Ownership Structure and Corporate Governance: Has an Increase in Institutional Investors Ownership Improved Business Performance?

Transcription

1 361 Ownership Structure and Corporate Governance: Has an Increase in Institutional Investors Ownership Improved Business Performance? * Hideaki Miyajima Professor, Graduate School of Commerce, Waseda University ( miyajima@waseda.jp) Takaaki Hoda Associate Professor, Faculty of Global Business, Showa Women s University ( hoda@swu.ac.jp) Abstract A dramatic increase in shareholder ownership by domestic and foreign institutional investors, along with the decline of the main bank system, is one of the greatest changes in the governance structure of Japanese companies in recent years. This paper aims to identify the effects of the rapid change in the shareholder ownership structure on corporate governance in Japan. Based on a comprehensive database concerning major shareholders in fiscal that specifies the attributes of shareholders to the maximum possible extent, this paper indicated the following points. First, by examining the selecting stocks among institutional investors, we found that both domestic and foreign institutional investors not only make the selection based on companies size and stock liquidity but also prefer high-quality stocks in terms of profitability, stability and financial soundness. In contrast, banks and insurance companies continue to invest in low-quality companies. Furthermore, compared with domestic institutional investors, foreign ones have a strong tendency to attach importance to formal governance features (size of the board of directors and presence or absence of outside directors). Foreign institutional investors also consistently have a strong home bias. Second, by testing the relationship between ownership and performance, we found that, whereas shareholder ownership by banks and insurance companies has negative effects on enterprise value and corporate earnings, share ownership by domestic and foreign institutional investors has positive effects. This finding indicates that even if a rise in domestic and foreign institutional investors ownership ratio is based on an institutional investor bias or a home bias, or if the impact on the stock price is due to a demand shock, an increase in share ownership by domestic and foreign institutional investors will enable them to exercise the monitoring effect through actions such as indicating the possibility of an exit (possibility of selling off their holdings of shares) and voicing of complaints. Keywords: Ownership structure, investment criteria, corporate governance, corporate performance, institutional investors JEL Classification: G11, G32, G34

2 362 T Hoda, H Miyajima / Public Policy Review I. Introduction The dramatic shift in ownership structure, along with the decline of the main bank system, was one of the most important changes that took place in corporate governance in Japan between the end of the 20th century and the beginning of the 21st century. From the early 1970s to mid-1990s, shareholder ownership in Japanese companies was characterized by the prominent presence of corporate shareholders represented by banks and business corporations. However, following the onset of the banking crisis in 1997, this corporate-dominant ownership structure underwent a dramatic transformation. Companies and banks swiftly dissolved their cross-shareholdings, while domestic and foreign institutional investors increased their shares, foreign ones in particular. 1 By the mid-2000s, it was no longer uncommon to see institutional investors holding a majority of total shares. This brings us to a question: Does the rise in institutional investors presence imply that a new mechanism for management discipline is being created in Japan? This paper examines this question by revealing how domestic and foreign institutional investors, and financial institutions (banks and insurance companies) select stocks, and how such selections affect corporate performance. 2 From the aspect of corporate governance, foreign institutional investors who are increasing their shareholdings and whose aim is to maximize their return on investment may replace former main banks as central players to create a new management discipline, since they are more independent from their investee compared to domestic institutional investors (Ferreira and Matos, 2008). 3 In fact, it is often pointed out that there is a positive correlation between the ownership ratio or the increase in the ratio of foreign institutional investors, and the increase in stock prices. As a drawback, however, foreign institutional investors face serious issues in the asymmetry of information, and thus may not sufficiently produce pre-in- * The data presented in this paper is based on a database developed by Miyajima and Nitta (2011). We would like to thank Keisuke Nitta (Nippon Life Insurance Company) for willingly allowing us to use the data. We are also grateful to Takuya Kawanishi (Prefectural University of Kumamoto) and Kaori Ubukata (Waseda University) for providing their counsel on our study. We received valuable advice from Yasuhiro Arikawa (Waseda University), Ryoichi Arai (Arai Capital Management; Doctoral Program, Graduate School, Waseda University), and Ryo Ogawa (Doctoral Program, Graduate School, Waseda University). We also received invaluable feedback from the following: Toshio Serita (Aoyama Gakuin University; commentator at the Japanese Economic Association 2012 Spring Meeting), Naoki Watanabe (Ritsumeikan University; commentator at the Nippon Finance Association 20th Annual Meeting), participants of Research Institute of Economy, Trade & Industry (RIETI) Frontiers of Analysis in Corporate Governance: Growth, value creation and corporate governance workshop, and participants of a symposium organized by the Tokyo Stock Exchange. In preparing this paper, Japan Society for the Promotion of Science provided us with grants-in-aid for scientific research (Kakenhi and ). Please note that any errors in this paper are attributed to the authors. 1 For details on the process of the unwinding of cross-shareholdings, refer to Miyajima and Kuroki (2007), Miyajima and Kuroki (2002), and Miyajima and Nitta (2011). 2 In this paper, we focus on the impact that the change in ownership structure has on corporate performance; however, we need to also consider the impact that change in ownership structure has on e.g., management reshuffles, dividend policies, business reorganization, and functions of market for corporate control. It is also important to consider the cause and effect between investors stock selections, decisions made by the board of directors, and the corporate actions described above. 3 Aoki (2010) also pointed out this possibility, and suggested that the answer lies in whether or not institutional investors can evaluate the quality of business models.

3 363 vestment monitoring (screening) effects. In addition, while the aggregate of their ownership ratio may be high, each investor s ownership could be low, and therefore respective investors may not have strong incentives to carry out post-investment monitoring. Moreover, foreign investors are known to have a strong behavioral bias when making investments, so it is also possible that the positive correlation between the increase in institutional investors ownership ratio and rise in stock price is the result of such investment bias serving as a demand shock and thereby raising stock prices, rather than the result of pre- or post-investment monitoring. In contrast, domestic institutional investors, whose number increased largely from the end of the 1990s, do not face as many issues in asymmetry of information as foreign institutional investors. Hence, they may be just as capable as, or perhaps even more capable than foreign institutional investors in imposing discipline on corporate management. Generally, institutional investors mainly consist of domestic institutional investors managing investment trusts or pension funds. They increased their ownership ratio from 2000, which was when they entered full-scale into stock investments for pension fund management. Domestic institutional investors most likely strengthened their post-investment monitoring roles not only because of the spread in pension fund mandates, but also due to stronger external pressure for fulfilling fiduciary responsibility. At the same time, many domestic institutional investors are group companies of large financial institutions (banks and insurance companies), e.g., an insurance company s subsidiary specializing in investment advisory, a trust bank affiliated with a bank-oriented group or in a keiretsu group. Therefore, it has been pointed out that such investors (i.e., grey investors) opt for investments that benefit the bank or insurance company that is their parent or affiliate company. 4 Japan is not the only country that experiences this issue where banks and insurance companies invest not only for economic benefits, but also with the intention to gain advantages in business relationships; the same issue can be seen in the U.S. and other countries. Therefore, it may be slightly shortsighted to claim that an increase in institutional investors will lead directly to adequate corporate governance. 5 Finally, it has generally been understood that banks and insurance companies whose ownership ratio declined largely after 1999 make investments with intentions other than to gain economic benefits. For example, banks held client companies shares to ensure credit collections, while life insurance companies held shares mainly to acquire new insurance contracts or maintain existing ones. Consequently, from the standpoint of corporate governance, bank and insurance company shareholders were expected to serve as an entrenchment that protects the company s management from capital market pressure (Morck and Na- 4 Flath (1993), Hiraki et al. (2003), and Ahmadjian (2007). Also, in 2004, Tomomi Yano, former Managing Director of the Employees Pension Fund Association, suggested that Japanese institutional investors lack independence since they are associated with a keiretsu or group (SHOJIHOMU Co., Ltd., 2004). 5 An empirical research conducted in the U.S. regarded this as an issue of disparity in investment criteria between independent and non-independent (i.e., grey) institutional investors (Brickley et al. 1988; Borokhovichi et al. 2006; Chen et al. 2007; Ferreira and Matos, 2008). For example, Ferreira and Matos (2008) indicated that compared to grey institutional investors, independent institutional investors tended to select stocks purely to maximize return on investment.

4 364 T Hoda, H Miyajima / Public Policy Review kamura, 2000). After 2000, however, the banking sector, which had to dispose of non-performing loans, and life insurance companies, which faced serious competition due to deregulations, drastically reduced their shareholdings. In the process, banks and insurance companies may have changed the way they selected stocks. For example, the banking sector, which was required to reduce shareholdings, may have focused on the performance of the investees when selecting stocks, and similarly, insurance companies may have raised their awareness toward fiduciary responsibility when making investments. However, these assumptions have not yet been fully validated in past research. The main objective of this paper is to clarify which of the aforementioned contrasting views regarding domestic and foreign institutional investors and financial institutions (banks and insurance companies) is valid. As with Miyajima and Nitta (2011), the analysis is based on a comprehensive database covering a long period, compiled from companies lists of major shareholders and specifying the attribute of such shareholders to a maximum possible extent. To unravel this question, we broke the analysis down into two steps. First, we analyzed the features of stock preference by shareholder type. According to analyses, both domestic and foreign institutional investors not only select stocks based on company size and stock liquidity, but also tend to select high-quality stocks, as defined in Del Guercio (1996), in terms of profitability, stability and financial soundness, and no gap was detected between domestic and foreign investors in such preferences. In contrast, financial institutions (banks and insurance companies) continued to invest in low-quality stocks with low liquidity up until This implies that Japanese banks and insurance companies continued to hold shares to sustain business ties, rather than to maximize return on investment. In terms of investment criteria, foreign institutional investors are strongly and consistently biased toward companies that have a high overseas sales ratio and whose stocks are incorporated in the MSCI Japan Index, meaning they are widely and well known. Meanwhile, domestic institutional investors include relatively smaller companies in their investment scope. Thus, there is moderate segmentation between the Japanese stocks that domestic and foreign institutional investors invest in. Also, foreign institutional investors apparently prefer companies with strong governance schemes: up until the 2000s, they placed value on the size of the board, and from 2000 onward, they valued the ratio of outside directors. Such preferences could not be confirmed among domestic investors. These findings are consistent with the view that foreign institutional investors are key players in driving governance system reforms. The second step validates whether the rapid change in ownership structure produced monitoring effects, thereby boosting managements efforts and enhancing companies financial results. We will focus on whether the increase in the ownership ratio of institutional investors serves to raise pressure through exits or voicing complaints, and in turn enforces management discipline, regardless of the reason for the increase. According to analyses, there was a positive correlation between firm value, measured by Tobin s Q and financial performance (ROA), and the ownership ratio level of foreign and domestic institutional in-

5 365 vestors, even when considering the reverse causality that institutional investors prefer companies with high financial performance. Comparing the 1990s to the 2000s, the correlation is stronger, especially for domestic institutional investors during the 2000s, which was when the ownership ratio of domestic and foreign institutional investors increased drastically. Although the exact cause and effects behind this phenomenon are yet to be analyzed, it is highly probable that once shareholder ownership by domestic and foreign institutional investors increases, whether it be the result of biased investments, such increase raises the pressure of exits and voices, and thereby has an impact on corporate governance. In contrast, the ownership ratio of banks and insurance companies had a significantly negative impact on companies performance indices, and this impact is once again more apparent during the 2000s. Accordingly, we cannot support the view that banks and life insurance companies started exercising management discipline through shareholding. They remain to act as an entrenchment that guards the management from stock market pressure. 6 This paper is composed of the following sections: Section 2 provides an overview of the changes in ownership structure in Japanese companies after the 1990s. Section 3 analyzes the stock selections of institutional investors, and banks and insurance companies. Section 4 analyzes the impact that the change in ownership ratio of institutional investors has on the price to earnings ratio. Section 5 analyzes the impact that the change in ownership structure has on firm value and financial performance. Section 6 presents our conclusion. II. Changes in Ownership Structure of Japanese Companies and its Background: Outline of Facts II-1. Unwinding of Cross-Shareholding and Increase in Institutional Investors The ownership structure of listed companies in Japan changed dramatically from the 1990s. The ownership ratio of domestic financial institutions (banks and insurance companies) decreased rapidly, and in their place, the ownership ratio of domestic and foreign institutional investors increased substantially. It is worth bringing to attention that the data published by the Tokyo Stock Exchange that is often sited to validate this point is weighted by market capitalization. Thus, the overall trend is determined by the trend of companies with large market capitalization, and the data does not capture the actual diversification in ownership structure in individual companies. It is therefore necessary to grasp the trend of ownership structure in respective companies; however, precautions need to be taken when referring to the shareholder composition presented in companies securities reports. As pointed out in Miyajima and Nitta (2011), classifications used in the shareholder composition are purely formal and do not duly address the differences in investors attributes or their objectives. For example, foreigners include foreign business corporations and foreign institutional investors despite the fact that 6 For example, Tachibanaki and Nagakubo (1997) and Isagawa (2003) pointed out that financial institutions who are stable shareholders may prevent companies from strengthening their management discipline, or damage shareholder value in the long run by not carrying out actions that maximize return on investment.

6 366 T Hoda, H Miyajima / Public Policy Review business corporations and institutional investors have different investment behavior and objectives. Even more problematic is financial institutions, as it consolidates shares held by institutional investors, who aim to maximize return on investment through investment trusts and pension funds, with shares held by banks and insurance companies, who have different objectives. 7 Taking such issues into account, we developed our own data to match the context of this paper. The data captures ownership structure classified more explicitly into foreign institutional investors, domestic institutional investors, and financial institutions (banks and insurance companies). The data was developed by drawing on shareholder compositions and lists of major shareholders released by individual companies from 1990 and onwards, and combining that with a database identifying cross-shareholding between companies. 8 In our data, ownership by foreign institutional investors was calculated by deducting shareholdings by foreign-registered business corporations and major foreign individual shareholders holding 3% or more of total shares, from the foreigner ownership ratio reported in each company s securities report. Thus, this category is mainly represented by foreign pension funds and investment funds whose beneficiaries are large account investors. Ownership by domestic institutional investors is the sum of the ownership ratio of investment trusts, pension trusts, and special accounts for life insurance (entrusted investment of life insurance companies), and was calculated by aggregating accounts from lists of major shareholders. It should be noted that ownership by domestic institutional investors does not include trust bank safe deposits that are managed by investment advisory companies whose beneficiaries are public or private pension funds, since obtaining and identifying such data is difficult. 9 Finally, ownership of banks and insurance companies was calculated by aggregating ownership by banks (excluding trust accounts), life insurance companies (excluding special accounts) and non-life insurance companies. As a benchmark, Table 1 presents the trend of ownership ratio by investor sector based on the number of shares, drawing on Tokyo Stock Exchange data. Table 2 summarizes the simple average and standard deviation of ownership ratio by the institution types defined in the preceding paragraphs, for companies listed on the First Section of the Tokyo Stock Ex- 7 TSE discloses the number of investment trusts and pension trusts included in the aggregate ownership ratio of banks and trust banks; however, it does not disclose respective data of banks and trust banks that deducts the number of such shares. Therefore, we cannot confirm the standalone ownership ratio of banks, which were the main players in cross-shareholding. 8 More specifically, the figures were compiled based on the following steps: A shareholder lists was created for respective non-financial institutions listed on the First Sections of Tokyo Stock Exchange, Osaka Securities Exchange and Nagoya Stock Exchange, by combining Major Shareholders Data (Toyo Keizai Inc.) with the schedule of securities and shareholder distribution contained in companies securities reports (Nikkei Economic Electronic Database System). Based on the recreated shareholder lists, we distinguished foreign institutional investors, domestic institutional investors, and banks and insurance companies. The data is the same as those used in Miyajima and Nitta (2011). We would like to thank Keisuke Nitta for willingly allowing us to use the data. 9 Shares held by trust banks can be classified into trust accounts and own accounts; however, the breakdown of trust accounts cannot be confirmed. Roughly speaking, from the late 1980s to late 1990s, securities trusts, also called tokkin trust funds (specified money in trusts) or fund trusts accounted for the majority of trust bank shareholder ownership; however, from 2000 and onward, it seems that safe deposits of funds set up by investment advisories became dominant.

7 367 Table 1 Trend in Ownership Ratio by Investment Sector Source: Compiled by the authors based on Tokyo Stock Exchange, Shareownership Survey Data. Notes: Unit: %. Share unit-based. Figures for fiscal 2004 to 2009 include shares of companies listed on JASDAQ Securities Exchange. Figures for other fiscal years represent shares of companies listed on all stock exchanges in Japan aside from JASDAQ. Figures are based on data as of each fiscal year-end. Table 2 Average and Standard Deviation of Ownership Ratio by Institution Type Source: Based on Toyo Keizai Inc., Major Shareholders Data, and schedule of securities and shareholder distribution contained in companies securities reports (Nikkei Economic Electronic Database System). Notes: The table aggregates data of companies (excluding financial institutions) listed in the First Sections of the three stock exchanges. As with Table 1, figures are simple averages of the numbers of shares held. Treasury shares are excluded from the total shares used to calculate ownership ratio. Values are as of each fiscal year-end. Figures for foreign institutional investors are values deducting foreign-registered business corporations and foreign-registered large-account individuals holding 3% or more of total shares from the ownership ratio of foreign investors reported in company securities reports. Figures for domestic institutional investors are the sum of pension trusts, investment trusts, and special accounts for life insurance. Figures for banks and insurance companies are values deducting trust accounts from the reported ownership ratio of banks and life insurance companies, then adding non-life insurance and domestic securities holdings. Please note that figures for domestic institutional investors, and banks and insurance companies are calculated by aggregating accounts from lists of major shareholders, and therefore do not completely cover the actual ownership ratio, or cover as many companies as Table 1.

8 368 T Hoda, H Miyajima / Public Policy Review change. Please note that in Table 2, figures for domestic institutional investors, and banks and insurance companies are computed by aggregating accounts extracted from lists of major shareholders, and therefore do not completely cover the actual ownership ratio, and also have a smaller coverage compared to Table 1. We describe the key characteristics of recent changes in listed companies ownership structure below. (1) Ownership structure was consistent until the mid-1990s, characterized by the dominance of financial institutions and business corporations. For companies listed on the First Section of TSE, the average ownership ratio for the two sectors combined exceeded 60%. This structure shifted substantially, triggered by the 1997 banking crisis. The ownership ratio of business corporations, however, did not change all that much even after In contrast, many companies and banks dissolved their cross-shareholdings, and life insurance companies also sold off their shares. Thus, the ownership ratio of banks and life insurance companies in TSE listed companies dropped from 30.4% in fiscal 1996 to 9.7% in fiscal 2006 (Table 1). Although sales of bank stocks by companies leveled out around 2002, sales (selloffs accompanying dissolved cross-shareholdings) by banks continued at a high level. Table 2 shows that the ownership ratio of banks and insurance companies declined some 0.5% to 1.0% per year between fiscal 1996 and (2) The following two factors need to be considered with respect to the process of such unwinding of cross shareholding. First, an important factor driving the rapid sales of stocks was the enforcement of the Act on Limitation on Shareholding by Banks and Other Financial Institutions in Up until then, the government s stance toward shareholding by banks was neutral, so this was a drastic change in policy. Banks were required to reduce their stock holdings to 8% of total assets by September Second, stock purchasing institutions and the Bank of Japan started to purchase stocks in order to ease the impact that the massive amount of selloffs accompanying dissolved cross-shareholdings had on the stock market; however, their purchases were restricted to stocks with credit ratings of BBB or higher. Consequently, banks and insurance companies concentrated on selling stocks that had high liquidity and were issued by companies that fulfilled the conditions above, and as a result, their stock portfolio ended up consisting mostly of companies with high risks (Miyajima and Nitta, 2011: Tables 2-4). (3) Foreign institutional investors were the most active in purchasing the stocks sold through dissolved cross-shareholdings. Table 2 shows that the ownership ratio of foreign institutional investors was 5% or lower in the 1980s, but gradually increased during the 1990s, and reached 7.1% in March This hike was driven by the expansion of pension funds in the U.S. and Europe, as well as rising investments in the Japanese market by U.S. and U.K. institutional investors on the back of increasing oil money in the Middle East (Shirota, 2002). 10 The ownership ratio of foreign institutional investors leveled out during the late 1990s, when the financial crisis occurred, but jumped again from fiscal 2002, and during the four years up until fiscal 2006 their ownership ratio (simple average) increased 8.1%. The second largest group to purchase stocks sold through dissolved cross-shareholdings was domestic institutional investors who procured funds from pension funds. This was

9 369 mainly owing to deregulations in the realm of pension management, as well as institutional investors gearing up stock-based pension managements. In December 1997, the Regulation pertaining to asset management of employees pension fund was abolished, allowing distribution of all funds at one s own discretion and without restrictions. In addition, the pension scheme revision in 2000 caused a fundamental shift in the handling of accumulated funds for employees pensions and national pensions; from entrustment to the former Trust Fund Bureau, to market investments. Furthermore, the Government Pension Investment Fund (GPIF) was established in 2000, prompting the diffusion of managing funds through domestic stock investments. Such investments served as an alternative to foreign institutional investors investments, albeit to a small extent, and the ownership ratio of domestic institutional investors increased 3.2% between fiscal 1999 and II-2. Diversification in Ownership Structure It should be highlighted that the change in ownership structure during the 1990s and 2000s described above did not progress uniformly among listed companies. Figure 1 shows the transition in the ownership ratios of foreign institutional investors, domestic institutional investors, and banks and insurance companies based on a comparison between fiscal 1991, 1997, 2003, and First, at the end of fiscal 1991, foreign institutional investors accounted for less than 3% of total ownership in the majority of companies, but many companies raised this ratio to 10% or more thereafter. By the end of fiscal 2003, the group with foreign institutional investors accounting for 20% to 32% peaked slightly, and as of the end of fiscal 2006, this group represented the mode. Second, the ownership ratio of domestic institutional investors dispersed as can be seen at the end of fiscal 2003, when the group with the lowest ratio of 0% to 9%, which was initially the mode, decreased, while the group with a ratio of 10% or more increased rapidly. Finally, for banks and insurance companies, the group with a ratio of 10% or more, which was initially the mode, decreased drastically by the end of fiscal 2003, and as of the end of fiscal 2006, the group with a ratio of less than 3% accounted for 45% of all listed companies. At the same time, the group with a ratio of 10% or more reached 12.2%. These results indicate that each institution type s investment behavior is affected by some kind of preference or bias. Generally, institutional investors are believed to have a strong preference for large companies. We therefore examined ownership by market capitalization, as shown in Table 3. According to Panel 1, the increase in ownership by foreign institutional investors from the early 1990s occurred mainly in the group with high market capitalization (the fourth and fifth quintiles), and this trend continued through the 2000s. The ownership ratio of foreign 10 For example, according to Jacoby (2009), CalPERS (California Public Employees Retirement System), the largest public pension in the U.S., showed increasing interest in foreign investments from the early 1990s, focusing on Japan alongside the U.K. as investment markets. In 1993, the ratio of foreign stocks incorporating Japanese stocks reached 45%. This caused 300 companies to exercise dissenting votes at the shareholder s meeting that same year.

10 370 T Hoda, H Miyajima / Public Policy Review Figure 1 Distribution (frequency) of Ownership Ratio of Domestic Institutional Investors, Foreign Institutional Investors, and Banks and Insurance Companies: in 1991, 1997, 2003, and 2006 Source: Compiled by combining Toyo Keizai Inc., Major Shareholders Data, and schedule of securities and shareholder distribution contained in companies securities reports (Nikkei Economic Electronic Database System). Note: Data is comprised of companies listed in the First Sections of the three stock exchanges (excluding financial institutions). As with Table 1, figures are simple averages of the numbers of shares held. Treasury shares are excluded from the total shares used to calculate the ownership ratio. Values are as of each fiscal year-end. Figures for foreign institutional investors are values deducting foreign-registered business corporations and foreign-registered large-account individuals holding 3% or more of total shares from the ownership ratio of foreign investors as reported in company securities reports. Figures for domestic institutional investors are the sum of pension trusts, investment trusts, and special accounts for life insurance. Figures for banks and insurance companies are values deducting trust accounts from the reported ownership ratio of banks and life insurance companies, then adding non-life insurance and domestic securities holdings.

11 371 Table 3 Trend of Ownership Ratio by Market Capitalization Quantile Source: Same as Table 2. Notes: The table aggregates data of companies (excluding financial institutions) listed in the First Sections of the three stock exchanges. As with Table 1, figures are simple averages of the numbers of shares held. Treasury shares are excluded from the total shares used in calculating the ownership ratio. Foreign represents foreign institutional investors, and Domestic represents domestic institutional investors. The total is the sum of the two values as of each fiscal year-end. Quantiles were created for companies listed on the First Section of TSE, based on the market capitalization as of each fiscal year-end (the fifth quantile has the largest market capitalization), and the average ownership ratio was calculated for each quantile. The median market capitalization of each quantile in fiscal 2006, when the ownership ratio of institutional investors was the highest, was billion yen in the fifth quantile, billion yen in the fourth quantile, 52.3 billion yen in the third quantile, 27.3 billion yen in the second quantile, and 12.2 billion yen in the first quantile. The threshold for each quantile was billion yen between the fourth and fifth quantiles, 81.2 billion yen between the third and fourth quantiles, 37.4 billion yen between the second and third quantiles, and 19.2 billion yen between the first and second quantiles. Figures for foreign institutional investors are values deducting foreign-registered business corporations and foreign-registered large-account individuals holding 3% or more of total shares from the ownership ratio of foreign investors as reported in company securities reports. Figures for domestic institutional investors are the sum of pension trusts, investment trusts, and special accounts for life insurance. Figures for banks and insurance companies are values deducting trust accounts from the reported ownership ratio of banks and life insurance companies, then adding non-life insurance and domestic securities holdings. institutional investors in the first quintile, the group with the lowest market capitalization, remained at a low level of 4.8% as of the end of fiscal 2006, while the same ratio for the fifth quintile, the group with the largest market capitalization, amounted to 26.0%. This indicates that foreign institutional investors consistently place emphasis on company size as a criterion when selecting stocks. This also reflects the fact that foreign institutional investors investing in Japanese stocks tend to invest in stocks incorporated in the MSCI Japan Index. 11 Meanwhile, the ownership ratio of domestic institutional investors increased substantial- 11 Since 2000, MSCI Japan Index consists of stocks for some 300 to 400 companies with high market capitalization. According to interviews that we conducted with foreign institutional investors between July and December 2011, from the perspective of liquidity, many investors set a floor of 100 billion yen in market capitalization when making investments. This is because low market capitalization can cause issues, for example, the investor s trades will affect the stock price, or selling shares will consume time. So even if investors want to invest in such small companies, they are prevented from doing so.

12 372 T Hoda, H Miyajima / Public Policy Review ly in companies belonging to the fourth and fifth quintiles around The aforementioned pension entrustment business initially expanded with a focus on large companies. However, from the late 2000s, their ownership increased in small- and medium-sized companies (the first and second quintile), and decreased in the fourth and fifth quintile. As a result, the ratio gap contracted between companies of different sizes by the end of fiscal Finally, let us look at the trend in the ownership ratio of banks and insurance companies by business size during the same period (Panel 2). The decline in ratio for the fourth and fifth quintile was relatively sharp between the end of fiscal 1996 and 2006, when cross-shareholdings dissolved. This trend suggests that during the time when cross-shareholdings were dissolved, banks and insurance companies prioritized selling stocks of companies with high liquidity and market capitalization (stocks that were easy to sell, or that had relatively low effects on the company s stock prices even if they were sold). From the end of fiscal 2006, when sales of stocks by banks and insurance companies leveled out, the ownership ratio of banks and insurance companies bottomed out in large companies, while the same ownership ratio in companies in the first and second quintile also decreased, and the gap in ratio between companies of different sizes contracted. In the next section we will unveil the factors that guided such changes in the ownership ratios of banks and insurance companies, which were distinct between companies of different sizes. III. Stock Selection by Institutional Investors, and Banks and Insurance Companies III-1. Investment Criteria In this paper, we are mainly concerned with whether corporate governance in Japan was affected by the drastic change in ownership structure, which was played out differently among companies of different sizes during the 1990s and 2000s. First, let us focus on pre-investment monitoring (screening). We can deduce that institutional investors, whose ownership ratio increased drastically, elicit pre-investment governance effects if they: adequately evaluate companies business models; invest in stocks of companies that are expected to improve efficiency and achieve high growth; and sell stocks of companies that are not expected to improve profitability or achieve growth. Unfortunately, there is no forthright method to assess the levels of such pre-investment monitoring. We therefore examine this issue by looking at the stock preference of each shareholder type. Institutional Investors Bias and the Prudent Man Rule Two criteria determine the stock selection of institutional investors, who aim to maximize return on investment while also conforming to their duty of diligence: a quantitative criterion based on stock price valuation, and a qualitative criterion such as the management s initiatives in corporate governance. As for the former quantitative criterion, institu-

13 373 tional investors do not necessarily share a uniform criterion for valuation. For example, even with basic indices such as PER (Price Earnings Ratio) and PBR (Price Book-value Ratio; or its reciprocal Book to Market Ratio), it cannot unanimously be determined whether a higher or lower PER and PBR means that pre-investment monitoring is adequate. Meanwhile, empirical analysis shows that there are several views on what kind of stocks institutional investors prefer. For example, based on a study of Swedish companies, Dahlquist and Robertsson (2001) suggested that institutional investors prefer stocks with high market liquidity, offered by companies that are large, and have low dividend payments and high cash positions. They called this institutional investors bias. Based on an analysis of U.S. companies between 1980 and 1996, Gompers and Metrick (2001) revealed that institutional investors preferred large-cap stocks with high liquidity and low shareholder returns for the previous year. These features were different from those preferred by other types of investors. 12 The institutional investors that we interviewed mentioned that they more or less agreed to the above statements. At the same time, institutional investors are entrusted with funds from an ultimate beneficiary, and are thus required to make investments conforming to fiduciary responsibilities. According to previous research, institutional investors therefore tend to prefer high-quality stocks, i.e., stock of companies with high profitability, stability and financial positions. Del Guercio (1996) titled this fiduciary motives as prudence, and suggested that institutional investors, who are required to take the utmost caution as fiduciaries, prefer stocks that are low-risk and can definitely yield profits. Preference for Strong Governance Institutional investors tend to attach weight to the investee s level of corporate governance as a qualitative criterion. To meet their fiduciary responsibilities, domestic institutional investors established guidelines for exercising voting rights among others around the beginning of the 2000s, clarifying that they will actively invest in companies that disclose sufficient information and assign outside directors. Previous empirical research, for example, Ferreira and Matos (2008) also pointed out that institutional investors prefer stocks of companies not only that are large yet have adequate governance schemes in place (companies with dispersed ownership and that take progressive measures in information disclosure). 13 Also, Leuz et al. (2009) suggested that foreign investors who face restrictions in their investments due to asymmetry of information and higher monitoring costs tend to 12 The intention of their research was to answer a question: Why did small-cap stock premium (i.e., small-cap stocks yielding a higher return on investment than large-cap stocks) vanish from the U.S. market? They discovered that one of the factors was because the ownership ratio of institutional investors accounted for an increasingly larger portion of the stock market, and boosted return on investment for large-cap stocks. 13 More specifically, the study analyzed the role that institutional investors played in monitoring corporate management based on data of 27 countries (for the six years from January 2000 through December 2005 inclusive) against the backdrop of the increasing volume of assets managed by institutional investors worldwide. As of the time of the analysis, it was estimated that institutional investors held close to 40% of the total global market capitalization, and half the floating stocks.

14 374 T Hoda, H Miyajima / Public Policy Review avoid investing in companies that do not disclose enough information, or that pose governance-related risks. Giannetti and Simonov (2006) stated that outsiders (domestic and foreign institutional investors and individual investors) whose only objective is to maximize returns on investment prefer to invest in companies with adequate governance schemes in place. 14 Asymmetry of Information and Home-Bias Even under the premise that foreign institutional investors are motivated by the principle of maximizing return on investments and by fiduciary responsibilities as independent parties, this does not ensure that they will fulfill the role of (pre-investment) monitoring. The most profound restriction that foreign institutional investors face is asymmetry of information; i.e., they cannot obtain sufficient information about Japanese companies. Foreign institutional investors that cannot fully assess the profitability or risks of Japanese companies will end up selecting stocks of companies that are well known among the ultimate fund provider (public pensions and/or investors), based on such indices as whether they have a high overseas sales ratio or issue ADRs. However, foreign institutional investors may be prevented from fulfilling their roles in adequate pre-investment monitoring if they rely heavily on such formal indices to assess the quality of corporate governance mentioned above, since the influence of outside directors depends on each company s business characteristics and is not necessarily valid in all cases. Much previous research has been conducted concerning such home bias of foreign investors. For example, Dahlquist and Robertsson (2001) indicated the presence of such bias based on data of Swedish companies. Also, Ahearne et al. (2004) analyzed U.S. investors shareholding of offshore stocks, and concluded that one of the factors preventing them from holding offshore stocks is the information cost required in dealing with the asymmetry of information. 15 It has been pointed out that home bias by foreign institutional investors also applies to Japanese stocks. Based on research for 1975 through 1991, Kang and Stulz (1997) reported that the ownership ratio of foreign investors is high in Japanese companies that have high liquidity (turnover ratio) and are listed on ADR. Based on data from 1985 to 1998, Hiraki et al. (2003) claimed that foreign institutional investors investing in Japanese stocks had a strong tendency to invest in companies with high export ratios, and that their stock selections were impacted substantially by levels of recognition and familiarity. Also, based on analysis focusing on 1990 through 2008, Miyajima and Nitta (2011) confirmed that foreign institutional investors valued company size and overseas sales ratio when selecting 14 Aggarwal et al. (2005) analyzed investments in developing countries by institutional investors managing U.S. investment trusts, and reported that they favored countries with established accounting standards, shareholder ownership, and laws and regulations, while also favoring companies that conduct transparent disclosure of accounting information, and issue an ADR. The study claims that establishing a robust governance scheme at both the company and country level will contribute to attracting more foreign investors. 15 Other research, e.g., Merton (1987), French and Poterba (1991), and Brennan and Cao (1997) also confirmed the presence of home bias.

15 375 stocks, and that in that sense they were inclined toward home bias. What previous research targeting Japan did not fully take note of was the impact that the MSCI Japan Index had on determining investment scope. According to an interview with foreign institutional investors, as a general rule, funds that carry out offshore investment instructions only invest in stocks that are incorporated in the MSCI Japan Index. 16 It is highly likely that an increase in the ratio of foreign institutional investors is confined to stocks incorporated in the MSCI Japan Index; therefore, it is necessary to find out to what extent this factor determines the stock selection of foreign institutional investors. Banks and Insurance Companies: Smart vs. Grey Investors While foreign institutional investors are most likely to confront issues in asymmetry of information, domestic institutional investors are immune from such issues. Domestic institutional investors can select stocks of companies that have high potential for growth, regardless of whether or not the company is listed in overseas stock markets, or is widely and well known, e.g., its stocks are a constituent of the MSCI Japan Index. If domestic institutional investors are free from home bias factors when selecting stocks, it implies that they have strong smart investor features. However, if domestic banks and insurance companies are bound by their business relationships (e.g., loan contracts, insurance contracts, keiretsu relationship) when selecting stocks as earlier mentioned, they still maintain strong grey investor features. It is necessary to analyze which feature is more pronounced. III-2. Estimation Model In this section we analyze the determinant of ownership ratio by shareholder type (foreign institutional investors, domestic institutional investors, and banks and insurance companies). The analysis targets non-financial business corporations that are listed in the First Section of the three stock markets, and whose fiscal year ends in March. The analysis focuses on the 19 fiscal years between fiscal 1990 (fiscal year ending March 31, 1991) and 2008 (fiscal year ending March 31, 2009) inclusive. We used Quick Astra Manager to extract companies incomes, financial data, and stock price-related data. The model below was used as an estimation model, drawing from Gompers and Metrick (2001), Ferreira and Matos (2008), and Miyajima and Nitta (2011). 16 The breakdown of foreign investors ownership cannot be determined explicitly; however, asset managements that are completely passive are few, and the breakdown consists of (1) funds centered on MSCI Japan Index stocks (representing some 300 to 400 stocks since 2000) that carry out investment instructions in Japan and offshore; and (2) funds and entrustments of asset management set up as Japanese stocks by overseas subsidiaries of Japanese Companies. (1) can be categorized into (1)-A, passive asset management linked to the weight of market capitalization for stocks incorporated in MSCI to the extent possible; and (1)-B, proactive asset management that selects some 50 to 100 stocks out of the MSCI stocks. Meanwhile, (2) consists of funds and entrustments of asset management for foreign pensions or large account customers, and there is a floor to the market capitalization and liquidity of the investee. Chronologically, (1)-A was initially mainstream; (1)-B and (2) increased around 1999, when foreign investors investments were on an upward trend; and (1) and (2) increased simultaneously from 2003 onwards in line with the rise in foreign investors investments.

16 376 T Hoda, H Miyajima / Public Policy Review SH jit = F (INSB it, QS it, GOV it, HB it, GOV it, CONT it, YEAR) (1) The dependent variable SH jit is the ownership ratio of company i belonging to shareholder type j, and is adjusted by deducting the ratio of floating shares. 17 Independent variable IN- SB it is a variable indicating the basic preference of institutional investors (institutional investors bias), and tests whether institutional investors tend to prefer large companies with high liquidity. 18 The logarithmic value of market capitalization and the trading turnover ratio are used respectively as proxy variables. The coefficients of each variable are expected to be positive against the ownership ratio of institutional investors. Also, following previous research, we used the PBR. Coefficients should be positive if investors prefer value stocks. QS it is a variable indicating the degree to which each investor is conscious of fiduciary responsibilities. According to Del Guercio (1996), institutions with strong fiduciary motives as prudence prefer companies of high quality in terms of profitability, stability, and financial soundness among other aspects (high-quality stock). In this paper, we used stock price volatility, dividend yield, ROA, investing opportunity (growth rate of net sales for the past two years), debt ratio and the ADR dummy (U.S. stocks that issue ADR) as the proxy variable. If institutional investors perform their fiduciary responsibilities with the utmost precautions, it is expected that they will prefer stocks with low risks that ensure definite returns. Hence, stock price volatility and debt ratio are expected to be negative, while ROA, investing opportunity, dividend yield and the ADR dummy are expected to be positive. HB it is a variable that captures home bias factors. More specifically, we used companies overseas sales ratio. Kang and Stulz (1997), Hiraki, Ito and Kuroki (2003), and Miyajima and Nitta (2011), reported the presence of home bias in Japan as well, and therefore, the overseas sales ratio is expected to be positive. In addition, as mentioned earlier, according to interviews with foreign institutional investors, the MSCI Japan Index is extremely important in determining investment scope, so its impact is expected to be large. We therefore applied a dummy variable that is a value of one when the stock is incorporated in the MSCI Japan Index. 19 GOV it is a variable that represents corporate governance. We used an objective index; namely, the number of executive officers, and the ratio of outside directors excluding those affiliated to banks or the parent company. 20 If both domestic and foreign institutional inves- 17 Gompers and Metrick (2001), Ferreira and Matos (2008), and Miyajima and Nitta (2011) all used the ownership ratio of institutional investors as the dependent variable, and incorporated the ratio of floating shares as the independent variable. Japan is characterized by the presence of cross-shareholdings, and the ratio of floating shares is expected to affect the ownership ratio of cross-shareholding by institutional investors or banks and insurance companies to a certain degree. Therefore, rather than using the ownership ratio of each shareholder type, we believe that it is more adequate to base the analysis on ownership ratio adjusted by the floating share ratio. In order to assure the liability of our results, we also conducted the analysis using the ownership ratio not adjusted by the floating share ratio, but the results were roughly the same. Toshio Serita (Aoyama Gakuin University) pointed out the endogeneity of the ratio of floating shares to us. 18 Gompers and Metrick (2001). This was also confirmed through interviews with domestic and foreign institutional investors conducted by the authors. 19 To the authors knowledge, there is no previous analysis on foreign institutional investors home bias with respect to Japan that considers the element of MSCI Japan Index stocks. 20 The above-mentioned ADR dummy, which is included in fiduciary responsibility, can also be considered an objective variable for corporate governance.