Printer Friendly

Firm performance: do non-executive directors have minds of their own? Evidence from UK panel data.

I investigate the relation between firm performance and both ownership structure and board composition. Use of the GMM methodology permits simultaneous control of both endogeneity of the independent variables and fixed effects. The data comprise an original, large, hand-collected panel dataset of UK firms for the period 1991-2001. Results indicate that the direction of causality runs from ownership and board composition to performance. I find a cubic relation between performance and ownership by executive directors. The proportion of non-executives on the board, but not their proportional ownership, is significantly and positively related to firm performance. Finally, the relation between performance and blockholdings by institutional and non-institutional owners is negative. Thus, results indicate that only non-executive directors are effective monitors.

**********

The relation between firm performance and both ownership structure and board composition is a widely debated and controversial issue. Two questions lie at the heart of this controversy: is there a causal link between firm performance and both ownership structure and board composition and what are the conditions that produce better performance? My empirical analysis addresses these core questions. First, I analyze whether a causal relation exists using a technique that controls simultaneously for endogeneity of the independent variables and for fixed effects. Second, I investigate in depth the incentives associated with improved firm performance for company insiders (usually defined as managers or directors) and external blockholders both institutional and non-institutional. The focus is on the monitoring role of non-executive directors and institutional blockholders. Results indicate that outside directors are effective monitors, but that institutional blockholders are not.

First, I address the question of whether a relation exists between ownership structure and firm performance. Previous empirical studies report mixed results on the direction of causality and the shape of the relation between ownership structure and performance. McConnell and Servaes (1990, 1995) observe an inverse U-shaped relation between insider ownership and Tobin's q; Morck, Shleifer, and Vishny (1988), Hermalin and Weisbach (1991), and Short and Keasey (1999) report a cubic relation; and, in a recent study, Davies, Hillier, and McColgan (2005) test a quintic relation. However, an increasing number of studies fail to detect any evidence that ownership affects performance (among others, Loderer and Martin, 1997; Cho, 1998; and Demsetz and Villalonga, 2001).

One possible explanation for these inconsistencies is that they are driven by methodological differences. Specifically, not all studies have controlled for the endogeneity of the explanatory variables and for endogeneity due to fixed effects, both of which may result in spurious inferences (Demsetz, 1983). For example, if endogeneity of the explanatory variables is not controlled for, a positive link between managerial ownership and firm performance can be interpreted in two ways. One interpretation is that owning ordinary shares aligns the interests of managers with those of external shareholders, and that this has a positive effect in performance. However, it is also reasonable to infer that managers of better performing firms are more willing to accept shares as part of their compensation. Further, endogeneity may arise due to the presence of unobserved firm heterogeneity (fixed effects). If ownership is correlated with unobserved firm-specific characteristics, studies that fail to control for this condition yield biased and inconsistent estimates (Himmelberg, Hubbard, and Palia, 1999).

To address these econometric issues, I use the GMM methodology (Arellano and Bond, 1991). Using data that has been transformed by taking first differences and using suitable lagged levels of the dependent variables as instruments, this technique produces results that are robust to both types of endogeneity.

The second important issue that I address is whether non-executive directors are effective monitors on behalf of shareholders. Although the board of directors is supposed to limit management's self-serving behavior, directors who are also executives are obviously not objective monitors. Non-executive directors, on the other hand, are "delegated monitors" charged by the shareholders with overseeing management's use of firm resources (Hart, 1995). In recent years, regulators have increasingly emphasized the importance of non-executive directors on company boards. For example, the Higgs Report (2003) recommended that boards be composed of a majority of independent non-executive directors in the UK. A similar requirement has been adopted recently in the US by the NYSE and NASDAQ (Holmstrom and Kaplan, 2003). Researchers maintain that concern about reputation and future career opportunities should suffice to make outside directors effective monitors (Fama and Jensen, 1983). As a test of this hypothesis, several studies evaluate the relation between the proportion of non-executives on the board and firm performance, with generally inconclusive results (among others, Agrawal and Knoeber, 1996; Yermack, 1996; and Dahya and McConnell, 2006).

On the other hand, many researchers have questioned the effectiveness of non-executive directors as monitors, arguing that they, like executives, are motivated to act in the interest of other shareholders only if they have a significant investment in the firm (Morck et al, 1988 and Jensen, 1993). Only two papers have attempted to assess separately the influences of share ownership by executive and non-executive directors. Both of these studies use US data, and they reach opposite conclusions (Morck et al., 1988 and Bhagat and Black, 2002).

To disentangle these various potential influences, I first test the relation between firm performance, measured by Tobin's q, and total board ownership. I then divide the board into executives and non-executives to assess whether a relation exists for each group separately. Also, I include a proxy for board composition to test whether the proportion of outside directors on the board, as distinct from their ownership position, has a significant effect on firm performance. This is an important aspect of the study which has a direct bearing on policy decisions being made by regulators in many countries, as they formulate codes of best practices. (1)

The third important issue investigated in this study is whether the presence of either institutional or non-institutional blockholders is relevant to firm performance. This is an important issue, given the rise in the size of institutional equity positions around the world. The Office for National Statistics (2001), reports that the average equity position owned by UK institutional investors increased from 29% in 1963 to 56.2% in 2000. Binay (2005) reports that in the US market, the percentage of equity owned by domestic institutional owners increased from 35% in 1981 to 58% in 2002. (2) Many researchers maintain that large investors have greater incentives than small investors to monitor. It is argued that they can more easily bear the cost and that they have more to gain from a rise in stock prices (Stiglitz, 1985 and Shleifer and Vishny, 1997). However, external shareholders, particularly institutional investors, are expected to hold well-diversified portfolios, and they are not tied to specific firms. In very liquid markets, the trade-off between the benefits of monitoring and the benefits of risk-sharing may motivate a strategy of selling rather than of monitoring and holding (Maug, 1998). Moreover, some authors argue that different types of shareholders, specifically institutional as opposed to non-institutional owners, may have different incentives and costs associated with monitoring (Brickley, Lease, and Smith Jr., 1988 and Pound, 1988). Therefore, they may not be equally vigilant monitors. Consequently, I first test the impact of the total proportion of shares held in blocks on Tobin's q, and then I test the proportional ownership of institutional and non-institutional blockholders separately.

My analysis uses a large, original, hand-collected panel dataset of UK firms for the period 1991-2001. During this period, the UK market provides an ideal setting in which to study the effects on firm performance of ownership structure and board composition. Following the corporate scandals of the early nineties, the effectiveness of non-executive directors as monitors has come under increasing scrutiny. The codes of best practice adopted in the UK, beginning with the Cadbury Report in 1992, provide a set of recommendations for the board's composition and responsibilities. Implementation of the codes created a natural experiment for testing my hypotheses. The data suggest that the codes have had an effect on both board composition and board ownership. The proportional representation of non-executive directors increased steadily during the sample period. Furthermore, although executive directors' share ownership declined steadily, non-executive directors' ownership rose.

Further, an investigation of the role of blockholding and institutional ownership is also particularly relevant in the UK. There has been an intense debate about the passive strategy typically adopted by UK institutional investors, who are the largest group of owners in the market. All UK codes of best practice have explicitly expressed concern about the failure of institutional owners to deal with underperformance in companies in which they invest.

The results of my analysis indicate that the direction of causality runs from proportional ownership to firm performance, and that different categories of investors affect performance differently. I find that the cubic effect of board ownership, which has been reported in previous research, is significant only for executive directors. Although the size of non-executive directors' ownership positions is not related to firm performance, results indicate that the proportion of non-executives on the board has a positive and significant effect on firm performance. On the other hand, the proportion of shares owned in block positions, and in particular those owned by institutional investors, is negatively related to firm performance. Thus, results suggest that non-executive directors are effective monitors on behalf of shareholders but that owners of block positions are not.

The remainder of the paper is organized as follows. Section I develops the hypotheses tested in this analysis. In Section II, the data and the methodology are described. The empirical results are presented in Section III, and Section IV discusses the conclusions to be drawn from the study.

I. Hypotheses

A. Board Ownership and Firm Performance

In their seminal paper, Jensen and Meckling (1976) derive a model in which the distribution of shares between firm insiders and outsiders influences the firm's market value. The rationale of this model is that executives' natural tendency is to allocate firm resources in their own best interests. When they own more shares, they are less inclined to divert resources away from firm value maximization because their interests are more closely aligned with those of other investors (the alignment effect). Accordingly, one would expect a positive relation between managerial ownership and performance.

Subsequently, Fama and Jensen (1983) and Demsetz and Lehn (1985) proposed that increased ownership also gives managers greater voting power and control, which they may use to expropriate company resources (entrenchment effect). Therefore, at higher levels of ownership, a negative relation with firm value would be anticipated. Furthermore, at very high levels of management ownership, the relation may become positive again since, at some point, the manager is effectively a manager-owner.

On the basis of these arguments, I test for a cubic relation between Board Ownership and firm performance, as measured by Tobin's q. Board Ownership is defined as the total percentage of shares outstanding held by board members. (3) The cubic functional form is confirmed by preliminary data inspection. Figure 1 shows Tobin's q plotted against Board Ownership, in five-percent intervals. At low levels of director ownership, the average q-ratio increases; beyond 5% to 10% ownership, the association is negative; and at very high levels of managerial ownership, between 30% and 40%, the relation is positive again.

B. The Separate Effects of Executives and Non-Executives on Firm Performance

As discussed earlier, a number of researchers have noted that outside directors have different incentives than inside directors. Non-executive directors are responsible for reviewing the performance of both the board and executive directors (Cadbury, 1992, Morck et al., 1988). Their positions are usually part time, they often sit on many boards, and they are typically paid less than executive directors (see Davies, 2002, for the UK, and see Morck et al., 1998, for the US). Consequently, one may wonder what incentives they have to actively monitor executive directors. Morck et al. (1988) argue that since monitoring requires both time and effort, outside directors should be given significant economic incentives to motivate them to monitor. Stock ownership is the usual mechanism used to provide these incentives.

Therefore, in the second step of this analysis, I investigate the separate impacts of inside and outside board ownership on firm performance. The two ownership variables are Executive Ownership, the percentage of ordinary shares owned by directors who are executives, and Non-Executive Ownership, the percentage of shares owned by non-executive directors.

[FIGURE 1 OMITTED]

To the best of my knowledge, only two previous studies have addressed this issue. Both use US data, and they reach different conclusions. Morck et al. (1988) report that non-executive ownership is related to firm performance. They find that the relation is positive between 0 and 5%, suggesting an alignment effect, and is negative between 5% and 25%, suggesting an entrenchment effect. On the other hand, Bhagat and Black (2002) find that the relation between the percentage ownership of non-executive directors and performance is generally insignificant. (4)

Theory does not offer predictions regarding the form of the relation between non-executive director ownership and performance. Therefore, I test three different functional forms: cubic, quadratic and linear.

Many researchers have proposed that non-executive directors may provide effective monitoring, irrespective of the level of their shareholding. Fama and Jensen (1983) contend that concern about their own reputations and future careers might create enough incentive to make outside directors effective monitors. In this case, one would expect a positive relation between firm performance and the proportion of non-executive directors on the board. However, it is also possible that outside directors' concern with their reputations and future careers might have the opposite effect. For example, Hart (1995) argues that "non-executives may owe their position to management." This may be particularly relevant in the UK market where, according to the Higgs Report (2003), "A high level of informality surrounds the process of appointing non-executive directors. Almost half of the non-executive directors surveyed for the Report, were recruited to their role through personal contacts or friendships. Only 4% had a formal interview and 1% had obtained the job through answering an advertisement." The implication of this observation is that British non-executive directors may not be independent; as Hart (1995) suggests, it may be that only "quiet non-executives" are selected for board positions.

Jensen (1993) suggests two additional reasons why there may not be a positive relation between the proportion of outside directors and firm performance. He maintains that non-executive directors' relative lack of expertise may inhibit their ability to monitor. He also suggests that large boards are easier for the CEO to control, and that therefore, in the presence of large boards, the CEO's influence may swamp that of the outside directors.

As discussed in the Introduction, the sample period offers a unique setting in which to investigate whether non-executive directors have an influence on firm performance. Starting with the Cadbury Report (1992), several new codes of best practices make specific recommendations regarding the responsibilities of non-executive directors as monitors. UK data is also of interest in view of the differences in board composition between the US and the UK. American boards often have more non-executive than executive directors. However, non-executives may not dominate the board; in the US, the CEO is often also the Chairman, and is able to control the board (Charkham, 1994). By contrast, until the early nineties, UK boards usually had more executive than non-executive directors. At that time, the newly issued codes of conduct started to encourage the appointment of more outside directors. Specifically, the Cadbury Report (1992) required that at least three non-executives be on the board; the Hampel Report (1998) required that at least one third of the board be non-executives; and most recently, the Higgs Report (2003) required at least 50%. These codes of best practice work on a "comply or explain" basis. That is, although adoption of these recommendations is not compulsory, the London Stock Exchange requires companies to explain in their annual reports the reasons for non-compliance. Since adoption of the codes of best practice, the number of outside directors has increased steadily, as has their average shareholding.

Another feature specific to the UK market is that voting at shareholder meetings is usually done by a show of hands (Goergen and Renneboog, 2001). Unless an issue is controversial, each shareholder has one vote, regardless of the percentage of shares owned. This may make it easier for non-executives (or small shareholders), to oppose managers, even though they hold few shares.

Several empirical studies have used US data to analyze the effect of board composition, as opposed to non-executive ownership, on firm performance, with mixed results. Rosenstein and Wyatt (1990), who used event study methodology, reported a small positive change in stock price following the appointment of an additional outside director. In an analysis using panel data, Yermack (1996) found a positive association between the proportion of outside directors and Tobin's q. However, Agrawal and Knoeber (1996) and Bhagat and Black (2002) observed a negative association between the percentage of board seats held by outsiders and Tobin's q. Finally, Hermalin and Weisbach (1991) and Mehran (1995) found no relation between the proportion of outside directors and firm performance.

Similarly, studies using UK data show mixed results. Faccio and Lasfer (2000) find little evidence that firm value is affected by the combination of managerial ownership and board structure. However, Weir, Laing and McKnight (2002) show that UK firms that perform well have a greater proportion of outsiders on both the board and the audit committee. Dahya, McConnell, and Travlos (2002) report a significant increase in management turnover following adoption of the Cadbury recommendations. They also find that turnover is more sensitive to performance when there are more outsiders on the board.

To test the hypothesis that having a larger proportion of outside directors on the board is associated with more effective monitoring and better performance, I use the variable Ratio, the proportion of the board who are non-executives. Moreover, to test whether the effect of Ratio on Tobin's q depends on the level of non-executive ownership, I also include an interaction term, Non-Exec Ownership*Ratio.

C. Outside Blockholders and Firm Performance

Several theoretical papers have investigated the effectiveness of monitoring by large shareholders. Admati, Pfleiderer, and Zechner (1994) argue that there is a trade-off between the benefits of monitoring, which are more likely to occur in the presence of concentrated ownership, and the benefits of risk-sharing, which is associated with diffuse ownership. In their model, the shareholding level reflects a commitment to monitor; a larger share position indicates a greater commitment to monitor because the owner receives more of the surplus that monitoring produces. Khan and Winton (1998) analyze the relation between liquidity and monitoring by large shareholders. They propose a theoretical model in which the choice between monitoring and selling is determined by the relative payoffs of these strategies. Maug (1998) notes that monitoring involves a free-rider problem; the costs are borne by one owner, but the benefits are enjoyed by all firm shareholders. Maug (1998) argues that a higher degree of liquidity in the market makes it more attractive to sell rather than to stay and monitor. Thus, higher market liquidity may, ceteris paribus, reduce monitoring by large shareholders.

Stiglitz (1985) argues that shareholders with larger positions are more likely to become involved in monitoring because they can more easily bear the cost of collecting information. Or, putting it differently, they are expected to be more active since they have more to win or lose from firm performance (Shleifer and Vishny, 1997). These arguments suggest that block positions are associated with better performance. On the other hand, a large shareholder may seek to maximize his own wealth, at the expense of other investors (Shleifer and Vishny, 1997). Additionally, Burkart, Gromb, and Panunzi (1997) suggest that even if tight control by shareholders is ex-post efficient, it may constitute an ex-ante threat to managers that they will receive fewer benefits from the firm, which would reduce their initiative. Consequently, blockholder monitoring may have a negative effect on firm performance.

The empirical literature on this issue is extensive, and it generally suggests that blockholders do not have a significant impact on performance. For instance, Holderness and Sheehan (1988) find no difference between the performance of firms with concentrated and dispersed ownership. McConnell and Servaes (1990, 1995) report no relation between firm performance and several proxies for block positions: the size of a firm's largest single block position, the total percentage of shares held in block positions, and a dummy indicating the presence of a blockholder. Likewise, both Shome and Singh (1995) and Agrawal and Knoeber (1996) fail to detect any evidence that blockholders play an important role in better performance. In contrast, Lasfer (2002) finds a significant negative association between blockholding and performance in his sample of UK companies. Similarly, Davies et al. (2005) find a strong negative association between firm value and total blockholder ownership in UK firms.

In my tests of the blockholder hypotheses, I use two different proxies: Blockholding, which is the total proportion of stock held in block positions by all non-managerial owners, where a block position is more than 3% of shares; and Largest Non-Managerial Ownership, which is the stock position of the largest non-manager owner.

However, it is important to distinguish among different types of outside shareholders, since they often have different interests. According to Pound (1988), institutional investors can be more efficient monitors than other shareholders because of their greater expertise (the efficient monitoring hypothesis). However, Pound also proposes that institutional investors may not monitor, either because they find it profitable to cooperate with managers (the strategic alignment hypothesis), or because they must cooperate with managers in order to protect other business relationships (the conflict of interest hypothesis).

The role of institutional investors in the UK is of special interest for several reasons. First, institutions have held more shares than any other group over the last decade (see Table I). Second, UK institutional investors face no legal restrictions on stock ownership, which is not the case in all markets. For example, US insurance companies may not invest more than 2% of their assets in a single company. Finally, UK institutions face no legal constraints on active participation in firm management. Again, this freedom is not universal. In the US, for example, the Schedule 13D filing requirement obliges shareholders owning more than 5% of shares to disclose in advance their plans regarding the company. Although UK institutional investors are free to participate actively in the firm, they tend not to do so. Plender (1997) reports that UK institutional investors seldom exercise their voting power, and Goergen and Renneboog (2001) provide evidence that this passive strategy increases "the already significant power of directors." In addition, Faccio and Lasfer (2000) observe that UK pension funds are not effective monitors. Cosh and Hughes (1997) find that institutional investors do not have a significant influence on decisions regarding either executive pay or dismissal policies. Similarly, Dahya et al. (2002) find no evidence that ownership by UK institutions has an effect on management turnover, while Short and Keasey (1999) find that UK institutional ownership plays no role in determining firm value. The Cadbury (1992), Hampel (1998) and Higgs (2003) codes of best practice all contain specific recommendations calling for institutional investors to increase their participation in corporate governance. Furthermore, the Myners Report (2001) proposed that institutional shareholders should be legally obligated to give proper consideration to the exercise of their voting rights. In view of this concern about UK institutional investors' passivity, I would expect to find either an insignificant or a negative relation between institutional ownership and firm value.

Previous evidence of a relation between institutional ownership and performance in US firms is mixed. Agrawal and Knoeber (1996) and Duggal and Millar (1999) find no evidence that the presence of institutional owners is associated with improved performance in US firms. In contrast, McConnell and Servaes (1990) report not only that the US institutional ownership is positively and significantly related to firm value, but they also report that the ownership level at which the entrenchment effect starts to dominate the alignment effect increases when this proxy is included in the model. This result is interpreted as supporting Pound's (1998) efficient monitoring hypothesis. Almazan, Hartzell, and Starks (2005) report that the presence of active institutions is associated with better performance, suggesting that these institutions are effective monitors. Finally, Jennings (2005) and Seifert, Gonenc, and Wright (2005) report negative associations between ownership by US institutional investors and Tobin's q.

To investigate whether different types of blockholders have a different impact on firm performance, I divide Blockholding into two categories: Institutional Ownership, comprising, among others, banks, pension funds and fund managers, and Non-Institutional Ownership, comprising private individuals and non-financial companies. As a robustness check, I create two additional variables based on the size of the largest individual block position (as a percent of shares outstanding): Largest Institutional Ownership and Largest Non-Institutional Ownership.

D. Control Variables

I include several control variables identified in the literature as being likely to influence performance. I control for the effect of firm size, since larger firms may find it easier both to generate funds internally and to access external capital. Larger firms may also benefit from the economies of scale and entry barriers associated with size. Further, since the average managerial shareholding is larger in smaller firms, managerial entrenchment may be more of a problem for them (McConnell and Servaes, 1990; 1995). Size is the natural logarithm of total assets in 1991 prices.

Following Morck et al. (1988), McConnell and Servaes (1990) and Cho (1998), I include two proxies for the firm's investments: one for investment in fixed capital Capital Expenditures, which is the ratio of fixed assets to total assets and one for the firm's investment in intangible assets R&D Expenditures, which is the ratio of R&D expenses to total assets. Starting with Jensen and Meckling (1976), it has frequently been argued that investment has a positive effect on firm performance. Consistent with this argument, researchers have found evidence that the US stock market reacts positively to announcements of both plans for new capital expenditures (McConnell and Muscarella, 1985) and increased R&D expenses (Chan, Martin, and Kensinger, 1990). In view of these findings, I expect that increases in investment will have a positive impact on firm performance.

Jensen (1986) argued that agency conflicts between managers and shareholders may be greater when managers have discretion over a larger quantity of liquid assets. If this is so, then cash flow would be a negative predictor of firm performance. On the other hand, it can be argued that greater cash flow may permit firms to finance their investments internally, which reduces underinvestment and bankruptcy risks. If this is so, then one would expect to find a positive association between cash flow and performance. The inclusion of a proxy for the availability of internal funds is appropriate in its own right, aside from the agency conflict issue, because it has been found to be strongly correlated with investment (Fazzari, Hubbard, and Petersen, 1988). My proxy for cash flow, Cash Flow, is the ratio of pre-tax profits plus depreciation to total assets.

Jensen (1986) also argued that higher levels of debt limit the agency problems associated with managers having access to large amounts of liquid assets, since managers will have less cash available after servicing the debt. Further, Stiglitz (1985) maintains that effective monitoring is provided mainly by lenders rather than shareholders, and that their presence should be associated with better performance. Modigliani and Miller (1963) also predict a positive correlation between leverage and performance due to the value of tax shields. Finally, Ross (1977) proposes that increased leverage may convey positive news, since it indicates that management is confident that the firm will be able to service more debt.

On the other hand, debt may have a negative relation with firm performance. Existing leverage may limit the firm's ability to raise new debt, and consequently it may force the firm to pass up valuable investment opportunities (Myers, 1977). Furthermore, higher leverage increases the risk of bankruptcy, and hence it increases equity risk. This risk will be priced in the market and may result in lower stock prices. I define the variable Leverage as the ratio of total debt to total assets.

I also include a proxy for dividend payments in the analysis. Some researchers contend that dividend payments, like leverage, reduce the agency problems associated with excess liquid assets (Easterbrook, 1984 and Jensen, 1986). Thus, one would expect a positive association between dividends and performance. Others argue that firms with few profitable investment opportunities pay higher dividends, rather than undertaking negative net present value projects (Smith and Warner, 1979). In this case, one would expect a negative relation. The measure of dividend payout, Dividends, is the ratio of ordinary dividends minus the Advance Corporation Tax to total assets.

II. Data and Methodology

A. Data Collection and Sampling

Initially, I selected a random sample of 1100 listed non-financial firms from Datastream. Ownership data were hand-collected from the Price Waterhouse Corporate Register (December issues for each year). Economic and market data were downloaded from Datastream. Much care was taken to track all name changes and de-listings for these companies. Most of this information was collected from the London Stock Exchange Yearbook, which systematically records name changes, delistings, and additionally, the status of companies in liquidation, receivership, and under court administration. The LSE Yearbook information was checked against information on the Companies House website (http://www.companieshouse.gov.uk).

The dataset was thoroughly inspected for errors. For example, I checked to be sure that the sum of all shares did not exceed 100%. If the sum was greater than 100%, I then tried to cross check the information in other issues of the Corporate Register (the September edition of the same year or the March edition of the following year) and in the London Stock Exchange Yearbook, which includes some ownership information. If the problem could not be resolved, the observation was excluded from the sample.

I then checked the economic variables for outliers, and trimmed the data to the 99% percentile. To assure that trimming did not distort the data, descriptive statistics for the trimmed data were checked against statistics for comparable data reported in other papers. Firms in the broadcasting and public utility sectors were excluded, because of their unique operating and regulatory conditions. Firms with dual class shares were also excluded, since they violate the "one share one vote" rule. (5) If a firm had missing data for any variable in a firm-year, that company was excluded. Finally, only firms with at least five consecutive years of data were retained. This step is necessary for computing asymptotically efficient, second-order serial correlation tests (Arellano and Bond, 1991). Further, the survivorship bias may be reduced by requiring a minimum number of consecutive years of data (Yermack, 1996). After this screening, there remained an unbalanced panel of 672 firms and 6340 observations.

I have gone to considerable lengths to check for consistency in the data and, in particular, for attrition bias. Analysis of comparative descriptive statistics shows that the firms used in the empirical analysis are not systematically different from those excluded.

B. Ownership Structure

In the share ownership data, the threshold for external blockholders is 3% of ordinary shares outstanding. Required disclosure of block positions began in 1967, with a threshold of 10% of shares outstanding. In 1976, the threshold was lowered to 5%, and in 1988 it was further reduced to 3%. However, even though the threshold had been lowered, the Corporate Register for 19891991 reports data based on a 5% threshold. Therefore, the data for 1991 is based on a 5% threshold, and the rest of the data is based on a 3% threshold. I categorized external blockholders into four groups: financial firms, non-financial firms, individuals and the state. Ownership by directors must always be disclosed, even if they own no shares. This makes it possible to identify them on the board.

Preliminary descriptive statistics are shown in Table I. Figures in the first four rows indicate that there has been little change in the percentage of shares held by outside blockholders during the sample period. On average, Blockholding is around 30% of shares outstanding; approximately 20% of shares are owned by institutions, and the remaining 10% are owned by non-institutions. However, it is interesting to note that, while the average position of institutional investors as a group is around 20%, this stake is held by an average of three institutions (see Institutional Investors in Table II). Thus, the average position of each institution is approximately 6.67%. For non-institutional blockholders, the average position is about 6.92% (Table I), and is therefore comparable to a single institutional block position. (6) Also, the average percentage position of the largest non-managerial owner appears to be increasing over time.

On the other hand, Board Ownership, the average percentage position owned by board members, has decreased by approximately 5% over ten years; this decline is due entirely to a decline in ownership by executives on the board. Half of this reduction occurred between 1991 and 1993, immediately after the Cadbury Report was issued (1992).

Table II shows that while the average board size has been relatively stable over time, the composition of the board has changed significantly. In 1991, the average board included 4.73 executives and 2.44 non-executives, but by 2001, non-executives constituted almost half of the average board. These figures are consistent with Lasfer (2002) and Marchica and Mura (2005). The same trend occurs in the subset of board members who own shares (cure shares).

C. Economic Performance

In many previous studies, researchers have used Tobin's q to measure firm performance. Tobin's q is the ratio of the market value of a firm's assets to their replacement value. Tobin's (1969) original idea was that the replacement cost of assets would be a logical measure of their alternative use. If the market values the firm's assets at less than replacement cost, management is not using them efficiently. However, calculating the replacement cost of assets is problematic. As Claessens, Djankov, Fan, and Lang (2002) noted, "the data required to calculate the replacement values are generally not available." To deal with this problem, researchers have used different approximations of the market-to-book ratio. (7) I use the following proxy: Tobin's q is the ratio of the book value of total assets minus the book value of equity plus the market value of equity to the book value of assets. (8) A brief summary of descriptive statistics for the variables used in this study is reported in Table III.

D. Methodology

As mentioned earlier, it is important to control for endogeneity of regressors and unobservable heterogeneity (fixed effects) when estimating the relation between managerial ownership and firm performance. All regressors are potentially endogenous because shocks that affect corporate value are also likely to affect other regressors such as investment, leverage and dividends. Furthermore, cross causality may complicate relations among variables. For example, managers' equity ownership may both have an impact on and be influenced by firm value. Another source of endogeneity is the presence of unobservable firm-specific characteristics that are correlated with the regressors. Controlling for fixed effects is important as it may limit omitted variable bias (Chi, 2005).

To control for both types of endogeneity, I estimate the following model using the GMM technique:

[Q.sub.it] = [[alpha].sub.1][Q.sub.it-1] + [summation.sup.k.sub.k=1] [[beta].sub.k][X.sub.it]+ [[epsilon].sub.it] [[epsilon].sub.it] = [[eta].sub.t] [[eta].sub.i][[mu].sub.it], (1)

where [[epsilon].sub.it] is a composite error term (Wooldridge, 2002) that includes [[eta].sub.t], a macroeconomic shock term for events that affect all firms; [[eta].sub.i], a firm specific component of the error term; and [[u.sub.it], an idiosyncratic error term, which I assume, for the moment, to be homoschedastic and serially uncorrelated. A dynamic model was selected because preliminary autocorrelation tests indicated misspecification of the mean function with a static model. As Bond (2002) observed, "even when coefficients on lagged dependent variables are not of direct interest, allowing for dynamics in the underlying process may be crucial for recovering consistent estimates of other parameters." The dynamic specification seems to yield consistent estimates in my analysis.

As I discussed earlier, Himmelberg et al. (1999) emphasize the importance of firm fixed effects ([[eta].sub.i]) when estimating the relation between value and ownership. They provide the following example. Assume that the effectiveness of monitoring mechanisms differs among firms. Managers of firms that have better monitoring cannot divert as many firm resources. Further, better monitoring reduces the need to align managers' incentives with shares and options in compensation packages. If this is the case, executive ownership will be negatively correlated with monitoring effectiveness. Consequently, the relation between executive ownership and market value will be negative because the monitoring effectiveness variable has been omitted. As a result, OLS estimates would be biased and inconsistent. Finally, as noted in Bond (2002), the dynamic specification implies that at least the lagged dependent variable is positively correlated with the fixed effects. This correlation would cause the OLS estimate of the autoregressive parameter in Equation (1) to be biased upwards.

The usual solution to this problem is to transform the data in order to remove the fixed effects. The standard transformation of the Within Groups (WG) estimator is to subtract the firm's mean for the sample interval from each period observation. Since [[eta].sub.i] is constant over time, the WG transformation eliminates it. However, a major disadvantage of this procedure is that unless all right hand side variables are strictly exogenous, it introduces a non-negligible correlation between non-exogenous variables and the time-demeaned error term. Consider the case in which managerial ownership is predetermined: E([Man.sub.it] ; [u.sub.it-1]) [not equal to] 0 but E([Man.sub.it] ; [u.sub.it]) = 0. In this case, the time-demeaned variable is [bar.[Man.sub.it]] = [Man.sub.it] - 1/T-1 ([Man.sub.i2] + [Man.sub.i3] +.... [Man.sub.iT]) and the error term is [bar.[mu].sub.it] = [[mu].sub.it] - 1/T-1 ([[mu].sub.i2] + [[mu].sub.i3] +.... [[mu].sub.iT]).

[[bar.Man].sub.it] and [bar.[mu]].sub.it] are correlated since all [Man.sub.it] terms are correlated with the corresponding [[mu].sub.it-1] terms. Also, in this case, the estimate of the autoregressive parameter is biased downwards, due to its negative correlation with the time-demeaned error term (Bond, 2002). Because the OLS and WG estimators are biased in opposite directions, a consistent estimator is expected to lie between them (Bond, 2002). (9)

Another technique for dealing with this problem is Arellano and Bond's (1991) Generalized Method of Moments (GMM-DIFF) estimator. They suggest taking first differences and then using suitable lagged levels of the dependent variables as instruments. Taking the first difference of Equation (1), we obtain:

[DELTA][Q.sub.it] = [DELTA][[alpha].sub.1][Q.sub.it-1] + [k.summation over (k=1)] + [DELTA][[beta].sub.k] + [X.sub.it] + [DELTA][[eta].sub.t] + [DELTA][[mu].sub.it]. (2)

Taking first differences effectively transforms predetermined variables into endogenous ones. However, after the transformation the error term does not include all realizations of the disturbances, and thus the WG correlation problem is avoided. Consequently, under the assumption of no serial correlation in the error term, the second (and earlier) lagged levels can be used as instruments for the endogenous variables. If the error term in Model (1) is MA(1), then the first differenced error term is MA(2). In this case the second lag is not a valid instrument, but the third and earlier lags are. Because the validity of the GMM relies heavily on the absence of higher order serial correlation, I test for autocorrelation of order one and two (test statistics ml and m2).

The choice of an appropriate set of instruments is crucial in this type of analysis. The validity of the instrument set can be tested with the Sargan test of over-identifying restrictions, which tests the null hypothesis of zero correlation between the instruments and the error term. Rejection of the null casts doubt on the validity of the instruments.

A high degree of persistence in the data may create another problem. (10) Under this condition, lagged levels have a low correlation with first differences and the standard linear GMM-DIFF estimator displays poor finite sample properties (Blundell, Bond, and Windmeijer, 2000). Blundell and Bond (1998) note that the GMM-DIFF finite sample bias is likely to be in the direction of the WG estimator when weak instruments are present.

To deal with the GMM-DIFF finite sample bias, Arellano and Borer (1995) propose the GMM-SYS estimator, which includes level equations. Lagged first-differences are used as instruments for level equations and lagged level terms are used as instruments for equations in first differences. Blundell and Bond (1998) examined this procedure in detail, and demonstrate that this estimator shows a significant gain in asymptotic efficiency. Therefore, I use the GMM-SYS methodology for this analysis.

III. Empirical Results

The results shown in Table IV are consistent with expectations. The OLS and WG estimated autoregressive parameters (Tobins 'q (-1)) define the range of values estimated for the autoregressive term. The coefficients in the two GMM-DIFF models are close to the WG estimate, as expected, since they are biased in the same direction. The GMM-SYS estimate is closer to the midpoint of the range, which is consistent with my expectation that it is less biased than the alternative estimates.

The Wald (joint) test indicates that the regressors are jointly significant, and the Wald (time) test provides support for including the time dummies. The Sargan statistic test for Model 3 indicates that instruments lagged at t-2 appear to be correlated with the error term. Therefore, the GMM estimations use the lagged levels at t-3 and t-4 for all variables in the first-difference equations, and use the first difference lagged at t-2 for all variables in the level equations. This parsimonious specification is preferred to using all available moment conditions, as this could result in a severe overfitting bias (Bond, 2002). Additionally, using too many moment conditions reduces dramatically the power of the Sargan statistic to detect invalid instruments (Bowsher, 2002). In Model 5, the insignificant Sargan statistic confirms the validity of the instrument set, and the insignificant Sargan Difference statistic validates the extra moment restrictions imposed by the use of the level equations in the GMM-SYS specification. As expected, the ml and m2 statistics indicate serial correlation of order one, but not of order two. (11)

A. Board Ownership and Firm Performance

First, I analyze whether a causal relation exists between board ownership and firm performance, and what the shape of the relation is. In the GMM-SYS model (Table IV, Model 5), the coefficients on the three board ownership terms (of powers one, two and three) are all significant at the .05 level.

Given the properties of the GMM estimator, these results can be interpreted in terms of causality, and they clearly indicate that the direction of the relation runs from ownership to performance. Also, this result confirms that the cubic function is appropriate. The implied turning points are the ownership levels at which performance first reflects the entrenchment effect, and then reflects the alignment effect. They are approximately 15% and 45% for this model. Data inspection reveals that Board Ownership is less than 15% of shares outstanding for about 67% of firms in the sample. At these low levels, the relation between ownership and performance is positive. This result is consistent with the alignment hypothesis. Board Ownership is in the range of 15% to 45% for around 22% of firms, where the relation between ownership and performance is negative. This result is consistent with the entrenchment hypothesis. For the remaining 11% of firms, with Board Ownership of more than 45%, the relation is positive again. At this level, board members may be viewed as owners of the firm, so that the conflict between management and shareholders fades. My findings are consistent with previous research results, which have indicated that UK directors become entrenched at higher ownership levels than their US counterparts (Lasfer, 2002).

B. The Separate Effects of Executives and Non-executives on Firm Performance In the second set of tests as reported in Table V, the sample of board members is divided into executives and non-executives. Since theory offers no prediction regarding the functional form of the relation between outside directors' ownership and firm performance, I test three functional forms: cubic, quadratic and linear (Models 1-3). Further, I include an interaction term, Non-Exec Ownership x Ratio, to account for the possibility that the proportion of non-executives on the board has an impact on firm performance conditional on their level of ownership (Model 4).

Results for all four models in Table V indicate that the relation between performance and ownership by executive directors has the same cubic form as that identified for ownership by the full board. The stability of the estimated turning points across the four models lends support to the conclusion that the relation has a cubic form. However, it is noteworthy that the relation between performance and non-executive ownership is insignificant in all models. This result confirms Bhagat and Black's (2002) findings.

The figures in my analysis are consistent with two different interpretations. On the one hand, they may indicate that UK outside directors are truly independent, and thus are not influenced by the alignment and incentive effects associated with share ownership. On the other hand, as suggested by Hart (1995), the results may also be interpreted as evidence that non-executive board members do not have sufficient financial incentives to motivate them to monitor. As reported in Table I, during the sample period, ownership by non-executives was substantially lower than ownership by executives, especially in the early nineties.

However, the significant positive coefficients on the Ratio variable (proportion of non-executives on board) indicate that having a larger percentage of outside directors is associated with better firm performance. It may be that when there are more non-executive directors, they are more likely to overrule executives, whether it is with their votes or by speaking up when they have objections. Further, the presence of more outsiders may constitute a credible threat to executives that may be enough to limit their self-serving behavior.

Several robustness checks were performed, both to confirm these findings and to further distinguish between the separate effects of non-executive directors' ownership and their percentage representation. First, all estimations were run using the percentage of non-executive directors cum shares on the board, rather than the percentage of all non-executive directors. The values in Table VI for Model 1 show that while the results are similar to the basic findings, one result is noticeably different in magnitude. The estimated coefficient for the percentage of non-executives cure shares (Ratio cure shares) is almost half of that for all non-executives reported in Table V, Model 4. (12) This result suggests that the market interprets an increase in the percentage of non-executive directors as an indication of good governance, irrespective of the size of their shareholding.

As a further test, the estimations were run with Ratio (cum shares) and Non-Executive Ownership separately (Models 2 and 3). The coefficient on the proportion of non-executive directors (cum shares) is significant, whereas the coefficient on their total shareholding is not.

I also investigated whether the results for non-executive ownership depend on the specific functional form of these models. Dummy variables corresponding to several threshold values of non-executive ownership were generated; for example, Dum.Non-Exec Own>5% equals one when non-executive ownership is greater than or equal to 5%. The dummies are combined with the percentage non-executive ownership in an interaction term. This procedure isolates firms in which non-executive ownership is high, and therefore permits determination of whether the lack of statistical significance is due to non-executive directors typically having small investments in their firms. For brevity, Table VII reports results for a selection of the robustness checks (5%, 15% and 25%). (13) The coefficients on the three dummy interaction terms in Models 1-3 are insignificant, including the coefficient on the interaction term for the largest non-executive directors' ownership position (>25% of shares outstanding). Thus, tests of various functional forms confirm that there is not a statistically significant relation between non-executive ownership and performance, while the significant positive relation between proportional representation and performance persists.

In an unreported test, I also investigated whether it is possible to identify an "optimal" board structure. One potential motivation for an optimal proportion of non-executive directors is that beyond some level, the responsibility for monitoring might be too diffuse, and each non-executive would have an incentive to free ride. Thus, monitoring would decline with additional non-executive representation. To test for an optimum board composition, I transformed the variable Ratio into a percentage and squared it. I tested for significance of the squared term by adding it to the Models in Tables IV through VII. Results indicate that Tobin's q increases with any increase in non-executive proportional representation, thus indicating that the relation is not non-linear and that an optimal board composition does not exist.

These results have significant policy implications. As discussed earlier, the codes of best practice adopted in the UK have all recommended increasing the representation of non-executives on boards of directors. The rationale behind this policy is that non-executives are expected to monitor managers' use of firm resources in order to protect shareholder interests. My results indicate that this has been an effective policy. It is reasonable to infer that the increased attention given to the role of non-executives may have amplified market scrutiny of their actions and decisions and may have provided stronger incentives to build reputations as efficient monitors (Fama and Jensen, 1983). Ultimately, this, may have lead to boards making better decisions.

This interpretation is in line with recent evidence on the UK market. For instance, one of the primary board tasks is to hire and fire the CEO. Dahya and McConnell (2005) report that boards that are in compliance with the codes are more likely to appoint outside CEOs. Further, Dahya et al. (2002) find that the sensitivity of CEO turnover to performance is greater when non-executives comprise a larger percentage of the board.

Other UK studies have reported that in the post-Cadbury period, the increase in the proportion of non-executive directors has been associated with improved corporate performance. Lasfer (2004) reports that in the post-Cadbury period, the positive relation between the proportion of non-executive directors and performance is stronger. Dahya and McConnell (2006) find that when UK companies increase the proportion of non-executive directors in order to comply with the Cadbury recommendations, they show significant improvement in operating performance. My finding is also consistent with previous US research results (Yermack, 1996, and Baysinger and Butler, 1985).

C. Outside Blockholders and Firm Performance

My analysis of the effect of outside blockholders on firm performance yields two noteworthy results. First, in all models in Tables IV to VII, the relation between Blockholding, the total ownership position of all external blockholders, and performance is negative and significant. This is consistent with findings reported in recent studies of the UK market (Lasfer, 2002 and Davies et al., 2005). Franks, Mayer, and Renneboog (2001) fail to detect a relation between ownership concentration and monitoring activity in the UK. Second, when the largest non-managerial ownership position Largest Non-Managerial Ownership, is substituted for Blockholding (Table VIII, Model 1), the negative relation with performance persists. The combination of these results indicates that outside blockholders, whom one would expect to be independent of management, do not seem to be effective monitors.

The literature offers several potential explanations for these results. First, too much block ownership may overly constrain managers, thus limiting their ability to make value maximizing decisions (Burkart et al., 1997). Alternatively, the position of each blockholder may be too small to motivate the owner to actively monitor. A third explanation involves the trade-off between liquidity and control discussed earlier. The highly liquid UK market may often make it more attractive economically for blockholders to sell their stock rather to hold and monitor (Maug, 1998, Khan and Winton, 1998).

In an attempt to shed more light on this matter, I performed an unreported test based on a procedure from McConnell and Servaes (1990); Blockholding was excluded from Model 1, Table V. The estimation was rerun, and the result was compared to the original Model 4 result. In the new model (without Blockholding), the estimated turning points are 14.99% and 41.13%, compared to 13.56% and 42.57% in the original model (with Blockholding). The decline in the lower turning point is consistent with a negative relation between Blockholding and performance; it indicates that, in the presence of blockholdings, managers become entrenched at lower levels of ownership; in other words, blockholding increases the range of ownership over which entrenchment dominates alignment, which has a negative effect on firm performance.

I also test the separate impacts of institutional and non-institutional blockholders on performance. In Table VIII, Model 2 tests the size of total institutional and total non-institutional blockholdings, and Model 3 tests the largest single blockholding of an institutional and a non-institutional owner. The coefficients on all four terms are negative and statistically significant. Thus, it appears that both groups have a negative impact on performance and that neither has a better effect on performance than the other.

The results for institutional shareholders support the thesis that UK institutional owners are passive investors (Faccio and Lasfer, 2000, Goergen and Renneboog, 2001). Their "absentee landlord" attitude gives management more opportunity for self-serving behavior, and thus has a negative effect on firm value. This result is inconsistent with the "efficient monitoring" hypothesis. As I argued earlier, institutional blockholders' lack of engagement may be due, at least in part, to the UK market's high liquidity, which makes it more attractive economically for all blockholders, including institutional owners, to sell rather than to hold and monitor.

An alternative interpretation is provided by Pound's (1988) hypotheses that institutional investors may find it profitable to cooperate with managers (strategic alignment hypothesis) or that they may be forced to cooperate with management in order to protect other business relationships with the firm (conflict of interest hypotheses). For instance, if the value of the business relationships with the firm are greater than its expected loss in equity value due to managerial agency problems, it would be rational for the institution to vote with incumbent management. (14) This interpretation is consistent with Brickley et al. (1988). They look at the benefits an institutional owner receives from business activities directly under top management's control. Their results indicate that institutions receiving such benefits are less likely to oppose management proposals than other institutions. (15)

The negative relation between institutional ownership and performance is also consistent with results reported in recent studies of the US market. Among others, Jennings' (2005) Granger causation tests indicate that although quality firms attract institutional investment, firm value declines following investment by quality institutional owners. Similarly, Seifert et al. (2005) find a negative relation between ownership by institutional investors and Tobin's q.

D. Control Variables

The results from tests on the coefficients of the firms' economic variables are similar to results reported in previous studies. Capital expenditures, cash flow and dividends all have highly significant positive impacts on firm value (see Tables IV to VIII). The positive association of cash flow and firm value suggests that internally generated funds allow firms to reduce the risk of underinvestment and to pursue positive net present value projects. On the other hand, the positive relation between dividends and firm value is consistent with Easterbrook's (1984) hypothesis; he proposed that dividends mitigate manager-shareholder conflict by making fewer liquid assets available for managers to divert to self-serving purposes, and hence market performance improves. In all of the models, there is a positive relation between investment expenditures and firm value, although only the relation with tangible asset expenditures (Capital Expenditures) is statistically significant. (16) My results are consistent with those reported by Davies et al. (2005) for UK firms. In contrast, Morck et al. (1988) and McConnell and Servaes (1990; 1995) find that in their samples of US firms, R&D expenditures are positively and significantly related to firm value. Finally, the coefficients on size and leverage are insignificant in all models.

IV. Conclusions

This study investigates whether a causal link exists between firm performance and both ownership structure and board composition. Three specific issues are analyzed in the study. First, is there a relation between ownership structure and firm performance, and if there is one, what are the direction of causality and the shape of the relation? Second, what is the effect of non-executive directors on firm performance, and is their effect associated with their proportional ownership or with their proportional representation on the board? And third, what are the separate effects on performance of the proportional ownership of institutional and non-institutional blockholders.

Two technical factors contribute to making the test results interesting. First, the GMM-SYS methodology that I use is able to control simultaneously for endogeneity of the regressors and for endogeneity due to fixed effects, both of which may confound inferences about causality. Second, I use an original, large panel dataset of UK firms for the 1991-2001 period. UK data is particularly appropriate for the tests because UK boards have traditionally included a majority of executive directors, and because UK institutions have a reputation for being passive investors. During the sample period, several codes of best practice were adopted in the UK, all of which draw attention to the importance of independent monitoring by non-executive directors and to the negative effect of passive institutional investors. The codes of best practices had a clear effect on the proportional representation of non-managers on boards and on the level of their ownership over the sample period, providing a unique opportunity to study the effect of these variables on firm performance.

Results indicate that there is a relation between directors' proportional ownership and firm performance, and that the direction of causality runs from ownership to performance. Results also confirm the cubic relation predicted by the alignment/entrenchment hypothesis, and found in previous studies. That is, an increase in directors' share ownership is associated with improving performance at low and at very high levels of ownership (alignment), and it is associated with deteriorating performance at moderate levels (entrenchment). Additional tests indicate that this result is driven by executive directors' ownership.

The relation between the proportional ownership of non-executive directors and performance is insignificant. This result is robust to 1) the introduction of an interaction term for non-executive directors' ownership and the proportion of non-executives on the board, 2) the use of different proxies for board composition, and 3) the specification of different functional forms.

Since the analysis indicates that there is no relation between the size of non-executive directors' share position and firm performance, the results cast doubt on the efficacy of equity ownership as an alignment mechanism for outside directors. One possible interpretation of this result is that non-executive directors are truly independent in the UK, and consequently are not sensitive to the alignment and entrenchment effects. On the other hand, another possible interpretation is that the levels of non-executive ownership are too low to affect their motivation to monitor (Hart, 1995). This interpretation is particularly interesting in view of the trend toward greater share ownership by non-executive directors in UK firms, which I document in this study. However, results provide mixed support for this explanation, since the relation between non-executive directors' ownership and performance is insignificant even for non-executive directors who own large positions.

However, results indicate that the proportion of non-executives on the board has a positive impact on firm performance. As I discussed earlier, this finding may have a direct bearing on policy decisions being made by regulators in many countries, as they formulate codes of best practices similar to the Cadbury Report in the UK. This result suggests that the boards of UK firms have been more effective monitors on behalf of other shareholders since the Cadbury and similar codes were issued in the UK.

In my evaluation of the association between performance and the proportion of shares held in block positions, I confirm the negative association found in previous studies of UK firms. Additional analyses suggest that the size of the ownership positions held by institutional and non-institutional investors separately, and the position of the largest non-managerial owner in each of these categories, all have negative associations with firm performance. Thus, owners of large positions appear not to monitor management effectively.

Special thanks are due to James Seward, Lemma Senbet, Alexander Triantis (the Editors) and the anonymous referee whose detailed and valuable suggestions have greatly improved the quality of the paper. I am indebted to Mara Faccio, Antoine Faure-Grimaud, Jana Fidrmuc, Leslie Godfrey, Andrew Jones, Arif Khurshed, Meziane Lasfer, Samuel Lee, Maria-Teresa Marchica, Aydin Ozkan, Enrico Perotti, Martin Walker, and Frank Windmeijer for helpful discussions. I also thank all the participants at the 2005 FMA Siena Meeting, and the EFMA 2005 Milan Meeting, for their useful comments. Kind help from Charles Bridge of the Higgs Report Review Team, Derek Rouse of Hemscott Plc Support Team, and David Roodman from the Center for Global Development is also acknowledged. I am also grateful to Suzanne Bellezza for editorial help. The usual disclaimer applies.

References

Admati, A.R., R Pfleiderer, and J. Zechner, 1994, "Large Shareholder Activism, Risk Sharing and Financial Market Equilibrium," Journal of Political Economy 102, 1097-1130.

Agrawal, A. and C.R. Knoeber, 1996, "Firms Performance and Mechanisms to Control Agency Problems Between Managers and Shareholders," Journal of Financial and Quantitative Analysis 31, 377-397.

Almazan, A., J.C. Hartzell, and L.T. Starks, 2005, "Active Institutional Shareholders and Costs of Monitoring: Evidence from Executive Compensation," Financial Management 34, 5-34.

Arellano, M. and S. Bond, 1991, "Some Tests of Specification for Panel Data: Monte Carlo Evidence and an Application to Employment Equations," Review of Economic Studies 58, 277-297.

Arellano, M. and O. Bover, 1995, "Another Look at the Instrumental Variable Estimation of Error Component Models," Journal of Econometrics 68, 29-51.

Baysinger, R.D. and H.N. Butler, 1985, "Corporate Governance and the Board of Directors: Performance Effects of Changes in Board Composition," Journal of Law Economics' and Organization 1, 101-124.

Bhagat, S. and B.S. Black, 2002, "The Non-Correlation Between Board Independence and Long-Term Firm Performance," Journal of Corporation Law 27, 231-273.

Binay, M., 2005, "Performance Attribution of US Institutional Investors," Financial Management 34, 127-152. Blommestein, H. J. and N. Funke, 1998, "The Rise of the Institutional Investor," OECD Observer 12, 37-42.

Blundell, R. and S. Bond, 1998, "Initial Conditions and Moment Restrictions in Dynamic Panel Data Models," Journal of Econometrics 87, 115-143.

Blundell, R., S. Bond, and F. Windmeijer, 2000, "Estimation in Dynamic Panel Data Models: Improving on the Performance of the Standard GMM Estimators," Institute for Fiscal Studies Working Paper.

Bond, S., 2002, "Dynamic Panel Data Models: A Guide to Micro Data Methods and Practice," Portuguese Economic Journal 1, 141-162.

Bowsher, C., 2002, "On Testing Overidentifying Restrictions in Dynamic Panel Data Models," Economic Letters 77, 211-220.

Brickley, J.A., R.C. Lease, and C.W. Smith Jr., 1988, "Ownership Structure and Voting on Anti-takeover Amendments," Journal of Financial Economics 20, 267-291.

Burkart, M., D. Gromb, and F. Panunzi, 1997, "Large Shareholders, Monitoring and the Value of the Firm," Quarterly Journal of Economics 112, 693-728.

Cadbury Report, 1992, Report of the Committee on the Financial Aspects of Corporate Governance, London, GEE.

Chart, S.H., J.D. Martin, and J.W. Kensinger, 1990, "Corporate Research and Development Expenditures and Share Value," Journal of Financial Economics 26, 255-276.

Charkham, J., 1994, Keeping Good Company, Oxford, Oxford University Press.

Chi, J., 2005, "Understanding the Endogeneity Between Firm Value and Shareholder Rights," Financial Management 35, 65-76.

Cho, M. H., 1998, "Ownership Structure, Investment, and the Corporate Value: An Empirical Analysis," Journal of Financial Economics 47, 103-121.

Chung, K.H. and S.W. Pruitt, 1994, "A Simple Approximation of Tobin's Q," Financial Management 23, 70-74.

Claessens, S., S. Djankov, J.P.H. Fan, and H.P. Lang, 2002, "Disentangling the Incentive and Entrenchment Effects of Large Shareholdings," Journal of Finance 57, 2741-2771.

Cosh, A. and A. Hughes, 1997, "Executive Remuneration, Executive Dismissal and Institutional Shareholdings," International Journal of Industrial Organization 15, 469-492.

Dahya, J., J.J. McConnell, and N.G. Travlos, 2002, "The Cadbury Committee, Corporate Performance and Top Management Turnover," Journal of Finance 57, 461-483.

Dahya, J. and J.J. McConnell, 2005, "Outside Directors and Corporate Board Decisions," Journal of Corporate Finance 11, 37-60.

Dahya, J. and J.J. McConnell, 2006, "Corporate Performance, and the Cadbury Committee Recommendation," Journal of Financial and Quantitative Analysis (Forthcoming).

Davies, J.R., D. Hillier, and P. McColgan, 2005, "Ownership Structure, Managerial Behavior and Corporate Value," Journal of Corporate Finance 11, 645-660.

Davies, P., 2002, Introduction to Company Law, Clarendon Law Series, Oxford University Press.

Demsetz, H., 1983, "The Structure of Ownership and the Theory of the Firm," Journal of Law and Economics 26, 375-390.

Demsetz, H. and K. Lehn, 1985, "The Structure of Corporate Ownership: Causes and Consequences," Journal of Political Economy 93, 1155-1177.

Demsetz, H. and B. Villalonga, 2001, "Ownership Structure and Corporate Performance," Journal of Corporate Finance 7, 209-233.

Duggal, R. and J.A. Millar, 1999, "Institutional Ownership and Firm Performance," Journal of Corporate Finance 5, 103-117.

Easterbrook, F.H., 1984, "Two Agency Costs Explanations of Dividends," American Economic Review 74, 650-59.

Erickson, T. and T.M. Whited, 2006, "On the Accuracy of Different Measures of Q," Financial Management 35, 5-33.

Faccio, M. and A. M. Lasfer, 2000, "Do Occupational Pension Funds Monitor Companies in Which They Hold Large Stakes?" Journal of Corporate Finance 6, 71-110.

Fama, E. and M.C. Jensen, 1983, "Separation of Ownership and Control," Journal of Law and Economics 26, 301-325.

Fazzari, S.M., G.R. Hubbard, and B. Petersen, 1988, "Financing Constraints and Corporate Investment," Brooking Papers on Economic Activity 1, 141-195.

Franks, J., C. Mayer, and L. Renneboog, 2001, "Who Disciplines Management in Poorly Performing Countries?" Journal of Financial Intermediation 10, 209-248.

Goergen, M. and L. Renneboog, 2001, "Strong Managers and Passive Institutional Investors in the UK," in F. Barca and M. Becht, Ed., The Control of Corporate Europe, Oxford, Oxford University Press.

Hampel Report, 1998, Committee on Corporate Governance: Final Report, London, GEE.

Hart, O., 1995, "Corporate Governance: Some Theory and Implications," Economic Journal 105, 678-689.

Hermalin, B. and M. Weisbach, 1991, "The Effect of Board Composition and Direct Incentives on Firm Performance," Financial Management 20, 101-112.

Higgs Report, 2003, Review of the Role and Effectiveness of Non-Executive Directors, London, GEE, 39.

Himmelberg, P. C., R.G. Hubbard, and D. Palia, 1999, "Understanding the Determinants of Managerial Ownership and the Link between Ownership and Performance," Journal of Financial Economics 53, 353-384.

Holderness, C.G. and D.P. Sheehan, 1988, "The Role of Majority Shareholders in Publicly Held Corporations: An Exploratory Analysis," Journal of Financial Economics 20, 317-347.

Holmstrom, B. and S.N. Kaplan, 2003, "The State of US Corporate Governance," Journal of Applied Corporate Finance 15, 8-20.

Jennings, W.W., 2005, "Further Evidence on Institutional Ownership and Corporate Value," Advances in Financial Economics 11, 167-207.

Jensen, C.M., 1986, "Agency Costs of Free Cash Flow, Corporate Finance and Takeovers," American Economic Review 76, 323-329.

Jensen, C.M., 1993, "The Modern Industrial Revolution, Exit and Failure of Internal Control Systems," Journal of Finance 48, 831-880.

Jensen, C.M. and W. Meckling, 1976, "Theory of the Firm: Managerial Behaviour, Agency Cost and Capital Structure," Journal of Financial Economics 3, 305-360.

Kahn, C. and A. Winton, 1998, "Ownership Structure, Speculation and Shareholder Information," Journal of Finance 53, 99-130.

Lasfer, M., 2002, "Board Structure and Agency Costs," Cass Business School Working Paper.

Lasfer, M.A., 2004, "On the Monitoring Role of the Board of Directors: The Case of the Adoption of Cadbury Recommendations in the U.K," Advances in Financial Economics 9, 1287-1326.

Loderer, C.F. and K. Martin, 1997, "Executive Stock Ownership and Performance: Tracking Faint Traces," Journal of Financial Economics 45, 223-255.

Marchica, M.T. and R. Mura, 2005, "Direct and Ultimate Ownership Structures in the UK: An Intertemporal Perspective Over the Last Decade," Corporate Governance: An International Review 13, 26-45.

Maug, E., 1998, "Large Shareholders as Monitors: Is There a Trade-off Between Liquidity and Control?" Journal of Finance 53, 65-98.

McConnell, J.J. and C.J. Muscarella, 1985, "Corporate Capital Expenditures and the Market Value of the Firm," Journal of Financial Economics 14, 399-422.

McConnell, J.J. and H. Servaes, 1990, "Additional Evidence on Equity Ownership and Corporate Value," Journal of Financial Economics 27, 595-612.

McConnell, J.J. and H. Servaes, 1995, "Equity Ownership and the Two Faces of Debt," Journal of Financial Economics 39, 131-157.

Mehran, H., 1995, "Executive Compensation Structure, Ownership, and Firm Performance," Journal of Financial Economics 38, 163-184.

Modigliani, F. and M. Miller, 1963, "Corporate Income Taxes and the Cost of Capital: A Correction," American Economic Review 53, 433-443.

Morck, R., Shleifer, A. and R.W. Vishny, 1988, "Management Ownership and Market Valuation: An Empirical Analysis," Journal of Financial Economics 20, 293-315.

Myers, S.C., 1977, "Determinants of Corporate Borrowing," Journal of Financial Economics 5, 147-75.

Myners Report, 2001, Institutional Investment in the United Kingdom, London, HM Treasury.

Nenova, T., 2003, "The Value of Corporate Voting Rights and Control: A Cross-country Analysis," Journal of Financial Economics 68, 325-351.

Office for National Statistics, 2001, Share Ownership: A Report on the Ownership of Shares at 31 December 2000, The Stationary Office, London.

Plender, J., 1997, A Stake in the Future. The Stakeholding Solution, London, Nicholas Brealey Publishing.

Pound, J., 1988, "Proxy Contests and the Efficiency of Shareholder Oversight," Journal of Financial Economics 20, 237-265.

Price Waterhouse Coopers, December Issues, Corporate Register, London, UK, HS Financial Publishing.

Rosenstein, S. and J.G. Wyatt, 1990, "Outside Directors, Board Independence and Shareholder Wealth," Journal of Financial Economics 26, 175-191.

Ross, S.A., 1977, "The Determination of Financial Structure: The Incentive Signalling Approach," Bell Journal of Economics 8, 23-40.

Seifert, B., H. Gonenc, and J. Wright, 2005, "The International Evidence on Performance and Equity Ownership by Insiders, Blockholders, and Institutions," Journal of Multinational Financial Management 15, 171-191.

Shleifer, A. and R. Vishny, 1997, "A Survey of Corporate Governance," Journal of Finance 52, 737-783.

Shome D.K. and S. Singh, 1995, "Firm Value and External Blockholdings," Financial Management 24, 3-14.

Short, E. and K. Keasey, 1999, "Managerial Ownership and the Performance of Firms: Evidence from the UK," Journal of Corporate Finance 5, 79-101.

Smith, C. and J. Warner, 1979, "On Financial Contracting: An Analysis of Bond Covenants," Journal of Financial Economics 7, 117-161.

Stiglitz, J.E., 1985, "Credit Markets and Control of Capital," Journal of Money, Credit, and Banking 17, 133-152.

Tobin J., 1969, "A General Equilibrium Approach to Monetary Theory," Journal of Money, Credit, and Banking 1, 15-29.

Weir, C., D. Laing, and P. McKnight, 2002, "Internal and External Governance Mechanisms: Their Impact on the Performance of Large UK Public Companies," Journal of Business Finance & Accounting 29, 579-611.

Wooldridge J.M., 2002, Econometric Analysis of Cross Section and Panel Data, Cambridge, MA, MIT Press.

Yermack, D., 1996, "Higher Market Valuation of Companies with Small Board of Directors," Journal of Financial Economics 40, 185-213.

(1) See Dahya and McConnell (2005) for an exhaustive and detailed list of countries.

(2) Similar figures are reported by Blommestein and Funke (1998) for all OECD countries.

(3) All ownership information is expressed as a percentage of ordinary shares outstanding.

(4) Mehran (1995) discusses unreported tests in which he analyzes the relation separately for executive and non-executive directors' ownership. He finds a linear association between outside director's share ownership and Tobin's q.

(5) Only 26 firms in the sample have dual class shares, which is comparable to the number reported in Nenova (2003).

(6) These statistics appear lower than the figures reported by the Office for National Statistics (2001). Their data are standardized over the disclosed ownership only (i.e. above 3%), while my data consider the FLOAT as well. This issue is also discussed in Faccio and Lasfer (2000).

(7) See Chung and Pruitt (1994) and Erickson and Whited (2006) for detailed discussions.

(8) Results are essentially the same using different approximations of Tobin's q. Results are available upon request.

(9) See Bond (2002) for a more technical treatment of these issues.

(10) An analysis of persistency in the data reveals that most of the autoregressive coefficients are close to 0.8.

(11) By construction, the process of taking first differences introduces serial correlation of order one.

(12) A similar reduction in the coefficient is evident in unreported results.

(13) Several additional thresholds, all multiples of 5%, were tested. The results are virtually identical to those reported in the table and are available upon request.

(14) For example, suppose that an institution is a shareholder in the company and it is also its main insurer.

(15) However it should be noted that UK companies are required to disclose the identity of their major suppliers of financial services, so that potential conflicts of interests would be revealed (Davies, 2002).

(16) One explanation for the insignificance of expenditures on intangible assets is that a large number of firms report zero R&D expenditures. Consequently, the data may not have sufficient variability to reveal a relation even if it exists.

Roberto Mura *

* Roberto Mura is a Lecturer in Finance at Manchester Business School, England, UK.
Table I. Average Percentage of Ordinary Shares Held by Insiders and
Outsiders

All values indicate the percentage of total shares outstanding held by
a particular group. In 1991, the disclosure hurdle was 5%, and thus
some values for 1991 are not comparable to those for later years; these
are marked with asterisks. Board Ownership is the total percentage of
shares owned by all directors. Executive Ownership is the total
percentage of shares owned by executive directors. Non-Executive
Ownership is the total percentage of shares owned by non-executive
directors. Blockholding is the total percentage of shares held in
positions of greater than 3% owned by external investors. Institutional
Ownership is the percentage of shares owned in blocks by financial
institutions. Non-Institutional Ownership is the percentage of shares
owned in blocks by private individuals and other non-financial
companies. Largest Non-Managerial Ownership is the largest share
position owned by a non-managerial shareholder. Float is the percentage
of shares held in positions of less than 3%.

 1991 1992 1993 1994

Board Ownership 15.26 13.71 13.13 11.88
Executive 13.34 11.93 11.30 9.97
Ownership
Non-Executive 1.92 1.77 1.83 1.90
Ownership
Blockholding 20.72 * 31.61 33.21 29.11
Institutional 12.63 * 22.31 23.22 20.52
Ownership
Non-Institutional 8.10 * 9.30 9.99 8.59
Ownership
Largest 8.53 * 9.10 9.67 10.03
Non-Managerial
Ownership
Float 64.02 * 54.69 53.66 59.01
Total sample 565 583 611 645
firms
 1995 1996 1997 1998

Board Ownership 12.24 10.99 10.80 10.28
Executive 9.94 8.94 8.54 8.16
Ownership
Non-Executive 2.29 2.06 2.26 2.11
Ownership
Blockholding 29.87 31.09 32.39 33.13
Institutional 20.96 22.84 23.38 24.51
Ownership
Non-Institutional 8.91 8.25 9.01 8.62
Ownership
Largest 10.32 10.78 11.01 11.42
Non-Managerial
Ownership
Float 57.90 57.92 56.81 56.59
Total sample 666 667 661 606
firms
 1999 2000 2001

Board Ownership 10.51 10.19 10.70
Executive 8.15 7.43 7.61
Ownership
Non-Executive 2.36 2.76 3.09
Ownership
Blockholding 33.80 33.40 32.43
Institutional 24.52 24.00 22.30
Ownership
Non-Institutional 9.28 9.40 10.13
Ownership
Largest 11.00 11.83 11.45
Non-Managerial
Ownership
Float 55.69 56.41 56.87
Total sample 501 440 395
firms

Table II. Ownership Composition: Average Number of Insiders and
Outsiders

This table shows the average number of company insiders and outsiders
belonging to several categories of interest. Board Size is the total
number of directors. Executive Directors is the total number of
directors who are firm executives. Non-Executive Directors is the
total number of directors who are not firm executives.
Ratio is the proportion of board members who are non-executives.
Executive and Non-Executive Directors(cum shares) are the numbers of
directors who own shares in the company. Ratio (cum shares) is the
proportion of board members who are non-executives and own shares in
the company. Blockholders is the total number of external shareholders
who hold positions of at least 3% of shares outstanding. Institutional
Investors is the total number of Blockholders who are financial
institutions. Non-Institutional investors is the total number of
Blockholders who are private individuals and non-financial companies.
In 1991, the disclosure hurdle was 5%, and thus some values for 1991
are not comparable to those for later years. These are marked with
asterisks.

 1991 1992 1993 1994 1995

Board Size 7.17 7.13 7.18 7.17 7.33
Executive 4.73 4.59 4.44 4.31 4.24
Directors
Non-Executive 2.44 2.54 2.74 2.86 3.10
Directors
Ratio 0.34 0.36 0.38 0.40 0.42
Executive 4.04 3.91 3.81 3.68 3.69
Directors
(cum shares)
Non-Executive 1.90 1.92 2.02 2.13 2.37
Directors
(cum shares)
Ratio 0.32 0.33 0.35 0.37 0.39
(cum shares)

Blockholders 2.15 * 4.84 4.95 4.20 4.27
Institutional 1.38 * 3.63 3.60 3.16 3.00
Investors

Non- 0.77 * 1.20 1.36 1.05 1.27
Institutional
Investors

 1996 1997 1998 1999 2000 2001

Board Size 7.32 7.28 7.33 7.38 7.39 7.35
Executive 4.15 4.05 4.00 3.94 3.92 3.82
Directors
Non-Executive 3.18 3.23 3.32 3.44 3.47 3.53
Directors
Ratio 0.43 0.44 0.45 0.47 0.47 0.48
Executive 3.60 3.55 3.56 3.50 3.47 3.39
Directors
(cum shares)
Non-Executive 2.47 2.49 2.64 2.73 2.78 2.81
Directors
(cum shares)
Ratio 0.41 0.41 0.43 0.44 0.44 0.45
(cum shares)

Blockholders 4.37 4.45 4.63 4.46 4.57 4.16
Institutional 3.36 3.02 3.51 3.13 3.45 2.79
Investors

Non- 1.01 1.43 1.12 1.33 1.12 1.37
Institutional
Investors

Table III. Economic Variables: Descriptive Statistics

This table shows the descriptive statistics for the economic variables
used in this work. Tobins q is the ratio of book value of total assets
minus the book value of equity plus the market value of equity to book
value of assets. Size is the natural logarithm of total assets in 1991
prices, reported in thousands of pounds. Capital Expenditures is the
ratio of total capital expenditures to total assets. R&D Expenditures
is the ratio of total research and development expenditures to total
assets. Cash Flow is the ratio of pre-tax profit plus depreciation to
total assets. Leverage is the ratio of total debt to total assets.
Dividends is the ratio of ordinary dividends net of the Advance
Corporation Tax to total assets.

 Mean Median Std. Dev. Min

Tobin's q 1.561 1.295 0.949 0.3628
Size 560,411 64,271 2,448,007 312
Capital Expenditures 0.072 0.045 0.109 -0.1134
R&D Expenditures 0.008 0 0.025 0
Cash Flow 0.070 0.080 0.128 -1.4431
Leverage 0.172 0.160 0.132 0
Dividends 0.028 0.025 0.033 0

 Max Obs

Tobin's q 9.732 6340
Size 67,400,000 6340
Capital Expenditures 2.130 6340
R&D Expenditures 0.381 6340
Cash Flow 0.713 6340
Leverage 0.940 6340
Dividends 1.285 6340

Table IV. Firm Performance and Board Ownership: Multivariate Regression
Models Using OLS, WG, GMM-DIFF and GMM-SYS

This table presents the estimated relation between firm performance
and ownership structure using different estimation techniques. The
dependent variable is Tobins q, defined as the ratio of book value
of total assets minus the book value of equity plus the market value
of equity to book value of assets. Board Ownership is the total
percentage of shares owned by all directors. Ratio is the proportion
of board members who are non-executives. Blockholding is the total
percentage of shares held in positions of greater than 3% owned by
external investors. Size is the natural logarithm of total assets in
1991 prices. Capital Expenditures is the ratio of total capital
expenditures to total assets. R&D Expenditures is the ratio of total
research and development expenditures to total assets. Cash Flow is
the ratio of pre-tax profit plus depreciation to total assets. Leverage
is the ratio of total debt to total assets. Dividends is the ratio of
ordinary dividends net of the Advance Corporation Tax to total assets.
Turning points are the ownership levels at which first the entrenchment
effect, and then the alignment effect, start to have an effect on
performance. Time dummies were included in all estimations. Model 1
is estimated using OLS in levels. Model 2 is estimated with the WG
methodology. Models 3 and 4 are the GMM-DIFF estimations using first
differences; GMM-DIFF(a) Model 3 uses levels dated [t-2] of all the
regressors as instruments, while GMM-DIFF(b) Model 4 uses levels dated
[t-3, t-4]. Model 5 uses the GMM-SYS method. For the first difference
equations, levels dated [t-3, t-4] are employed as instruments; in the
level equations, first differences dated [t-2] are used as instruments.
Regression coefficients are shown with heteroskedasticity robust
standard errors. P-values are provided in parentheses. The Wald (joint)
statistic tests the joint significance of all regressors. Wald (time)
statistic tests the joint significance of the time dummies. The Sargan
statistic tests for over-identifying restrictions, and is
asymptotically distributed as [chi square] under the null hypothesis of
valid instruments. The Sargan Difference statistic tests the extra
moment restrictions imposed by the use of the level equations in
the GMM-SYS specification. The ml and m2 statistics test the absence of
first-order and second-order correlation in the residuals. They are
asymptotically distributed as N (0,1) under the null hypothesis of no
serial correlation.

 OLS WG GMM-DIFF(a)

Independent Variables (1) (2) (3)

Tobin'sq(-1) 0.688 *** 0.393 *** 0.436 ***
 (0.000) (0.000) (0.000)
Board Ownership 0.001 -0.007 0.002
 (0.818) (0.318) (0.943)
Board Ownership (2) -5.39E-06 0.0004 -9.48E-05
 (0.977) (0.208) (0.920)
Board Ownership (3) 3.53E-07 -3.99E-06 1.71E-07
 (0.873) (0.248) (0.987)
Ratio 0.191 *** 0.047 0.032
 (0.002) (0.590) (0.924)
Blockholding -0.002 *** -0.002 ** -0.004
 (0.000) (0.039) (0.225)
Size -0.024 *** -0.206 *** -0.329 *
 (0.001) (0.000) (0.053)
Capital Expenditures 0.328 ** 0.299 * 0.968
 (0.016) (0.073) (0.320)
R&D Expenditures 2.414 *** 0.799 2.426
 (0.000) (0.572) (0.530)
Cash Flow 0.794 *** 0.991 *** 1.707 ***
 (0.000) (0.000) (0.010)
Leverage 0.146 0.347 ** 0.124
 (0.123) (0.030) (0.791)
Dividends 2.338 *** 2.301 *** 5.073
 (0.000) (0.001) (0.178)
Implied Turning Points
Wald (joint) 2412 *** 431.3 *** 146.6 ***
 (0.000) (0.000) (0.000)
Wald (time) 141.1 *** 107.6 *** 19.55 **
 (0.000) (0.000) (0.012)
Sargan 120.2 **
Sargan Difference
m1 -1.952 ** -0.657 -6.082 ***
 (0.051) (0.511) (0.000)
m2 0.582 -3.548 *** -0.629
 (0.560) (0.000) (0.529)

 GMM-DIFF(b) GMM-SYS

Independent Variables (4) (5)

Tobin'sq(-1) 0.431 *** 0.514 ***
 (0.002) (0.000)
Board Ownership -0.021 0.040 **
 (0.440) (0.016)
Board Ownership (2) 0.0003 -0.0018 **
 (0.771) (0.042)
Board Ownership (3) -2.13E-06 1.97E-05 *
 (0.863) (0.079)
Ratio -0.469 0.458
 (0.258) (0.108)
Blockholding -0.006 -0.006 **
 (0.111) (0.039)
Size -0.826 *** -0.035
 (0.000) (0.409)
Capital Expenditures 0.572 1.738 **
 (0.415) (0.029)
R&D Expenditures 4.456 1.918
 (0.224) (0.191)
Cash Flow 1.719 *** 1.289 **
 (0.005) (0.025)
Leverage 0.062 0.104
 (0.916) (0.716)
Dividends 1.618 * 4.056 ***
 (0.087) (0.005)
Implied Turning Points (15.06; 45.43)
Wald (joint) 90.72 *** 597.2 ***
 (0.000) (0.000)
Wald (time) 34.92 *** 20.78 **
 (0.000) (0.023)
Sargan 178.40 255.70
 (0.165) (0.511)
Sargan Difference 77.3
 (0.919)
m1 -4.391 *** -6.750 ***
 (0.000) (0.000)
m2 -0.615 -0.879
 (0.538) (0.380)

No. of observations: 6,340

*** Significant at the 0.01 level.

** Significant at the 0.05 level.

* Significant at the 0.10 level.

Table V. Firm Performance and Share Ownership by Executive and
Non-Executive Directors

This table presents the estimated relation between firm performance and
the percentage ownership of various groups using GMM-SYS. The dependent
variable is Tobin's q defined as the ratio of book value of total
assets minus the book value of equity plus the market value of equity
to book value of assets. Board ownership is divided into two groups:
Executive Ownership is the total percentage of shares owned by
executive directors and Non-Executive Ownership is the total percentage
of shares owned by non-executive directors. Ratio is the proportion of
board members who are non-executives. The interaction variable
Non-Executive Ownership*Ratio tests for the presence of an effect of
Non-Executive Ownership conditional on the level of the Ratio.
Blockholding is the total percentage of shares held in positions of
greater than 3% owned by external investors. Size is the natural
logarithm of total assets in 1991 prices. Capital Expenditures is the
ratio of total capital expenditures to total assets. R&D Expenditures
is the ratio of total research and development expenditures to total
assets. Cash Flow is the ratio of pre-tax profit plus depreciation to
total assets. Leverage is the ratio of total debt to total assets.
Dividends is the ratio of ordinary dividends net of the Advance
Corporation Tax to total assets. Turning Points are the levels of
executive ownership at which the effect of alignment (entrenchment)
starts to dominate. Time dummies were included in all estimations.
In all models, levels lagged at [t-3, t-4] of all the regressors are
used as instruments for the equations in first differences. In the
level equations, first differences lagged at [t-2] are used as
instruments. Regression coefficients are shown with heteroskedasticity
robust standard errors. P-values are provided in parentheses.
The Wald (joint) statistic tests the joint significance of all
regressors. Wald (time) statistic tests the joint significance of the
time dummies. The Sargan statistic tests for over-identifying
restrictions, and is asymptotically distributed as [chi square] under
the null hypothesis of valid instruments. The m1 and m2 statistics
test the absence of first-order and second-order correlation in the
residuals. They are asymptotically distributed as N (0, 1) under the
null hypothesis of no serial correlation.

Independent Variables (1) (2)

Tobin's q (-1) 0.519 *** 0.516 ***
 (0.000) (0.000)
Executive Ownership 0.0479 ** 0.0432 **
 (0.011) (0.025)
Executive -0.0023 ** -0.002*
 Ownership (2) (0.023) (0.059)
Executive 2.68E-05 ** 2.29E-05 *
 Ownership (3) (0.035) (0.086)
Non-Executive 0.017 -0.026
 Ownership (0.618) (0.179)
Non-Executive -0.002 0.0007
 Ownership (2) (0.291) (0.293)
Non-Executive 5.43E-05
 Ownership (3) (0.121)
Ratio 0.836 ** 0.842 ***
 (0.016) (0.005)
Non-Executive
 Ownership*Ratio
Blockholding -0.006 ** -0.006 **
 (0.034) (0.040)
Size -0.049 -0.046
 (0.195) (0.237)
Capital Expenditures 1.809 ** 1.990 ***
 (0.039) (0.010)
R&D Expenditures 0.920 1.520
 (0.492) (0.285)
Cash Flow 1.184 ** 1.373 **
 (0.041) (0.019)
Leverage -0.078 0.001
 (0.838) (0.997)
Dividends 4.129 ** 4.135 **
 (0.011) (0.015)
Implied Turning (14.53; 42.46) (14.62; 42.77)
 Points
Wald (joint) 538.8 *** 574.2 ***
 (0.000) (0.000)
Wald (time) 21.52 ** 28.65 ***
 (0.018) (0.001)
Sargan 350 331.9
 (0.215) (0.167)
m1 -6.804 *** -6.899 ***
 (0.000) (0.000)
m2 -0.796 -0.826
 (0.426) (0.408)

Independent Variables (3) (4)

Tobin's q (-1) 0.514 *** 0.521 ***
 (0.000) (0.000)
Executive Ownership 0.0437 ** 0.0447 **
 (0.027) (0.015)
Executive -0.0021 * -0.0022 **
 Ownership (2) (0.051) (0.028)
Executive 2.44E-05 * 2.57E-05 **
 Ownership (3) (0.068) (0.039)
Non-Executive -0.009 0.006
 Ownership (0.248) (0.747)
Non-Executive
 Ownership (2)
Non-Executive
 Ownership (3)
Ratio 0.872 *** 0.963 ***
 (0.003) (0.002)
Non-Executive -0.0241
 Ownership*Ratio (0.515)
Blockholding -0.006 ** -0.006 **
 (0.028) (0.019)
Size -0.055 -0.058
 (0.177) (0.197)
Capital Expenditures 1.689 ** 1.536 *
 (0.023) (0.057)
R&D Expenditures 1.801 1.804
 (0.215) (0.207)
Cash Flow 1.325 ** 1.211 **
 (0.022) (0.025)
Leverage 0.091 0.199
 (0.798) (0.562)
Dividends 4.914 ** 5.243 **
 (0.029) (0.028)
Implied Turning (14.05; 42.26) (13.56; 42.57)
 Points
Wald (joint) 530.9 *** 528.7 ***
 (0.000) (0.000)
Wald (time) 28.5 *** 34.31 ***
 (0.001) (0.000)
Sargan 302.8 323.8
 (0.236) (0.257)
m1 -6.830 *** -6.815 ***
 (0.000) (0.000)
m2 -0.967 -0.964
 (0.334) (0.335)

No. of observations: 6,340

*** Significant at the 0.01 level.

** Significant at the 0.05 level.

* Significant at the 0.10 level.

Table VI. Firm Performance and the Proportion of Non-executives Cum
Shares on the Board: Robustness Tests

This table presents a robustness test of the estimated relation between
firm performance and ownership structure using GMM-SYS. In this table,
I substitute the group of non-executives who own shares in the firm
(cum shares) for all non-executive directors in Ratio, the measure of
proportional representation. The dependent variable is Tobins q defined
as the ratio of book value of total assets minus the book value of
equity plus the market value of equity to book value of assets.
Executive Ownership is the total percentage of shares owned by
executive directors. Non-Executive Ownership is the total percentage
of shares owned by non-executive directors. Ratio(cum shares) is the
proportion of board members who are non-executives and own shares in
the company. The interaction variable Non-Executive Ownership*Ratio(cum
shares) tests for the presence of an effect of Non-Executive Ownership
conditional on the level of the Ratio(cum shares). Blockholding is the
total percentage of shares held in positions of greater than 3% owned
by external investors. Size is the natural logarithm of total assets
in 1991 prices. Capital Expenditures is the ratio of total capital
expenditures to total assets. R&D Expenditures is the ratio of total
research and development expenditures to total assets. Cash Flow is
the ratio of pre-tax profit plus depreciation to total assets. Leverage
is the ratio of total debt to total assets. Dividends is the ratio of
ordinary dividends net of the Advance Corporation Tax to total assets.
Turning Points are the levels of executive ownership at which the
effect of alignment (entrenchment) starts to dominate. Time dummies
were included in all estimations. In all models, levels lagged at [t-3,
t-4] of all the regressors are used as instruments for the equations
in first differences. In the level equations, first differences lagged
at [t-2] are used as instruments. Regression coefficients are shown
with heteroskedasticity robust standard errors. P-values are provided
in parentheses. The Wald (joint) statistic tests the joint significance
of all regressors. Wald (time) statistic tests the joint significance
of the time dummies. The Sargan statistic tests for over-identifying
restrictions, and is asymptotically distributed as a [chi square] under
the null hypothesis of valid instruments. The m 1 and m2 statistics
test the absence of first-order and second-order correlation in the
residuals. They are asymptotically distributed as N (0,1) under the
null hypothesis of no serial correlation.

Independent Variables (1) (2)

Tobins'q(-1) 0.543 *** 0.543 ***
 (0.000) (0.000)
Executive Ownership 0.044 ** 0.032 *
 (0.017) (0.084)
Executive Ownership (2) -0.002 ** -0.002 *
 (0.025) (0.079)
Executive Ownership (3) 2.67E-05 ** 2.21E-05 *
 (0.031) (0.075)
Non-Executive Ownership -0.017
 (0.417)
Ratio (cum shares) 0.592 ** 0.497 **
 (0.032) (0.040)
Non-Exec Ownership* 0.034
 Ratio(cum shares) (0.526)
Blockholding -0.004 * -0.004 *
 (0.060) (0.090)
Size -0.042 -0.048
 (0.292) (0.210)
Capital Expenditures 1.577 ** 1.428 *
 (0.045) (0.070)
R&D Expenditures 1.289 0.877
 (0.297) (0.510)
Cash Flow 0.950 * 1.075 **
 (0.055) (0.040)
Leverage 0.076 0.036
 (0.807) (0.920)
Dividends 4.876 ** 5.045 **
 (0.025) (0.010)
Implied Turning Points (12.98; 42.71) (11.97; 40.73)
Wald (joint) 512.9 *** 491.1 ***
 (0.000) (0.000)
Wald (time) 37.71 *** 35.57 ***
 (0.000) (0.000)
Sargan 322 269.6
 (0.265) (0.393)
m1 -6.791 *** -6.773 ***
 (0.000) (0.000)
m2 -0.883 -0.805
 (0.377) (0.421)

Independent Variables (3)

Tobins'q(-1) 0.542 ***
 (0.000)
Executive Ownership 0.032 *
 (0.056)
Executive Ownership (2) -0.002 *
 (0.060)
Executive Ownership (3) 2.16E-05 *
 (0.084)
Non-Executive Ownership -0.025
 (0.216)
Ratio (cum shares)

Non-Exec Ownership* 0.061
 Ratio(cum shares) (0.229)
Blockholding -0.005*
 (0.074)
Size -0.046
 (0.250)
Capital Expenditures 1.593 **
 (0.026)
R&D Expenditures 1.002
 (0.418)
Cash Flow 0.958 **
 (0.048)
Leverage 0.118
 (0.725)
Dividends 5.029 **
 (0.018)
Implied Turning Points (11.70; 42.82)
Wald (joint) 563.8 ***
 (0.000)
Wald (time) 37.48 ***
 (0.000)
Sargan 292.6
 (0.532)
m1 -6.801 ***
 (0.000)
m2 -0.825
 (0.378)

No. of observations: 6,340

*** Significant at the 0.01 level.

** Significant at the 0.05 level.

* Significant at the 0.10 level.

Table VII. Firm Performance and Non-Executive Directors' Ownership:
Robustness Tests of the Functional Form of the Relation

This table presents a robustness test of the estimated relation between
firm performance and ownership structure using GMM-SYS. In this table,
I specify several functional forms to further test for the presence
of a relation between non-executive ownership and performance. The
dependent variable is Tobins g defined as the ratio of book value of
total assets minus the book value of equity plus the market value of
equity to book value of assets. Executive Ownership is the total
percentage of shares owned by executive directors. Non-Executive
Ownership is the total percentage of shares owned by non-executive
directors. Dum. Non-Exec own>5% (15%, 250) is a dummy equal to 1 if
non-executive directors' ownership is greater than or equal to 5%
(15%, 25%) of shares. Blockholding is the total percentage of shares
held in positions of greater than 3% owned by external investors.
Ratio is the proportion of board members who are non-executives.
Size is the natural logarithm of total assets in 1991 prices. Capital
Expenditures is the ratio of total capital expenditures to total
assets. R&D Expenditures is the ratio of total research and development
expenditures to total assets. Cash Flow is the ratio of pre-tax profit
plus depreciation to total assets. Leverage is the ratio of total debt
to total assets. Dividends is the ratio of ordinary dividends net of
the Advance Corporation Tax to total assets. Turning Points are the
levels of executive ownership at which the effect of alignment
(entrenchment) starts to dominate. Time dummies were included in all
estimations. In all models, levels lagged at [t-3, t-4] of all the
regressors are used as instruments for the equations in first
differences. In the level equations, first differences lagged at [t-2]
are used as instruments. Regression coefficients are shown with
heteroskedasticity robust standard errors. P-values are provided in
parentheses. The Wald (joint) statistic tests the joint significance
of all regressors. Wald (time) statistic tests the joint significance
of the time dummies. The Sargan statistic tests for over-identifying
restrictions, and is asymptotically distributed as a [chi square]
under the null hypothesis of valid instruments. The m1 and m2
statistics test the absence of first-order and second-order correlation
in the residuals. They are asymptotically distributed as N (0,1) under
the null hypothesis of no serial correlation.

Independent Variables (1) (2)

Tobin's q (-1) 0.533 *** 0.527 ***
 (0.000) (0.000)
Executive Ownership 0.040* 0.0423**
 (0.053) (0.029)
Executive Ownership 2 -0.002* -0.002**
 (0.064) (0.040)
Executive Ownership 3 2.49E-05* 2.70E-05**
 (0.068) (0.044)
Non-Executive Ownership 0.012 0.0181
 (0.800) (0.379)
Non-Exec Own* Dum. -0.02
 Non-Exec Own>5% (0.666)
Non-Exec Own* Dum. -0.026
 Non-Exec Own>15% (0.000) (0.354)
Non-Exec Own* Dum.
 Non-Exec Own>25% (0.000) (0.000)
Ratio 0.853 *** 0.658**
 (0.003) (0.020)
Blockholding -0.005* -0.006**
 (0.073) (0.027)
Size -0.057 -0.047
 (0.209) (0.243)
Capital Expenditures 1.466** 1.499**
 (0.035) (0.037)
R&D Expenditures 1.725 1.121
 (0.187) (0.398)
Cash Flow 1.09** 1.168**
 (0.044) (0.037)
Leverage 0.036 -0.097
 (0.910) (0.767)
Dividends 4.827** 3.802**
 (0.020) (0.014)
Implied Turning Points (12.90; 41.11) (12.70; 40.91)
Wald (joint) 624.1 *** 533 ***
 (0.000) (0.000)
Wald (time) 27.69 *** 26.77 ***
 (0.002) (0.003)
Sargan 319.2 323.3
 (0.318) (0.263)
m1 -6.704 *** -6.660 ***
 (0.000) (0.000)
m2 -0.885 -0.956
 (0.376) (0.339)

Independent Variables (3)

Tobin's q (-1) 0.517 ***
 (0.000)
Executive Ownership 0.035*
 (0.068)
Executive Ownership 2 -0.002*
 (0.088)
Executive Ownership 3 2.10E-05*
 (0.071)
Non-Executive Ownership -0.018
 (0.388)
Non-Exec Own* Dum.
 Non-Exec Own>5%
Non-Exec Own* Dum.
 Non-Exec Own>15%
Non-Exec Own* Dum. 0.019
 Non-Exec Own>25% (0.348)
Ratio 0.792 ***
 (0.007)
Blockholding -0.007**
 (0.020)
Size -0.07
 (0.116)
Capital Expenditures 1.504**
 (0.037)
R&D Expenditures 1.297
 (0.288)
Cash Flow 1.245**
 (0.014)
Leverage 0.011
 (0.973)
Dividends 4.199**
 (0.014)
Implied Turning Points (13.30; 42.23)
Wald (joint) 584.7 ***
 (0.000)
Wald (time) 37.48 ***
 (0.000)
Sargan 317.2
 (0.347)
m1 -6.600*
 (0.000)
m2 -0.824
 (0.410)

No. of observations: 6,340

*** Significant at the 0.01 level.

** Significant at the 0.05 level.

* Significant at the 0.10 level.

Table VIII. Firm Performance and External Shareholders: Institutions
and Non-Institutions

This table presents a test of the estimated relation between firm
performance and ownership structure using GMM-SYS. The dependent
variable is Tobin's q defined as the ratio of book value of total
assets minus the book value of equity plus the market value of equity
to book value of assets. Executive Ownership is the total percentage
of shares owned by executive directors. Non-Executive Ownership is
the total percentage of shares owned by non-executive directors. Ratio
is the proportion of board members who are non-executives. The
interaction variable Non-Executive Ownership*Ratio tests for the
presence of an effect of Non-Executive Ownership conditional on the
level of the Ratio. Largest Non-Managerial Ownership is the largest
share position owned by a non-managerial shareholder. Institutional
Ownership is the percentage of shares owned in blocks by financial
institutions. Non-Institutional Ownership is the percentage of shares
owned in blocks by private individuals and other non-financial
companies. Largest Institutional Ownership is the largest percentage
share position owned by a financial institution. Largest Non-
Institutional Ownership is the largest percentage share position owned
by private individuals and other non-financial companies. Size is the
natural logarithm of total assets in 1991 prices. Capital Expenditures
is the ratio of total capital expenditures to total assets. R&D
Expenditures is the ratio of total research and development
expenditures to total assets. Cash Flow is the ratio of pre-tax profit
plus depreciation to total assets. Leverage is the ratio of total debt
to total assets. Dividends is the ratio of ordinary dividends net of
the Advance Corporation Tax to total assets. Turning Points are the
levels of executive ownership at which the effect of alignment
(entrenchment) starts to dominate. Time dummies were included in all
estimations. In the equations in first differences, levels lagged at
[t-3, t-4] of all the regressors are used as instruments. In the
level equations, first differences lagged at [t-2] are used as
instruments. Regression coefficients are shown with heteroskedasticity
robust standard errors. P-values are provided in parentheses. The Wald
(joint) statistic tests the joint significance of all regressors. Wald
(time) statistic tests the joint significance of the time dummies. The
Sargan statistic tests for over-identifying restrictions, and is
asymptotically distributed as a [chi square] under the null hypothesis
of valid instruments. The m1 and m2 statistics test the absence of
first-order and second-order correlation in the residuals. They are
asymptotically distributed as N (0, 1) under the null hypothesis of no
serial correlation.

Independent Variables (1) (2)

Tobin's q (-1) 0.542 *** 0.525 ***
 (0.000) (0.000)
Executive Ownership 0.0402 ** 0.0414 **
 (0.045) (0.017)
Executive Ownership (2) -0.002 * -0.002 **
 (0.059) (0.032)
Executive Ownership (3) 2.40E-05 * 2.46E-05 **
 (0.064) (0.044)
Non-Executive Ownership 0.0055 0.0034
 (0.759) (0.847)
Ratio 0.906 *** 0.967 ***
 (0.001) (0.002)
Non-Exec -0.0185 -0.0167
 Ownership*Ratio (0.642) (0.644)
Largest Non-Managerial -0.0054 *
 Ownership (0.081)
Institutional Ownership -0.0061 **
 (0.023)
Non-Institutional -0.0077 *
 Ownership (0.025)
Largest Institutional
 Ownership
Largest Non-
 Institutional Ownership
Size -0.02 -0.059
 (0.639) (0.168)
Capital Expenditures 1.652 ** 1.364 *
 (0.031) (0.082)
R&D Expenditures 1.991 1.756
 (0.175) (0.217)
Cash Flow 1.215 ** 1.211 **
 (0.025) (0.017)
Leverage 0.214 0.055
 (0.526) (0.854)
Dividends 5.095 ** 4.831 **
 (0.016) (0.012)
Implied Turning Points (13.50; 40.60) (13.39; 42.01)
Wald (joint) 556.5 *** 558 ***
 (0.000) (0.000)
Wald (time) 34.97 *** 42.56 ***
 (0.000) (0.000)
Sargan 326.6 329.4
 (0.343) (0.499)
m1 -6.889 *** -6.773 ***
 (0.000) (0.000)
m2 -0.849 -0.953
 (0.396) (0.341)

Independent Variables (3)

Tobin's q (-1) 0.547 ***
 (0.000)
Executive Ownership 0.0439 **
 (0.044)
Executive Ownership (2) -0.0021
 (0.055)
Executive Ownership (3) 2.52E-05 *
 (0.059)
Non-Executive Ownership 0.0043
 (0.801)
Ratio 0.899 ***
 (0.002)
Non-Exec -0.018
 Ownership*Ratio (0.623)
Largest Non-Managerial
 Ownership
Institutional Ownership

Non-Institutional
 Ownership
Largest Institutional -0.0057
 Ownership (0.133)
Largest Non- -0.0058 **
 Institutional Ownership (0.045)
Size -0.015
 (0.695)
Capital Expenditures 1.434 *
 (0.072)
R&D Expenditures 1.635
 (0.260)
Cash Flow 1.176 **
 (0.021)
Leverage 0.109
 (0.732)
Dividends 4.565 **
 (0.011)
Implied Turning Points (13.96; 40.63)
Wald (joint) 579.3 ***
 (0.000)
Wald (time) 36.12 ***
 (0.000)
Sargan 363.3
 (0.205)
m1 -6.84 ***
 (0.000)
m2 -0.876
 (0.381)

No. of observations: 6,340

*** Significant at the 0.01 level.

** Significant at the 0.05 level.

* Significant at the 0.10 level.
COPYRIGHT 2007 Financial Management Association
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2007 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Mura, Roberto
Publication:Financial Management
Article Type:Report
Geographic Code:4EUUK
Date:Sep 22, 2007
Words:17350
Previous Article:Lockup and voluntary earnings forecast disclosure in IPOs.
Next Article:What's in the news? Information content of S&P 500 additions.
Topics:

Terms of use | Privacy policy | Copyright © 2018 Farlex, Inc. | Feedback | For webmasters