Full Model
ODTD
7 1 2
FS _ OS
3
FOROWN
4
DISP
5
LNASSETS
6
EST _ DUM
SOLV
u
Coefficients and significance levels for the variables are presented in Table 5. Two variables, FS_OS and FOROWN, were significant at the 5 percent level. Using the same methodology as with the board size regression above, we removed the insignificant variables one by one and compared the adjusted Rsquares for the different models. With ODTD as the dependent variable, it proved that the removal of SOLV gave us the highest adjusted R-square. Even though DISP is found insignificant, we keep it for the sake of improved adjusted R-square. The final model is presented below. ODTD Final Model
ODTD FS _ OS FOROWN DISP LNASSETS EST _ DUM u
1
2
3
4
5
6
26
Table 5 Outside directors of total directors
Dependent Variable: ODTD Constant
Full Model -.197 (.206)
Final Model -.179 (.234) .071** (.017) .231*** (.007) -.002 (.169) .017 (.132) .062* (.077)
FS_OS
.067** (.035)
FOROWN
.221** (.012)
DISP
-.002 (.198)
LNASSETS
.018 (.121)
EST_DUM
.062* (.079)
SOLV
.000 (.651)
N R-square Adjusted R-square
236 .105 .082
236 .105 .085
p-values: ***<1%, **<5%, *<10%. Significance levels in brackets. Also in this model, the R-square is fairly low at 0.105, but as mentioned above, similar studies get Rsquares in the same range. Both explanatory variables, FS_OS and FOROWN, have the predicted positive signs and are significant at the 5 percent level and the 1 percent level, respectively. EST_DUM is positively related to ODTD as expected and significant at the 10 percent level. The coefficient of LNASSETS has a positive sign as predicted, but is not significant.
5.3.
Outside Directors
We have not hypothesized about the relation between internationalization and the total number of outside directors. Nevertheless, since the ratio of outside directors on the board (ODTD) is related to board size, we want to test whether internationalization has an effect on the absolute number of outside directors. A
27
high ODTD could be the result of a small board size and does not necessarily imply that there are many outsiders on the board. We run the same regression for OUTDIR as we did with ODTD. OUTDIR Final Model
OUTDIR FS _ OS FOROWN DISP LNASSETS EST _ DUM u
1
2
3
4
5
6
Table 6 Outside directors
Dependent Variable: OUTDIR Constant
Full Model -5.255*** (.002)
Final Model -5.304*** (.001) .549* (.093) 1.918** (.038) -.017 (.217) .441*** (.000) .451 (.237)
FS_OS
.562 (.104)
FOROWN
1.947** (.042)
DISP
-.017 (.217)
LNASSETS
.439*** (.000)
EST_DUM
.452 (.237)
SOLV
.000 (.908)
N R-square Adjusted R-square
236 .121 .098
236 .121 .101
p-values: ***<1%, **<5%, *<10%. Significance levels in brackets. As for the model with ODTD as the dependent variable, the removal of SOLV gave the model with best adjusted R-square. The final model has an R-square of 0.121. FS_OS is positive and significant at the 10 percent level, while FOROWN is positive and significant on the 5 percent level. Both internationalization variables thus have the expected signs.
28
5.4.
Quality of Models
We have controlled the quality of the regression models with a number of tests. White’s estimated standard errors do not change much from the OLS standard errors indicating low heteroscedasticity. Multicollinearity does not seem to be a problem, since we have low VIF-values in combination with low pair-wise correlations and low R-square values. Finally our histograms and Jarque-Bera statistics indicate that the residuals might not be normally distributed, but this is not perceived as a major issue since our sample size is quite large. For further explanation of the quality tests, see Appendix B.
29
6.
Analysis
The regression results do not support Hypothesis 1a, i.e. that foreign presence is negatively related to board size as we had predicted. Nevertheless, the result is still interesting since we get the predicted negative sign, which indicates that internationalization in the form of foreign presence is at least not positively related to board size as was the case in U.S. companies according to previous studies (Sanders and Carpenter 1998). A possible interpretation of this result is that foreign presence has dual effects on board size in Japanese companies, since Japanese boards tend to be quite large from the outset. On the one hand, the increased complexity that foreign presence implies requires companies to add more directors to their boards in order to bring more knowledge to the boards and effectively monitor management as argued by Sanders and Carpenter (1998). On the other hand, higher foreign presence means that the company is more exposed to global competition and thus needs to conform to global practices regarding for example corporate governance, i.e. smaller boards. From our results we conclude that the latter effect is the dominating one. We predicted that foreign presence would be positively related to the ratio of outsiders on Japanese boards (Hypothesis 1b), which is supported by our results. This is in line with the literature that has investigated the effects of foreign presence on outside directors. We think that Japanese companies are adding outsiders to their board for two reasons as foreign presence increases. Firstly, the argument of outsiders being more able to deal with the complexity that foreign presence brings probably holds for Japanese companies as well. In addition to this, Japanese companies with a high degree of foreign presence are more likely to adhere to Anglo-American practices of more outsider-dominated boards. We therefore see a double effect of foreign presence working to push up the ratio of outsiders. In order to make sure that a high ratio of outside directors is not only explained by a small board size, we also ran a regression with the absolute number of outside directors as dependent variable. The result shows that the total number of outside directors is also positively correlated to foreign presence, and therefore corroborates our findings that the ratio of outside directors being high is not merely due to a small board size, but is in fact a result of a larger number of outside directors as well. One of our strongest and most interesting results is the effect that foreign ownership has on the size of Japanese boards. The variable is negative and significant on the 1 percent level, which supports Hypothesis 2a. This contradicts the TSE Whitepaper (Tokyo Stock Exchange 2007), which finds a positive relationship between foreign shareholdings and board size. It concludes that this might be
30
explained by the fact that foreign investors tend to invest in larger companies that have relatively large boards. Our correlation analysis confirms that large companies generally have large boards, and that foreign ownership is positively related to firm size. However, when controlling for firm size, we find a negative correlation between foreign ownership and board size. By this we can conclude that companies with a high degree of foreign ownership do in fact adopt more Anglo-American corporate governance practices in terms of board size. We also find strong support for Hypothesis 2b, that the foreign shareholding ratio is positively related to the percentage of outsiders on the board. Foreign ownership is also positively correlated to the absolute number of outside directors. From our cross-sectional analysis we cannot prove the direction of causality between foreign ownership and the adoption of Anglo-American board practices. Foreign investor might have been able to push companies to change their practices, but it is also possible that foreign investors tend to target companies that already have reformed their board structure. However, despite the endogenous nature of the relations that make it difficult to establish direct causality, foreign investor can be said to have influence over corporate governance either way. On the one hand, companies may be motivated to improve their corporate governance in order to attract foreign capital, and on the other hand, increased investment by foreign institutional investors may provide the companies with the power to enforce governance change (Gillan and Starks 2003). In our small sample of companies with data for the period 2002-2007, we can see that the ratio of foreign ownership and the ratio of outside directors increase almost simultaneously on average (Figure 1). Graphs of the trends for individual companies show somewhat disperse outcomes (see Appendix A). In some cases, however, the increase in foreign ownership clearly precedes the increase in the ratio of outside directors, which indicates that foreign investors may have influenced the companies to reform their boards after becoming shareholders. This was the case in, for example, KDDI Corporation, Honda Motor Co Ltd and NEC Corporation. Internationalization in the form of foreign ownership seems to have a greater impact on board size and board composition than foreign presence. One explanation for this might be the contradicting effects of foreign presence on board structure discussed above. A more plausible explanation, though, is that foreign investors are a much stronger force pushing the companies to adopt Anglo-American governance standards. The influence of foreign investors either through voice or through the threat of exit is more tangible for companies, than the need to conform to these standards due to a presence on foreign markets. We think that the influence through foreign presence might be strong in the longer run, but nevertheless it is a more indirect pressure compared to the power that shareholders wield. An alternative explanation is
31
that Japanese companies’ tend to have a low degree of local adaption (Bartlett and Ghoshal 1998). Japanese subsidiaries have traditionally not been very good at utilizing local competence, as they were mainly using Japanese expatriates for manager position. This could indicate that Japanese companies are, to some extent, immune to the impact that overseas subsidiaries have. Our dummy variable for companies established after 1997 is negatively related to both board size and positively related to the ratio of outside directors. In other words companies established after 1997 tend to have smaller boards with a higher ratio of outside directors. One reason for the negative relationship with board size is that younger companies are generally smaller and therefore not as prone to have large boards. However, another explanation is that companies founded after 1997, when corporate governance reform had already begun, are more likely to adopt Anglo-American board practices from the outset. Our dispersion measure is found insignificant and it does not have the predicted sign. It can be the case that the companies with a high dispersion only have sales offices in many countries without being particularly internationalized. Although not included in our results, we tested to split the two variables foreign sales of total sales and overseas subsidiaries of total subsidiaries, but these two measures separately gave us inconclusive results and we therefore chose to use the composite measure as suggested by the literature. This implies that we can not conclude what form of foreign presence influence what, but that the total effect is as expected.
32
7.
Conclusion and Discussion
This study has added some insights to the understanding of corporate governance in Japan and the impact of internationalization on corporate governance. Our results show that foreign presence and foreign ownership are correlated to board structure in Japanese companies. Based on this we draw the conclusion that internationalization does have an impact on corporate governance in Japan. No previous study on Japanese corporate governance has used the foreign presence measures that we use in this thesis. For our sample of Japanese companies, we find that foreign presence is negatively correlated to board size (even though the results are not statistically significant), which contradicts earlier findings for American companies (Sanders and Carpenter 1998), but is in line with our hypothesis. We also find a negative relationship between foreign ownership and board size, when controlling for firm size, which contradicts the study by Tokyo Stock Exchange (2007). Although we cannot make general statements about convergence of the Japanese system as a whole, the fact that we have observed changes in the board structure is an indication that the Anglo-American system is gaining ground in Japan. Coffee (1999) argues that firms aiming to compete on a global scale need to adopt global best-practices. In the Japanese annual reports that we used when collecting data for the econometric analysis, we have found some illustrating quotes on this line: “Ideally, companies must adapt their management supervision and business execution functions to global standards, while retaining sound corporate philosophies and traditions. If it is deemed best to hire external directors and other U.S.- style systems then such steps should be taken.” (Minolta 2002) “IHI conducted a study on the best management practices to establish a corporate governance structure befitting a global corporation and a system able to flexibly respond to sudden changes in the operating environment.” (IHI 2003)
The quotes above point to the kind of functional rather than formal convergence that, for example, Gilson (2001) is referring to. This shows that even if the corporate laws are not changing very quickly in Japan, companies find it essential to adjust their systems to more efficient forms of governance.
33
In the future, as sufficiently many Japanese companies reform their corporate governance practices, we could expect a spill-over effect on domestic companies although they are not directly exposed to the pressure of internationalization. Smaller boards with higher ratios of outside directors would then become the norm, or “best-practice”, even in Japan. A question that remains is why corporate governance reforms in Japan started when it did. We think that a possible explanation is that the financial crisis and following economic stagnation revealed some inherent faults in the Japanese system, faults that made the system inapt of handling an ailing economy. Because of their embeddedness in the existing institutional environment, it has been proved hard for Japanese companies to change their behavior even when their practices are obviously inefficient. As we have seen in the theory of path dependence as described by for example Bebchuk and Roe (1999), there might be rent-seeking behavior by parties that are favored by the existing system. Therefore, we think that the financial crisis was a prerequisite for change, but in itself it was not enough to push the Japanese corporate governance reform in the right direction. An external impetus was needed in order for reform to gain momentum and the increased internationalization of Japanese companies worked as part of this pushing force. There are of course several shortcomings in our thesis. One limitation is that we, due to time and data constraints, have conducted a cross-sectional study that focuses on only one point in time. Thus we have not been able to statistically establish causality. For future research it would therefore be interesting to look at panel data for a large sample of companies to be able to establish the influence of internationalization on corporate governance over time. The study could also have been corroborated by adding a qualitative approach such as interviews in order to investigate how internationalization influences. We have also concentrated our study to one specific governance mechanism. In order to get a better understanding of the influence of internationalization on the corporate governance system as a whole, studies including more governance mechanisms are needed. Furthermore, the study is limited to Japanese companies. Therefore, it would be interesting for future studies to look at other countries and see if there exists a relationship between internationalization and corporate governance structure in these countries as well. As this thesis is being written, we are in the midst of a financial crisis that many categorize as the worst one since the great depression of the 1930s. The reasons for the current crisis are plenty, but the chasing of short-term shareholder value creation is by many considered as one of the causes of the current economic downturn. Although completely outside the scope of our thesis, the authors find it interesting to
34
see how history will judge the current Anglo-American system. Clearly, a full-fledged stakeholder system of corporate governance, as the one that prevailed in Japan, with its slow maneuvering, cozy interlocking shareholdings and inability to change, has proven not to be the optimal solution. On the other hand, the Anglo-American system, focusing primarily on short-term gains has recently revealed its deficiencies and come under criticism. Like the rest of the economy, corporate governance systems are in a constant flux, and what is considered good today, might be deemed detrimental tomorrow. Maybe, as the world economy starts to recover and gets back on track, a new and modified system of corporate governance will emerge as best-practice. After all, the ability of capitalism to reinvent itself and adapt to environmental changes is one of the reasons why it is dominating economic system of the world today.
35
8.
References
Abe, M. & Hoshi, T. 2007, "Corporate Finance and Human Resource Management in Japan" in Corporate Governance in Japan: Institutional Change and Organizational Diversity [Electronic resource], eds. M. Aoki, G. Jackson & H. Miyajima, Oxford University Press, Oxford. Abe, N. & Jung, T. August, 2004, Cross-Shareholdings, Outside Directors, and Managerial Turnover: The Case of Japan, Institute of Economic Research, Hitotsubashi University. ACCJ 2009, ACCJ Viewpoint Reform Corporate Governance by Requiring Listed Companies to Have Independent Outside Directors. Available: http://www.accj.or.jp/user/211/Advocacy/#Viewpoints [2009, March 10]. Ahmadjian, C. 2007, "Foreign Investors and Corporate Governance in Japan" in Corporate Governance in Japan: Institutional Change and Organizational Diversity [Electronic resource], eds. M. Aoki, G. Jackson & H. Miyajima, Oxford University Press, Oxford. Ahmadjian, C.L. & Robbins, G.E. 2005, "A Clash of Capitalisms: Foreign Shareholders and Corporate Restructuring in 1990s Japan", American Sociological Review, vol. 70, no. 3, pp. 451-471. Aoki, M. 1988, Information, incentives, and bargaining in the Japanese economy, Cambridge Univ. Press, Cambridge. Bartlett, C.A. & Ghoshal, S. 1998, "Managing Across Borders: The Transnational Solution, 2nd Edition (Hardcover)", Harvard Business School Press Books, pp. 1. Bathala, C.T. & Rao, R.P. 1995, "The Determinants of Board Composition: An Agency Theory Perspective", Managerial & Decision Economics, vol. 16, no. 1, pp. 59-69. Beamish, P.W. & Inkpen, A.C. 2001, "Japanese firms and the decline of the Japanese expatriate" in Japanese subsidiaries in the new global economy, eds. P.W. Beamish, A. Delios & S. Makino, Elgar, Cheltenham. Bebchuk, L.A. & Roe, M.J. "A Theory of Path Dependence in Corporate Ownership and Governance", Stanford Law Review, vol. 52, November, 1999, pp. 127. Blomström, M., Konan, D. & Lipsey, R.E. 2001, "FDI in the Reconstructuring of the Japanese Economy" in Japan's New Economy: Continuity and Change in the Twenty-First Century, eds. M. Blomström, B. Gangnes & S.J. La Croix, Oxford University Press, New York. Charkham, J.P. 2005, Keeping better company: corporate governance ten years on, Oxford University Press, Oxford. Coffee, J.C., Jr. 1999, The Future as History: The Prospects for Global Convergence in Corporate Governance and Its Implications, SSRN.
36
Coombes, P. & Watson, M. 2000, "Three surveys on corporate governance", McKinsey Quarterly, no. 4, pp. 74-77. de Andres, P., Azofra, V. & Lopez, F. 2005, "Corporate Boards in OECD Countries: size, composition, functioning and effectiveness", Corporate Governance: An International Review, vol. 13, no. 2, pp. 197-210. Denis, D.K. & McConnell, J.J. 2003, "International Corporate Governance", The Journal of Financial and Quantitative Analysis, vol. 38, no. 1, pp. 1-36. Dicken, P. 2007, Global shift: mapping the changing contours of the world economy, 5. ed., SAGE, London. Economist 2007, "JapAnglo-Saxon capitalism", Economist, vol. 385, no. 8557, pp. 17-18. Edlund, P. 1997, SPSS för Windows 95: multipel regressionsanalys: [version 7.5], Studentlitteratur, Lund. Flath, D. 2005, The Japanese Economy, 2:a uppl. edn, Oxford University Press, New York. Gillan, S.L. & Starks, L.T. 2003, "Corporate Governance, Corporate Ownership, and the Role of Institutional Investors: A Global Perspective", Journal of Applied Finance, vol. 13, no. 2, pp. 4-22. Gilson, R.J. 2001, "Globalizing Corporate Governance: Convergence of form and function", American Journal of Comparative Law, vol. 49, pp. 329. Gujarati, D.N. 2002, Basic econometrics, Boston: McGrawHill. Hansmann, H. & Kraakman, R. 2001, "The End of History for Corporate Law", Georgetown Law Journal, vol. 89, no. 2, pp. 439. Hennart, J. & Park, Y. 1994, "Location, Governance, and Strategic Determinants of Japanese Manufacturing Investment in the United States", Strategic Management Journal, vol. 15, no. 6, pp. 419-436. IHI 2003, IHI Annual Report 2003. Available: http://www.ihi.co.jp/ihi/ir/annual/annual20030331.pdf [2009, May 22]. Itoh, M. 2000, The Japanese economy reconsidered, Palgrave, Basingstoke. Jackson, G. 2003, "Corporate Governance in Germany and Japan: Liberalization Pressures and Responses during the 1990s" in The end of diversity?: prospects for German and Japanese capitalism, eds. K. Yamamura & W. Streeck, Cornell University Press, Ithaca. Jackson, G. & Miyajima, H. 2007, "Introduction: The Diversity and Change of Corporate Governance in Japan" in Corporate Governance in Japan: Institutional Change and Organizational Diversity [Electronic resource], eds. M. Aoki, G. Jackson & H. Miyajima, Oxford University Press, Oxford.
37
Jacoby, S.M. 2007, "Principles and Agents: CalPERS and corporate governance in Japan", Corporate Governance: An International Review, vol. 15, no. 1, pp. 5-15. Jacoby, S.M. 2005, The embedded corporation: corporate governance and employment relations in Japan and the United States, Princeton University, Princeton, N. J. Jacoby, S.M. 2001, Corporate Governance in Comparative Perspective: Prospects for Convergence, SSRN. JETRO 2009, Foreign Investment - JETRO. Available: http://www.jetro.go.jp/en/jetro/activities/fdi/ [2009, May 22]. Kester, C.W. 1996, "American and Japanese Corporate Governance: Convergence to Best Practice?" in National diversity and global capitalism, eds. S. Berger & R.P. Dore, Cornell University Press, Ithaca, N.Y. Knickerbocker, F.T. 1973, Oligopolistic reaction and multinational enterprise, Boston. Konica Minolta 2002, Annual Report 2002. Available: http://www.konicaminolta.com/about/investors/pdf/ar/minolta/ar2002/ar2002.pdf [2009, May 22]. Li, J. 1994, "Ownership Structure and Board Composition: A Multi-Country Test of Agency Theory Predictions", Managerial and Decision Economics, vol. 15, no. 4, Special Issue: Aspects of Corporate Governance, pp. 359-368. Melin, L. 1992, "Internationalization as a Strategy Process", Strategic Management Journal, vol. 13, Special Issue: Fundamental Themes in Strategy Process Research, pp. 99-118. METI 2009, Foreign Direct Investment into Japan - Policy Information - Ministry of Economy, Trade and Industry. Available: http://www.meti.go.jp/english/policy/index_FDI_into_Japan.html [2009, May 22]. Mitchell, J. 1997, "Representation in Government Boards and Commissions", Public administration review, vol. 57, no. 2, pp. 160-167. Muramatsu, S. 2001, "Foreign Capital and the Recent M&A Environment in Japan" in Can Japan globalize?: studies on Japan's changing political economy and the process of globalization: in honour of Sung-Jo Park, ed. A. Holzhausen, Physica-Verl., Heidelberg. NASDAQ 2009, Listing Standards & Fees, April 2009. Available: http://www.nasdaq.com/about/nasdaq_listing_req_fees.pdf [2009, May 22]. NYSE 2009, NYSE, New York Stock Exchange > Regulation > NYSE > Listed Companies. Available: http://www.nyse.com/regulation/nyse/1101074746736.html [2009, May 22]. Orbis Database [Electronic resource], Bureau van Dijk Electronic.
38
O'Sullivan, M. 2003, "The Political Economy of Comparative Corporate Governance", Review of International Political Economy, vol. 10, no. 1, pp. 23-72. Pearce II, J.A. & Zahra, S.A. 1992, "Board Composition from a Strategic Contingency Perspective", Journal of Management Studies, vol. 29, no. 4, pp. 411-438. Rhee, M. & Ji-Hwan Lee 2008, "The Signals Outside Directors Send to Foreign Investors: Evidence from Korea", Corporate Governance: An International Review, vol. 16, no. 1, pp. 41-51. Sanders, W.M.G. & Carpenter, M.A. 1998, "Internationalization and Firm Governance: the Roles of CEO Compensation, Top Team Composition, and Board Structure", Academy of Management Journal, vol. 41, no. 2, pp. 158-178. Sheard, P. 1996, "Interlocking Shareholdings and Corporate Governance in Japan" in The Japanese firm: sources of competitive strength, eds. R. Dore & M. Aoki, Clarendon, Oxford. Sullivan, D. 1994, "Measuring the Degree of Internationalization of a Firm", Journal of International Business Studies, vol. 25, no. 2, pp. 325-342. Thomsen, S. 2008, An introduction to corporate governance: mechanisms and systems, 1. ed., DJØF Publishing, Copenhagen. Tokyo Stock Exchange 2005, Fact Book 2005. Available: http://www.tse.or.jp/english/market/data/factbook/fact_book_2005.pdf [2009, May 22]. Tokyo Stock Exchange 2007, TSE-Listed Companies White Paper of Corporate Governance 2007. Available: http://www.tse.or.jp/english/rules/cg/white_paper.pdf [2009, May 22]. Tokyo Stock Exchange 2009, Fact Book 2009. Available: http://www.tse.or.jp/english/market/data/factbook/fact_book_2009.pdf [2009, May 22]. UNCTAD 2008, World Investment Report 2008, Japan. Available: http://www.unctad.org/sections/dite_dir/docs/wir08_fs_jp_en.pdf [2009, May 22]. Yafeh, Y. 2000, "Corporate governance in Japan: past performance and future prospects", Oxford Review of Economic Policy, vol. 16, no. 2. Yin, R.K. 1994, Case study research: design and methods, Sage, Thousand Oaks, CA.
39
9.
Appendix A
The graphs below show the development of foreign ownership, board size, number of outside directors and percentage of outside directors (y-axis) over time for a sample of 13 companies.
KDDI CORPORATION
FUJITSU LIMITED
HONDA MOTOR CO LTD
NEC CORPORATION
KANSAI ELECTRIC POWER CO INC
MARUBENI CORPORATION
40
MITSUI CHEMICALS INC
OLYMPUS CORP.
MITSUBISHI CORPORATION
COSMO OIL CO LTD
MITSUBISHI HEAVY INDUSTRIES LTD
SOFTBANK CORPORATION
YAMAHA MOTOR CO LTD
41
10.
Appendix B
10.1. Heteroscedasticity
We have controlled for heteroscedasticity by using White’s procedure, through which heteroscedasticitycorrected standard errors are estimated. For some variables White’s heteroscedasticity-corrected standard errors are slightly higher than the OLS standard errors and for others they are slightly lower. Overall the heteroscedasticity-corrected standard errors do not change much from the OLS standard errors, and therefore the p-values do not change much either. Most importantly, no variable becomes insignificant when we control for heteroscedasticity. Our conclusion from these results is that heteroscedasticity is not a problem in our models. BOARD SIZE
------------------- White's estimated standard errors ------------------b se(b) wse(b) wt wp Constant -10,67477 4,27071 5,17476 -2,06285 ,04025 FS_OS -1,03016 ,75182 ,73800 -1,39587 ,16410 ForOwn -6,88439 2,59997 2,53208 -2,71887 ,00705 LnAssets 1,74213 ,30499 ,36690 4,74823 ,00000 AgeDummy -2,69244 1,03804 ,67545 -3,98615 ,00009 Solvency ,02371 ,01963 ,02006 1,18195 ,23845 b = estimated coefficient, se(b) = OLS standard error wse(b) = White's standard error, wt = White's t value, wp = White's p value -------------------------------------------------------------------------
ODTD
------------------- White's b Constant -,17930 FS_OS ,07141 Disp -,00171 ForOwn ,23084 LnAssets ,01680 AgeDummy ,06178 estimated standard errors ------------------se(b) wse(b) wt wp ,15014 ,15106 -1,18689 ,23650 ,02978 ,03482 2,05089 ,04141 ,00124 ,00148 -1,15957 ,24743 ,08405 ,08260 2,79468 ,00563 ,01112 ,01079 1,55751 ,12072 ,03480 ,03000 2,05927 ,04059
b = estimated coefficient, se(b) = OLS standard error wse(b) = White's standard error, wt = White's t value, wp = White's p value -------------------------------------------------------------------------
42
OUTDIR
------------------- White's estimated standard errors ------------------b se(b) wse(b) wt wp Constant -5,30376 1,64111 1,85824 -2,85419 ,00471 FS_OS ,54939 ,32550 ,37840 1,45187 ,14790 Disp -,01679 ,01357 ,01671 -1,00484 ,31603 ForOwn 1,91795 ,91873 ,90757 2,11328 ,03565 LnAssets ,44102 ,12154 ,13223 3,33527 ,00099 AgeDummy ,45118 ,38039 ,33280 1,35572 ,17652 b = estimated coefficient, se(b) = OLS standard error wse(b) = White's standard error, wt = White's t value, wp = White's p value -------------------------------------------------------------------------
10.2. Normality
The histograms below show that the residuals of our models follow a normal distribution fairly well, but not perfectly. We also tested the normality assumption using Jarque-Bera tests, see results below. The Jarque-Bera statistics we obtain exceeds the critical value, and therefore we have to reject the nullhypothesis of a normally distributed error term. However, for large samples it may be possible to relax the normality assumption.
43
Jarque-Bera test of normality We test the null-hypothesis of a normally distributed error term.
The Jarque-Bera statistic (JB) asymptotically follows the chi-square distribution with 2 degrees of freedom. Thus, the critical value is 5.9915.
N Statistic Res_BSize Valid N (listwise) 236 236
Mean Statistic ,0000000
Std. Deviation Statistic 4,17784326
Skewness Statistic 1,193 Std. Error ,158
Kurtosis Statistic 1,917 Std. Error ,316
44
N Statistic Res_ODTD Valid N (listwise) 236 236
Mean Statistic ,0000000
Std. Deviation Statistic ,13989755
Skewness Statistic ,956 Std. Error ,158
Kurtosis Statistic 1,130 Std. Error ,316
N Statistic Res_OutDir Valid N (listwise) 236 236
Mean Statistic ,0000000
Std. Deviation Statistic 1,52918005
Skewness Statistic 1,267 Std. Error ,158
Kurtosis Statistic 4,392 Std. Error ,316
The observed JB-values exceed the critical value, and the null-hypothesis of a normally distributed error term is therefore rejected for all models.
10.3. Multicollinearity
High R-square together with few significant variables is the classic symptom of multicollinearity. The Rsquare for our models is rather low and most variables are significant, indicating that multicollinearity is not a problem in our sample. Another sign of multicollinearity is high pair-wise correlations among independent variables. As shown in the correlation matrix in Table 3 we have no correlation coefficients exceeding 0.8 and only one correlation coefficient exceeding 0.5, which supports that there is no multicollinearity (Edlund 1997). We also estimated the variance-inflating factor (VIF) for our variables to test for multicollinearity. As rule of thumb, if the VIF of a variable exceeds 10, the variable is said to be highly collinear (Gujarati 2003). The VIF for all our variables is below 2, and thus far below the critical value of 10. We conclude that our sample does not suffer from multicollinearity.
45
Coefficients
a
Collinearity Statistics Model 1 FS_OS ForOwn LnAssets AgeDummy Solvency Tolerance ,796 ,738 ,853 ,995 ,791 VIF 1,257 1,355 1,172 1,005 1,263
a. Dependent Variable: BSize
a
Coefficients
Collinearity Statistics Model 1 FS_OS Disp ForOwn LnAssets AgeDummy Tolerance ,569 ,507 ,792 ,720 ,993 VIF 1,758 1,971 1,263 1,389 1,007
a. Dependent Variable: ODTD
Coefficients
a
Collinearity Statistics Model 1 FS_OS Disp ForOwn LnAssets AgeDummy Tolerance ,569 ,507 ,792 ,720 ,993 VIF 1,758 1,971 1,263 1,389 1,007
46
Coefficients
a
Collinearity Statistics Model 1 FS_OS Disp ForOwn LnAssets AgeDummy Tolerance ,569 ,507 ,792 ,720 ,993 VIF 1,758 1,971 1,263 1,389 1,007
a. Dependent Variable: OutDir
47
doc_965551924.pdf