Ownership Structure, Managerial Behavior and Corporate Value

Journal of Corporate Finance 11 (2005) 645 – 660 www. elsevier. com/locate/econbase Ownership structure, managerial behavior and corporate value J. R. Daviesa, David Hillierb,T, Patrick McColganc a University of Strathclyde, UK b University of Leeds, UK c University of Aberdeen, UK Received 21 November 2002; accepted 6 July 2004 Available online 20 April 2005 Abstract The nonlinear relationship between corporate value and managerial ownership is well documented. This has been attributed to the onset of managerial entrenchment, which results in a decrease of corporate value for increasing levels of managerial holdings. We propose a new structure for this relationship that accounts for the effect of conflicting managerial incentives, and external and internal disciplinary monitoring mechanisms. Using this specification as the basis for our analysis, we provide evidence that the managerial ownership–corporate value relationship is co-deterministic. This finding is at odds with recent work which reports that corporate value determines managerial ownership but not vice-versa. D 2005 Elsevier B. V. All rights reserved. JEL classification: G32 Keywords: Ownership structure; Capital expenditure; Corporate value; Tobin’s Q 1. Introduction In a market without agency problems, corporate managers will choose investments that maximise the wealth of shareholders. In practice, competing objectives which are incompatible with the shareholder wealth-maximising paradigm may also be pursued. T Corresponding author. Leeds University Business School, University of Leeds, Maurice Keyworth Building Leeds, LS2 9JT, UK. Tel. : +44 113 3434359; fax: +44 113 3434459. E-mail address: d. j. [email  protected] c. uk (D. Hillier). 0929-1199/$ – see front matter D 2005 Elsevier B. V. All rights reserved. doi:10. 1016/j. jcorpfin. 2004. 07. 001 646 J. R. Davies et al. / Journal of Corporate Finance 11 (2005) 645–660 Following Jensen and Meckling (1976), a large literature has developed that examines how managerial behavior impacts upon firm performance. A vibrant strand of this literature concerns the relationship between managerial ownership levels, the direct investment decisions made by management and the inherent value of the firm, as proxied by Tobin’s Q ratio. Morck et al. 1988), McConnell and Servaes (1990), and Hermalin and Weisbach (1991) provide evidence of a significant nonlinear relationship between corporate value and managerial ownership. Specifically, value increases with managerial holdings for low levels of ownership. At some level, managers become entrenched within the firm resulting in a decrease in firm value. However, whereas Morck et al. (1988) and Hermalin and Weisbach (1991) document further changes in the corporate value–managerial holdings relationship at high levels of equity ownership, McConnell and Servaes (1990) report no such change. Recent work has built upon the findings of Demsetz and Lehn (1985) who argue that levels of managerial ownership will be determined endogenously in equilibrium. Moreover, Cho (1998) and Himmelberg et al. (1999) have shed doubt upon the earlier findings of Morck et al. (1988) and McConnell and Servaes (1990) by controlling for the effects of endogeneity and unobservable (to the econometrician) firm characteristics in their analysis. After controlling for the effects of endogeneity in the corporate value– managerial holdings relationship, they showed that managerial ownership had little or no effect on corporate value and investment. Short and Keasey (1999) and Faccio and Lasfer (1999) utilize a cubic specification to model the corporate value–managerial holdings relationship and both report a significant nonlinear functional form, similar to Morck et al. (1988), for British companies. However, neither study fully examines the misspecifying impact of endogeneity on their results. In this paper, we propose a new structure to the managerial ownership–corporate value relationship which captures a more complex characterisation of the evolving behavior of managers. We argue that at high levels of managerial ownership when external market discipline becomes neffective, there will be a resurgence of entrenchment behavior. With equity holdings around 50%, managers will have implicit control of their company, but still do not have objectives completely aligned to external shareholders. Only at very high levels of managerial holdings are incentives akin to other shareholders. When this model is applied to a l arge sample of firms incorporated in the UK, managerial ownership is seen to have a significant impact on corporate value. This relationship is endogenous, and consistent with Cho (1998) and Himmelberg et al. (1999), corporate value has a corresponding effect on managerial holdings. We also find that although ownership levels are affected by firm level investment, there is no evidence of the reverse occurring. In the next section we outline our model of the managerial ownership–corporate value relationship. We present empirical results in Section 3 and conclude in Section 4. 2. The model In this section, we propose an alternative structure to the managerial holdings–corporate value relationship and argue that the cubic, or simpler representations, used in earlier J. R. Davies et al. / Journal of Corporate Finance 11 (2005) 645–660 647 studies1 are unnecessarily restrictive and misspecified. The model that is presented here captures further nonlinearities in this relationship at high levels of managerial holdings and has a quintic specification. Management is faced with both negative and positive incentives to ensure that they follow objectives which maximise shareholder wealth. The effectiveness of these incentives is potentially a function of the level of managerial ownership in the firm. We view the propensity of management to maximise shareholder wealth to be a function of three unobserved factors: external market discipline, even if it is weak, internal controls and convergence of interests. Moreover, the strength of each factor can be viewed as a function of the level of managerial ownership in the firm. 2 2. 1. Low levels of managerial ownership For low levels of managerial ownership, external discipline and internal controls or incentives will dominate behavior (see Fama, 1980; Hart, 1983; Jensen and Ruback, 1983). Empirically, Morck et al. (1988), McConnell and Servaes (1990) and Hermalin and Weisbach (1991) report results consistent with this behavior for the relationship between managerial holdings and corporate value. However, there is also the possibility that lower levels of ownership within this range have endogenously arisen from performance related compensation packages, such as stock options and stock grants rather than increased ownership in itself leading to higher Q ratios. 2. 2. Intermediate levels of managerial ownership At intermediate levels of managerial ownership, management interests begin to converge with those of shareholders. However, with greater ownership comes greater power in the form of voting rights. Managers may, at this level of holdings, maximise their personal wealth through increasing perquisites and guaranteeing their employment at the expense of corporate value. In addition, while low managerial ownership levels may have arisen through the vesting of compensation plans, it is unlikely that such plans will provide management with a moderate ownership stake in the firm. Moreover, even though external market controls are still in place, these and the effect of convergence of interests are not strong enough to align the behavior of management to shareholders. Managerial labour markets operate on the principal that poorly performing 1 See Morck et al. (1988), McConnell and Servaes (1990), Hermalin and Weisbach (1991), Cho (1998) and Himmelberg et al. (1999) for US companies and Short and Keasey (1999) and Faccio and Lasfer (1999) for UK companies. 2 For example, since compensation packages such as stock options are a transfer of wealth from shareholders to management, their value will lessen as managerial ownership increases. External market discipline is also a function of managerial ownership. Large shareholdings by top management act as a deterrent for takeovers because of the greater ability to oppose a hostile bid or drive up premiums to the point where bidders no longer view the target company as a positive net present value investment Stulz (1988). Finally, internal controls in the form of monitoring from large shareholders and corporate boards should reduce the scope for managers to diverge greatly from the interests of shareholders. Again, however, such discipline is likely to be inversely related to managerial control Denis et al. (1997). 648 J. R. Davies et al. / Journal of Corporate Finance 11 (2005) 645–660 anagers can be removed and appropriately disciplined. Studies by Denis et al. (1997) in the US and Dahya et al. (2002) in the UK both find an inverse relation between topmanagement turnover and managerial ownership. This lack of discipline provides evidence of a deficiency in incentives for managers to maximise shareholder value at this level of owners hip. Franks and Mayer (1996) also report that hostile takeover targets in the UK are not poorly performing firms, which is in contrast to the findings of a disciplinary role for corporate takeovers in the US by Martin and McConnell (1991). In this context, Franks and Mayer (1996) provide significant evidence that takeovers in the UK may not act to remove a self-serving board even when they are performing poorly. This lack of disciplinary control over poorly performing management may strengthen management’s ability to pursue sub-optimal corporate policies at intermediate ownership levels. 2. 3. High levels of managerial ownership (less than 50%) As levels of managerial equity ownership grow, objectives converge further to those of shareholders. At ownership levels, below 50% management do not have total control of the firm and external discipline still exists. While perhaps no longer being subject to any major discipline from external takeover markets, it is likely that even at these levels of ownership, managers are still subject to discipline from external block shareholders. This is particularly true in the UK, where because of strong informal ties between institutions (Short and Keasey, 1999), a lax regulatory environment concerning the ownership of listed companies (Roe, 1990) and low monitoring costs (Faccio and Lasfer, 1999), institutional activism is stronger than in the US. This view is also consistent with Franks et al. (2001) contention of strong minority protection laws in the UK, whereby large shareholders cannot transact with related companies without the consent of the firm’s minority shareholders. The UK regulatory framework stands in contrast to US corporate law which limits minorities to seeking redress after the related party transaction has taken place. Combined with monitoring from UK institutions, this may allo w external shareholders to impose some form of control on management even at elatively large levels of managerial ownership. 2. 4. High levels of managerial ownership (greater than 50%) At levels above 50% ownership, management has complete control of the company. Although atomistic shareholders are unlikely to have been able to in influence managers at far lower levels of ownership than this, there is always a possibility that a cartel of blockholders, allied with minority shareholder’s rights under UK company law, may be able to mount a challenge to management if they fail to make decisions in shareholders’ best interests. For a more in-depth discussion of the institutional differences and similarities between the United Kingdom and United States, see Short and Keasey (1999) and Faccio and Lasfer (1999). 3 J. R. Davies et al. / Journal of Corporate Finance 11 (2005) 645–660 649 At greater than 50% managerial ownership, this is no longer likely to be a serious issue to management. Furthermore, with majority ownership, the probability of a hostile takeover effectively becomes zero. The failure of external discipline combined with a lack of blockholder incentives above 50% may result in a decrease in corporate value for a small window of managerial holdings above this level. This fall in corporate value is consistent with the theoretical predictions of Stulz (1988). 2. 5. Very high levels of managerial ownership Finally, as managerial shareholdings rise to very high levels, management effectively become sole owners of the company. This would lead to value-maximising behavior as predicted by Jensen and Meckling (1976). Consistent with Morck et al. 1988), Short and Keasey (1999) and Faccio and Lasfer (1999) at above a certain level of ownership, corporate managers are faced with such severe financial penalties for failing to maximise the value of their companies that they are forced to make decisions which will maximise firm value, regardless of how this affects their private benefits of control. 2. 6. Summary Our characterisation of a highly nonlinear relationshi p between managerial equity holdings and corporate value is in contrast to earlier studies (Morck et al. , 1988; McConnell and Servaes, 1990; Hermalin and Weisbach, 1991; Cho, 1998; Himmelberg et al. 1999)4, which posit fewer turning points in their analysis. There is little theoretical basis on which the individual turning points can be determined, and the findings of Kole (1995) suggest that these will be in influenced by the size of the firms in the sample. However, it is expected that the second local maximum will be in the region of 50% managerial ownership reflecting the stage at which management gain total control of the company. In the next section, the main tests of our hypotheses will be carried out. 3. Empirical results 3. 1. Description of the data We use data on managerial and external block ownership for 1995 from the MacMillan London Stock Exchange Yearbook for 1996 and 1997. The Yearbook provides summary accounting data including a consolidated balance sheet, information on company directors, legal information on the company’s lawyers, auditors and stockbrokers, principle activities, company history, capital and dividend payments, and industrial sector for the McConnell and Servaes (1990) modelled the corporate value–managerial ownership relationship as a quadratic function, which by construction has only one turning point. 650 J. R. Davies et al. / Journal of Corporate Finance 11 (2005) 645–660 vast majority of all quoted companies and securities. 5 We restrict our attention to nonfinancial companies only and require that each firm has complete managerial and external ownership data for 1995, which leaves 802 industrial companies in our sample. 6 Data on capital expenditures, to tal assets employed, after tax profits, depreciation, leverage, equity market values, and research and development costs are collected from Datastream. We estimate Tobin’s Q ratio (our proxy for corporate value) using the formula below: Q? MVEQ ? PREF ? DEBT BV ASSETS ? 1? where: MVEQ=the year-end market value of the firm’s common stock; PREF=the yearend book value of the firmTs preference shares (preferred stock); DEBT=the year-end book value of the firmTs total debt; and BV ASSETS=the total assets employed by the firm, which is measured as total assets minus current liabilities. Our measure is consistent with the modified version of the formula as used by Chung and Pruitt (1994) who find that 96. 6% of the variability in the popular Lindenberg and Ross (1981) algorithm of Tobin’s Q is explained by their approximation. Our method also avoids the data availability problems which arise from using the more rigorous algorithms proposed by Lindenberg and Ross (1981) and Lewellen and Badrinath (1997) in order to estimate the replacement cost of assets. We use book values of preferred stock and long-term debt, rather than the market values proposed by Lindenberg and Ross (1981) and Lewellen and Badrinath (1997). In the UK, there is a far less active market for the trading of corporate debt than that which exists in the US, forcing us to rely on book values for these variables. In a final stratification of our sample, we mitigate the problem of potential outliers and trim 25 firms with the largest and smallest Tobin’s Q measure, leaving a final sample of 752 firms. 7 Table 1 presents descriptive statistics for our sample data. The mean managerial ownership stake of all board members is 13. 02%, which is similar to comparable US studies, but slightly lower than Faccio and Lasfer (1999) who report mean ownership of 16. 7%. Tobin’s Q is slightly higher than that reported for related US work with a mean value of 1. 96. The standard deviation of Tobin’s Q is 1. 21, which is also greater than other studies. However, it is substantially less than the mean of 2. 47 reported by Doukas et al. (2002) and is relatively similar to the mean value of 1. 86 that Short and Keasey (1999) report for their market valuation ratio. 8 The mean blockholder ownership is 37. 34% and is on a par with that reported for US firms by McConnell and Servaes (1990) (32. 4%) and 34. 57% reported by Faccio and Lasfer (1999) for UK firms. The full range of firm sizes is included in the sample with the 5 To establish the reliability of the summary ownership data, we carried out a correlation analysis of a subsample of 422 firms from he original data set of 802 companies (52. 62%) for which we were able to obtain company annual reports. The yearbook data and company accounts data exhibited a correlation of 0. 90, with a pvalue of 0. 00. We also establish the robustness of our data by re-estimating the model using data for 1997. This result is discussed later in this section. 6 Recently listed, merged or acquired firms are not included. 7 This is a larger sample than that used by Morck et al. (1988)—371 firms, Cho (1998)—326 firms and Himmelberg et al. (1999)—maximum 427 firms in any 1 year. Measured as the market value of equity divided by the book value of equity, minus any intangibles. J. R. Davies et al. / Journal of Corporate Finance 11 (2005) 645–660 Table 1 Descriptive statistics Variable Management ownership Blockholder ownership Largest stakeholder Capital expenditures Total assets employed After tax profits less depreciation/assets employed Debt/assets employed Market value of equity Research and development Tobin’s Q Mean 13. 02% 37. 34% 18. 82% 21,221 255,642 0. 1425 0. 1411 335 2918 1. 9647 S. D. 18. 06% 23. 57% 21. 64% 75,317 1,583,274 0. 4763 0. 252 1399 44,108 1. 2092 Minimum 0. 00% 0. 00% 0. 00% 7 268 A10. 977 0. 0000 0. 68 0 0. 4502 651 Maximum 79. 90% 100. 00% 100. 00% 1,024,200 37,774,000 3. 4207 4. 8358 26,224 1,198,988 7. 0997 Managerial own ership data measures the total level of holdings held by company management that are greater than 0. 5% of a company’s equity. Blockholder data measures the total level of holdings by outside blockholders that are greater than 3% of a company’s equity. Largest stakeholder is the largest single outside blockholder that holds at least 3% of company’s outstanding equity. Capital expenditures (thousands), total assets employed (thousands), after tax profits, depreciation, leverage, equity market values (millions) and research and development costs (thousands) are collected from Datastream. Tobin’s Q is measured as the ratio of the market value of equity and book values of debt and preferred equity to the book value of assets in the firm minus current liabilities. Shareholdings data is taken from the London Stock Exchange Yearbook for 1996 and 1997. All data are for industrial companies quoted on the London Stock Exchange in 1995. mallest company having an equity market capitalization of o680,000 and the largest company’s equity valued at approximately o26 billion. The mean market capitalization of firms in the sample is o335 million. Table 2 provides the distribution of sample statistics grouped by managerial ownership. A very large proportion of the sample (62%) have managerial ownership levels less than or equal to 10%. However, a larg e fraction of companies (11%) also in the sample had boards Table 2 Breakdown of sample by managerial ownership Manager level Ownership Number of firms 464 87 75 41 34 26 21 4 Blockholder ownership, % 43. 34. 5 34. 4 24. 0 22. 7 13. 0 12. 7 5. 8 Tobin’s Q 1. 952 2. 033 1. 736 2. 109 2. 113 2. 257 1. 933 1. 808 Total assets employed 393,861 44,093 26,186 34,322 35,864 28,190 14,234 10,127 Capital expenditures/ assets employed 0. 106 0. 161 0. 124 0. 117 0. 114 0. 100 0. 099 0. 114 Liquidity 0. 130 0. 129 0. 157 0. 194 0. 194 0. 177 0. 169 0. 239 0VMOb10% 10VMOb20% 20VMOb30% 30VMOb40% 40VMOb50% 50VMOb60% 60VMOb70% 70VMOb100% Managerial ownership (MO) data measures the total level of holdings held by company management that are greater than 0. 5% of a company’s equity. Blockholder ownership measures the total level of holdings by outside blockholders that are greater than 3% of a company’s equity. Capital expenditure (thousands), total assets employed (thousands), after tax profits and equity market values (millions) are collected from Datastream. Liquidity is measured as cashflow divided by total assets employed. Tobin’s Q is measured as the ratio of the market value of equity and book values of debt and preferred equity to the book value of assets in the firm minus current liabilities. Shareholdings data is taken from the London Stock Exchange Yearbook for 1996 and 1997. All data are for industrial companies quoted on the London Stock Exchange in 1995. 652 J. R. Davies et al. / Journal of Corporate Finance 11 (2005) 645–660 Table 3 Regression results for Tobin’s Q on managerial ownership Variable Coefficient t-Statistic Adj. R 2 Intercept 1. 85 28. 14 0. 017 MO 0. 12 3. 23 MO2 A0. 013 A3. 08 F MO3 4. 63A10 2. 82 2. 651 A4 MO4 A6. 73A10 A2. 53 A6 MO5 3. 36A10A8 2. 24 The following equation was estimated using data for 752 firms listed on the London Stock Exchange during 1995. Q ? a0 ? a1 MO ? a2 MO2 ? a3 MO3 ? a4 MO4 ? a5 MO5 ? e where Q is Tobin’s Q and MO is managerial ownership. Ownership data is taken from the London Stock Exchange Yearbook and Tobin’s Q is calculated from Datastream. which owned at least 40% of all outstanding equity. As would be expected, outside blockholder ownership decreases with managerial ownership. At managerial ownership levels of 30%, blockholder ownership is slightly less at 24%. It is probable that external discipline, as provided by blockholders, would still be strong at these levels of managerial holdings, particularly where informal coalitions among blockholders are more prominent (Short and Keasey, 1999). At higher levels of managerial holdings, blockholder ownership decreases sharply leading to a collapse in the power of blockholders. Managerial ownership is a decreasing function of company size, which is consistent with Demsetz and Lehn (1985). Although firm sizes in the UK are considerably smaller than US firms, the ratios in Table 2 are similar to summary statistics provided in Morck et al. (1988), McConnell and Servaes (1990), Cho (1998) and Himmelberg et al. (1999). Table 2 also illustrates the nonlinear relationship between Tobin’s Q and managerial holdings. Visual inspection indicates two maximum points in the region of 10% to 20% and 50% to 60%, respectively. The convergence of managerial interests to those of shareholders at very high levels of ownership is not apparent at this stage because of the small number of companies with managerial holdings above 70%. However, the statistics for all other groupings are consistent with our theoretical motivation. 3. 2. Estimation of ownership breakpoints In order to model the Tobin’s Q–managerial ownership (MO) function as having two maximum and two minimum turning points, we specify a quintic function, as follows: Q ? 0 ? a1 MO ? a2 MO2 ? a3 MO3 ? a4 MO4 ? a5 MO5 ? e ? 2? For the nonlinear relationship discussed in Section 2 to be valid, the coefficients in Eq. (2) must have the following signs: a 0N0; a 1N0; a 2b0; a 3N0; a 4b0; a 5N0. The estimated values of the coefficients in Eq. (2) are given in Table 3. 9 The intercept coefficient, which is an estimate of Tobin’s Q i n firms with no managerial holdings, is 1. 85. Each slope coefficient is of the correct sign and statistically significant at the 5% level. Although the It is clear that Tobin’s Q will be in influenced by more than just managerial ownership. However, the objective of this paper is to investigate whether the standard quadratic and cubic specifications used in previous studies are too simplistic. To maintain parsimony, we therefore omit other factors from this specific model. Other relevant factors are incorporated into the analysis in a later table. 9 J. R. Davies et al. / Journal of Corporate Finance 11 (2005) 645–660 653 Estimated Relationship between Tobin's Q and Managerial Ownership 2. 40 2. 20 2. 00 1. 80 1. 60 1. 40 1. 20 0 0. 1 0. 2 0. 3 0. 4 0. 5 0. 6 0. 7 0. 8 0. 9 Tobin's Q Insider Ownership Fig. 1. Estimated relationship between Tobin’s Q and Managerial Ownership. Tobin’s Q was modelled as a quintic function of insider ownership using ordinary least squares regression. The estimated regression line is: Q=1. 85+0. 12IOA0. 013OI2+4. 63A10A4IO3A6. 73A10A6IO4+3. 36A10A8IO5. adjusted R 2 is low, it is similar to that found in comparable US studies. The use of this model as a basis to estimate managerial ownership turning points leads to four critical values: 7. 01%, 26. 0%, 51. 4%, 75. 7% and is illustrated in Fig. 1. To establish the robustness of our regression model, the spline approach as applied by Morck et al. (1988), Cho (1998) and Himmelberg et al. (1999) to estimate breakpoints was carried out using our generated turning points. Table 4 presents the coefficients resulting from the piecewise linear regression. Similar to Table 3, each coefficient has the expected sign and all but one variable is statistically significant at the 5% level. The only variable that is not significant, MOover 76% , has the correct sign. The probable cause for the lack of significance is the small number of firms in this managerial ownership grouping. An examination of these results suggests that Tobin’s Q increases in firms for managerial ownership levels up to 7% and then declines to ownership levels of 26%. This is almost identical to the turning points in Morck et al. (1988) and Himmelberg et al. (1999) (5% and 25%, respectively) and is comparable to Cho (1998), who uses breakpoints of 7% and 38%. However, it differs from the UK studies of Short and Keasey (1999) and Faccio and Lasfer (1999) who each reports two turning points of 12. 99% and 41. 99%, and 19. 68% and 54. 12%, respectively. Earlier studies limited the turning points to two but in our extension, it is clear that there are another two turning points at much higher levels of managerial ownership. It also appears that market discipline has an influence on managerial objectives up to the point where the board takes complete control (51%). Tobin’s Q then decreases until ownership levels reach 76%, after which Q increases. Denis and Sarin (1999) argue that cross-sectional studies may be subject to bias, whereby they fail to account for events with potentially large valuation consequences. 10 10 Examples of such events may include receiving a takeover bid, top management turnover, etc. 654 J. R. Davies et al. / Journal of Corporate Finance 11 (2005) 645–660 Table 4 Spline regression results for Tobin’s Q on managerial ownership Variable Coefficient t-Statistic Adj. R 2 Intercept 1. 854 28. 38 0. 012 MOup 0. 056 2. 93 to 7% MO7% to 26% MO26% 0. 0187 2. 57 2. 769 to 51% MO51% A0. 053 A1. 99 to 76% MOover 0. 624 1. 12 76% A0. 020 A2. 62 F The following equation was estimated using data for 752 firms listed on the London Stock Exchange during 1995. Q ? a0 ? a1 MOup to 7% ? a2 MO7% to 26% a3 MO26% to 51% ? a4 MO51%to 76% ? a5 MOover 76% ?e where Q is Tobin’s Q and MOup to 7%=managerial ownership if managerial ownership b7%, =7% if managerial ownershipN7%. MO7% to 26%=0 if managerial ownership b7%, =managerial ownership minus 7% if 7%bmanagerial ownershipb26%, =26% if managerial ownershipN26%. MO26% to 51%=0 if managerial ownershipb26%, =managerial ownership m inus 26% if 26%bmanagerial ownershipb51%, =51% if managerial ownershipN51%. MO51% to 76%=0 if managerial ownership b51%, =managerial ownership minus 51% if 51%bmanagerial ownershipb76%, =76% if managerial ownership N26%. MOover 76%=0 if managerial ownershipb76%, =managerial ownership minus 76% if managerial ownershipN76%. Ownership data is taken from the London Stock Exchange Yearbook and Tobin’s Q is calculated from Datastream. As a further test of robustness, we carried out the quintic analysis for managerial ownership and Tobin’s Q for the same sample of available firms in 1997. 11 Again, each coefficient was significant with the correct signs and the turning points from the estimated model were relatively stable at 7. 9%, 26. 5%, 55. 2% and 86. 2%. . 3. Endogeneity of managerial equity ownership, investment and corporate value To analyse the effects of endogeneity in the managerial ownership, investment and corporate value relationship, we follow Cho (1998) and carry out a simultaneous equations analysis using two-stage least squares. Cho (1998) and Himmelberg et al. (1999) showed that once endogeneity was controlled, the perceived impact of managerial ownership on corporate value d isappeared. Moreover, corporate value was found to positively affect levels of managerial ownership. It is possible that if the model specification employed by these studies is wrong, what appears to be a lack of statistical significance in the endogenous variables in the simultaneous equations analysis may actually be due to errors in variables arising from the intermediate regressions. We re-run the two-stage least squares analysis of Cho (1998) using our more complex specification. 12 The control variables in our regression are the same as in Cho (1998). Namely, managerial ownership, investment and corporate value are Some firms fell out of the sample because of mergers, delisting, and being taken over. Cho (1998) also attempts to control for specification error by re-estimating his simultaneous regression analysis using managerial ownership as a linear variable and again finds no relationship between managerial ownership and corporate value. However, if indeed there is a nonlinear relationship between ownership and corporate value, such an approach would fail to capture this. 12 11 J. R. Davies et al. / Journal of Corporate Finance 11 (2005) 645–660 655 defined to be endogenously determined by each other as well as some additional relevant exogenous variables. That is: Managerial Ownership ? ? market value of firm0s common equity; corporate value; investment; volatility of earnings; liquidity; industry? Corporate Value ? g? managerial ownership; investment; leverage; asset size; industry; block ownership; largest stakeholder? Investment ? h? managerial ownership; corporate value; volatility of earnings; liquidity; industry? For comparability, we define each of the above vari ables as in Cho (1998). For each company, industry dummy variables are set equal to one for each Financial Times Industry Classification (FTIC) grouping that sample firms lie within, and zero otherwise. In addition to the variables used by Cho (1998), we include blockholder ownership and largest stakeholder in the corporate value regressions to reflect the potential impact of blockholder discipline in the UK and the role of a founding or dominant individual on corporate value. All accounting and market variables are taken at the financial year-end from Datastream. In Table 5, we report results from the simultaneous equations analysis. Taking the managerial ownership regression first, all variables with the exception of investment have coefficients with the expected sign. Managerial ownership is negatively related to the market value of equity, which reflects the fact that wealth constraints and risk-aversion will prevent managers from holding substantial stakes in large firms. Firm level liquidity is shown to be positively related to managerial ownership, which is a stronger result than Cho (1998) who reported no significance for this variable. Importantly, Tobin’s Q is found to be significant and positively related to the level of managerial ownership. This is consistent with Cho (1998) but is opposed to Demsetz and Villalonga (2001), who find the opposite effect. This result suggests that managers tend to hold larger stakes in firms that are successful or have higher corporate value. This may also be indicative of successful managers benefiting from equity-related compensation policies. The investment variable, which has a negative impact on managerial ownership is surprising as theory predicts that firm level investment will be positively related to managerial ownership. Himmelberg et al. (1999) contend that firms with high investment spending will have high managerial ownership to alleviate the monitoring problem caused by discretionary managerial spending. However, Jensen (1986) argued that firms may overinvest as a result of an earnings retention conflict, rather than underinvest as Jensen and Meckling’s (1976) moral hazard theory would predict. When a firm is in this situation, managers may be able to maximise their size-related compensation by overinvesting, but are aware that this may ultimately reduce the value of their shareholdings. Although tentative, this could in part explain the negative relation between investment and ownership. Cho (1998) also finds a negative (but insignificant) coefficient on the investment variable using both capital and research and development expenditures. 56 J. R. Davies et al. / Journal of Corporate Finance 11 (2005) 645–660 Table 5 Simultaneous equations analysis of managerial ownership, corporate value and investment Variable MVEQ Tobin’s Q Volatility Liquidity Investment Leverage Asset size Largest stakeholder Blockholder ownership MO MO2 MO3 MO4 MO5 Industry dummies Adj. R 2 F Managerial ownership A1. 8A10 (A3. 74) 0. 127 (4. 63) A1. 0A10A6 (A0. 74) 0. 035 (2. 24) A1. 314 (A2. 67) A5 Corporate value Investment 0. 073 (2. 35) 3. 89A10A6 (A2. 86) 0. 013 (1. 01) Yes 0. 045 8. 014 5. 136 (2. 23) 1. 088 (4. 36) 3. 33A10A8 (1. 17) A0. 20 (A0. 06) A0. 837 (A2. 60) 1. 588 (3. 07) A0. 395 (A2. 22) 0. 037 (1. 64) A0. 001 (A1. 14) 1. 9A10A5 (0. 76) Yes 0. 033 3. 497 A0. 035 (A0. 46) 0. 018 (0. 72) A0. 003 (A0. 92) 1. 72A10A4 (1. 03) A3. 12A10A7 (A1. 07) Yes 0. 009 2. 497 Results from a simultaneous equations analysis of managerial ownership, corporate value and investment for 752 firms, using the two-stage least squares method to estimate the following equations: Managerial Ownership ? f ? market value of firm0s common equity; corporate value; investment; volatility of earnings; liquidity; industry? CorporateValue ? g? anagerial ownership; investment; financial leverage; asset size; industry; block ownership; largest stakeholder? Investment ? h? managerial owner ship; corporate value; volatility of earnings; liquidity; industry? In the above equations, managerial ownership measures the total level of holdings held by company management that are greater than 0. 5% of a company’s equity. Blockholder data measures the total level of holdings by outside blockholders that are greater than 3% of a company’s equity. Largest stakeholder is the largest single outside blockholder that holds at least 3% of company’s outstanding equity. Investment is defined as capital expenditure divided by total assets employed, leverage is the ratio of total debt to total assets employed and liquidity is measured as cashflow divided by total assets employed. Capital expenditure, total assets employed, after tax profits, depreciation, leverage, equity market values and profit volatilities are collected from Datastream. Tobin’s Q is measured as the ratio of the market value of equity and book values of debt and preferred equity to the book value of assets in the firm minus current liabilities. Shareholdings data is taken from the London Stock Exchange Yearbook for 1996 and 1997. All data are for industrial companies quoted on the London Stock Exchange in 1995. t-Statistics are in parenthesis. The estimated coefficients from the corporate value regression are given in the second column of Table 5. Corporate value is shown to be positively related to investment and leverage. While the investment coefficient is as expected, the sign of the leverage variable requires more discussion. Morck et al. 1988) find that leverage has a negative but insignificant impact on corporate value and attribute this to the possibility of managers in highly levered firms holding a higher than average level of ownership. However consistent with our results, McConnell and Servaes (1990) report a positive significant coefficient for leverage. Leverage can have various effects on firm value. The notion that high debt levels lead to greater corporate value has been argued by Modigliani and Miller (196 3) with respect J. R. Davies et al. / Journal of Corporate Finance 11 (2005) 645–660 57 to valuable tax shields, Ross (1977) and Myers (1977) with respect to a signalling hypothesis and Jensen’s (1986) free cashflow hypothesis. Ultimately, leverage is one way of imposing external discipline on management and if it is effective, will lead to increased corporate value. Alternatively, Demsetz and Villalonga (2001) interpret a negative association between leverage and firm value as being due to relative inflation between the current time period and the earlier time period where companies had issued much of their debt. We view the most important result from the corporate value regression as being the significance of the managerial ownership variables. Our results indicate that although managerial ownership levels are determined by corporate value, corporate value itself is determined in part by managerial ownership. This finding is at odds with Cho (1998) and Himmelberg et al. (1999) but consistent with the classical view of Jensen and Meckling (1976) and empirical work by Morck et al. (1988) and McConnell and Servaes (1990). An interesting result is that blockholder ownership is shown to negatively impact Tobin’s Q. This result is consistent with Faccio and Lasfer (1999, 2000). McConnell and Servaes (1990) suggest that this could be due to a conflict of interests, which results from blockholders being forced into aligning themselves with managers so as not to jeopardize their other dealings with the firm. Alternatively, the negative coefficient may be explained by the strategic alignment hypothesis, which argues that blockholders and managers find it mutually beneficial to cooperate with each other. Finally, such findings may be consistent with the arguments of Burkart et al. 1997) in that too much block ownership will overly constrain management and reduce their ability to take value-maximising investment decisions. The investment regression coefficients presented in column three of Table 5 show a significant positive effect of corporate value on investment and a negative effect of profit volatility on investment. The finding that corporate value has a positive effect on investment is consisten t with the arguments of Cho (1998) that highly valued firms will have large investment opportunities. Also, firms with variable earnings will be reluctant to invest if future income is uncertain. Managerial ownership is found to have no impact on firm level investment. However, this may reflect optimality in that investment policy may be one way in which managers affect value, but not the only means. Ultimately we view our findings of a causal relation between ownership and firm value as being of greater significance than the lack of a relation between ownership and investment. These results are consistent with Cho (1998) but slightly stronger, in that volatility of earnings is significant in our regressions but insignificant in Cho (1998). . Conclusions Debate as to the relationship between corporate value and managerial ownership in the US is still unresolved. Studies such as Morck et al. (1988), McConnell and Servaes (1990), and Hermalin and Weisbach (1991) document a nonlinear relation between these two variables. More recent work by Cho (1998), Himmelberg et al. (1999), and Demsetz and Villalonga (2001) shows that when controlling for endogeneity, managerial ownership is determined by corporate value but not vice-versa. 658 J. R. Davies et al. Journal of Corporate Finance 11 (2005) 645–660 We argue that even accepting that corporate value and managerial ownership are endogenously related to each other, misspecification of the managerial holding–corporate value relationship may lead to spurious conclusions concerning the direction of causality. Applying a quintic structure, we present results which suggest that the correct form of this relationship is a double humped curve. This is in contrast to other studies that have assumed a cubic or quadratic specification and by construction only one hump. The second hump or local maximum is attributed to a collapse in external market discipline at or around the point where managers take overall control of their firm. At this point, which is around 50% ownership, the management is not sufficiently akin to owners but have sufficient power to disregard any form of external monitoring or discipline. This has a detrimental affect on corporate value for a short window of managerial holdings. At high levels of managerial ownership, managers are effectively majority owners of their firm leading to a convergence of interests with other outside shareholders. Utilizing the quintic specification for managerial ownership, we show that even when controlling for endogeneity, not only is corporate value a determinant of managerial ownership but managerial ownership is also a determinant of corporate value. This finding is consistent with the classical work of Jensen and Meckling (1976), as well as the early empirical work of Morck et al. (1988) and McConnell and Servaes (1990) who do not control for endogeneity in their analysis of corporate value and managerial ownership. We believe our analysis to have several important contributions to the literature on the relationship between managerial ownership and corporate value. First, our quintic specification extends previous work in this area and successfully captures the complex nonlinear relationship between corporate value and managerial ownership. Second, by analysing a completely different market which is similar in structure to the United States, we strengthen the power and insights gained from earlier comparable US studies. Third, we provide evidence that corporate value, firm level investment and managerial holdings are interdependent with each other. This has implications for the debate on the effectiveness of compensation policies involving stock options for top managers. Moreover, our findings suggest that some levels of managerial ownership may not be beneficial to outside shareholders even when these levels are high. At the very least, this paper has served to add to the debate concerning the importance of managerial ownership on corporate value by providing evidence that even controlling for endogenous effects, managerial ownership and stock compensation schemes do have a significant influence on corporate value. Our research has provided an initial step towards a more accurate characterisation of the corporate value–managerial ownership relationship. While we do not posit that our specification can be applied to every given data set, we argue that previous research may be misspecified where it has failed to fully explore alternative specifications of the managerial ownership–corporate value relationship. Future work in this area may focus on other structural forms, which more effectively reflect the interdependence of managerial ownership and corporate prospects. The nonlinear endogenous impact of blockholders on corporate value and managerial ownership would also provide interesting insights on the external discipline that is faced by firm managers and the impact this has on corporate value. J. R. Davies et al. / Journal of Corporate Finance 11 (2005) 645–660 659 Acknowledgements The authors would like to thank John Capstaff, Scott Linn, Andrew Marshall, James Wansley and seminar participants at the Financial Management Association International (2001), European Financial Management Association (2002), Dublin Economics Workshop, the University of Strathclyde and an anonymous referee for their valuable comments on earlier versions of the paper. The normal caveat applies. References Burkart, M. , Gromb, D. , Panunzi, F. , 1997. Large shareholders, monitoring, and the value of the firm. Quarterly Journal of Economics 112, 693 – 728. Cho, M. H. , 1998. Ownership structure, investment, and the corporate value: an empirical analysis. Journal of Financial Economics 47, 103 – 121. Chung, K. H. , Pruitt, S. W. , 1994. A simple approximation of Tobin’s Q. Financial Management 23, 70 – 74. Dahya, J. , McConnell, J. J. , Travlos, N. G. , 2002. The Cadbury committee, corporate performance and top management turnover. Journal of Finance 57, 461 – 483. Demsetz, H. , Lehn, K. , 1985. The structure of corporate ownership: causes and consequences. Journal of Political Economy 93, 1155 – 1177. Demsetz, H. , Villalonga, B. , 2001. Ownership structure and corporate performance. Journal of Corporate Finance 7, 209 – 233. Denis, D. J. , Sarin, A. , 1999. Ownership and board structures in publicly traded corporations. Journal of Financial Economics 52, 187 – 223. Denis, D. J. , Denis, D. K. , Sarin, A. , 1997. Ownership structure and top executive turnover. Journal of Financial Economics 45, 193 – 221. Doukas, J. A. , McKnight, P. J. , Pantzalis, C. , 2002. Security analysis, agency costs and UK firm characteristics. Working Paper. Faccio, M. , Lasfer, M. A. , 1999. Managerial ownership, board structure and firm value: the UK evidence. Working Paper. Faccio, M. , Lasfer, M. A. , 2000. Do occupational pension funds monitor firms in which they hold large stakes? Journal of Corporate Finance 6, 71 – 110. Fama, E. F. , 1980. Agency problems and the theory of the firm. Journal of Political Economy 88, 288 – 307. Franks, J. , Mayer, C. , 1996. Hostile takeovers and the correction of management failure. Journal of Financial Economics 40, 163 – 181. Franks, J. , Mayer, C. , Renneboog, L. , 2001. Who disciplines management in poorly performing companies? Journal of Financial Intermediation 10, 209 – 248. Hart, O. D. , 1983. The market mechanism as an incentive scheme. Bell Journal of Economics 14, 366 – 382. Hermalin, B. Weisbach, M. , 1991. The effects of board composition and direct incentives on firm performance. Financial Management 20, 101 – 112. Himmelberg, C. P. , Hubbard, R. G. , Palia, D. , 1999. Understanding the determinants of managerial ownership and the link between ownership and performance. Journal of Financial Econ omics 53, 353 – 384. Jensen, M. C. , 1986. Agency costs of free cashflow, corporate finance and takeovers. American Economic Review 76, 323 – 329. Jensen, M. C. , Meckling, W. H. , 1976. Theory of the firm: managerial behavior, agency costs and ownership structure. Journal of Financial Economics 3, 305 – 360. Jensen, M. C. , Ruback, R. S. , 1983. The market for corporate control: the scientific evidence. Journal of Financial Economics 11, 5 – 50. Kole, S. , 1995. Measuring managerial equity ownership: a comparison of sources of ownership data. Journal of Corporate Finance 1, 413 – 435. Lewellen, W. G. , Badrinath, S. G. , 1997. On the measurement of Tobin’s Q. Journal of Financial Economics 44, 77 – 122. 660 J. R. Davies et al. / Journal of Corporate Finance 11 (2005) 645–660 Lindenberg, E. , Ross, S. , 1981. Tobin’s Q ratio and the industrial organization. Journal of Business 54, 1 – 33. Martin, K. J. , McConnell, J. J. , 1991. Corporate performance, corporate takeovers, and management turnover. Journal of Finance 46, 671 – 687. McConnell, J. J. , Servaes, H. , 1990. Additional evidence on equity ownership and corporate value. Journal of Financial Economics 27, 595 – 612. Modigliani, F. , Miller, M. H. , 1963. Corporate income taxes and the cost of capital: a correction. American Economic Review 53, 433 – 443. Morck, R. , Shleifer, A. , Vishny, R. W. , 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 – 175. Roe, M. J. , 1990. Political and legal restraints on ownership and control of public companies. Journal of Financial Economics 27, 7 – 42. Ross, S. A. , 1977. The determination of financial structure: the incentive-signalling approach. Bell Journal of Economics 8, 23 – 40. Short, H. , Keasey, K. , 1999. Managerial ownership and the performance of firms: evidence from the UK. Journal of Corporate Finance 5, 79 – 101. Stulz, R. E. , 1988. Managerial control of voting rights: financing policies and the market for corporate control. Journal of Financial Economics 20, 25 – 54.