The simultaneous determination of managerial ownership, corporate performance and financial analysts coverage in the United Kingdom

Transcription

1 The simultaneous determination of managerial ownership, corporate performance and financial analysts coverage in the United Kingdom By Patrick McColgan, Department of Accounting & Finance, University of Strathclyde, 100, Cathedral Street, Glasgow. G4 0LN. United Kingdom. Ph: (0) Fax: (0) Draft: 24 June 2001

2 The simultaneous determination of managerial ownership, corporate performance and financial analysts coverage in the United Kingdom Acknowledgements I would like to thank participants at the BAA Doctoral Colloquium 2001 for their helpful comments in the writing of this paper. The normal caveat applies.

3 The simultaneous determination of managerial ownership, corporate performance and financial analysts coverage in the United Kingdom Abstract Various monitoring and incentive mechanisms interact within the context of organizations to encourage management to maximise the value of their firm. Within the context of this paper I explore the interaction between insider ownership, financial analyst coverage and corporate value. I propose that any relationship between these variables is endogenous. However, due to managerial entrenchment and the diminishing increase in value of additional analyst coverage, it is found that such relationships are non-linear. Within the context of cubic and logarithmic functions I find that these three variables interact to jointly affect one another in a non-linear manner.

4 The simultaneous determination of managerial ownership, corporate performance and financial analysts coverage in the United Kingdom 1. Introduction Finance theory has developed to explore the role of equity ownership by inside company managers 1 in providing incentives to maximise the value of their company. Research exploring this issue documents both incentive and entrenchment effects that point to a non-linear relationship between insider ownership and corporate value. Jensen and Meckling (1976) propose that structures could be set up to monitor the activities of management, thus reducing the agency conflicts within organisations which ultimately lead to value losses for shareholders. Such monitoring may come from both internal and external sources. Insider ownership provides an internal monitoring device, providing management with a financial incentive to maximise the value of their company. External monitoring may come from market mechanisms [Fama (1980)], or also, coverage by financial analysts [Jensen and Meckling (1976)]. Kole (1995) argues that sample heterogeneity will determine the exact relationship between various firm characteristics. While competing theories may hypothesise how these variables should interact with one another, the true relationship is a cross-sectional phenomenon that must be estimated within the framework of data being considered, Hermalin and Weisbach (1991). The importance of each of these mechanisms in reducing monitoring costs and increasing company value is likely to be interdependent. Recent work by Chen and Steiner (2000) has suggested that these variables inside ownership, analyst coverage and corporate performance may be endogenously determined by one another. 1 The terms inside and managerial ownership are used interchangeably within this paper.

5 The next section provides a review of the relevant literature to the topic. Section three discusses features of the UK stock market which justify an examination of this market. Section four provides a discussion of the models of insider ownership and analyst coverage. A description of the data used and empirical results will be provided in section five and section six shall provide conclusions and discussion. 2. Review of Relevant Literature The literature within this section can be classified into three broad categories, that which focuses on ownership and corporate performance, that which deals with the link between analyst coverage and performance, and that which deals with ownership and analyst coverage. Finally, we can look at the brief literature which attempts to combine these theories. 2.1 Managerial Ownership and Corporate Value Jensen and Meckling (1976) predict that as managerial equity stakes increase, the incentive of management to consume private perquisites at the expense of investing in positive NPV projects will diminish. Managers receive the full amount of such private perquisites and bear only a fraction of the cost (being the fraction of equity they hold in the company). Demsetz and Lehn (1985) made an unsuccessful attempt at estimating a linear relation between managerial ownership and accounting performance. However, Fama and Jensen (1983) point to the problem of managerial entrenchment, suggesting both positive and negative effects arising from managerial ownership in companies. Entrenchment arises where share ownership provides management with the power to

6 insulate themselves from the pressures of external discipline and internal monitoring devices that may reduce the extent of agency conflicts within organisations. Morck et al. (1988) find a non-monotonic relationship between inside ownership and corporate value (as proxied by Tobin s Q) in a sample of 371 Fortune 500 companies. Specifically, they find that inside ownership has a positive effect on value up to 5% ownership and above 25% managerial holdings. However, between 5 and 25% they find negative corporate value, indicating the existence of managerial entrenchment. They argue that both entrenchment and a convergence of interests are present at all times, where the strength of the entrenchment effect is dominant at levels of ownership between 5 and 25%. McConnell and Servaes (1990) provide some support for Morck et al. (1988). In a more heterogeneous sample, they find a curvilinear relationship between ownership and value. Value increases originally with ownership by top officers of the company, but declines at higher levels due to the onset of managerial entrenchment. Hermalin and Weisbach (1991) find three turning points and 1, 5 and 20%. They argue that any relationship between ownership and corporate value is determined by the sample and is likely to be determined by cross-sectional characteristics of the firms used. In UK studies both Short and Keasey (1999) and Faccio and Lasfer (2000) document similar findings to those of Morck et al. (1988). Specifically they report breakpoints of 13 and 42% and 19.7 and 54.1% respectively. They argue that entrenchment arises at higher ownership levels due to institutional differences between the UK and US in terms of investor types, increased potential for informal blockholder collusion in the UK, and takeover related differences. 2 2 These differences will be discussed further in section 3.

7 Recent research in this area by Kole (1995) has built upon the finding of Demsetz and Lehn (1985) that ownership levels will be determined endogenously in equilibrium, where they find ownership to be inversely related to the volatility of a company s stock. Managers with larger shareholdings face difficulty in diversifying their personal portfolio risk, and will be reluctant to hold high stakes in high-risk companies. Kole (1995) develops this to argue for reverse causality, suggesting that managers in high performance firms are likely to prefer higher equity stakes, which may arise in part through stock based compensation schemes. Cho (1998) extends this to examine the interdependence of insider ownership, investment and corporate value. Initially he replicates the findings of Morck et al. (1988) for ownership and Q and documents similar findings. However, by applying simultaneous equations analysis, he finds that investment determines corporate value, but not insider ownership. Also, he finds that Q has a positive effect on inside ownership, confirming Kole s (1995) predictions. Himmelberg et al. (1999) provide further evidence of Cho s (1998) findings of endogenity. However, rather than reverse causality, they interpret such findings as evidence of companies choosing amongst various corporate governance devices based upon unobservable (to the econometrician) aspects of their contracting environment. 3 The above literature is largely inconclusive on the effects of inside ownership by company managers on corporate performance. While the theoretical arguments of Jensen and Meckling (1976) are certainly valid, empirically they have failed to hold up to the empirical testing by Morck et al. (1988) amongst others. Recent work by Cho (1998) and Himmelberg et al. (1999) would suggest more complex analysis than that provided by ordinary least squares regressions (OLS) is required.

8 2.2 Analyst Coverage and Corporate Performance Jensen and Meckling (1976) hypothesise that the monitoring activities of security analysts will reduce the scope for managerial opportunism within the firms they evaluate. To the extent that this is the case, such benefits will be reflected in a higher market capitalisation of such companies. Chung and Jo (1996) contend that any such relationship between analyst coverage and performance is likely to be endogenous. Analysts will be attracted to higher quality firms because they are easier for analysts to market. In turn, analyst coverage is likely to reduce the monitoring costs of such companies, which should be reflected in a higher market value. Jensen and Meckling (1976) hypothesise that analysts will generally be employed on behalf of financial institutions, investment banks, brokers, etc. Similarly to Chung and Jo (1996), Allen et al. (2000) present a model of dividends where institutional investors are attracted to high performance companies, paying high dividends, in a model of tax clienteles. They argue that such institutional investor s greater ability in monitoring management will then contribute to higher performance in such companies. 2.3 Managerial Ownership and Analysts Coverage Drawing upon the arguments of Jensen and Meckling (1976), the internal monitoring provided by managerial equity ownership and the external monitoring from security analysts would appear to be substitutes for one another. Himmelberg et 3 However, panel data approach and firm fixed effects employed by Himmelberg et al. (1999) have been subject to heavy criticism by Zhou (2001). This may call into question the findings of their paper.

9 al. (1999) argue that, in equilibrium, companies will utilise various monitoring mechanisms depending upon their contracting environment. 4 Drawing on this, Moyer et al. (1989) find that analyst coverage is negatively impacted by insider ownership. Chen and Steiner (2000) contend that the OLS regressions used may produce spurious results, where analyst coverage may be a determinant of managerial ownership, rather than the reverse. They also contend that Moyer et al. s (1989) linear specification ignores the diminishing marginal value of managerial ownership that should arise from managerial entrenchment. Similarly in the UK, both Faccio and Lasfer (2000) and Peasnell et al. (2000) document a U-shaped relationship between the percentage of managerial equity holdings and proportion of outside directors on the company s board of directors. They suggest that initially increased ownership creates a convergence of interests effect where there is less demand for outside directors. However, the entrenchment brought on at higher levels of ownership creates an internally generated demand for higher levels of non-executive directors to increase monitoring of top management. The findings of Chen and Steiner (2000), Faccio and Lasfer (2000) and Peasnell et al. (2000) suggest that simple assumption of a linear substitution effect will not be sufficient to properly estimate the exact nature of any causal relationship between managerial ownership and analyst coverage, and vice-versa. 2.4 The Joint Determination of Ownership, Analyst Coverage and Performance Cho (1998) and Himmelberg et al. (1999) find that managerial ownership is endogenously determined by corporate performance, while Chen and Steiner (2000) 4 For example, the size of larger companies generally acts as a deterrent to high managerial equity stakes [Morck et al. (1988), Cho (1998)], but larger companies are likely to attract a higher level of analyst coverage, providing them with an optimal level of monitoring.

10 find that these variables are interdependent. Also, Chung and Jo (1996) find that corporate performance and analyst coverage are interdependent. Chen and Steiner (2000) find that ownership, analyst coverage and performance are interdependent with one another. Employing the quadratic function for the ownership performance relationship proposed by Stulz (1988) and tested by McConnell and Servaes (1990), they find that increased ownership affects performance in this curvilinear manner. Higher corporate value linearly leads to higher managerial ownership. 5 Additionally, they find a diminishing substitution effect for the relationship between analyst coverage and managerial ownership. As ownership increases the level of analyst coverage declines at a diminishing rate. They present a U-shaped function for the relationship between the number of analysts and increasing levels of managerial ownership, similar to that presented by Peasnell et al. (2000) for the relationship between the demand for outside directors and managerial ownership. Finally, they find a diminishing increase in Tobin s Q as levels of analyst coverage increases. At the same time, higher Q s are linearly associated with higher analyst coverage. 3. UK Governance and Institutional Characteristics While providing an alternative framework for this analysis to the US markets analysed in previous studies, UK companies provide many similarities to US based companies. 5 Chen and Steiner find a turning point of 28.8% after which corporate value begins to decline. Theories of managerial incentives would suggest that this is a very low point and further testing may reveal that any relationship is better explained within a more complex model of the ownership corporate value relationship.

11 The London Stock Exchange (LSE) is one of the world s major stock markets, with roughly 2,500 domestic and international companies listed. Governance structures in the UK are largely similar to the US, in terms of corporate board structures, active takeover markets, and diffuse corporate ownership amongst different classes of shareholders. Conyon and Murphy (2000) propose that UK governance and market structures are largely similar to those in the US, in relation to market based contracting environments. This is in contrast to the relationship-based systems more prominent in German and Far-Eastern markets, where banks tend to contribute high levels of debt financing, take active roles in corporate governance and there is often a lack of discipline by external markets, Franks and Mayer (2000). Dahya et al. (2001) find that outsider-dominated corporate boards act to discipline management for poor performance in the same way as Weisbach (1988) reports in the US. In contrast, Kang and Shivdasani (1995) find no such relation in a sample of Japanese companies. Short and Keasey (1999) argue that institutional investors in the UK are likely to play a more prominent role in corporate governance than they do in the US. They claim that the geographical clustering of investment companies in London allows for more informal coalitions between block investors, increasing their ability to monitor company management. Conyon and Murphy (2000) find that executive compensation levels in the UK are lower than those received by the US peers. They find that CEO s in the US receive 46% higher cash compensation and 150% higher total compensation even after controlling for firm characteristics. 6 A lack of incentives provided by executive 6 These characteristics included size, industry, human capital, growth opportunities and performance.

12 compensation would suggest that insider ownership may take on greater importance in UK companies. In addition they find pay-performance sensitivities 7 in the US of 1.48%, compared to 0.25% in the UK, indicating that US executives receive a higher proportion of the increases in shareholder wealth they have helped to create. Their results also indicated that US CEO s on average hold 2.5% more of their company s stock than their UK counterparts. In contrast to the above, this would suggest lower incentives provided by stock ownership in the UK, but also, lower levels of managerial voting power. 8 In contrast to the US, Franks and Mayer (1996) find a lack of discipline provided by external takeover markets. They fail to find a relationship between poor prior performance and top management turnover following companies being acquired in hostile takeovers. Martin and McConnell (1991) find such a relationship in their US study. This points to comparisons between the UK and Germany, where Franks and Mayer (2000) discuss the lack of an external takeover market to discipline company management. In this setting, managers in UK companies may become more entrenched due to this lack of external discipline. 4. Models of Ownership, Corporate Value and Analysts Coverage The exact nature of the relationship between each of these variables is determined by heterogeneity within the sample firms, Hermalin and Weisbach (1991). It is argued that managerial entrenchment occurs from managerial power and the failure of both market and internal control mechanisms. Financial analyst coverage 7 Calculated using data on share ownership, options held and Long-Term Incentive Plans (LTIP) shares held. 8 Which is generally seen as the cause of managerial entrenchment.

13 represents an additional external monitoring device which limits the extent to which managers may become entrenched. Therefore, the true nature of insider ownership s affect on corporate value becomes ambiguous once again. 4.1 Managerial Ownership and Corporate Performance Management are faced with both positive and negative incentives to maximise the value of their firms. Incentives may arise from managerial interests converging with those of their shareholders, which should arise from financial incentives generated by their equity stakes. Additionally, external monitoring provided by financial analyst coverage should bring attention to the actions of management, and encourage them to take value maximising decisions. Finally, incentives may come from external market discipline provided by product markets, managerial labour markets and takeover markets. The strength of each of these factors may be viewed as a function of managerial ownership. 9 Once the additional monitoring from financial analyst coverage is taken into account, the managerial entrenchment effect diminishes. Analyst monitoring provides a further incentive for managers to maximise firm value. 4.2 Financial Analyst Coverage and Corporate Value Theory predicts that higher levels of analyst coverage will improve the monitoring of the firm, thus lowering agency costs, and leading to higher corporate values. Therefore a positive relationship between corporate value and the level of 9 The arguments of Jensen and Meckling (1976) and Stulz (1988) and the findings of Chen and Steiner (2000), Dahya et al. (2001), Martin and McConnell (1991), Weisbach (1988) all suggest that financial

14 analysts coverage is predicted. However, since the additional contribution to value of one additional analyst is likely to be diminishing, we predict that this relationship will be concave. The above relationship is likely to be cyclical. While higher analyst s coverage should lead to higher corporate value, analysts should be attracted to higher performance companies in the first instance. Analysts may find higher performance companies easier to market, and as such, will be able to attain greater fees from devoting their research time to these companies. As such we predict a positive relationship between corporate value and analysts coverage. 4.3 Managerial Ownership and Financial Analyst Coverage The nature of the relationship between insider ownership and coverage by financial analysts is also uncertain. Theory predicts that these devices may act as substitutes in the monitoring of managerial decision making. However, simple linear models fail to capture the full nature of possible interactions once the problem of managerial entrenchment is taken into account. Therefore, it is predicted that increasing levels of managerial ownership have a non-linear effect on analyst coverage. The findings of Chen and Steiner (2000) and Peasnell et al. (2000) suggest that convergence of interests at initially increasing levels of ownership will result in a decline in analysts coverage. However, at higher levels of ownership, entrenchment may lead to higher levels of analyst coverage to control for problems arising through managerial entrenchment. Finally, following Jensen and Meckling s (1976) arguments we predict that analyst coverage is negatively related to the level of managerial ownership, consistent incentives and the effectiveness of both external and internal monitoring are functions of the level of

15 with a substitution effect. However, the marginal value of additional analyst coverage is likely to be diminishing, and as such, we predict this relationship to be concave. 4.4 Control Variables In addition to insider ownership, corporate value and analysts coverage, a number of control variables are used. With respect to analyst coverage, we include the market value of equity to control for analysts being attracted to larger companies whose stocks are traded more frequently. Research and development to assets employed is included to control for growth opportunities. To the extent that R&D intensive companies have high growth prospects, and therefore low free-cash flow [Jensen (1986)], less monitoring from financial analysts is needed. However, there may alternatively be a positive relationship between these variables, since higher R&D firms may be perceived as being of higher quality. Consistent with Denis et al. (1997) amongst others, we infer R&D to assets based upon the FT-SE actuaries industry classification system to prevent any bias on our results arising from nonreporting companies. Higher risk companies (as modelled by standard deviation of stock returns) may attract analyst s coverage, since analyst s information is likely to be of greater value to such companies. Alternatively, analysts may avoid such companies if they are unable to generate sufficient demand for information high-risk companies. Finally, we use dummy variables based on FT-SE actuaries industry classifications to control for such industry effects. For the managerial ownership equation, we again include the market value of equity to control for the fact that managers find it difficult to own large stakes in larger companies. However, Conyon and Murphy (2000) find a positive relationship managerial ownership.

16 between firm size and managerial ownership within US companies. While apparently contradictory to wealth constraints forbidding large stakes in larger companies, such a finding may be explained if good past performance has lead to increased firm size, whilst at the same time has rewarded management through higher levels of equitybased compensation. A leverage ratio is included to control for possible relationships between debt and ownership as predicted by Jensen and Meckling (1976) and Jensen (1986) 10. The standard deviation of stock returns is included to control for the finding of Demsetz and Lehn (1985) that ownership is inversely related to stock price risk. Following the arguments of Cho (1998), I include a liquidity variable since liquidity may interact with investment to affect insider ownership. Finally, I include an R&D variable to control for the effect of investment opportunities on managerial ownership. Free cash-flow arguments would predict that high levels of investment opportunities reduce the need for higher managerial equity stakes to reduce the scope for perquisite consumption. Again industry dummies are used to control for cross-sectional differences in managerial ownership arising from industry effects. Finally, previous studies have found that Tobin s Q is inversely related to the value of the company s assets, and the log of this variable is included in the analysis. The arguments of Jensen (1986) amongst others predict a positive relationship between leverage and corporate value. R&D to assets is again included as a measure of firms quality and growth prospects, which should lead to higher corporate values. Finally, again we control for industry effects. This leads to the following functional relationships: 10 Jensen and Meckling (1976) argue that higher levels of debt reduces the amount of equity within the firm, thus enabling management to hold larger stakes. Jensen (1986) predicts that higher levels of debt serve as a monitoring substitute for inside equity. Alternatively, theories of managerial risk aversion predict a negative relationship between leverage and managerial ownership.

17 Analysts Coverage = f(insider ownership, corporate value, market value of common equity, R&D, stock price risk, industry) [1] Insider Ownership = g(analysts coverage, corporate value, market value of firm s common equity, leverage, stock price risk, R&D, liquidity, industry) [2] Corporate Value = h(analysts coverage, insider ownership, leverage, asset size, R&D, industry) [3] 5. Data and Empirical Testing 5.1 Data collection Data on ownership by inside managers and external blockholders for 1995 is gathered from the Macmillan Stock Exchange Yearbooks for 1996 and The Yearbook provides summary information similar to that found in FT Extel cards such as director information, summary accounts, dividend information, details of substantial and managerial shareholdings, and legal details. 11 The sample is limited to non-financial companies. Data on financial analyst coverage is collected from IBES. I do not restrict this analysis to companies that are covered by IBES in the year to be analysed. Instead, the sample also includes companies that IBES reports on outside of the sample period of Recently listed, merged or acquired firms are not included. 12 This has the effect of producing a minimum value of 0 for the number of analysts, rather than 1, which is the case in the study of Chen and Steiner (2000). Companies which IBES has no information on in any year of the sample 1997 CD are excluded. The result of this is that the analyst variable becomes log of number of analysts plus one to allow for a diminishing function.

18 Finally, the remaining data for the construction of other variables is gathered from Datastream. This leads to a final sample of 627 companies with which to conduct the analysis. 5.2 Variables Tobin s Q ratio is estimated using the formula below: MVEQ + PREF + DEBT Tobin' s Q = [4] BVASSETS Where MVEQ = the year-end market value of the firm s common stock; PREF = the year-end book value of the firm s preference shares (preferred stock); DEBT = the year-end book value of the firm s total debt; and BVASSETS = the total assets employed by the firm, assets minus current liabilities. This measure is consistent with the modified version suggested by Chung and Pruitt (1994). They find that 96.6% of the variability in the commonly used Lindenberg and Ross (1981) method can be explained by their formula. This method is also convenient in that it avoids the data availability problems inherent in the more rigorous Lindenberg and Ross (1981) and Lewellen and Badrinath (1997) algorithms that require several year s past data. In such a case, this analysis would otherwise be restricted to more mature firms, which Jensen (1986) categorises as those with the greatest agency problems. Differences in financial reporting standards between the UK and US also make Q problematic for UK studies. UK companies have generally had greater

19 freedom in the revaluation of their fixed assets than their US counterparts. The rigorous Lewellen and Badrinath (1997) and Lindenberg and Ross (1981) algorithms both have an implicit assumption that no revaluations have occurred. Where assets have previously been revalued, further inflationary adjustments to the value of such assets would overstate the value of the firm s assets. Doing so produces a downward bias in measures of Q. While Datastream does provide separate historical cost figures for fixed assets, these are only available for roughly half of the companies in this sample. Excluding firms that don t report this variable may result in a bias for larger firms that may be required to report this for international exchange listing requirements. Book values of preferred stock and long-term debt are included, rather than market values suggested by Lindenberg and Ross (1981) and Lewellen and Badrinath (1997). The UK market for trading corporate debt is largely inactive in comparison to the US market. This creates difficulty is obtaining market values, even using relatively simple bond ratings as suggested by Lindenberg and Ross (1981). Preference shares account for a tiny proportion of the traded capital in the UK, 13 as such, we believe it is unnecessary to update this for market values. 5.3 Descriptive Statistics Table 1 provides summary statistics for the sample as a whole. The mean level of managerial ownership is 12.95%, which is largely consistent with that reported in Cho (1998) and Chen and Steiner (2000), but slightly lower than Faccio and Lasfer who report 16.14% for mean ownership. Tobin s Q is higher than reported in other studies, with a mean value of 2.1 and has a standard deviation of The % of the total capital of the firms included within this sample.

20 greater heterogeneity within our sample is likely to be the main cause for this result. Additionally, UK firms tend to be smaller than their US counterparts, and Q is negatively related to firm size [Morck et al. (1988), Cho (1998)]. The mean blockholder ownership is 34.04%, which is largely similar to the findings of McConnell and Servaes. 14 R&D to assets employed is 4.79%, which is approximately equal to that reported by Cho (1998). Our sample captures a full range of firm sizes, with equity values ranging from 1.05million to 12.2billion, where the average firm has a market capitalization of 297million. Table 2 provides a breakdown of the sample by various levels of managerial ownership. It can be seen that most of the sample (61.5%) lies in the between 0 and 10% ownership categorisation. However, 11% of the sample do have ownership of 40% or greater. Analyst coverage is relatively high in the first ownership decile, with a mean (median) of 6.12 (5). At ownership above 60%, analyst coverage is very low. Unsurprisingly, block ownership decreases as inside ownership increases. Finally, there is an apparently inverse relationship between levels of insider ownership and both assets employed and the market value of equity. This is consistent with managerial wealth constraints preventing high ownership stakes in larger companies. This table also points to a non-linear relationship between managerial ownership and Tobin s Q. However, the exact nature of this relationship is apparently ambiguous. Table 3 provides a breakdown of the sample by levels of analyst coverage. The distribution is initially relatively even, however, a slight skewness is apparent at the highest levels. It is immediately apparent that there is a positive relationship between Tobin s Q and analyst coverage. A negative relationship between analyst coverage and insider ownership is also apparent.

21 5.4 Results The results of this analysis are presented in table 4. Firstly we find that analyst coverage is non-linearly related to levels of insider ownership. Specifically, the cubic function presents one turning point and a point of inflection. However, it can be illustrated graphically that the point of inflection is before zero and these findings essentially reduce to a U-shaped function. This is consistent with the findings of Chen and Steiner, Faccio and Lasfer (2000) and Peasnell et al. (2000). Higher managerial incentives at initially increasing levels of ownership reduce the need for outside monitoring from analysts. However, the managerial entrenchment brought on by higher levels of managerial control leads to an increased need for analyst monitoring at higher levels of ownership. Additionally, Tobin s Q is positively related to analysts coverage. This is consistent with Chung and Jo (1996) and Chen and Steiner (2000). Such findings reflect that fact that analysts are attracted to high performance companies since they will be easier for the analysts to market and justify their research. All of the control variables are significant for at the last the 10% level. Analyst coverage is positively related to size as predicted. It is also negatively related to R&D intensive firms. This may jointly reflect that such companies will tend to be smaller and to the extent that the scope for managerial opportunism is smaller in firms with high growth prospects, there is less need for analyst monitoring. Finally, analysts appear to shy away from higher risk companies. Such firms may be more difficult for analyst to market. Significantly, both the R 2 and the F-statistic are very 14 However, US firms must only disclose shareholders with a stake of greater than 5%, whereas UK firms must disclose those with a stake of greater than 3%.

22 high indicating that the joint hypothesis is significant and the model has high explanatory power. The second column of table 4 presents the results of the insider ownership function. As predicted, insider ownership is negatively related to financial analysts coverage within the diminishing functional relationship. This is consistent with a diminishing substitutional effect as proposed by Chen and Steiner (2000). At higher levels of insider analyst coverage there is less need for managerial incentives provided by higher levels of ownership. Consistent with Cho (1998) and Chen and Steiner (2000), ownership is positively related to corporate value, indicating that managers prefer to hold higher equity stakes in high performance companies. The control variables indicate that ownership is positively related to firm size. While consistent with Conyon and Murphy (2000) for US executives this is very surprising. Further testing reveals that this result is entirely driven by the inclusion of the analyst coverage variable and it appears that these two variables interact in some unknown manner to influence managerial ownership. The control variables indicate that insider ownership is positively related to liquidity and negatively related to leverage, stock price risk and R&D. Managerial risk aversion predicts that insider holdings will be negatively related to stock price risk and leverage. Debt may also be regarded as a substitute for insider ownership. The negative relationship between R&D indicated that lower ownership levels are required to control agency problems in firms with high growth opportunities. Again the R 2 and F-statistic indicate that the model has high explanatory power. Finally, the corporate value equation indicated that analyst coverage has a positive effect on Tobin s Q, within the diminishing marginal value function proposed. The non-linear relationship between insider ownership and corporate value is captured within a cubic model consistent with Morck et al. (1988), Short and

23 Keasey (1999) and Faccio and Lasfer (2000). Specifically corporate value increase with managerial ownership but begins to decline at higher levels of ownership. Finally, corporate value increases at very high levels of managerial ownership. These findings are consistent with both a convergence of interests and managerial entrenchment effects at increasing levels of managerial ownership. As predicted, growth opportunities (as proxied by R&D intensity) are positively related to Tobin s Q. Also, Q is positively related to corporate value as predicted by Jensen (1986) and asset size is negatively related to Q, which is consistent with Morck et al. (1988) amongst others. While the R 2 declines in comparison to the previous functions it is still reasonably high in comparison to at least some of the previous research of this type. Despite this, the F-statistic is again highly significant. 6. Conclusions The above analysis would indicate the within the framework proposed corporate value, insider ownership and coverage by financial analysts are jointly determined. The non-linear framework proposed accounts for both the effects of managerial entrenchment and the diminishing value of additional analysts coverage in reducing the agency problems inherent within an organisation. It is not proposed that this paper offers a definitive conclusion upon the way in which these three variables will interact within any given company. Rather, it is proposed that what is documented here is a model of the complex relationship that exists within the sample used. Hermalin and Weisbach (1991) argue that any relation between insider ownership and corporate value is likely to be a cross-sectional

24 phenomenon, and I would expand this to include the role of financial analysts coverage also. Kole (1995), amongst others, points to the importance of sample heterogeneity in estimating the exact form of these relationships. Future research in this area should seek to explore both whether other functional forms can explain the data and what it is within the data that drives the choice of a particular function. With respect to what can be concluded from this paper, I propose the following. It is evident that analysts are attracted to high performance companies and, in turn, analyst coverage contributes to the higher valuation of such companies. Additionally, corporate value is non-linearly related to insider ownership, consistent with both convergence of interests and entrenchment effects. At the same time, managers prefer to hold higher equity stakes in high performance companies. Finally, analyst coverage is non-linearly related to insider ownership, perhaps due to the onset of managerial entrenchment and insider ownership is negatively related to analyst coverage, where this effect is attributable to a diminishing substitution effect of analyst coverage on managerial ownership.

25 References Allen, F., A.E. Bernardo and I. Welch. (2000), A Theory of Dividends Based on Tax Clienteles Journal of Finance, forthcoming. Brennan, M., and P. Hughes. (1991), Stock prices and the supply of information. Journal of Finance, Vol. 46, pp Chen, C.R., and T.L. Steiner, Tobin s Q, managerial ownership, and analyst coverage. Journal of Economics and Business, Vol. 52, pp Cho, M.H., (1998). Ownership Structure, Investment, and the Corporate Value: an empirical analysis. Journal of Financial Economics, Vol. 47, pp Chung, K.H., and H. Jo. (1996), The impact of security analysts monitoring and marketing functions on the market value of firms. Journal of Financial and Quantitative Analysis, Vol. 31, pp Chung, K.H., and S.W. Pruitt, (1994), A Simple approximation of Tobin s q. Financial Management, Vol. 23, pp Conyon, M.J. and K.J. Murphy. (2000), The Prince and the Pauper? CEO Pay in the US and UK, Economic Journal, forthcoming. Demsetz, H., and K. Lehn, (1985), The Structure of Corporate Ownership: Causes and Consequences. Journal of Political Economy, Vol. 93, pp Denis, D.J., D.K. Denis and A. Sarin, (1997), Agency problems, Equity Ownership and Corporate Diversification. Journal of Finance, Vol. 52, pp Dahya, J., J.J. McConnell, and N.G. Travlos, (2001), The Cadbury Committee, Corporate Performance and Top Management Turnover. Journal of Finance, forthcoming. Faccio, M., and M.A. Lasfer, (2000), Managerial Ownership, Board Structure and Firm Value: The UK Evidence Working Paper. Fama, E.F., (1980), Agency Problems and the Theory of the Firm. Journal of Political Economy, Vol. 88, No. 2, pp Fama, E.F., and M.C. Jensen, (1983), Separation of ownership and control Journal of Law and Economics, Vol. 88, No. 2, pp Franks, J., and C. Mayer, (1996), Hostile takeovers and the correction of management failure. Journal of Financial Economics, Vol. 40, pp Franks, J., and C. Mayer, (2000), Ownership and Control in German Corporations. Working Paper Hermalin, B.E. and M. S. Weisbach, (1991), The effects of board composition and direct incentives on firm performance. Financial Management, Vol. 20, pp

26 Himmelberg, C.P., R.G. Hubbard, and D. Palia, (1999), Understanding the determinants of Ownership and the link between Ownership and Performance. Journal of Financial Economics, Vol. 53, pp Jensen, M.C., (1986), Agency Costs of Free Cash Flow, Corporate Finance and Takeovers. American Economic Review, Vol. 76, No. 2, pp Jensen, M.C., and W.H. Meckling, (1976), Theory of the Firm: Managerial Behavior, Agency Costs and Ownership Structure. Journal of Financial Economics, Vol. 3, No. 4, pp Kang, J., and A. Shivdasani, (1995), Firm Performance, Corporate Governance, and Top Executive Turnover in Japan. Journal of Financial Economics, Vol. 38, pp Kole, S., (1995), Measuring Managerial Equity Ownership: a comparison of sources of Ownership Data. Journal of Corporate Finance, Vol. 1, pp Lewellen, W.G., and S.G. Badrinath, (1997), On the Measurement of Tobin s q. Journal of Financial Economics, Vol. 44, pp Lindenberg, E., and S. Ross, (1981), Tobin s q ratio and the industrial organization. Journal of Business, Vol. 54, pp Martin, K.J., and J.J. McConnell, (1991), Corporate Performance, Corporate Takeovers, and Management Turnover. Journal of Finance, Vol. 46, No. 2, pp McConnell, J.J., and H. Servaes, (1990), Additional Evidence on Equity Ownership and Corporate Value. Journal of Financial Economics, Vol. 27, pp Morck, R., A. Shleifer, and R.W. Vishny, (1988), Management Ownership and Market Valuation: An Empirical Analysis. Journal of Financial Economics, Vol. 20, No.1-2, pp Moyer, C.R., R.E. Chatfield and P.M. Sisneros, (1989), Security analysts monitoring activity: Agency costs and information demands. Journal of Financial and Quantitative Analysis, Vol. 24, pp Peasnell, K.V., P.F. Pope and S. Young. (2000), Managerial Equity Ownership and the Demand for Outside Directors. Working Paper. Short, H., and K. Keasey, (1999), Managerial Ownership and the Performance of Firms: Evidence from the UK. Journal of Corporate Finance, Vol. 5, pp Stulz, R.M. (1988), Managerial Control of Voting Rights: Financing Policies and the Market for Corporate Control. Journal of Financial Economics, Vol. 20, No. 1, pp

27 Weisbach, M.S., (1988), Outside Directors and CEO Turnover. Journal of Financial Economics, Vol. 20, No. 1-2, pp White, H., (1980), A hetroskedasticity-consistent covariance matrix estimator and a direct test for hetroskedasticity. Econometrica, Vol. 48, pp Zhou, X., (2001), Understanding the determinants of Ownership and the link between Ownership and Performance. Journal of Financial Economics, forthcoming.

28 Table 1 Descriptive Statistics Variable Mean Median Std. Dev. Minimum Maximum Insider Holdings 12.95% 4.50% 17.79% 0.00% 79.90% Blockholder Holdings 34.04% 34.10% 19.03% 0.00% 92.3% Largest Stakeholder 14.46% 11.7% 11.95% 0.00% 89.9% No. of Analysts Assets Employed m m m , m R&D Intensity Tobin s Q Debt/Assets Employed MVEQ m m m m 12, m Annual SD 22.59% 19.28% 13.50% 3.87% 92.80% Note Insider holdings is the percentage of the company s total equity capital that is held by the company s directors. Blockholder ownership refers to the percentage of the company s capital that is held by outside investors with a stake of greater than 3%, as disclosed in the annual report. Largest stakeholder represents the stake of the largest individual blockholder with an equity stake of greater than 3%. The number of financial analysts reporting on the firm is taken from 1997 IBES tapes. Data on assets employed, research and development, debt equity market values and standard deviations are taken from Datastream. Where no research and development to assets employed ratios are available, we follow Denis et al. (1997) and infer these variables based on FT-SE actuaries industry classifications. Tobin s Q is calculated as the market value of equity, plus the book values of debt and preferred equity divided by the book value of assets minus current liabilities. Shareholdings information is taken from the MacMillan London Stock Exchange Yearbooks for 1996 and All data are for non-financial companies quoted on the London Stock Exchange in 1995.

29 Table 2 Breakdown of Sample by Insider Ownership Insider Ownership 0 I/O < 10% 10 I/O < 20% 20 I/O < 30% 30 I/O < 40% 40 I/O < 50% 50 I/O < 60% 60 I/O < 70% 70 I/O < 100% Number of Firms Tobin s Q (1.7208) (1.9465) (1.5482) (1.6705) (1.6886) (2.4178) (1.6124) (1.2347) No. of Analysts (5) (3) (2) (3) (2) (2) (0.5) 1 (0) Blockholder Ownership 38.23% (37.70%) 34.13% (33.90%) 34.50% (33.00%) 23.14% (24.80%) 20.84% (21.30%) 12.24% (11.60%) 12.36% (12.80%) 4.66% (5.20%) Assets Employed m ( m) ( m) m ( m) m ( ) m ( m) m ( m) m ( m) m ( m) Market Value of Equity m ( m) m ( m) m ( m) m ( m) m ( m) m ( m) m ( m) m ( m) Note Table reports mean (median) values of each variable for different levels of insider ownership. Insider holdings is the percentage of the company s total equity capital that is held by the company s directors. Blockholder ownership refers to the percentage of the company s capital that is held by outside investors with a stake of greater than 3%, as disclosed in the annual report. Largest stakeholder represents the stake of the largest individual blockholder with an equity stake of greater than 3%. The number of financial analysts reporting on the firm is taken from 1997 IBES tapes. Data on assets employed and equity market values Datastream. Tobin s Q is calculated as the market value of equity, plus the book values of debt and preferred equity divided by the book value of assets minus current liabilities. Shareholdings information is taken from the MacMillan London Stock Exchange Yearbooks for 1996 and All data are for non-financial companies quoted on the London Stock Exchange in 1995.

30 Table 3 Breakdown of Sample by Analyst Coverage No. of Analysts No. of Firms Tobin s Q (1.3781) (1.3781) (1.3858) (2.0983) (2.0100) (1.8644) (2.0299) (1.8500) (2.1006) (2.2303) (1.5732) (1.8490) (1.5469) Insider Ownership 19.42% (10.00%) 19.90% (20.10%) 17.55% (11.60%) 16.46% (7.90%) 12.91% (8.00%) 8.65% (3.55%) 10.42% (2.80%) 10.20% (3.60%) 4.56% (0.00%) 3.43% (0.00%) 4.99% (0.00%) 0.46% (0.00%) 0.00% (0.00%) Blockholder Ownership 37.96% (37.00%) 36.93% (37.65%) 38.42% (39.60%) 36.36% (36.40%) 35.00% (34.50%) 38.98% (37.35%) 31.25% (32.5%) 30.47% (32.50%) 31.64% (31.60%) 31.18% (30.40%) 22.86% (19.35%) 19.29% (14.20%) 22.03% (22.00%) Assets Employed m ( m) m ( m) m ( m) m ( m) m ( m) m ( m) m ( m) m ( m) m ( m) m ( m) m ( m) 1, m ( 1, m) 2, m ( 2, m) Market Value of Equity m ( m) m ( m) m ( m) m ( m) m ( m) m ( m) m ( m) m ( m) m ( m) m ( m) m ( m) 2, m ( 2, m) 4, m ( 4, m) Note Table reports mean (median) values of each variable for different levels of financial analysts coverage. Insider holdings is the percentage of the company s total equity capital that is held by the company s directors. Blockholder ownership refers to the percentage of the company s capital that is held by outside investors with a stake of greater than 3%, as disclosed in the annual report. Largest stakeholder represents the stake of the largest individual blockholder with an equity stake of greater than 3%. The number of financial analysts reporting on the firm is taken from 1997 IBES tapes. Data on assets employed and equity market values Datastream. Tobin s Q is calculated as the market value of equity, plus the book values of debt and preferred equity divided by the book value of assets minus current liabilities. Shareholdings information is taken from the MacMillan London Stock Exchange Yearbooks for 1996 and All data are for non-financial companies quoted on the London Stock Exchange in 1995.

31 Table 4 Joint determination of analyst coverage, insider ownership and corporate value Variable Analyst Coverage Insider Ownership Corporate Value Analyst Coverage (-39.05) (8.04) IO (-2.00) (3.63) IO (-2.32) (-2.67) IO (2.82) (2.47) Tobin s Q (9.84) (5.83) MVEQ 9.8 x 10 8 (1.73) 2.05 x 10 2 (33.40) R&D (-7.99) (-11.16) (8.70) SD (-12.31) (-22.79) Liquidity (6.09) Leverage (-3.877) (2.72) Asset Size (-2.93) Industry Dummies Yes Yes Yes Adj. R F-statistic Note Table reports from a simultaneous equations analysis of financial analysts coverage, insider ownership and corporate value using two stage least squares to estimate the following functional relationships: Analysts Coverage = f(insider ownership, corporate value, market value of common equity, R&D, stock price risk, industry) Insider Ownership = g(analysts coverage, corporate value, market value of firm s common equity, leverage, stock price risk, R&D, liquidity, industry) Corporate Value = h(analysts coverage, insider ownership, leverage, asset size, R&D, industry) Insider holdings are the percentage of the company s total equity capital that is held by the company s directors. The number of financial analysts reporting on the firm is taken from 1997 IBES tapes, where the variable used is the log of analysts plus one. Tobin s Q is calculated as the market value of equity, plus the book values of debt and preferred equity divided by the book value of assets minus current liabilities. MVEQ is the market value of common equity (in thousands), R&D is the ratio of research and development expenditures to assets employed, where missing data is inferred from FTSE industry classifications. SD is the annual standard deviation of a company s stock returns, liquidity is the ratio of after tax profits plus depreciation to assets employed and leverage is the ratio of long-term debt to assets employed. Shareholdings information is taken from the MacMillan London Stock Exchange Yearbooks for 1996 and All data are for non-financial companies quoted on the London Stock Exchange in White (1980) hetroskedasticity controlled t-statistics in parentheses.