Beyond the Biggest: Do Other Large Shareholders Influence Corporate Valuations?

Transcription

1 Beyond the Biggest: Do Other Large Shareholders Influence Corporate Valuations? Luc Laeven and Ross Levine* This Draft: March 13, 2005 Abstract: This paper examines the relationship between corporate valuations and the presence, size, incentives, and identity of multiple large shareholders. Using data on a large number of firms across Western Europe, we find that about one-third of all firms, and over 40 percent of firms with one large owner, have two or more owners that each holds more than 10 percent of the voting rights. We find that large shareholders beyond the biggest boost corporate valuations, but this result holds conditional on the characteristics of these other large owners. Specifically, we find that the distribution of voting and cash flow rights across large shareholders matters. For example, only when the gap in voting rights between the first and the second largest shareholders is small does the second largest shareholder increase corporate valuations. Furthermore, the evidence stresses that only when large shareholders have sound incentives as measured by cash-flow rights do we observe a strong, positive relationship between corporate valuations and the existence of large owners. Keywords: Corporate Governance, Corporate Finance, Ownership Structure JEL Classification: G32, G34 * Laeven: World Bank and CEPR; Levine: University of Minnesota and NBER. We would like to thank Stijn Claessens, Armando Gomez, Daniel Wolfenzon, and seminar participants at the University of Minnesota for helpful comments. Ying Lin provided expert research assistance. This paper s findings, interpretations, and conclusions are entirely those of the authors and do not necessarily represent the views of the World Bank, its Executive Directors, or the countries they represent.

2 I. Introduction Much of corporate finance theory examines the agency problems that arise from two extreme ownership structures: (1) 100 percent dispersed ownership or (2) one large, controlling owner combined with small, diffuse shareholders. 1 In the first case, each small shareholder lacks the incentives, organizational abilities, and contractual mechanisms to align the interests of powerful managers with those of shareholders. Consequently, managers may exert substantial discretion over firm decisions and divert corporate resources for their private gain. 2 At the other extreme, a large shareholder internalizes the benefits from either closely monitoring managers or directly running the firm, which reduces the diversion of resources and boosts corporate valuations. However, the existence of a large shareholder creates a different agency problem: Conflicts arise between the controlling shareholder and other shareholders. The controlling owner may expropriate resources or provide jobs and generous business deals to related parties. 3 The incentives and ability of the largest owner to divert corporate resources for private gain motivates an inquiry into how the presence and characteristics of other large stockholders influence agency problems within firms. Some theoretical models consider more complex ownership structures, characterized by corporations with two or more large owners. Pagano and Roell (1998) emphasize that other large shareholders beyond the biggest shareholder will monitor the controlling shareholder and reduce 1 For influential examinations, see Jensen and Meckling (1976), Grossman and Hart (1980, 1986, 1988), Shleifer and Vishny (1986), Harris and Raviv (1988), Stulz (1988), Hart and Moore (1990), and Burkart, Gromb, and Panunzi (1997, 1998). For broader treatments, see the book by Hart (1995) and the review article by Shleifer and Vishny (1997). 2 The ability of small shareholders to mitigate agency problems depends on a multitude of factors, including laws and enforcement mechanisms. See Coase (1937), Jensen and Meckling (1976), Hart (1995), and La Porta et al. (1998). 3 Concentrated ownership may involve other complications besides the agency problem between the controlling owner and other shareholders. Demsetz and Lehn (1985) argue that concentrated ownership implies that large shareholders will not be optimally diversified. The heavy firm-specific exposure will lead risk-averse controlling owners to act excessively prudently. Furthermore, while there are tensions between debt and equity holders, the existence of large equity holders will change the nature of this conflict and potentially influence corporate governance. 1

3 the diversion of corporate resources. 4 Bloch and Hege (2001) stress that only when the second largest owner is sufficiently large relative to the largest owner can the second large owner contest control and reduce diversion. 5 Bennedsen and Wolfenzon (2000) emphasize that (i) large shareholders form coalitions and (ii) diversion falls as the controlling coalition s cash flow rights increase. Moreover, their model predicts that controlling collations with substantial cash flow rights tend to emerge when there is either one large shareholder or when large shareholders are of similar sizes. Alternatively, when cash flow rights are distributed unevenly among many shareholders, this increases the likelihood that a coalition of shareholders with sufficient voting power, but low cash flow rights, will form and control the firm. This type of collation will have the incentives (low cash flow rights) and ability (sufficient voting rights) to divert corporate resources for private gain. In this paper, we empirically examine the predictions of these theories by assessing the impact of complex ownership structures on corporate valuations. In particular, the first part of the paper examines whether complex ownership structures are common in Western Europe. Some research questions the relevance of examining firms with more than one large shareholder. In a sample of the largest, publicly traded firms, La Porta et al. (1999) find that firms with a controlling shareholder (the largest shareholder with a minimum of 20% of the voting rights) have another large 4 Pagano and Roell (1998) stress that a controlling owner may sell shares to other large shareholders to reduce the controlling owner s ability to divert corporate resources for private gain, which will lower the cost of raising capital. Bolton and Von Thadden (1998) show that there is a trade-off between the benefit of corporate control and the costs of reduced liquidity associated with concentrated ownership. Maug (2003) presents a model showing that liquid stock markets help increase the incentives of large shareholders to monitor when monitoring costs are high. Noe (2002) analyzes situations under which monitoring by strategic shareholders affect firm value. His model predicts that increases in strategic investors shareholdings increase firm value, because future expected capital gains on such holdings increase the incentives of such informed investors to monitor. 5 Burkart, Gromb, and Panunzi (1997) note that a controlling owner monitors the manager, but this could lead to opportunistic behavior by the controlling owner or to collusion with the manager. With multiple controlling owners, a similar problem arises: The controlling owner could collude with other owners. Zwiebel (1995) presents a model in which investors allocate their wealth across firms and form controlling coalitions that share the private benefits of control. In turn, Gomes and Novaes (2001) study the ex-post bargaining among large shareholders and how this affects small shareholders. These models do not examine multiple large owners in an environment with liquid equity markets. 2

4 shareholder (with voting rights of greater than 10%) in 25% of firms in their sample of 27 countries. We look beyond the biggest corporations and examine a broad range of firms in Western Europe. The paper s second part examines the relationship between corporate valuations and complex ownership structures, where we pay particular attention to (i) studying large shareholders beyond the biggest and (ii) testing whether particular characteristics of large shareholders improve corporate governance. In terms of characteristics, we examine cash flow rights to gauge the incentives of large shareholders to divert corporate resources. As the controlling owner s cash flow rights increase, expropriation involves a greater reduction in the owner s own cash flow from the corporation, which moderates incentives to divert corporate assets (Jensen and Meckling, 1976; Burkart, Gromb, Panunzi, 1998; Shleifer and Wolfenzon, 2002). We also examine whether the presence, size, and identity of other large shareholders influence corporate valuations. As suggested by theory, the presence of other large shareholders may improve monitoring, reduce diversion, and boost valuations (Pagano and Roell, 1998). Other models suggest that corporate governance only improves when large owners are of comparable sizes (Bennedsen and Wolfenzon, 2000; Bloch and Hege, 2001). Finally, the operation of ownership coalitions may depend on the identity of the owners. We test if corporate values differ depending on whether the two largest owners are both families, whether they are a family and a bank, or whether the state is a large owner. Our work builds on recent empirical research that examines the effects of multiple large owners. Lehmann and Weigand (2000) show that the existence of a second large owner increases the profitability of German firms. 6 Faccio, Lang, and Young (2001) find that the existence of multiple large shareholders increases dividend payouts in Europe, but lowers them in Asia. In an examination of Finnish firms, Maury and Pajuste (2004) show that the relationship between 6 In related work, Volpin (2002) shows that firm valuations are higher in Italy when the controlling owner is a syndicate, rather than a single shareholder. 3

5 corporate valuations and the presence of multiple large shareholders depends on the comparative sizes of the large shareholders. We follow Maury and Pajuste s (2004) extension of earlier studies by looking beyond the mere presence of other large shareholders and also examining the relative sizes of these owners. We extend Maury and Pajuste s (2004) Finnish study by (i) examining the relationship between corporate valuations and multiple large owners across a broad cross-section of countries and (ii) considering other characteristics, besides voting rights, which may influence the impact of large shareholders on corporate governance. To conduct our analyses, we identify large owners and compute their voting and cash flow rights for a cross-section of manufacturing firm in Western Europe. Following La Porta et al. (2002), a large owner is a legal entity that directly or indirectly controls at least 10% of the voting rights. If more than one entity holds more than 10% of the voting rights, the controlling owner is defined as the legal entity with the greatest number of voting rights. In computing each large owner s voting rights we also determine the identity of the large owners, i.e., whether they are legal persons (which we call family ), banks, or the state. For each large owner, we also compute the cash flow rights, which may differ substantively from control rights when there are indirect chains of control. As demonstrated below, large owners frequently structure their shareholding so that they have large control rights, but comparatively low cash flow rights. The theories discussed above suggest that low cash flow rights relative to voting rights intensify incentives to divert corporate resources, with adverse implications on corporate valuations. The first part of the paper finds that many firms have complex ownership structures. This confirms the results in Faccio and Lang (2002), who show that of 5232 firms across Western Europe, 39 percent have at least two owners with more than 10% of the voting rights and 16% have three such owners. Building on the Faccio and Lang (2002) data, we focus on publicly traded 4

6 manufacturing firms. In particular, we examine almost 900 manufacturing firms across 13 countries. The data indicate that the existence of a second and even a third large shareholder is not rare. We find that 35% of all firms and 41% of firms with a controlling owner have at least one additional owner with more than 10% voting rights. Put differently, firms with more than one large owner account for 16% of the market capitalization of all firms in the sample and 31% of the market capitalization of firms with a controlling owner. These findings advertise the importance of considering the impact of other large shareholders on corporate valuations. The second part of the paper finds that large shareholders influence corporate valuations, but this result holds conditional on the characteristics of these owners. Specifically, the mere existence of multiple large owners with more than ten percent voting rights is not related to firm valuations. Rather, it is only when the second large shareholder has sound incentives as measured by cashflow rights or the ratio of cash flow rights to control rights that we observe a positive relationship between the second large owner and corporate valuations. Moreover, we find that the size of the second largest owner relative to the largest owner is important for accounting for cross-firm differences in valuations. Specifically, consistent with theory, we find a non-linear relationship. The cash-flow rights of the second largest shareholder boost corporate valuations when the second largest shareholder s voting rights are not much smaller than the largest owner s. More generally, the analyses show that when a few, similarly-sized large shareholders hold substantial cash flow rights, corporate valuations tend to be higher than ownership structures where (i) cash flow rights are less concentrated or where (ii) cash flow rights are distributed unevenly across large shareholders. In terms of the identity of owners, we do not find that the identity of the second largest owner relative to the largest owner is important once we condition on incentives. 5

7 The paper relates to an expanding empirical literature on corporate ownership around the world. Examining twenty-seven economies and a sample of almost 600 of the largest publiclytraded firms, La Porta et al. (1999) find that two-thirds of large corporations have an owner with 10% or more of the voting rights. 7 In a large sample of Asian corporations, Claessens et al. (2000) show that only 3% of firms lack at least one owner with a minimum of 10% of the voting rights. Faccio and Lang (2002) compile ownership data on more than 5,000 firms in Western Europe. They show that 86% of the companies have an owner with more than 10% of the voting rights. We contribute to this literature by assessing whether companies have more than one large shareholder and whether these other large shareholders influence corporate valuations. We also contribute to studies that examine the impact of the largest shareholder on firm valuations. La Porta et al. (2002) find that cash flow rights concentration is positively associated with firm valuations in a sample of the largest publicly-traded firms in 27 countries. This finding is supported by detailed studies of East Asia (Claessens, Djankov, Fan, and Lang, 2002; Lemmon and Lins, 2003), Korea (Joh, 2003), and emerging market economies (Johnson, Boone, Breach, and Friedman, 2000; Lins, 2003). Our work instead focuses on Western European countries and considers control rights, cash flow rights, the ratio of cash flow to control rights, and the identify of the controlling owner in assessing corporate valuations. We find that the existence of a large controlling owner boosts valuations if the controlling owner has sufficiently large cash flow rights. Moreover, we build on these analyses and assess the impact of other large shareholders. The remainder of the paper is organized as follows. After discussing the data and defining key variables, Section II presents summary statistics that advertise the importance of examining 7 Even in the United States, research demonstrates the importance of concentrated ownership (Eisenberg, 1976; Demsetz, 1983; Morck, Shleifer, and Vishny, 1988; Holderness and Sheehan, 1988; and Holderness, Kroszner, and Sheehan, 1999). 6

8 complex ownership structures. Section III first assesses the impact of the presence, size, identity, and incentives of large owners on the corporate valuations. Section IV concludes. II. Data and the Importance of Complex Ownership Structures To examine the relationship between ownership structure and valuations, we need data on corporate valuations, firm performance, and ownership structure. We use data on almost 900 manufacturing firms across 13 countries in Western Europe (Austria, Belgium, Finland, France, Germany, Ireland, Italy, Norway, Portugal, Spain, Sweden, Switzerland, and the United Kingdom). We start with the Faccio and Lang (2002) database and augment this data in ways that we discuss while defining individual variables. We study manufacturing firms to enhance comparability. This section has two purposes. First, the beginning three subsections define the key variables. These include information on (a) the presence, identity, and size of any stockholders with greater than ten percent of the voting rights of the corporation, (b) the cash-flow rights of the large shareholders, which are inversely related to their incentives to divert corporate resources for private gain, and (c) the market valuation of firms, as defined by Tobin s Q, and firm performance indicators that we use as control variables. Second, we present information on the empirical relevance of studying complex ownership structures. We show that a substantial proportion of manufacturing firms in Western Europe have complex ownership structures. Hence, it is important to examine the potential impact of complex ownership structures on corporate governance. 7

9 A. Control Rights: Presence, Identity, and Size of Large Owners We classify a firm as having a controlling owner if at least one shareholder has direct and indirect voting rights that sum to 10 percent or more. If no shareholder holds 10 percent of the voting rights, we classify the firm as widely held. If the firm has more than one shareholder with control over 10 percent or more of the voting rights, then each of these shareholders is classified as a large shareholder. Since 10 percent voting rights is frequently sufficient to exert control, this cutoff is used extensively (e.g., LLS, 1999; LLSV, 2002). When we use a 20 percent criterion, however, we obtain the same conclusions as those discussed below. While direct ownership involves shares registered in the shareholder s name, indirect ownership involves shares held by entities that the ultimate shareholder controls. Since the principal shareholders of corporations are sometimes financial institutions or corporate entities, we find the major shareholders in these financial institutions or corporate entities. Often, we need to trace this indirect ownership chain backwards through numerous corporations to identify the ultimate controllers of the votes. Thus, to construct data on control rights, we follow LLS s (1999) and LLSV s (2002) procedure. Mechanically, we first identify all major shareholders who control over 5 percent of the votes. We use 5 percent because (1) it provides a significant threshold and (2) most countries do not mandate disclosure of ownership shares below 5 percent. Given these major shareholders, we then begin our search for indirect chains of control. Next, if these major shareholders are themselves (financial or non-financial) corporations, we find the major shareholders of these financial or non-financial corporations. We continue this search until we find the ultimate owners of the votes. For example, a shareholder has x percent indirect control over firm A if she controls directly firm B (i.e., if she holds at least 10 percent of the 8

10 voting rights of firm B) that, in turn, directly controls x percent of the votes of firm A. As another example, a shareholder has x percent indirect control over firm A if she controls directly firm C that, in turn, controls directly firm B, which directly controls x percent of the votes of firm A. The control chain from firm A to firm C can be a long sequence of firms, each of which has control (greater than 10 percent voting rights) over the next one. If there are several chains of ownership between a shareholder and the firm, we sum the control rights across all of these chains. Control-1 equals the control rights of the largest shareholder with control of ten percent or more of the voting rights and zero if the corporation is widely held. Control-2 equals the control rights of the second largest shareholder with control of ten percent or more of the voting rights and zero if the corporation is widely held, or if there is only one shareholder with control of ten percent or more of the voting rights. We create similar variables for the third largest shareholder and so forth. In terms of the identity of each large shareholder (if any), we determine whether the owner is an individual (or family), a bank (or financial institution), or the state. We use this information on the identity of large shareholders in the regression analyses. B. Cash-Flow Rights: Incentives of Large Owners We also compute the direct and indirect cash-flow rights of each of the large shareholders. Shareholders may hold cash-flow rights directly and indirectly. For example, if the large shareholder of firm A holds the fraction y of cash-flow rights in firm B and firm B in turn holds the fraction x of the cash-flow rights in firm C, then the controlling shareholder s indirect cash-flow rights in firm C are equal to the product of x and y. If there is a chain of controlling ownership, then we use the products of the cash-flow rights along the chain. To compute the shareholder s total cash-flow rights we sum direct and all indirect cash-flow rights. 9

11 There can be important differences between cash-flow rights and control rights when there are indirect chains of control. As a simple example, consider a shareholder who owns 10 percent of the voting rights and cash-flow rights of firm A, and firm A in turn holds 20 percent of the voting rights and cash-flow rights of firm B. Assume that this shareholder (i) does not own direct shares in firm B and does not have control or cash-flow rights of firm B through other indirect chains of control and (ii) is the largest equity holder of firm A. In our calculations, this shareholder has 20 percent control rights of firm B because the shareholder controls firm A and firm A has 20 percent of the voting rights of the firm B. This shareholder s cash-flow rights, however, equals 2 percent because the shareholder only receives 2 percent of the firm s dividends (20% * 10%). This may provide perverse incentives. The large owner maybe able to use her controlling position to expropriate, for example, 1$, but she only loses $0.02 in lower dividends. Specifically, Cash-Flow-1 equals the cash-flow rights of the largest shareholder that has control of ten percent or more of the voting rights and zero if the corporation is widely held. Cash-Flow-2 equals the cash-flow rights of the second largest shareholder that has control of ten percent or more of the voting rights and zero if the corporation is widely held, or if there is only one shareholder with control of ten percent or more of the voting rights. We create similar variables for the third largest shareholder and so forth. Besides examining cash-flow rights, we also measure the ratio of cash-flow rights to control rights as a mechanism for examining the incentives of large shareholders to expropriate corporate resources. If an owner has substantial control rights, this increases the ability of the owner to divert corporate resource for private gain. But, this diversion reduces dividends (and may entail other costs as well). If this owner has low cash-flow rights, however, then the costs of diversion are 10

12 lower. Thus, the ratio of cash-flow rights to control rights provides an indication of the costs and ability to expropriate firm resources. Higher values of this ratio signal better incentives, i.e., less of an inclination to divert corporate resources. Specifically, Cash/Control-1 equals the ratio of cash-flow rights to control rights of the largest shareholder that controls ten percent or more of the voting rights. When using this variable in the analyses, we restrict the sample to firms with a controlling owner. Cash/Control-2 equals the ratio of cash-flow rights to control rights of the second largest shareholder that controls ten percent or more of the voting rights. We create similar variables for the third largest shareholder and so forth. C. Valuation and Firm Performance To measure corporate valuations, we use Tobin s Q. Tobin s Q is the traditional measure of valuation and is calculated as the ratio of the book value of assets minus the book value of common equity and deferred taxes plus the market value of common equity to the book value of assets. Table I presents summary statistics by country of Tobin s Q. The average Tobin s Q across the 865 firms in the sample is 1.73, and varies from 1.12 on average in Portugal to 2.28 in Switzerland. To capture recent firm performance, we use the firm s average sales growth rate. We use sales growth rather than earnings growth because earnings are more volatile and sensitive to accounting manipulation. Growth equals the firm s average sales growth rate computed over the past three years. The Tobin s Q and Growth variables are calculated using firm-level data obtained from the Worldscope 11

13 database maintained by Bureau Van Dijk. Worldscope is a commercial database that provides data on large, listed firms around the world. D. Are Complex Ownership Structures Relevant? A prerequisite for a study of complex ownership structures is the existence of complex ownership structures. If publicly-traded firms are overwhelmingly characterized by either completely dispersed ownership or the existence of a single large owner, then this reduces the importance of examining the impact of multiple large owners on firm valuations. However, if many firms have multiple large owners, then it is important to understand the existence of complex ownership structures on corporate governance. Table II indicates that complex ownership structures are relevant. If we simply consider all firms in the sample, the data indicate that 35% of the firms have more than one shareholder with greater than ten percent of the voting rights. Similarly, when considering all firms in the sample, firms with multiple large shareholders account for 16% of market capitalization. This suggests that multiple large shareholders are less common in the largest corporations. Since La Porta et al. (1999) use a sample of the largest firms in each country, this helps explain why their findings on the importance of firms with more than one large owner differ from ours. If we exclude widely-held firms (i.e., if we exclude firms that do not have any owners with ten percent or more of the voting rights), then the existence of complex ownership looks even more important. The data indicate that 41% of firms with at least one controlling owner have multiple large owners. Moreover, firms with more than one large shareholder account for 31% of the market capitalization of firms with a controlling owner. Since over a third of the publicly listed manufacturing firms in Western Europe have more than one large shareholder, we believe this motivates considerably more research than has been 12

14 devoted thus far into better understanding the conditions under which these other large shareholders help or hurt corporate governance. Does the impact of these other large shareholders depend on the size of their voting and cash flow rights relative to those of the largest shareholder? Does the impact of multiple large shareholders depend on the identity of the shareholders, that is, do banks, or families, or the state influence governance differently? Does the impact of large shareholders on valuations depend primarily on voting or cash-flow rights? It is to these questions that we now turn. III. Ownership Structure and Corporate Valuations Given the importance of multiple large owners in Western European firms, the remainder of this paper examines the relationship between ownership structure and Tobin s Q. We first assess the connections between the presence, size, incentives, and identify of the largest controlling owner if any on corporate valuations. While a growing body of evidence provides similar assessments for different samples of firms and countries, we believe this is the first test involving firms in Europe of diverse sizes. The second part of this section turns to a less cultivated area of research: The role of other large shareholders. Does the presence, size, incentives, and identity of other large shareholders, relative to the largest shareholder, influence corporate governance as measured by Tobin s Q? 13

15 A. Valuation and the Largest Controlling Owner Table III presents simple regressions that employ different measures of the size and incentives of the largest shareholder. Thus, the regressions are of the form: Tobin s Q = αx + β (Growth) + γ (Control-1; Cash-Flow-1; Cash/Control-1) + u. The Greek symbols α, β, and γ are coefficient estimates, u is the error term, and X is a matrix that includes a constant term, both country and industry dummy variables, and sometimes a dummy variable for whether the controlling owner holds more than 50 percent of the voting rights. The equation is estimated using Ordinary Least Squares and the parentheses in Table III contain robust standard errors. The first two regressions include the voting rights of the largest shareholder (Control-1) and the cash-flow rights of the largest shareholder (Cash-Flow-1) without controlling for firm growth over the preceding three years (Growth) respectively. Control-1 and Cash-Flow-1 equal zero if the firm does not have an owner with at least 10 percent of the voting rights. The next two regressions control for firm growth. Then, regressions (5) and (6) also control for whether the controlling owner holds more than 50 percent of the voting rights (Majority). Finally, the last three regressions examine the ratio of cash-flow rights to voting rights of the largest owner (Cash/Control-1) to assess whether the incentives facing the largest owner influence corporation valuations. The three regressions differ in terms of the additional control variables. Regressions with Cash-Control-1 only include firms with a controlling shareholder since firms without a controlling shareholder have both cash-flow and control rights equal to zero. The simple message that emerges from the nine regressions in Table III is that there is a strong positive relationship between corporate valuations and the degree to which the controlling owner of a firm has substantial cash-flow rights. Cash-Flow-1 always enters positively and significantly at the five percent significance level, regardless of the other control variables 14

16 (regressions 2, 4, and 6). Thus, when there is a large owner with substantial cash-flow rights, corporate valuations tend to be higher. This is consistent with the argument that higher cash-flow rights reduce incentives to expropriate corporate resources. While Control-1 sometimes enters positively and significantly at the ten percent level (regressions 1 and 3), the results on control rights are not robust to controlling for firm growth and the existence of a majority owner (regression 5). The different results on cash-flow rights and control rights stress the role of incentives: control rights per se do not seem to improve corporate governance, but cash-flow rights are associated with higher valuations. Finally, note that Majority does not enter significantly. This suggests that while having an owner with greater than ten percent voting rights and large cash-flow rights is associated with higher corporate valuations, adding information on whether the owner has a majority stake does not improve the explanatory power of the model. These results are consistent with the common practice of using ten percent voting rights as a benchmark of corporate control. The Table III results emphasize the importance of the incentives of the largest owner, as measured by Cash/Control-1, on corporate valuations (regressions 7-9). As stressed above, a low ratio of cash-flow rights to voting rights creates the ability and incentives to divert corporate resources for private gain. Put bluntly, with a low ratio of cash-flow rights to voting rights, the largest owner has the ability to control corporate decisions (control rights) but is not penalized much by expropriation because of disproportionately low cash-flow rights. Consistent with theory we find a positive relationship between Cash/Control-1 and valuations. The Table III results are robust along a number of dimensions. First, the results hold when including firm growth, which helps control for differences of corporate opportunities. Second, the results hold when including a dummy variable that equals one if the largest owner holds more than 50 percent of the voting rights (i.e., if Control-1 > 0.50). Finally, the results hold when restricting 15

17 the sample to only firms with a controlling owner. Thus, in regressions 7-9 we exclude widely-held corporations. As an additional test, we confirm our results using the Shapley-1 index of the control rights of the largest shareholder. As defined in greater detail in the Appendix, the Shapley-1 index considers the voting rights of other blockholders in measuring the control of the largest shareholder. It ranges from zero (no controlling owners) to one (control rights of 50 percent or more). As shown in Table III regression (10), the Shapley-1 index indicates that firm value increases if the largest shareholder has a larger control stake. While confirming our earlier findings and the predictions by Bennedsen and Wolfenzon (2000) about coalition formation, the Shapley-1 index focuses on voting rights and does not consider cash flow rights. The relationship between corporate valuations and the incentives of the largest owner, as measured by Cash/Control-1, is economically large. If we take the column (7) results in Table III, then the regression results suggest that a one standard deviation increase in Cash/Control-1 (of 0.28) is associated with a 0.24 increase in Tobin s Q. This is substantial considering that the average Tobin s Q for the sample of 729 firms in this regression is B. Do Other Big Owners Matter? We now examine whether the presence, size, identity, and incentives of other big owners other owners with ten percent or more of the voting rights influence corporate valuations. We conduct this assessment using a series of tests that measure different characteristics of big owners and how these traits compare to the largest owner. First, we run simple regressions where we include the cash-flow rights of the second and third largest owners (Cash-Flow-2 and Cash-Flow-3 respectively) along with the cash-flow rights of the largest owner (Cash-Flow-1) and do the same analyses using the ratio of cash-flow to control rights. Second, we consider identity. Do the impact 16

18 of the cash-flow rights (Cash-Flow-2) and the ratio of cash-flow to control rights of the second large owner (Cash/Control-2) depend on the identity of this second owner and whether the identity is different from the largest controlling owner? Third, we study the size of the second large owner relative to that of the largest owner. Does the impact of the second large shareholder on corporate valuations depend on the size of the second owner s voting rights relative to the largest owner s? Finally, we study the comparative incentives of the first and second largest shareholders as measured by Cash/Control. Specifically, while we use Cash/Control-1 and Cash/Control-2 in assessing whether the second large owner influences corporate valuations, this final test examines the comparative value of Cash/Control-1 to Cash/Control-2, i.e., does the impact of a second large shareholder depend on whether the second large owner has better incentives than the largest owner? To examine the impact of complex ownership structures on valuations, we focus on a subset of manufacturing firms with a controlling owner, but without a majority owner. Since we are focusing on how the mixture of big owners influences corporate valuations, it is natural to examine firms with big owners. However, an owner with a majority stake can unilaterally make decisions. This raises two issues regarding the consistency of results across sub-samples. First, do the Table III findings on the impact of the largest owner on corporate valuations hold for this subset of firms? As we show in Table IV.A, the answer is yes. Second, is our conjecture about majority owners eliminating the importance of any other large owners true? The answer is yes. The results we present below on the importance of other big owners vanish when including majority owned firms. In firms with a majority owner, other large shareholders do not influence corporate valuations. Thus, we continue the analysis focusing on firms with a controlling owner, but not with a majority owner. Focusing only on firms with a controlling owner reduces the sample from 859 to 724, while eliminating majority-owned firms reduces the sample to 488 firms. 17

19 B.1. Second and Third Large Owners: Cash-Flow and Cash/Control Ratio Table IV.A indicates that the cash-flow rights and the ratio of cash-flow to control rights of a second large owner are positively associated with corporate valuations. We augment the standard regression as follows: Tobin s Q = αx + β (Growth) + γ (Cash-Flow-1; Cash/Control-1) + δ(cash-flow-2; Cash/Control-2) + κ(cash-flow-3; Cash/Control-3) + u. Using the sub-sample of firms with a controlling owner but without a majority owner, regression 1 in Table IV.A simply confirms the results in Table III: There is a positive relationship between the cash-flow rights of the largest owner and corporate valuations. Next, regressions 2 and 3 consider the influence of other large owners. As shown in regressions 2 and 3, Cash-Flow-2 enters positively and significantly at the ten percent level even when controlling for Cash-Flow-1 and Cash-Flow-3 (the cash-flow rights of the third largest shareholder with more than ten percent of the voting rights). In both of these regressions, the addition of Cash-Flow-2 drives out the independent significance of Cash-Flow-1 compared with regression 1, but the two cash-flow variables enter jointly significantly at the 5% level. The results on Cash/Control are stronger. Again, regression 4 confirms the Table III results using this smaller sample: Larger values of Cash/Control-1 are linked with higher corporate valuations. In regressions 5 and 6, both Cash/Control-1 and Cash/Control-2 enter significantly at the ten percent level and jointly at the one percent level. Furthermore, when only including the Cash/Control ratio of the second largest owner, it enters significantly at the five percent level. While ownership information on the third largest shareholder never enters significantly, the data suggest that the cash-flow rights of the second largest owner and especially the measure of the incentives of the second largest owner (Cash/Control-2) enters positively in the corporate 18

20 valuation equation. The results are consistent with the view that the second large shareholder matters for corporate valuation. B.2. Ownership Structure: Distribution of Shares We also examined the relationship between corporate valuations and the distribution of shares. This provides evidence on ownership structure in general, not simply the existence, size, and incentives of second or third large shareholders. In particular, Table IV.B uses five different measures of the distribution of shares. Control Concentration equals the combined voting rights of the three largest shareholders with voting rights of 10% or more. Control Concentration equals zero if there is no controlling shareholder. If there are only one or two owners with 10% or more of the voting rights, then only the voting rights of these owners are included in the Control Concentration variable. Cash Flow Concentration is defined similarly except that we use the combined cash flow rights of the three largest shareholders with 10% or more of the voting rights. The third measure of the distribution of shares is Control Herfindahl, which is the Hirschman-Herfindahl index of the control rights of all shareholders with voting rights of 10% or more. Similarly, Cash Flow Herfindahl is the Hirschman-Herfindahl index of the cash flow rights shares of all shareholders with 10% or more of the corporation s voting rights. Finally, we use a Shapley index of the power of small shareholders. This Shapley index, unlike the Shapley-1 index used above, measures the degree to which small shareholders are pivotal players in voting among all shareholders. Thus, if there is one majority shareholder, then the Shapley index equals zero. As ownership concentration falls and small shareholders become more important in establishing control rights, the Shapley index increases. The data appendix provides references and more detailed definitions of these ownership structure indexes. 19

21 These aggregate indexes of the distribution of shares confirm and extend our earlier findings. The Table IV.B results indicate that more concentrated ownership by large shareholders is associated with greater corporate valuations. Moreover, the results emphasize the importance of cash flow rights concentration. As shown, the cash flow rights proxies, Cash Flow Concentration and Cash Flow Herfindahl, enter significantly at the five percent significant level (regressions 4 and 5), but the voting rights measures of the distribution of shares only enter significantly at the ten percent level (regressions 1-3). B.3. Instrumental variables We are concerned about potential endogeneity between firm valuations and ownership structure. Valuable firms may select particular ownership structures. Or, some third factor may drive both corporate valuations and ownership structure. While we attempted to ease these concerns by controlling for firm growth and by including industry and country dummy variables, we also use instrumental variables. Table V presents instrumental variable results. We face considerable problems in identifying valid instruments that are sufficiently powerful to explain the nature of the voting and cash-flow rights of both the first and second largest shareholders. We follow Demsetz and Lehn (1985) and use the logarithm of total assets as an instrument for ownership structure. 8 Demsetz and Lehn (1985) argue that optimal ownership structure is firm-specific and depends on the size of the corporation. They note that larger firms are more likely to be widely-held. This, however, provides only one instrumental variable. Thus, we include the ownership structure variables one-at-a-time, 8 We experimented with other instrumental variables. We obtain very similar results when using the lagged value of the log of firm s assets as instrument for the ownership structure variables. Furthermore, we used information on shareholder protection laws from La Porta et al (1998) as an additional instrument. This has the disadvantage of being a country-specific variable and we are examining cross-firm differences in valuation, but including this as an instrument does not change the results or produce a rejection of the overidentifying restrictions. Finally, we used the average values of the ownership variables of all other firms in the country. Again, this produced very similar conclusions. 20

22 so that the first six regressions in each panel use Control-1, Cash-Flow-1, Cash/Control-1, Control- 2, Cash-Flow-2, and Cash/Control-2 respectively. Since Growth may also be endogenously determined, we omit this from the regression, but we obtain the same results on these ownership structure variables when including Growth. The instrumental variable results confirm the earlier findings: Each of the ownership structure variables that include the cash-flow rights of either the largest controlling owner or the second large owner enters significantly at the five percent level. Thus, higher values of Cash-Flow- 1, Cash/Control-1, Cash-Flow-2, or Cash/Control-2 boost market values. This first set of instrumental variable results, however, includes the ownership variables one-at-a-time (regression 1-6). They do not include ownership information on both the first and second largest owners in the same regression. Confirming the OLS results, the instrumental variable results indicate that the exogenous component of Control-2, Cash-Flow-2, and Cash/Control-2 boost market valuation. This holds when controlling for the exogenous component of Control-1, Cash-Flow-1, and Cash/Control-1 respectively. Furthermore, in unreported results, we continue to find that corporate valuations are positively linked with the instrumented component of Control-2, Cash-Flow-2, and Cash/Control-2 when including Control-1, Cash-Flow-1, and Cash/Control-1 directly (i.e., without instrumenting for Control-1, Cash-Flow-1, and Cash/Control-1). The instrumental variable results confirm this paper s conclusion that the second large shareholder matters for corporate valuations. B.4. The Identity of the Second Large Owner Relative to the Largest We next examine whether the identity of the second owner relative to the identity of the largest owner matters for corporate valuation. The intuition is as follows. Large owners may collude, or the second largest owner may exert a check on expropriation by the largest owner. The 21

23 evidence thus far suggest that on average the existence of a second large owner with substantial cash-flow rights relative to control rights is associated with higher market values. But, this aggregate assessment may miss important distinctions across different ownership structures. As noted in the data section, we distinguish whether each owner is an individual/family, a bank, or the state. If the largest and second largest owners have different identities, this may reduce collusion and improve corporate governance. Also, since the largest owner is most frequently a family, it may matter whether the second largest owner is a bank or the state. 9 Specifically, we augment the specification by including an interaction term: Tobin s Q = αx + β (Growth) + γ (Cash-Flow-1; Cash/Control-1) + δ(cash-flow-2; Cash/Control-2) + λ(cash-flow-2*identity; Cash/Control-2*Identity) + u. The variable Identity represents one of three dummy variables: (i) Different equals one if the first and second largest shareholders have different identities and 0 otherwise, (ii) Family-Bank equals one if the first shareholder is a family and the second is a bank and 0 otherwise, and (iii) Family- State equals one if the first shareholder is a family and the second is the state and 0 otherwise. Thus, for the cash-flow rights variable we run three regressions with the three different dummy variables for Identity. Next, for the ratio of cash-flow to control rights, we run three corresponding regressions. In terms of interpreting the results, consider one example with Different as the dummy variable for Identity and Cash-Flow-1 and Cash-Flow-2 as the two ownership structure variables. If a second large owner is particularly beneficial for corporate governance when the second large owner has a different identity from the first, then we should find a positive coefficient on the Cash- Flow-2*Different variable, i.e., λ should enter positively and significantly. If a difference in 9 In many countries, banks hold significant equity stakes in companies (see Gorton and Schmid (2000) for the case of Germany). Because of the private information acquired through their lending relationships, banks are generally in a good position to monitor firms. However, these same lending relationships can lead to a divergence between the interests of the bank and those of other equity holders. 22

24 identity is unimportant, then Cash-Flow-2*Different should enter insignificantly. Table VI presents the results of the six identity regressions. The results in Table VI do not provide support for the view that a difference in the identities of the largest and second largest shareholders matters for corporate valuations. The interaction term never enters significantly. Moreover, even when adding these interaction terms to the valuation equation, the results confirm the earlier findings: incorporating information on the cash-flow rights and the ratio of cash-flow rights to control rights of the second largest shareholders significantly improves the explanatory power of the valuation model. Specifically, Cash-Flow-2 and Cash/Control-2 enter all of the regressions positively and significantly at the ten percent level. B.5. The Size of the Second Large Owner Relative to the Largest In Table VII, we evaluate whether the comparative size of the second owner relative to the largest owner influences the relationship between corporate valuations and the presence of a second large owner. For Finnish firms, Maury and Pajuste (2002) examine the impact of the comparative size of large shareholders. We test this for a broad cross-section of manufacturing firms across Western Europe. Specifically, we run two sets of regression, one that employs Cash-Flow-2 and a second set that uses Cash/Control-2: Tobin s Q = αx + β (Growth) + γ (Cash-Flow-1; Cash/Control-1) + δ(cash-flow-2; Cash/Control-2) + λ(cash-flow-2*size21; Cash/Control-2*Size21) + u. The variable Size21 equals Control-2/Control-1, so that large values imply that the second large shareholder is close in voting rights to the largest shareholder. We first conduct the analyses using Cash-Flow-2 and Cash-Flow-2 interacted with Size21. Then, we examine Cash/Control-2 and Cash/Control-2 interacted with Size21. Again, we emphasize that all of the regressions include country and industry dummy variables. 23