Board remuneration, company performance, and ownership concentration: evidence from publicly...
I. Introduction
This study investigates the relationship between total board remuneration and corporate performance in a sample of firms listed on the Kuala Lumpur Stock Exchange (KLSE). Our objective is to examine the relationship between performance and executive salaries within a principal-agent
This paper makes two main contributions to the existing literature on executive compensation and corporate performance. First, there are several studies that investigate the relationship between executive compensation and corporate performance in Europe and the United States (see, for example, Murphy 1985; Main, Bruce, and Buck 1996; Brunello, Graziano, and Parigi 1997; Conyon 1997; Crespi and Gispert 1998). However, there has been little investigation of the relationship between corporate performance, performance criteria, and executive compensation in firms in Asian countries.
This is an important gap in the literature, given that there are differences in the level of ownership concentration between Asian and Anglo-American economies. Berle and Means' (1932) classic description suggests that ownership in the United States is dispersed among small shareholders with control concentrated in the hands of the managers. This description has been questioned in more recent research. La Porta, Lopez-De-Silanes, and Shleifer (1999) found that ownership in countries other than the United States is more concentrated than depicted in Berle and Means (1932). There is some evidence to suggest that even in the United States ownership is now more concentrated than when Berle and Means (1932) wrote. Studies such as Morck, Shleifer, and Vishny (1988a) and Shleifer and Vishny (1986) have found at least a moderate degree of ownership concentration in the United States. Nevertheless, the concentration of ownership is greater in most Asian countries than in either the United Kingdom or the United States. In Malaysia's case, the main reasons for high levels of ownership concentration are the prevalence of family control and the significant amount of state control in listed companies (see, for example, Claessens, Djankov, and Lang 1999, OECD 1999).
This has potentially important implications for the relationship between executive compensation and corporate performance. It suggests that the agency problems arising from the separation of ownership and control may not be as pronounced in Asian countries compared with the United Kingdom or the United States. If this is the case, it implies that the relationship between board remuneration and firm performance should be weaker than in the United Kingdom or the United States. The rationale behind this is that the level of ownership concentration is a proxy for the intensity of shareholder supervision. When ownership is diverse, the potential exists for sub-optimal monitoring because individual shareholders are unable to fully appropriate the gains from the monitoring function. Because monitoring has the features of a public good, an individual shareholder who carries out the monitoring function personally bears the full specification and enforcement costs, but in return appropriates just a fraction of the assumed total gain. As a result, the probability of free rider behaviour will increase in companies where ownership is dispersed, and the discretionary power of managers will be higher because shareholder supervision will be sub-optimal (Vickers and Yarrow 1988; Conyon and Leech 1994). Thus, higher levels of compensation will be needed to increase manager effort. On the other hand, when ownership is concentrated, shareholders have better information and are better placed to monitor manager effort. In these cases compensation can be tied more closely to actual or observable effort.
A second contribution of this paper is to use total board remuneration as a proxy for executive compensation. A few existing studies look at the relationship between total board remuneration and corporate performance (Main, Bruce, and Buck 1996; Brunello, Graziano, and Parigi 1997). Most studies, however, use compensation of the highest paid director or chief executive officer (CEO). Moreover, none of the studies of the relationship between executive compensation and performance in Asian countries use total board remuneration. As a result, in a recent literature review of research in this area Indjejikian (1999, p. 152) pleads for future researchers to pay much closer attention to performance evaluation and compensation practices of directors other than the CEO. Main, Bruce, and Buck (1996, p. 1634) suggest that in most cases researchers have used the remuneration of the CEO because data has not been available for the board as a whole. Where data is available, total board remuneration is a better proxy for executive compensation. This is because, in terms of agency it is the board collectively, rather than the CEO or individual directors, that act on behalf of the principals or shareholders (Gower 1992, p. 160; Main, Bruce, and Buck 1996, p. 1634).
The paper is set out as follows. The next section provides a conceptual framework for considering the factors which determine executive compensation, sets out several hypotheses based on previous research on firms in the United Kingdom and the United States, and considers the implications of these hypotheses for Asian countries where ownership is more concentrated. Section III discusses the data set and empirical specification used in the study. The results are examined in Section IV. The final section contains some concluding comments and considers implications for future research.
II. Conceptual Framework and Hypotheses
The design of compensation contracts for managers has been an important topic of investigation in the principal-agent literature. The objective is to formulate an optimal compensation scheme that motivates managers to maximize firm performance, taking into account governance arrangements. The specific governance conditions depend on the institutional rules and markets in which the firm operates, which are related, among other things, to factors such as industry or sector performance, firm size, and the structure and composition of ownership. These factors are able to influence the board-shareholder relationship, impacting on problems of information asymmetries. This, in turn, influences the manner in which control and supervision of managers is exerted (Crespi and Gispert 1998, p. 8).
Principal-agent models of the firm explain firm performance in terms of manager's effort and a set of randomly distributed variables that are outside the manager's control and not observable (see, for example, Tirole 1988 and Rosen 1992). These models suggest that in order to increase the manager's effort, compensation should be sensitive to performance. This assumes that effort and firm performance move in the same direction, otherwise the observability of performance would be irrelevant in motivating managers to increase effort (Crespi and Gispert 1998). While the existing literature is far from definitive on the point, several studies support the existence of a positive relationship between board or CEO compensation and performance (see, for example, Murphy 1985; Main, Bruce, and Buck 1996; Gibbons and Murphy 1990). Thus, the first hypothesis based on existing studies for firms in the United Kingdom and the United States is that there is a positive relationship between firm performance and board remuneration.
However, as discussed in the introduction, an important implication of higher levels of concentration in Malaysia, and Asia more generally, is that the agency problems arising from the separation of ownership and control are not as pronounced as in Anglo-American economies. Differences in ownership concentration are illustrated in Table 1, which presents comparable statistics for the United States and United Kingdom from La Porta, Lopez-De-Silanes, and Shleifer (1999) and several Asian countries, including Malaysia, from Claessens, Djankov, and Lang (1999) for 1996/97. The table shows who controls the twenty largest publicly traded companies in the United Kingdom, United States, and a sample of Asian countries and who controls all the firms in the Claessens, Djankov, and Lang (1999) sample for the Asian countries using a 20 per cent cut-off. Using a 20 per cent cut-off means that a corporation has a controlling shareholder (ultimate owner) if the shareholder's direct and indirect voting rights in the firm exceed 20 per cent. The number of firms for each Asian country in the Claessens, Djankov, and Lang (1999) sample is given in the table. In general, the sample covers about three-quarters of market capitalization, although the share of firms in the sample relative to the total number of listed firms is less than 50 per cent for Malaysia, South Korea, Taiwan, and Thailand.
Table 1 shows that all the shares in the largest 20 firms in the United Kingdom were widely held, and 80 per cent of the shares in the largest firms in the United States were widely held. A widely-held corporation is a corporation which does not have any owners who have significant control rights. These figures for the largest companies in the United Kingdom and United States might overstate the level of ownership dispersion in these countries because one would expect that the likelihood of wide ownership dispersion would be greater in very big firms (La Porta, Lopez-De-Silanes, and Shleifer 1999, p. 472). La Porta, Lopez-De-Silanes, and Shleifer (1999) also collect information on the ten smallest firms with market capitalization of at least US$500 million. Among these firms in the United States, 90 per cent of shares were widely held and 10 per cent were controlled by families at the end of 1995. In the United Kingdom, the comparable figures were 60 per cent widely held and 40 per cent controlled by families. Using these figures, ownership appears to be more concentrated in the United Kingdom, but the general impression that ownership in the United Kingdom is widely dispersed holds true.
A problem with Table 1 is that it compares Asian firms solely with U.S. and U.K. firms where, it could be argued, the size differences make this comparison inappropriate. Thus, Table 2 is presented to compare ownership concentration with a range of European countries for 1998, in addition to comparisons with U.K. and U.S. firms. Table 2 shows that the percentage of shares held by the largest shareholder in Malaysia, the Philippines, and Thailand is similar to those in European countries such as the Netherlands and Spain, but less than in countries such as Austria, Belgium, France, Italy, and Germany. Meanwhile, ownership concentration of the largest shareholder in China and Indonesia is similar to that in Italy and slightly less than in Belgium and France. However, if shares held by the largest five shareholders are considered, the figures are generally similar for most of the Asian and European countries shown in Table 2.
Table 2 shows that the five largest shareholders in Belgium own 63.3 per cent of the shares, and the largest five shareholders in Germany own 70.9 per cent of the shares. This compares with Indonesia (67.5 per cent), Malaysia (58.8 per cent), Philippines (60.2 per cent), and Thailand (56.6 per cent). The figure for Malaysian companies is similar to that for China and transitional economies in Central Europe such as the Czech Republic where there is or has been substantial amounts of state ownership. Xu and Wang (1998) find that the average concentration of the five biggest investors in a sample of 316 Chinese listed companies in 1995 was 58.1 per cent. Claessens and Djankov (1999) find that the average concentration of the five largest investors in a sample of 706 listed Czech firms in 1996 was 60.2 per cent. The two Asian countries which have much lower ownership concentrations are Japan and South Korea. According to Table 2, the five largest shareholders in South Korea owned 38.5 per cent of the shares. We do not have figures for 1998 for Japan, but Prowse (1992) found the corresponding number for Japanese firms in the 1980s was 32 per cent.
Table 2 reinforces the conclusions from Table 1 that shareholding in the United Kingdom and United States is dispersed relative to continental Europe and most of Asia. In 1998 the largest shareholder in the United Kingdom owned 14 per cent of the shares and the largest shareholder in the United States owned 22.8 per cent of the shares. The largest five shareholders in the United States owned 39.8 per cent of shares, which is similar to South Korea. This figure for the largest five shareholders in the United States is slightly higher than the comparable figure reported by Demsetz and Lehn (1985) which was 25 per cent. The conclusion that the United States and United Kingdom have low rates of ownership concentration is reinforced by other studies. For example, in a study of corporate governance and ownership in nine European countries (Belgium, France, Switzerland, Netherlands, Italy, Germany, Sweden, Spain, Austria and the United Kingdom), Wymeersch (1998) concluded that the United Kingdom had the lowest ownership concentration in Europe.
Differences in ownership concentration between the United States and the United Kingdom on one hand and continental Europe on the other can be partly explained in terms of different legal codes. La Porta et al. (1998) show that common-law countries generally have the strongest and French civil-law countries the weakest legal protection for investors with German and Scandinavian civil-law countries in the middle, holding per capita income constant. The common law is widespread in countries that originated as English colonies while the civil law codes are dominant in continental Europe (see La Porta et al. 1998, Table 2). La Porta et al. (1998) find that ownership concentration in the largest public companies is strongly negatively related to investor protection.
Tables 1 and 2 suggest that ownership concentration in Malaysia is indicative of most Asian countries. This conclusion is independently supported by a range of studies over a long period of time. These include Lim (1981) and the OECD (1999). Table 3 presents data on ownership concentration in the ten largest listed companies in Malaysia by market value in 1999. It indicates that following the Asian financial crisis ownership is still quite concentrated. On average, the largest shareholder in the five largest companies owns 46.11 per cent of shares and, on average the largest shareholder in the ten largest companies owns 41.27 per cent of shares.
One reason for the high levels of ownership concentration in countries such as Indonesia, Korea, Malaysia, Singapore, and Thailand is significant amounts of state control of listed firms. In Malaysia, the main political parties--United Malays National Organization (UMNO), Malaysian Chinese Association (MCA), and the (Malaysian Indian Congress (MIC)--have substantial business holdings. Claessens, Djankov, and Lang (1999) report that using the percentage of shares owned by the largest ten shareholders as a benchmark, the state controls 17.8 per cent of publicly listed companies in Malaysia. For the twenty largest publicly listed companies, this figure increases to 36.3 per cent. This is the second highest figure in the Claessens, Djankov, and Lang (1999) study for the Asian countries after Singapore, where the comparable figure is 45.8 per cent.
However, the most important reason for high levels of ownership concentration in most of Asia and Malaysia in particular is that family control of listed companies is more prevalent in Asia than in the United States or United Kingdom. Families controlled 67.2 per cent of the 238 publicly listed Malaysian firms in the Claessens, Djankov, and Lang (1999) study and 35 per cent of the largest twenty firms (see Table 1). According to Claessens, Djankov, and Lang (1999), in 1996, the top five families in Malaysia controlled 17.3 per cent of market capitalization; the top ten families controlled 24.8 per cent of market capitalization; and the top 15 families controlled 28.3 per cent of market capitalization on the KLSE.
Fama and Jensen (1983) and DeAngelo and DeAngelo (1985) argue that family relationships among owner-managers should reduce agency costs because when ownership and control rests with the same person the need for outside monitoring is reduced. Writing about the situation in India where family control and ownership is also high, Ramaswamy, Veliyath, and Gomes (2000) put forward a similar view. Their argument is that a high concentration of family ownership has important implications for CEO compensation because it functions as a built-in system of checks and balances, which operate to reduce managerial excesses. Ramaswamy, Veliyath, and Gomes (2000, p. 171) state:
[U]nlike many Western family companies that prefer a "hands off" approach by using professional managers, Indian firms seek more direct control. An offshoot of this form of control is the dampening effect on CEO pay. The process of informal family approval of the CEO provides a built-in check against excessive compensation. Since any "unspent" money reverts to the family, such firms do not emphasize CEO pay at all. [M]any family patriarchs leading family businesses are content with a token salary because their primary rewards accrue through increasing the value of the firm over time.
Several studies have found that firms which are managed by members of the founding family are less likely to experience managerial succession or be subject to hostile takeover relative to widely dispersed firms even in periods of low profitability (see, for example, Allen and Panian 1982; Morck, Shleifer, and Vishny 1988b). The empirical evidence on the effect of large family holdings on executive compensation is mixed. Goldberg and Idson (1995) find that large holdings by a family concern generally exert a positive influence on executive compensation, reflecting a founding father effect. However, consistent with their argument above, Ramaswamy, Veliyath, and Gomes (2000) find the proportion of family ownership to be negatively correlated with CEO compensation in India. Deckop (1988) finds that CEOs who are founders of the firm earn less than internally promoted CEOs who in turn earn less than those recruited from outside the firm.
There is a sizeable allied literature looking at the effect of family ownership on firm value. Using data for the United States, McConaughy et al. (1998) find that founding family controlled firms have more efficient operations as well as higher values relative to non-founding family controlled firms after controlling for differences in industry and size. Their explanation for this finding is that family relationships improve monitoring while providing incentives that are associated with better firm performance. In contrast, Yermack (1996) finds that Tobin's Q is lower in firms controlled by a founding CEO. Morck and Strangeland (1994) find that firms controlled by descendants of the founder have lower rates of profits than comparable firms controlled by outsiders. Morck, Shleifer and Vishny (1988a) get mixed results. On one hand they find that Tobin's Q for older firms is lower when the firm is run by a member of the founding family compared with an outsider. On the other hand, among newer firms the presence of a member of the founding family raises Tobin's Q. They explain this result on the basis that founders with long tenures and their descendants thwart value maximization, but newer founders possess added entrepreneurial talents.
Some studies have considered whether succession by family members is value enhancing. One argument suggesting that family succession might have a negative effect on firm value is that the market might see it as a form of nepotism particularly if there are other individuals who are more appropriate for the position. Another consideration is that there might be more "quality uncertainty" surrounding the appointment of a family member than a non-family insider or outsider (Smith and Amoako-Adu 1999, pp. 344-45). On the other hand family members can be regarded as having a special interest in the success of the firm and this might enhance firm value. Using Canadian data, Smith and Amoako-Adu (1999) found that stock prices initially declined when family successors were appointed, while there was no significant change when non-family insiders or outsiders were appointed. However, they also found that the fall in stock prices was related to the relatively young age of family successors, suggesting that their lack of managerial experience rather than family connections per se was the real reason for the fall in share prices.
Most of the arguments relating to the relative efficiency of family enterprises has been considered in an Asian context in relation to ethnic Chinese family-run enterprises (see Whyte 1996, pp. 3-5). In the 1960s, the Weber inspired view was that ethnic Chinese family-run enterprises were inefficient and an obstacle to economic development. More recently, primarily because of the economic success of many Southeast Asian nations with large emigrant Chinese populations over the last twenty-five years, the Chinese family-run firm has come to be seen as an engine for development in Southeast and East Asia. After a balanced review, Whyte (1996) concludes that both the obstacle and engine for growth arguments are oversimplified.
There is limited empirical evidence on the issue. Studies reviewed in Greenhalgh (1988) suggest that family-controlled small and medium-sized enterprises (SMEs) were as efficient, or more efficient, than large firms in many sectors of the economy in Taiwan in the 1960s and 1970s. Espy (1970) (cited in Whyte 1996, p. 13) found that there was no relationship between the rate of growth and whether family members were employed in the firm in a sample of Hong Kong firms. There are also several case studies of family-controlled conglomerates in Asia which argue that they add value as an important source of capital accumulation in late-industrializing countries. These include case studies for the Charoen Pokphand group in Thailand (Suehiro 1993; Ampalavanar-Brown 1998); the Lucky-Goldstar group in South Korea (Taniura 1993); and the Salim Group in Indonesia (Sato 1993). On the other hand, there is plenty of evidence to suggest that family obligations can interfere with the efficient operation of family-controlled Chinese firms (Silin 1976; Whyte 1996). Koike (1993, p. 365) argues that business families tend to cultivate political ties resulting in a patron-client relationship. This sort of relationship underscored accusations of crony capitalism in the fallout from the Asian financial crisis.
Table 4 shows the distribution of ownership concentration in the firms used in this study. It shows ownership concentration for the largest shareholder, largest five shareholders, and largest ten shareholders. The picture that emerges is that ownership was highly concentrated among firms in the sample. In 69.4 per cent of company/years the largest shareholder controlled 20 per cent or more of shares. The comparable figures for the largest five shareholders and largest ten shareholders were 98.4 per cent and 99.7 per cent of company/years, respectively. In 46 per cent of company/years the largest shareholder controlled 30 per cent or more of the shares. This compares with 94.6 per cent of companies for the largest five shareholders and 98.5 per cent of company/ years for the largest ten shareholders. If we use a 40 per cent cut-off the situation is similar. In 27.9 per cent of company/years the largest shareholder controlled 40 per cent or more of shares and this figure increased to 85.6 per cent of company/years for the largest five shareholders and 94.9 per cent of company/years for the largest ten shareholders. The average concentration ratio in the sample firms for the largest five shareholders is 61.4 per cent, which is similar to the comparable figure for the five largest shareholders for all Malaysian firms given in Table 2. (1)
Thus, our second hypothesis is that board compensation is decreasing in ownership concentration. This means that ownership concentration, which is higher in Asia, should lead to a lower pay-performance relationship. There are several empirical studies consistent with the view that large shareholders are more effective monitors. McConnell and Servaes (1990) find that shareholders who own large stakes in a firm are more effective monitors. Brickley, Lease, and Smith (1988) find that large block-holders vote more actively on anti-takeover amendments than non-block-holders. There are also empirical studies linking ownership concentration and shareholding structure to levels of compensation. Mehran (1995) finds that publicly traded firms with a high percentage of shares held by large non-management shareholders use less equity-based compensation than firms with a low percentage of shares held by large non-management shareholders. Ke, Petroni, and Safieddine (1999) compare the sensitivity of the pay-performance relationship in publicly held property liability insurers where there is diffusely held ownership with privately held property liability insurers where there is closely held ownership. Their hypothesis was that for closely held private insurers there should be more direct monitoring of management by owners and less reliance on contracts that link compensation to firm performance. Consistent with this hypothesis, their findings were that there is a significant positive association between return on assets and the level of executive compensation in publicly held insurers, while there is no such relationship in privately held insurers.
Our third hypothesis is that there is an inverse relationship between board remuneration and sector performance. Relative performance evaluation in incentive contracts is a major empirical prediction of agency theory (Holmstrom 1982; Murphy 1999, pp. 2535-39). The performance of other managers in the same sector provides additional information that can be used to decide appropriate compensation. Principal--agent models suggest that a more accurate picture of manager's effort decreases the variance of observable outcomes. This reduces the risk differential that needs to be built into the contract (see, for example, Holmstrom 1979; Zwiebel 1995). As Murphy (1999, p. 2535) describes it, "if the stochastic component of company performance contains an industry or market effect as well as an idiosyncratic effect then `taking out the noise' through relative performance evaluation is incrementally informative in assessing the actions taken by the CEO". Gibbons and Murphy (1990) and Conyon and Leech (1994) provide the strongest empirical support for the relative performance evaluation hypothesis for the Anglo-American economies. (2)
Murphy (1999, p. 2531) documents an inverse relationship between company size and pay-performance sensitivities. This reflects that agency problems are a cost of company size and must be weighed against the benefits derived from economies of scale and scope. Thus, when considering the effect of the degree of corporate control on executive compensation it is important to control for size-related heterogeneity. Rosen (1992) argues that the costs of wrong managerial decisions and the benefits of right managerial decisions are larger in bigger companies. Therefore, executives in bigger companies should be paid more. Another way of looking at it is that firm size can be considered a measure by which companies judge the performance of their managers. Since firm size is typically measured by turnover, managers who have increased company sales are viewed as more successful. There are a number of empirical studies which support the contention that there is a positive relationship between firm size and executive compensation (see, for example, Murphy 1985; Baker, Jensen, and Murphy 1988; Gregg, Machin, and Szymanski 1993; Main, Bruce, and Buck 1996). Given these studies, our fourth hypothesis is that there is a positive relationship between board remuneration and firm size.
III. Data, Methodology, and Specification
We used pooled cross-sectional and time series data for companies listed on the KLSE over the period 1989-2000. The number of companies and observations vary according to which specification is used, or which independent variables are employed, in the regressions with the maximum number of companies and observations being 223 and 1,092 respectively. Our sample only includes those companies for which we have data on remuneration, performance variables, and sales turnover for at least three consecutive years over the sample period. There were 307 companies listed on the KLSE in 1989, and this number increased to 795 by 2000.
Table 5 compares the size of firms in the sample to the KLSE as a whole. The average market value for firms in our sample is similar to the KLSE as a whole. Tables 6 and 7 provide information on foreign ownership in the sample. The percentage of foreign ownership in our sample is quite low with medians in the range of 0.54-1.24 and means in the range 2.63-8.35. The frequency distribution of foreign ownership indicates that foreigners own 5 per cent or less of shares in most of the cases (65 per cent to 84 per cent of the observed 225 company-years).
Financial data are available for most firms on the KLSE, but it was not possible to collect ownership data on all companies for pooled consecutive time series over the sample period. We also excluded firms that operate in the finance, plantation, and mining sectors from the sample. These sectors use different accounting procedures than the ones in other sectors. In the finance sector, the presentation of accounting data in balance sheets and income statements differs from that in the sectors included in the study. "The extractive industries also have unique accounting issues in such areas as revenue recognition, inventories, and arrangements that allow two or more entities to share the risks of exploring and developing mineral reserves." (3)
Limiting the sample to mainly manufacturing and service industries makes the sample more uniform and helps to reduce potential biases that would be due to a mix of relatively incompatible sectors. This practice is rather common in the literature. Claessens et al. (1999) exclude firms whose primary business segment is financial services in a study on Hong Kong, Indonesia, South Korea, Japan, Malaysia, Philippines, Singapore, Taiwan, and Thailand. Wiwattanakantang (1999) uses cross-sectional data for non-financial companies listed on the Securities Exchange of Thailand.
In addition, firms in the infrastructure (6), technology (11), hotels (6), and closed-end funds (1) sectors as well as trusts (4) have also been excluded since there are only a handful of companies listed on the KLSE board for these sectors. This was necessary in order to allow us to compute a meaningful or representative sector performance measure, which was used as an indicator of relative performance in the regressions. Thus our sample includes those companies operating in the consumer products, industrial products, construction, trading/services, and property sectors. Table 8 shows the representation of different industries in the sample. It shows that food/beverages (15), trading/retail (16), plastics/adhesives (13), and property development (34) are the most heavily represented industries in our sample.
Financial data such as sales turnover, last transacted price, (4) and market value are measured at financial year-ends. Ownership data are measured as of any month during the financial year (there is no fixed date at which to report these data). Financial year-ends are not the same for all firms. Since we define turnover and shareholder wealth consistently by financial year, this should not be a problem for these variables. However, for sector performance we use calendar years, that is, when we calculate the sector performance measure for a given firm we include not only the firms with the same financial year as the given firm, but all firms on which we have data in our calculations. (5) All data in ringgits are expressed in 1994 ringgits. The Consumer Price Index (CPI) has been used to convert the nominal values to real values using 1994 as the base year. In the cases where a firm is owned by a holding company, we have included only the holding company to prevent double counting as holding company accounts consolidate the accounts of subsidiaries. In cases where we have data on several subsidiaries and also on the holding company itself, we have excluded the holding company from the sample to increase the number of observations.
All data come from various KLSE publications. The main sources are the KLSE Annual Companies Handbook and KLSE on Disc (which is the electronic version of the handbook). For directors' remuneration we used the data provided on the website of the KLSE Research Institute of Investment Analysts Information Systems. (6) Data on this website as well as that in the KLSE handbook and KLSE on Disc are derived from the annual reports of companies listed on the KLSE.
We used total board remuneration as our measure of pay (denoted as REM). (7) This includes salaries and fees paid to all directors. In developed countries fees paid to non-executive directors are fairly constant across different groups of firms, although larger firms can afford to pay more to attract non-executive directors with better skill-sets and experience. A Price Waterhouse Coopers (2001) survey of 114 CEOs, executive directors, and non-executive directors in companies listed on the Main Board of the KLSE suggests that the situation is similar in Malaysia. While there is considerable variation, there are generally few executive directors on the board of directors in developed countries. Table 9 shows average board size and the percentage of non-executive directors for several representative countries taken from Tricker (1999) and Tenev, Zhang, and Brefort (2002). The average board size ranges from five in Switzerland and six in Brazil to thirteen in Canada, France, South Africa, and the United States and fifteen in Belgium. The percentage of non-executive directors ranges from 15 per cent in Hong Kong to 85 per cent in Sweden and 89 per cent in Switzerland.
In 1998, Price Waterhouse Coopers carried out a survey of corporate governance practices in Malaysian listed firms (see Price Waterhouse Coopers 1998). The survey was sent to 450 public companies listed on the Main Board and 272 companies listed on the Second Board of the KLSE. In total, 304 companies responded, representing 42 per cent of public companies listed on the KLSE at that time. The average board size among firms who responded was eight, of which the average number of non-executive directors was 5.2 constituting 65 per cent of the board (Price Waterhouse Coopers 1998, p. 19). The results from the Price Waterhouse Coopers survey suggest that board size in Malaysia is the same as in Australia and Hong Kong, and slightly less than in Sweden and China. In terms of the proportion of non-executive directors on the Board, the figure for Malaysia is in the mid-range of the countries in Table 9. It is higher than countries such as the United Kingdom and South Africa, but less than countries such as Australia, Canada, Sweden, Switzerland, and the United States.
We regressed our variable for board remuneration on measures of firm performance, sector performance, and firm size. We used two alternative measures of firm and sector performance. In the first set of regressions we defined firm and sector performance in terms of stockholder wealth (SHW) and sector stockholder wealth (SSHW), respectively. Stockholder wealth is defined as
[[([P.sub.t] + [d.sub.t])/[P.sub.t-1])].sup.*]M[V.sub.t-1]
where [P.sub.t] is the current period stock market price, [d.sub.t] is the dividend per share, [P.sub.t-1] is the previous period stock market price, and M[V.sub.t-1] is the market value of the firm. This definition is consistent with that used in Conyon and Leech (1994) for their study using data on U.K. listed companies and is similar to that used by Murphy (1985) and Jensen and Murphy (1990) for U.S. listed companies. The sector performance measure is stockholder wealth averaged over firms in each sector.
The advantage of using shareholder wealth rather than accounting profits to measure firm performance is that in principal-agent theories, shareholders are generally considered to be the principals (Murphy 1985, p. 21). This, of course, assumes that share price movements always dominate firm performance. Thus, for comparative purposes, in the second set of regressions we used return on assets and sector return on assets as alternative measures of firm and sector performance respectively. These variables were entered into the regressions as (1 + ROA) and (1 + SROA), which is the standard practice in the literature (see, for example, Crespi and Gispert 1998). Following the definition used by the KLSE, return on assets was calculated by dividing either the firm's profit or loss before extraordinary items by total assets. The sector performance measure in the second set of regressions was return on assets averaged over firms in each sector. We used the sector classification guidelines in KLSE publications to determine which firms are in each sector. In both sets of regressions, firm size was measured using sales turnover (TRN). LNREM, LNSHW, LNSSHW, LN(1 + ROA), LN(1 + SROA), and LNTRN are the natural logs of REM, SHW, SSHW, (1 + ROA), (1 + SROA), and TRN. Table 10 provides some basic descriptive statistics for REM, SHW, SSHW, (1 + ROA), (1 + SROA), and TRN as well as their natural logs.
Ownership concentration is employed to measure the effect of corporate governance on board remuneration. Shareholder concentration is an important dimension of corporate governance and has been treated as such in the literature (see, for example, Conyon and Leech 1994, p. 234). Following Crespi and Gispert (1998) we use the percentage of shares owned by the largest stockholder (CR1) to measure concentration. (8) Consistent with Crespi and Gispert (1998), a dummy variable (denoted by DCR1) is used, which takes the value of one if the concentration measure is above the median level and is otherwise zero. This dummy variable is entered as time invariant.
We needed to enter DCR1 as time invariant because it was not possible to find ownership data for all years. Thus, we used the latest ownership data available to construct this dummy variable. Conyon and Leech (1994) and Crespi and Gispert (1998) follow the same approach. Time dummies were included in all initial regressions (which are reported in Table 11) to control for the effects of economy-wide shocks. However, in later specifications (which are reported in Table 12) we examine the impact of the Asian financial crisis on remuneration. In these specifications, instead of using time dummies, we substituted a crisis dummy variable (denoted as DCRISIS), set equal to one for the period 1997-2000 and zero otherwise.
To test the hypotheses in Section II, we used the first-differences model, which is standard in the literature (see, for example, Conyon and Leech 1994; Crespi and Gispert 1998). Specifically, in the first set of regressions we regressed DLNRE[M.sub.i,t] on different combinations of DLNSHW, DLNSSHW, and DLNTRN in period t and period t - 1 and DCR1. In the second set of regressions we used DLN(1 + ROA) and DLN(1 + SROA) instead of DLNSHW and DLNSSHW as proxies for firm and sector performance, where for a variable X, DLN[X.sub.t] is LN[X.sub.t] - LN[X.sub.t-1] and DLNX(t - 1) is LN[X.sub.t-1] - LN[X.sub.t-2].
IV. Results
The results of various specifications using either stakeholder wealth or return on assets as a proxy for performance, with time dummies to control for shocks are reported in Table 11. With the exception of the fourth regression reported in Table 11, both the Wald (time) and Wald (joint) statistics are significant in each specification, which rejects the null hypothesis that the tested coefficients are simultaneously zero. All estimates were obtained using the Dynamic Panel Data (DPD) package of Doornik, Arellano, and Bond (1999). These are one-step estimates obtained by using the robust variance covariance matrix. In Table 11, tests for the absence of first-order and second-order serial correlation in the first-differenced residuals are reported as a first-order autoregression disturbance process (AR(1)) and a second-order autoregression disturbance process (AR(2)) respectively. In the second, third, fifth, and seventh specification, AR(1) is negative and significant and AR(2) is insignificant, meaning that we can accept the null hypothesis that the disturbance terms are not serially correlated in these specifications. (9) For this reason, we concentrate on the results in these specifications when interpreting our findings in the following discussion. (10)
The first hypothesis, based on studies in the United Kingdom and United States, is that there is a positive relationship between firm performance and board remuneration. However, we noted above that the relationship between board remuneration and firm performance might be weaker in Malaysia than in the United Kingdom or United States because ownership is more concentrated. Our results suggest that the relationship between board remuneration and firm performance is ambiguous in Malaysia. The first column of Table 11 reports the effect of the change in the log of shareholder wealth in period t - 1 on current change in the log of board remuneration. This is the same econometric methodology as that employed by Conyon and Leech (1994). It is also similar to the econometric approach used by Gibbons and Murphy (1990) who estimate the relationship between the change in CEO remuneration and the continuously accrued rate of return received by firms' shareholders. In Table 11, DLNSH[W.sub.i,t-1] is positive and significant at 10 per cent in the first specification, but here we cannot accept the null hypothesis that the disturbance terms are not serially correlated. Turning to measures of accounting performance, in the third specification in Table 11, which does satisfy the null hypothesis that the disturbance terms are not serially correlated DLN[(1 + ROA).sub.i,t] is significant at the 10 per cent level and it has an unexpected negative sign.
The descriptive statistics in Table 10 suggest that the negative sign on DLN[(1 + ROA).sub.i,t] may be due to outliers. The minimum ROA is -99 per cent. Such extreme values could easily result in a negative regression coefficient even when the true relationship is positive. Thus, the results were checked for robustness to outliers through rerunning all of the specifications in Table 11 excluding outliers. Following Xu and Wang (1999), an observation was treated as an outlier if it fell outside the range of the sample mean plus or minus two times the standard deviation. This entailed excluding between nine companies (specifications 4 and 6) and eighteen companies (specifications 3 and 5). We do not report the results excluding outliers because the disturbance terms were serially correlated in each instance and the results were very similar in terms of signs and significance of the coefficients. One change, however, is that while DLN[(1 + ROA).sub.i,t] continues to have a negative sign, it became insignificant in specifications 3 and 6.
Overall the results suggest that the nature of the relationship between board remuneration and company performance in Malaysia depends on the measure of performance. The relationship is positive for stock market performance measures, but negative for accounting measures. Jensen and Murphy (1990, p. 246) suggest a possible explanation for the negative sign on the coefficient of DLN[(1 + ROA).sub.i,t]:
Accounting profits, for example, may yield information that is valuable in assessing an executive's unobservable actions. But paying executives on the basis of accounting profits rather than changes in shareholder wealth not only generates incentives to directly manipulate the accounting system but also generates incentives to ignore projects with large net present values in favor of less valuable projects with larger immediate accounting profits.
If, despite the possible adverse effects on immediate profits, directors in the Malaysian companies under examination have undertaken projects to increase long-term returns rather than to increase short-term returns, then a negative relationship between pay and accounting performance measures might emerge. This is because shareholders, or the block-holders who are effectively in control of the company, may be rewarding directors on the basis of their value increasing actions rather than their performance with respect to returns. For this to happen there must be relatively strong negative correlation between immediate returns and undertaking projects with large net present values to increase distant returns. Such a relationship need not hold in general, but might have occurred in the period under consideration.
While the estimated coefficient on DLNSH[W.sub.i,t-1] is positive and significant, which is consistent with the perspective of agency theory that directors are rewarded for increases in past shareholder wealth, the coefficient is very small. The results for DLNSH[W.sub.i,t-1] suggest that the elasticity of board remuneration with respect to past shareholder wealth is around 0.06, which indicates that the pay-performance link is very weak. These pay-performance sensitivities, though, can still be regarded as consistent with agency theory. Haubrich (1994), using a static model, and Wang (1997), using a dynamic model, show that low pay-performance sensitivities are quite consistent with the prediction of agency theory for sufficiently risk-averse executives.
The elasticity here is much lower than in previous U.S. studies. Jensen and Murphy (1990) find that the elasticity of CEO compensation with respect to shareholder wealth is 0.14. Earlier studies such as Murphy (1985, 1986) find a pay-performance elasticity of around 0.10. More generally, Rosen (1992) surveys a number of the extant U.S. studies, which suggest that the elasticity of CEO remuneration with respect to shareholder return is in the range 0.10 to 0.17. Nevertheless, when comparing our results with the results of studies for U.S. firms, it is important to stress the caveat discussed earlier that there are large size differences between the firms in our sample and Dow or S&P 500 firms. When the results in this study are compared with those from U.K. studies the picture is not as clear. The results in this study are similar to findings of some studies for the United Kingdom, which have also found that the relationship between CEO remuneration and shareholder measures is fairly weak. Included in this category are studies such as Gregg, Machin, and Szymanski (1993) and Conyon and Leech (1994), which suggest an elasticity of CEO pay with respect to current shareholder wealth and past shareholder wealth of around 0.06 for the United Kingdom.
In fact, some studies for U.K. companies have suggested that the relationship between remuneration and performance has weakened over time. Gregg, Machin, Szymanski (1993) and Conyon and Gregg (1994) find that at the beginning of the 1990s the link between remuneration and performance in U.K. companies became statistically insignificant, meaning there was no relationship between remuneration of U.K. executives and stock market performance (see Conyon, Gregg, and Machin 1995 for a review). However, a more recent study by Main, Bruce and Buck (1996) using a broader definition of pay suggests that pay-performance sensitivity in the United Kingdom is more robust and is similar to the United States.
The second hypothesis is that there is an inverse relationship between board remuneration and ownership concentration. In the seventh specification in Table 11, DCR1 is negative and significant at 10 per cent. We have not reported the results with DCR1 from the first and second specification in Table 11, where the coefficient on DCR1 was also negative and significant because they are impaired by serial correlation. The results for the seventh specification suggest that ownership concentration has a negative and at least a weakly significant effect on remuneration.
This result differs from that of Conyon and Leech (1994) who found that ownership concentration has no effect on executive compensation in the United Kingdom. Our results, though, are not surprising, given the higher levels of ownership concentration in Malaysia than in the United Kingdom. Because ownership is more concentrated, one would expect that shareholders would have better information and be better placed to monitor manager effort in Malaysia than in the United Kingdom. Our findings are consistent with those of Crespi and Gispert (1998) for Spanish listed companies. In Spain, where ownership concentration is also much higher than in the United Kingdom, Crespi and Gispert (1998) found that ownership concentration had a negative and significant effect on remuneration. To the extent that DCR1 is a proxy for concentrated family ownership, our results for ownership concentration are also consistent with the findings of Ramaswamy, Veliyath, and Gomes (2000) for the effects of family ownership in India discussed earlier. The negative and significant sign on DCR1 is consistent with the view that high levels of family control of companies in Asia should have a dampening effect on compensation.
The third hypothesis is that there is an inverse relationship between board remuneration and sector performance. Our results reject this hypothesis. In Table 11, DLNSSH[W.sub.i,t] has a negative sign but is insignificant in the second specification, while DLN[(1 + SROA).sub.i,t] has a negative sign, but is insignificant in the third, fifth, and seventh specification. When we exclude outliers DLNSSH[W.sub.i,t] becomes significant in the second specification, but, as noted above, when outliers were excluded all results were impaired by autocorrelation. Thus, our results suggest that those who set board remuneration in Malaysia do not consider contemporaneous sector performance.
This result is consistent with having high ownership concentration to the extent that as shareholders are better placed to monitor managers there is less need to rely on industry or sector performance in determining compensation. Having said this, previous studies for the United States and United Kingdom only provide weak support for this hypothesis. After reviewing the literature, Brunello, Graziano, and Parigi (1997) suggest:
"The empirical evidence ... is mixed and is weakly in favor of relative performance evaluation." Gibbons and Murphy (1990) and Conyon and Leech (1994) find that pay setters take account of industry performance. However, Antle and Smith (1986) find, at best, mixed support for the relative performance evaluation thesis and Jensen and Murphy (1990) found that industry performance variables were insignificant. In one of the few existing studies to use sector performance rather than industry performance to measure relative performance, Main, Bruce, and Buck (1996) also found that contemporaneous sector performance had a statistically insignificant effect on board remuneration in U.K. listed companies.
The fourth hypothesis is that there is a positive relationship between remuneration and firm size. Our results provide strong evidence for this hypothesis. In the second, third, fifth and seventh specification in Table 11, DLNTR[N.sub.i,t] has a positive sign and is statistically significant at the 1 per cent level. The magnitude of the coefficient on DLNTR[N.sub.i,t] varies between 0.24 and 0.28, which is consistent with the findings of U.K. and U.S. studies that have repeatedly found the elasticity of pay with respect to sales to be around 0.3 (see Rosen 1992). Baker, Jensen, and Murphy (1998, p. 609) suggest that this is "the best documented empirical regularity regarding levels of executive compensation". The magnitude on DLNTR[N.sub.i,t-1] in the first and fourth specification is in the range of 0.05 to 0.07. This result is consistent with that of Conyon and Leech (1994) who estimate that the elasticity of the highest paid director with respect to lagged turnover is 0.07 for the United Kingdom. However, it is lower than Main, Bruce, and Buck (1996) who estimate the elasticity of total board remuneration with respect to lagged turnover to be 0.28 for the United Kingdom.
Table 12 examines the impact of the 1997 Asian financial crisis on board remuneration. It substitutes a crisis dummy variable (DCRISIS), which is set equal to 1 for the period 1997-2000, for the time dummies in Table 11. We hypothesize that after the financial crisis, change in board compensation should be lower than its pre-crisis value, which in turn means that the expected sign for the crisis dummy should be negative. This will occur if, in response to the financial crisis, firms restructure their management compensation contracts and change their remuneration policy, for instance by reducing pay, when they experienced severe financial difficulties. Gilson and Vetsuypens (1993) provide evidence to suggest that this is how distressed firms respond to financial difficulties. In a study of seventy-seven publicly traded firms in the United States that filed for bankruptcy or privately restructured their debt in the 1980s, Gilson and Vetsuypens (1993) found that one-third of CEOs were replaced, and those who kept their jobs experienced large reductions in bonuses and salaries.
Table 12 also considers the effect of interactive variables--DCRISI[S.sup.*]DLNSH[W.sub.i,t-1], DCRISI[S.sup.*] DLN[(1 + ROA).sub.i,t], and DCRISI[S.sup.*]DLNTR[N.sub.it]--on board remuneration. Interactive dummies test whether pay-performance sensitivities have changed due to the crisis. We hypothesize that following the crisis, firms will tie board compensation more strictly to performance and therefore pay-performance sensitivity should be higher. Gilson and Vetsuypens (1993) find that sixty-four of the firms in their sample of financially distressed firms changed their compensation policy to tie senior management wealth more closely to the value of the company's common stock.
The Wald (joint) test is significant at 1 per cent in each specification in Table 12, while AR(1) is negative and significant and AR(2) is insignificant in specifications three through to six, which means that we can accept the null hypothesis that the disturbance terms are not serially correlated in four of the six specifications. Focusing on these four specifications, the results in Table 12 are generally consistent with the findings with respect to the four hypotheses based on Table 11. The coefficient on the crisis dummy variable is negative and significant at 5 per cent in specifications four and six, but insignificant in specification five. In Table 12, none of the crisis interactive variables are significant. This suggests that consistent with our hypothesis, the Asian crisis had a statistically significant negative effect on board remuneration. The Asian financial crisis, though, did not affect the role of either stockholder wealth, return on assets or sales turnover as potential factors determining board remuneration.
As with Table 11, we checked whether the results in Table 12 are robust to the exclusion of outliers through rerunning each specification excluding outliers. We do not report the results excluding outliers because, similar to Table 11, we rejected the null hypothesis that the disturbance terms are not serially correlated for each specification plus the results with outliers excluded are very similar to those reported in Table 12. The ownership concentration variable is not included in Table 12 because when we entered DCR1 all results were impaired by serial correlation. In separate regressions, which are not reported because of the serial correlation problem, we entered DCR1 in each specification in Table 12 including and excluding outliers. The coefficient on DCR1 was negative and significant in specifications 1-3 and 5-6.
V. Conclusions and Future Research
Most existing studies examine executive compensation in the United Kingdom or United States. This study contributes to the literature through examining the determinants of executive compensation in Malaysia, where ownership is usually much more concentrated than in either of these countries. In addition, while most studies use the remuneration of the CEO or highest paid director, we use board remuneration which is more justifiable in terms of principal-agent theory because it is the board as a whole rather than the highest paid director that can be best regarded as the shareholders' agent. In Section IV, we tested four hypotheses based on the existing literature. Our results provide support for two of the four hypotheses. We found evidence of a statistically significant positive relationship between board remuneration and sales turnover and a statistically significant negative relationship between board remuneration and ownership concentration. Our results suggest that the relationship between board remuneration and firm performance is ambiguous, while there is no evidence of a significant relationship between board remuneration and sector performance.
Future research could extend this study in several ways. First, the reference period for the concentration dummies could be changed from later years to earlier years. If ownership has become more dispersed after the crisis the pay performance relationship should have strengthened. Second, by using data on share distribution it should be possible to determine, for example, whether or not the pay performance relationship differs between firms where ownership is concentrated in the hands of Malaysian owners and the firms where there is a high proportion of foreign ownership. Third, it might be desirable to use a dynamic model, in the sense this term is used in panel data estimation, employed in Main, Bruce, and Buck (1996) instead of the static one we use in order to account for the delay in the administration of pay. To do this it would be necessary to include a lagged dependent variable on the right-hand side.
TABLE 1
Control of Publicly Traded Companies in Asia, United States, and
United Kingdom Using a 20 per cent Cut-off
Country Widely Family State Widely Held Widely Held
Held Financial Corporation
United States
Largest 20 firms 80.0 20.0 0.0 0.0 0.0
United Kingdom
Largest 20 firms 100.0 0.0 0.0 0.0 0.0
Malaysia
238 firms in 10.3 67.2 13.4 2.3 6.7
Claessens,
Djankov, and
Lang (1999)
study
Largest 20 firms 30.0 35.0 30.0 0.0 5.0
Hong Kong
330 firms in 7.0 66.7 1.4 5.2 19.8
Claessens,
Djankov, and
Lang (1999)
study
Largest 20 firms 5.0 72.5 7.5 10.0 5.0
Indonesia
178 firms in 5.1 71.5 8.2 2.0 13.2
Claessens,
Djankov, and
Lang (1999)
study
Largest 20 firms 15.0 60.0 20.0 0.0 5.0
Japan
1240 firms in 79.8 9.7 0.8 6.5 3.2
Claessens,
Djankov, and
Lang (1999)
study
Largest 20 firms 90.0 5.0 5.0 0.0 0.0
South Korea
345 firms in 43.2 48.4 1.6 0.7 6.1
Claessens,
Djankov, and
Lang (1999)
study
Largest 20 firms 65.0 20.0 10.0 0.0 5.0
Philippines
120 firms in 19.2 44.6 2.1 7.5 26.7
Claessens,
Djankov, and
Lang (1999)
study
Largest 20 firms 40.0 40.0 7.5 7.5 5.0
Singapore
221 firms in 5.4 55.4 23.5 4.1 11.5
Claessens,
Djankov, and
Lang (1999)
study
Largest 20 firms 20.0 32.5 42.5 0.0 5.0
Taiwan
141 firms in 26.2 48.2 2.8 5.3 17.4
Claessens,
Djankov, and
Lang (1999)
study
Largest 20 firms 45.0 15.0 15.0 5.0 20.0
Thailand
167 firms in 6.6 61.6 8.0 8.6 15.3
Claessens,
Djankov, and
Lang (1999)
study
Largest 20 firms 10.0 57.5 20.0 7.5 5.0
NOTES: The largest twenty firms are the largest twenty firms based
on market capitalization. Data for Asian countries are for 1996. The
source is Claessens, Djankov, and Lang (1999). Data for the United
States and United Kingdom are for 1996 or 1997. The source is La
Porta, Lopez-De-Silanes, and Shleifer (1999). The Claessens, Djankov,
and Lang (1999) study is based on a data-set of 2,980 publicly traded
companies collated from Worldscope, Asian Company Handbook and other
sources such as Annual Reports of various stock exchanges (see
Claessens, Djankov, and Lang 1999, pp. 6-11 for details). The source
of most information in the La Porta, Lopez-De-Silanes, and Shleifer
(1999) study was Worldscope. Some information was also from the
Internet and annual reports (see La Porta, Lopez-De-Silanes, and
Shleifer 1999, pp. 474-76 for details).
TABLE 2
Ownership Concentration in Selected Countries in 1998
(percentage of shareholding)
2nd 3rd 4th-10th
Country Largest Largest Largest Largest
Austria 82.2 9.5 1.9 6.5
China 47.0 8.0 3.0 NA
France 56.0 16.0 6.0 5.0
Italy 52.3 7.7 3.5 5.1
Netherlands 28.2 9.2 4.3 7.1
Spain 38.3 11.5 7.7 10.3
U.K. 14.0 8.3 6.1 9.2
2nd 4th 6th-10th
Country Largest and 3rd and 5th Largest
Largest Largest
Belgium 55.8 6.9 0.6 0.2
Germany 59.7 8.6 2.6 0.3
U.S.A 22.8 9.5 7.5 3.8
1st-5th
Country Largest Largest
Indonesia 48.2 67.5
Malaysia 30.3 58.8
Philippines 33.5 60.2
South Korea 20.4 38.5
Thailand 28.5 56.6
SOURCES: ADB (1999), Tricker (1999), Tenev, Zhang,
and Brefort (2002).
TABLE 3
Ownership Concentration in the Ten Largest Malaysian Companies
(Based on Market Value) in 1999
Company Market Value Largest
(RM) Shareholder
1 TELEKOM 44,396,225,286 Khazanah Nasional
MALAYSIA Bhd (38.88)
2 MAYBANK 31,451,114,889 Permodalan Nasional
Bhd (35.27)
3 TENEGA 30,398,740,540 Khazanah Nasional
NATIONAL BHD. Bhd (36.24)
4 PETRONASGAS 16,209,737,872 RHB Nominees
(Tempatan) Sdn Bhd
(58.15)
5 MALAYSIA 12,275,431,034 RHB Nominees
INTERNATIONAL (Tempatan) Sdn Bhd
SHIPPING A/C Petroliam
Nasional Bhd (62.01)
6 RESORTS 11,901,092,341 Genting Bhd (55.49)
WORLD BHD
7 COMMERCEASSET- 11,248,236,256 Citicorp Nominees
HOLD. BHD (Tempatan) Sdn Bhd
(Pledged securities
account for Pitisan
Sdn Bhd) (19.39)
8 SIME DARBY 11,210,732,317 Amanah Raya Bhd
Sekim Amanah
Saham Bumiputera
(27.78)
9 GENTING 9,508,575,879 Kien Huat Realty
Sdn Bhd (33.88)
10 YTL 8,821,835,130 Yeoh Tiong Lay &
Sons Holdings Sdn
Bhd (45.56)
Company Second Largest Third Largest Fourth
Shareholder Shareholder Largest
Shareholder
1 TELEKOM Minister of
MALAYSIA Finance
(Incorporated)
(21.69)
2 MAYBANK Amanah Raya
Nominees
(15.81)
3 TENEGA Minister Of Bank Negara
NATIONAL BHD. Finance Malaysia
(Incorporated) (11.36)
(17.12)
4 PETRONASGAS Kumpulan Wang Employees
Amanah Pencen Provident Fund
(16.14) Board (9.28)
5 MALAYSIA
INTERNATIONAL
SHIPPING
6 RESORTS
WORLD BHD
7 COMMERCEASSET- Syarikat Nominee Employees
HOLD. BHD Bumiputra Provident Fund
(Tempatan) Sdn Board (10.13)
Bhd (Pledged
securities
account for
Fleet
Group Sdn Bhd)
(12.81)
8 SIME DARBY Permodalan Yayasan Employees
Nasional Pelaburan Provident
Bhd (10.30) Bumiputera Fund Board
(held through (10.38)
Permodalan
Nasional Bhd)
(10.30)
9 GENTING
10 YTL Lembaga Tabung
(14.17)
NOTE: Figures in parentheses are the percentage of shares
owned by each shareholder.
SOURCE: KLSE on Disc.
TABLE 4
Distribution of Ownership Concentration in the Whole Sample
(564 companies)
Ownership Largest Largest Five
Concentration Shareholder
(%)
Number * % Number * %
Less than 5 15 0.911 2 0.121
5-10 79 4.797 3 0.182
10-20 410 24.89 21 1.275
20-30 386 23.44 64 3.886
30-40 298 18.09 148 8.986
40-50 203 12.33 217 13.18
50-60 192 11.66 354 21.49
60-70 39 2.368 390 23.68
70-80 17 1.032 279 16.94
80-90 0 0 140 8.5
More than 90 8 0.486 29 1.761
Ownership Largest Ten
Concentration
(%)
Number * %
Less than 5 0 0
5-10 1 0.061
10-20 4 0.243
20-30 22 1.336
30-40 57 3.461
40-50 109 6.618
50-60 200 12.14
60-70 361 21.92
70-80 437 26.53
80-90 369 22.4
More than 90 87 5.282
NOTES:
Results are calculated for the whole 564 companies in the
sample.
* refers to number of observations. (The total number of
observations is 1,647.)
SOURCE: All data come from various KLSE publications.
The main sources are the KLSE Annual Companies Handbook
and KLSE on Disc.
TABLE 5
Market Value and Average Market Value of Listed Malaysian
Companies
Firms in the Sample
No of Firms Market Value Average Market Value
(RM million) (RM million)
1989 45 14,391.39 319.8087
1990 84 25,099.19 298.7998
1991 108 55,394.4 512.9111
1992 145 147,186.1 1,015.077
1993 175 3,201,988 18,297.07
1994 200 212,447.7 1,062.238
1995 222 212,800.2 958.5596
1996 222 278,360.6 1,253.877
1997 213 111,952.8 525.5998
1998 212 102,645.4 484.1766
1999 95 76,786.29 808.2767
89-99 1,721 4,439,052 2,579.344
All Firms on the KLSE
No of Firms Market Value Average Market Value
(RM million) (RM million)
1995 529 565,628 1,069.240076
1996 621 806,774 1,299.152979
1997 708 376,006 531.0819209
1998 736 374,517 508.8546196
1999 757 552,691 730.1070013
TABLE 6
Number of Companies in the Sample with Foreign Ownership
(Concentration Greater than Median in the Sample Used in
Regressions) (225 companies)
Number (%)
Individuals-foreign (0.6, 2.75) * 111 (49.11%)
Corporations-foreign (0.54, 2.63) * 119 (52.65%)
Nominees-foreign (1.235, 8.35) * 127 (56.19%)
NOTES:
* median and mean respectively (these have been calculated
by using the whole sample of 564 companies). Ownership
concentration has been calculated from recent data
(late 1990s).
TABLE 7
Foreign Ownership Concentration in the Sample
(225 companies observed over 1989-99)
Bin range Individuals Corporations Nominees
(% of shares owned
by foreigners)
Less than 5 556 592 434
5-10 65 65 57
10-15 20 20 46
15-20 13 13 17
20-25 5 5 14
25-30 0 0 11
30-35 5 5 15
35-40 0 0 12
40-45 1 1 10
45-50 0 0 9
50-55 0 0 16
55-60 0 0 11
60-65 3 3 6
65-70 0 0 3
70-75 0 0 4
75-100 0 0 0
Total 668 704 665
NOTES: Bin ranges do not include the first number in each range.
TABLE 8
The Frequencies of Different Industries Included in the Sample
Sectors and industries
Consumer Products
Food/Beverages 15
Animal Feed 2
Breweries 2
Tobacco 1
Textiles/Garments/Accessories 5
Footwear 1
Electrical Appliances 4
CP - Transport Equipment 6
Furniture & Fittings 1
Industrial products
Paper-based 1
Auto Parts (Used by Retailers) 1
Wood and Wood Products 11
Chemicals 5
Gas 2
Plastics/Adhesives 13
Oil 2
Rubber-based 1
Paper/Paper Products 4
Building Materials-Cement 11
Trading Services
Engineering 8
Media 5
Gaming 7
Trading/Retail 16
Transport 7
Utilities 2
Other Services 6
Construction
Building Materials 10
Basic Metals 4
Fabricated Metals 13
Electrical Machinery/Electronics 5
Transport Equipment 2
Auto Parts 1
Construction 17
Property Development
Property Development 34
TABLE 9
Average Board Size and the Proportion of Non-Executive
Directors, Selected Countries
Country Average Board Size Non-executive Directors (%)
Australia 8 75
Belgium 15 78
Brazil 6 60
Canada 13 80
China 10 50
France 13 82
Hong Kong 8 15
Italy 11 73
South Africa 13 60
Spain 12 71
Sweden 9 85
Switzerland 5 89
U.K. 12 50
U.S.A 13 77
SOURCES: Tricker (1999), Tenev, Zhang, and Brefort (2002).
TABLE 10
Descriptive Statistics for the Sample
(in RM thousands)
Mean Standard Deviation Minimum Maximum
REM 1606.621 2513.068 4 39351
LNREM 6.81924 1.121711 1.386294 10.58028
SHW 8578753 251369153.8 3023.758 8916627500
LNSHW 12.6332 1.456834 8.014256 22.91118
TRN 503854.9 1105199 176.3425 12043676
LNTRN 12.06379 1.443419 5.172428 16.30405
SSHW 8398000 48530039 113735.3 326188448.4
LNSSHW 13.51845 1.248863 11.64163 19.60299
(1 + ROA) 1.033362 0.102131 0.012501 1.999518
LN(1 + ROA) 0.02235 0.196443 -4.38194 0.692906
S(1 + ROA) 1.016664 0.102679 0.4997 1.092531
LNS(1 + ROA) 0.009219 0.132681 -0.69375 0.088497
NOTES:
LNREM, LNSHW, LNSSHW, LN(1 + ROA), LN(1 + SROA), and LNTRN are the
natural logs of board remuneration (REM), stockholder wealth (SHW),
sector stockholder wealth (SSHW), (1 + return on Assets) (1 + ROA),
(1 + sector return on assets) (1 + SOR), and sales (TRN). REM is
defined as salaries and fees paid to all directors. SHW is defined as
[([P.sub.t] + [d.sub.t])/[P.sub.t-1])]*M[V.sub.t-1], where [P.sub.t]
is the current period stock market price, [d.sub.t] is the dividend
per share, [P.sub.t-1] is the previous period stock market price, and
M[V.sub.t-1] is the market value of the firm. SSHW is stockholder
wealth averaged over firms in each sector. ROA was calculated by
dividing either the firm's profit or loss before extraordinary items
by total assets. SROA is return on assets averaged over firms in each
sector. TRN is defined as sales turnover.
TABLE 11
Ownership Concentration, Directors' Remuneration, and Company
Performance (Standard errors in parentheses)
Variable 1 2 3
Constant -0.0862139 0.0169897 0.0301743
(0.05757) (0.05676) (0.05397)
DLNSH[W.sub.i,t-1] 0.0620403 -- --
(0.036393) *
DLN[(1+ROA).sub.i,t] -- -- -0.186605
(0.09985) *
DLN[(1+ROA).sub.i,t-1] -- -- --
DLNTR[N.sub.i,t] -- 0.280049 0.257596
(0.07829) *** (0.07333) ***
DLNTR[N.sub.i,t-1] 0.0589581 -- --
(0.03272) *
DLNSSH[W.sub.i,t] -- -0.0324405 --
(0.01147)
DLNSSH[W.sub.i,t-1] 0.0128963 -- --
(0.01385)
DLN[(1+SROA).sub.i,t] -- -- -0.112983
(0.1448)
DLN[(1+SROA).sub.i,t-1] -- -- --
DCR1 -- -- --
Time Dummies YES YES YES
Wald (time) 19.61 (7) *** 18.91 (7) *** 20.79 (7) ***
Wald (joint) 7.930 (3) ** 19.21 (2) *** 15.49 (3) ***
AR(1) test -1.014 -2.055 ** -2.156 **
AR(2) test -2.679 *** -1.893 -1.724
Observations 909 966 1082
No. companies 223 223 222
Variable 4 5
Constant -0.0421716 0.0316285
(0.06570) (0.05449)
DLNSH[W.sub.i,t-1] -- --
DLN[(1+ROA).sub.i,t] -- --
DLN[(1+ROA).sub.i,t-1] 0.155958 --
(0.172)
DLNTR[N.sub.i,t] -- 0.245847
(0.07183) ***
DLNTR[N.sub.i,t-1] 0.0669184 --
(0.03797) *
DLNSSH[W.sub.i,t] -- --
DLNSSH[W.sub.i,t-1] -- --
DLN[(1+SROA).sub.i,t] -- -0.0900478
(0.1390)
DLN[(1+SROA).sub.i,t-1] 0.239857 --
(0.1637)
DCR1 -- --
Time Dummies YES YES
Wald (time) 28.00 (7) *** 19.91 (7) ***
Wald (joint) 5.761 (3) 11.84 (2) ***
AR(1) test -2.062 ** -2.172 **
AR(2) test -2.672 *** -1.783
Observations 1152 1092
No. companies 223 222
Variable 6 7
Constant 0.0498019 0.0520582
(0.05755) (0.05819)
DLNSH[W.sub.i,t-1] -- --
DLN[(1+ROA).sub.i,t] -0.183685 --
(0.1002) *
DLN[(1+ROA).sub.i,t-1] -- --
DLNTR[N.sub.i,t] 0.255703 0.244323
(0.07316) *** (0.07166) ***
DLNTR[N.sub.i,t-1] -- --
DLNSSH[W.sub.i,t] -- --
DLNSSH[W.sub.i,t-1] -- --
DLN[(1+SROA).sub.i,t] -0.115367 -0.0919229
(0.1456) (0.1398)
DLN[(1+SROA).sub.i,t-1] -- --
DCR1 0.0448294 -0.0468180
(0.02638) * (0.02623) *
Time Dummies YES YES
Wald (time) 20.65 (7) *** 19.80 (7) ***
Wald (joint) 16.45 (4) *** 12.94 (3) ***
AR(1) test -2.188 ** -2.213 **
AR(2) test -1.768 ** -1.822
Observations 1082 1092
No. companies 222 222
NOTES: The dependent variable is DLNRE[M.sub.i,t]. LNREM, LNSHW,
LNSSHW, LN(1 + ROA), LN(1 + SROA) and LNTRN are the natural logs of
board remuneration (REM), stockholder wealth (SHW), sector stockholder
wealth (SSHW), (1 + return on Assets) (1 + ROA), (1 + sector return on
assets) (1 + SOR) and sales (TRN). For a variable X, DLN[X.sub.t] is
LN[X.sub.t] - LN[X.sub.t-1] and DLNX(t - 1) is LN[X.sub.t-1]
- LN[X.sub.t-2]. DCR1 is a dummy variable measuring the concentration
ratio. It is set equal to 1 if the percentage of shares owned by the
largest shareholder is above the median level. Wald (time): Wald test
on time dummies only. Wald (joint): Wald test on all regressors except
the dummies. For both Wald tests we report the Chi square statistic and
degrees of freedom (in the parentheses next to each statistic). AR(1)
and AR(2) are tests for first and second-order serial correlation,
which are distributed as N(0,1) under the null of no autocorrelation.
*** Significant at 1 per cent level, ** Significant at 5 per cent
level, * Significant at 10 per cent level.
TABLE 12
Ownership Concentration, Directors' Remuneration, and Company
Performance Effects of the 1997 Crisis
(Standard errors in parentheses)
Variable 1 2
Constant 0.0716952 0.102772
(0.01695) *** (0.02542) ***
DLNSH[W.sub.i,t-1] 0.0689266 0.0794072
(0.02763) ** (0.04591) **
DLNTR[N.sub.i,t] -- --
DLNTR[N.sub.i,t-1] 0.0626759 0.0529860
(0.03336) ** (0.03342)
DLNSSH[W.sub.i,t-1] 0.0245268 0.0208876
(0.01183) ** (0.01210) **
DLN[(1 + ROA).sub.i,t] -- --
DLN[(1 + SROA).sub.i,t-1] -- --
DCRISIS -- -0.0667758
(0.03883) **
DCRISIS * -- -0.0324854
DLNSH[W.sub.i,t-1] (0.05394)
DCRISIS * - -- --
DLN[(1 + ROA).sub.i,t]
DCRISIS * -- --
DLNTR[N.sub.i,t]
Time Dummies NO NO
Wald (joint) 20.43 (3) *** 25.59 (5) ***
AR(1) test -0.9840 -0.9679
AR(2) test -2.806 *** -2.746 ***
Observations 909 909
No. companies 223 223
Variable 3 4
Constant 0.0576716 0.0932784
(0.01510) *** (0.02240) ***
DLNSH[W.sub.i,t-1] -- --
DLNTR[N.sub.i,t] 0.273727 0.260835
(0.07116) *** (0.07379) ***
DLNTR[N.sub.i,t-1] -- --
DLNSSH[W.sub.i,t-1] -- --
DLN[(1 + ROA).sub.i,t] -0.169166 -0.218103
(0.1014) * (0.1354)
DLN[(1 + SROA).sub.i,t-1] 0.00395148 -0.0658267
(0.1053) (0.1119)
DCRISIS -- -0.0827635
(0.03496) **
DCRISIS * -- --
DLNSH[W.sub.i,t-1]
DCRISIS * - 0.0524085
DLN[(1 + ROA).sub.i,t] (0.1872)
DCRISIS * -- --
DLNTR[N.sub.i,t]
Time Dummies NO NO
Wald (joint) 17.64 (3) *** 32.37 (5) ***
AR(1) test -2.075 ** -2.153 **
AR(2) test -1.722 -1.739
Observations 1082 1082
No. companies 222 222
Variable 5 6
Constant 0.0615030 0.100636
(0.01453) *** (0.02076) ***
DLNSH[W.sub.i,t-1] -- --
DLNTR[N.sub.i,t] 0.263780 0.185853
(0.06842) *** (0.07080) ***
DLNTR[N.sub.i,t-1] -- --
DLNSSH[W.sub.i,t-1] -- --
DLN[(1 + ROA).sub.i,t] -- --
DLN[(1 + SROA).sub.i,t-1] 0.00207437 -0.0599471
(0.1010) (0.1045)
DCRISIS -- -0.0855983
(0.03377) **
DCRISIS * -- --
DLNSH[W.sub.i,t-1]
DCRISIS * - -- --
DLN[(1 + ROA).sub.i,t]
DCRISIS * -- 0.127785
DLNTR[N.sub.i,t] (0.1279)
Time Dummies NO NO
Wald (joint) 16.57 (2) *** 33.59 (4) ***
AR(1) test -2.099 ** -2.076 **
AR(2) test -1.783 -1.896
Observations 1092 1092
No. companies 222 222
NOTES: The dependent variable is DLNRE[M.sub.i,t]. LNREM, LNSHW,
LNSSHW, LN(1 + ROA), LN(1 + SROA), and LNTRN are the natural logs of
board remuneration (REM), stockholder wealth (SHW), sector stockholder
wealth (SSHW), (1 + return on Assets) (1 + ROA), (1 + sector return on
assets) (1 + SOR), and sales (TRN). For a variable X, DLN[X.sub.t] is
LN[X.sub.t] - LN[X.sub.t-1] and DLNX(t - 1) is LN[X.sub.t-1]
- LN[X.sub.t-2]. (DCRISIS), is a dummy variable set equal to 1 for the
period 1997 to 2000. DCRISIS*DLNSH[W.sub.i,t-1], DCRISIS*DLN[(1 +
ROA).sub.i,t] and DCRISIS*DLNTR[N.sub.it] are interaction variables.
Wald (joint): Wald test on all regressors except the dummies. For the
Wald test we report the Chi square statistic and degrees of freedom
(in the parentheses next to each statistic). AR(1) and AR(2) are
tests for first and second-order serial correlation, which are
distributed as N(0,1) under the null of no autocorrelation.
*** Significant at 1 per cent level, ** Significant at 5 per
cent level, * Significant at 10 per cent level.
NOTES
Earlier versions of this paper were presented at "Whose Millenium?" the Thirteenth Biennial Conference of the Asian Studies Association of Australia (ASAA), University of Melbourne, July 2000 and the National Workshop on Capacity Building Towards Excellence in Economic Research and Policy Formulation, Universiti Utara, Malaysia, April 2001. We thank participants at these conferences, Mita Bhattacharya, Dietrich Fausten, and two referees for helpful comments that have improved the final version. Funding for this project was supplied by a grant from Monash University, Malaysia
(1.) Means and medians have been calculated by using the whole sample. (We have pooled all data available on 564 companies, which gave us 1,647 observations of company/year.)
(2.) In the regressions reported later in the paper we use sectors rather than industries. Companies listed on the KLSE board are grouped into twelve sectors: construction, consumer products, finance, industrial products, trading and services, hotel, infrastructure, mining, plantation, properties, trusts, and closed-end funds. Each sector is divided into a number of industries. Some previous studies use industries and others use sectors to get a comparative performance measure. Typically, the choice of measure depends on the availability of data, but industry data is not always available. In this study it was not possible to use industries because we needed data for at least three firms in one sector or industry to be able to calculate a comparative performance measure and in many cases we did not have that many at the industry level. By including a sector performance measure instead of an industry one we assume that when directors' salaries are set, the former is used.
(3.) International Accounting Standards website, <http://www.iasplus.com/agenda/extract.htm>.
(4.) For some companies the stock market data for earlier years provided in the latest issue of KLSE on Disc (Vol. 1, No. 4) are inconsistent with the ones provided in previous issues. For example, 1993 data on latest transacted price of some companies in Vol. 4 do not match the 1993 data in earlier discs. We used the latest data in these cases.
(5.) In our data set, data from, say, 1999-2000 financial year appears as year 2000 information. To calculate sector returns for a year we use all data available for that year, no matter when the financial year ends.
(6.) <http://www.klse-ris.com.my>.
(7.) The median level of directors' remuneration in real terms was RM679,536 in 1991, RM1,099,000 in 1995, and RM1,127,508 in 1999 for the companies in the sample.
(8.) We also experimented with alternative concentration measures such as the percentage of shares owned by the largest five and largest ten shareholders. The results were similar to those reported below using the largest shareholder.
(9.) According to the DPD manual: "If the disturbances [v.sub.it] are not serially correlated, there should be evidence of significant negative first order serial correlation in differenced residuals, (i.e [v.sub.i,t] - [v.sub.i,t-1]), and no evidence of second order serial correlation in the differenced residuals. These tests are based on the standardized average residual autocovariances which are asymptotically N(0; 1) variables under the null of no autocorrelation" (Doornik, Arellano, and Bond 1999, p. 8).
(10.) We experimented with a number of different lag structures to remove serial correlation in the first, fourth, and sixth specifications without success. We report the results even though they are impaired by serial correlation to allow the reader to compare our results with those in the previous literature.
REFERENCES
Allen, M. P. and S. K. Panian. "Power, Performance and Succession in the Large Corporation". Administrative Science Quarterly 27 (1982): 538-47.
Ampalavanar-Brown, R. "Overseas Chinese Investments in China--Patterns of Growth, Diversification and Finance: The Case of Charoen Pokphand". China Quarterly (1998): 610-36.
Antle, R. and A. Smith. "An Empirical Investigation of the Relative Performance Evaluation of Corporate Executives". Journal of Accounting Research 24 (1986): 1-32.
Asian Development Bank (ADB). Country Studies Under RETA. Manila: Asian Development Bank, 1999.
Baker, G., M. Jensen, and K. Murphy. "Compensation and Incentives: Practice vs Theory". Journal of Finance 43 (1988): 593-616.
Berle, A. and G. Means. The Modern Corporation and Private Property. New York: Harcourt, 1932.
Brickley, J. A., R. C. Lease, and C. W. Smith. "Ownership Structure and Voting on Anti-takeover Amendments". Journal of Financial Economics 29 (1988): 267-91.
Brunello, G., C. Graziano, and B. Parigi. "Executive Compensation and Firm Performance in Italy". Unpublished manuscript. Department of Economics, University of Udine, Udine, Italy, 1997.
Claessens, S. and S. Djankov. "Ownership Concentration and Corporate Performance in the Czech Republic". Journal of Comparative Economics 27 (1999): 498-513.
Claessens, S., S. Djankov, and L. Lang. "Who Controls East Asian Corporations?". Unpublished manuscript. World Bank and the University of Chicago, 1999.
Claessens, S., S. Djankov, J. Fan, and L. Lang. "Diversification and Efficiency of Investment by East Asian Corporations". World Bank Research Paper 2033, 1999.
Conyon, M. J. "Corporate Governance and Executive Compensation". International Journal of Industrial Organization 15 (1997): 493-509.
Conyon, M. J. and P. Gregg. "Pay at the Top: A Study of the Sensitivity of Top Director Remuneration to Company Specific Shocks". National Institute Economic Review 3 (1994): 83-92.
Conyon, M., P. Gregg, and S. Machin. "Taking Care of Business: Executive Compensation in the United Kingdom". Economic Journal 105 (1995): 704-14.
Conyon, M. J. and D. Leech. "Top Pay, Company Performance and Corporate Governance". Oxford Bulletin of Economics and Statistics 56 (1994): 229-47.
Crespi, R. and C. Gispert. "Board Remuneration, Performance and Corporate Governance in Large Spanish Companies". Unpublished manuscript. Department d'Economia de l'Empresa, Universitat Autonoma de Barcelona, 1998.
DeAngelo, H. and L. DeAngelo. "Managerial Ownership of Voting Rights: A Study of Public Corporations with Dual Classes of Common Stock". Journal of Financial Economics 14 (1985): 33-69.
Deckop, J. "Determinants of Chief Executive Officer Compensation". Industrial and Labor Relations Review 41 (1988): 215-26.
Demsetz, H. and K. Lehn. "The Structure of Ownership: Causes and Consequences". Journal of Political Economy 93 (1985): 1155-77.
Doornik, J. A., M. Arellano, and S. Bond. "Panel Data Estimation using DPD for Ox". Unpublished material, 1999.
Espy, J. L. "The Strategy of Chinese Industrial Enterprises in Hong Kong". DBA Dissertation, Harvard University, 1970.
Fama, E. and M. Jensen. "Separation of Ownership and Control". Journal of Law and Economics 26 (1983): 301-25.
Gibbons, R. and K. Murphy. "Relative Performance Evaluation for Chief Executive Officers". Industrial and Labor Relations Review 43 (1990): 30-49.
Gilson, S. and M. Vetsuypens. "CEO Compensation in Financially Distressed Firms: An Empirical Analysis". Journal of Finance 48 (1993): 425-58.
Goldberg, L. and T. Idson. "Executive Compensation and Agency Effects". Financial Review 30 (1995): 313-35.
Gower, I. G. B. Principles of Modern Company Law. London: Sweet & Maxwell, 1992.
Greenhalgh, S. "Families and Networks in Taiwan's Economic Development". In Contending Approaches to the Political Economy of Taiwan, edited by E. Winkler and S. Greenhalgh. Armonk, NY: Sharpe, 1988.
Gregg, P., S. Machin, and S. Szymanski. "The Disappearing Relationship Between Directors' Pay and Corporate Performance". British Journal of Industrial Relations 30 (1993): 1-10.
Haubrich, J. "Risk-Aversion, Performance Pay and the Principal-Agent Problem". Journal of Political Economy 102 (1994): 258-76.
Holmstrom, B. "Moral Hazard and Observability". Bell Journal of Economics 10 (1979): 74-91.
--. "Moral Hazard in Teams". Bell Journal of Economics 13 (1982): 324-40.
Indjejikian, R. J. "Performance Evaluation and Compensation Research: An Agency Perspective". Accounting Horizons 13 (1999): 147-57.
Jensen, M. and K. Murphy. "Performance Pay and Top Management Incentives". Journal of Political Economy 98 (1990): 225-64.
Ke, B., K. Petroni, and A. Safieddine. "Ownership Concentration and Sensitivity of Executive Pay to Accounting Performance Measures: Evidence from Publicly and Privately-held Insurance Companies". Journal of Accounting and Economics 28 (1998): 185-209
Koike, K. "Introduction to a Special Issue on Business Groups in Developing Economies". Developing Economies 31 (1993): 363-77.
La Porta, R., F. Lopez-De-Silanes, A. Shleifer, and R. Vishny. "Law and Finance". Journal of Political Economy 106 (1998): 1113-54.
La Porta, R., F. Lopez-De-Silanes, and A. Shleifer. "Corporate Ownership Around the World". Journal of Finance 54 (1999): 471-513.
Lira, M. H. Ownership and Control of the One Hundred Largest Corporations in Malaysia. Oxford: Oxford University Press, 1981.
Main, B., A. Bruce, and T. Buck. "Total Board Remuneration and Company Performance". Economic Journal 106 (1996): 1627-44.
McConaughy, D. L., M. C. Walker, G. V. Henderson, and C. S. Mishra. "Founding Family Controlled Firms: Efficiency and Value". Review of Financial Economics 7 (1998): 1-19.
McConnell, J. J. and Servaes, H. "Additional Evidence on Equity Ownership and Corporate Value". Journal of Financial Economics 27 (1990): 595-612.
Mehran, H. "Executive Compensation Structure, Ownership and Firm Performance". Journal of Financial Economics 38 (1995): 163-84.
Morck, R. and D. A. Strangeland. "Corporate Performance and Large Shareholders". Working Paper No. 4-94, Institute for Financial Research, Faculty of Business, University of Alberta, Edmonton, 1994.
Morck, R., A. Shleifer, and R. Vishny. "Management, Ownership and Market Valuation: An Empirical Analysis". Journal of Financial Economics 20 (1988a): 293-315.
--. "Characteristics of Targets of Hostile and Friendly Takeovers". In Corporate Takeovers, Causes and Consequences, edited by A. Auerbach. Cambridge, MA: NBER, 1988b.
Murphy, K. "Corporate Performance Pay and Managerial Remuneration: An Empirical Analysis". Journal of Accounting and Economics 7 (1985): 11-42.
--. "Incentives, Learning and Compensation: A Theoretical and Empirical Investigation of Managerial Labor Contracts". Rand Journal of Economics 17 (1986): 59-76.
--. "Executive Compensation". In Handbook of Labor Economics, edited by O. Ashenfelter and D. Card. Vol. 3. North Holland, Amsterdam: 1999.
Organization of Economic Co-operation and Development (OECD). Corporate Governance in Asia: A Comparative Perspective. Seoul: OECD and World Bank, 1999.
Price Waterhouse Coopers. Corporate Governance: 1998 Survey of Public Listed Companies. Kuala Lumpur: Price Waterhouse Coopers, 1998.
--. Board of Directors: A Survey of Remuneration Practices 2001. Kuala Lumpur: Price Waterhouse Coopers, 2001.
Prowse, S. "The Structure of Corporate Ownership in Japan" Journal of Finance 47 (1992): 1121-40.
Ramaswamy, K., R. Veliyath, and L. Gomes. "A Study of the Determinants of CEO Compensation in India". Management International Review 40 (2000): 167-91.
Rosen, S. "Contracts and the Market for Executives" In Contract Economics, edited by L. Werin and H. Wijkander, pp. 181-217 Oxford: Basil Blackwell, 1992.
Sato, Y. "The Salim Group in Indonesia: The Development and Behaviour of the Largest Conglomerate in Southeast Asia". Developing Economies 31 (1993): 408-41.
Shleifer, A. and R. Vishny. "Large Shareholders and Corporate Control". Journal of Political Economy 94 (1986): 461-88.
--. "A Survey of Corporate Governance". Journal of Finance 94 (1997): 461-88
Silin, R. Leadership and Values: The Organization of Large-scale Taiwanese Enterprises. Cambridge, MA: Harvard University Press, 1976.
Smith, B. and B. Amoako-Adu. "Management Succession and Financial Performance of Family Controlled Firms". Journal of Corporate Finance 5 (1999): 341-68.
Suehiro, A. "Family Business Reassessed: Corporate Structure and Late-Starting Industrialization in Thailand". Developing Economies 31 (1993): 378-407.
Taniura, T. "The Lucky-Goldstar Group in the Republic of Korea". Developing Economies 31 (1993): 465-84.
Tenev, S, C. Zhang, and L. Brefort. Corporate Governance and Enterprise Reform in China: Building the Institutions of Modern Markets. Washington, D.C.: World Bank and IFC, 2002.
Tirole, J. The Theory of Industrial Organization. Cambridge, MA: MIT Press, 1988.
Tricker, R. "Corporate Governance and the Challenges Ahead". Corporate Governance International 2 (1999): 64-81.
Vickers, J. and G. Yarrow. Privatization: An Economic Analysis. Cambridge, MA: MIT Press, 1988.
Wang, C. "Incentives, CEO Compensation and Shareholder Wealth in a Dynamic Agency Model". Journal of Economic Theory 76 (1997): 72-105.
Whyte, M. "The Chinese Family and Economic Development: Obstacle or Engine?". Economic Development and Cultural Change 45 (1996): 1-30.
Wiwattanakantang, Y. "An Empirical Study of the Determinants of the Capital Structure of Thai Firms". Pacific-Basin Finance Journal 7 (1999): 371-403.
Wymeersch, E. "A Status Report on Corporate Governance Rules and Practices in Some Continental European States". In Comparative Corporate Governance: The State of the Art and Emerging Research, edited by K. Hopt et al., pp. 1045-199. Oxford: Clarendon Press, 1998.
Xu, X. and Y. Wang. "Ownership Structure and Corporate Governance in Chinese Stock Companies"