Berle and Means (1932) claimed that ownership structure was widely dispersed, but this is not the case, particularly for firms located outside the US and the UK. There is increasing evidence of concentrated shareholding, with the majority of equity in a firm held by one shareholder or a shareholder group (La Porta, Lopez-deSilanes, & Shleifer, 1999 for developed country firms; Becht & Mayer, 2000 for European firms; Claessens, Djankov, & Lang, 2000 for East Asian developing firms). Indeed, we find that at least 50% of the firms in our sample, drawn from 37 countries, have one shareholder or a shareholder group owning at least 25% of the equity in the firm,1 with the largest shareholder holding more than 30% of total voting shares on average.
The largest shareholder is a quite distinctive class of shareholder. Their investment is sensitive to firm decisions and there are costs associated with maintaining such a large investment in a firm, particularly the cost of under-diversification. But, there are also benefits arising from the influence that such shareholders exercise over the firm (Claessens, Djankov, Fan, & Lang, 2002). While the ownership literature is substantial, there are some gaps in this literature and the role of large shareholders, especially the largest shareholder, is not well developed (Holderness, 2003). In this paper, we focus on the impact of the largest shareholder on firm dividend policy in an effort to gain a better understanding of what role the largest shareholder might play in setting dividend policy.2 Large shareholders could exert pressure on a firm to adopt a dividend policy that reduces private consumption by firm management yet they could also enforce a dividend policy that maximizes private benefits for them at the expense of minority shareholders. It is important to analyse the relationship that exists between large shareholders and dividend policy if we are to better understand the dividend policy decision.
Few studies have examined the effect of the largest shareholder on dividend policy, and these studies are largely single country-based, including Gugler (2003) for Austrian firms and Correia Da Silva, Goergen, and Renneboog (2004) for German firms. Research dealing with the relation between management ownership and dividend policy (Jensen, Solberg, & Zorn, 1992; Rozeff, 1982) is more apparent as is analysis of the relation that exists between institutional ownership and dividend policy (Barclay, Holderness, & Sheehan, 2006; Grinstein & Michaely, 2005; Short, Zhang, & Keasey, 2002). Agency theory suggests that the largest shareholder and dividend policy might be viewed as substitute monitoring devices (Easterbrook, 1984, Jensen, 1986 and Rozeff, 1982).
In examining firm dividend policy, there are two key decisions; (i) whether or not to pay dividends, and (ii) how much to pay. Following Fama and French (2002), we extend the model of Lintner (1956) assuming that the so-called “target” dividend payout rate is a function of firm leverage and investment opportunities. Unlike previous studies, we also examine the potential effects of heterogeneity in the largest shareholder on dividend policy. Different types of the largest shareholder (insider, financial institution or state) differ in terms of their preferences and desires as well as in their ability to influence firm management and hence dividend policy (Gugler & Yurtoglu, 2003; Renneboog & Trojanowski, 2005).
Dividend policy is observed to vary across countries and legal systems (La Porta, Lopez-deSilanes, Shleifer, & Vishny, 2000b) and so we choose cross-sectional data for 8,279 large firms drawn from 37 countries around the world, with annual reports available in 2004. We examine cross-sectional variations in dividend policy, and the impact of largest shareholder on this policy choice. We split the analysis between dividend paying and non-dividend paying firms, as well as specifically allowing for different legal systems (common law vs. civil law). Further, it is possible that while the largest shareholder could influence dividend policy, it is also feasible that dividend policy affects the ownership level that the largest shareholder chooses. For example, Grinstein and Michaely (2005) argue that dividend payout may affect firm ownership. It is feasible that firm management could design a dividend policy that induces the largest shareholder to maintain their ownership stake in the firm.3
We find that firms are more likely to pay dividends when profits are high, debt is low or where investment opportunities are low. This finding is generally consistent with the literature (Fama & French, 2001). It is also found that for firms that choose to pay dividends, the magnitude of dividend payment is increasing in profitability and investment opportunities, and decreasing in debt, consistent with previous findings (Fama & French, 2002; Jensen et al., 1992). We also note the existence of a convex relation between the largest shareholding and dividend payout ratios. It is observed that the shareholding of the largest shareholder is negatively related to dividend payout at low levels of shareholding, yet as the shareholding increases, this relation becomes positive. Where the largest shareholder has a comparatively small interest in the company, there will be a tendency for them to more actively monitor the firm's management, and so there is less need for the use of dividends in controlling agency costs. This is largely consistent with traditional agency theory where it is argued that ownership and dividends provide substitute monitoring devices (Easterbrook, 1984 and Rozeff, 1982). Yet, as ownership levels of the largest shareholder increase, this creates its own set of agency problems, resulting in the need for higher dividend levels and the external monitoring activity that accompanies them.4 Further, consistent with Grinstein and Michaely (2005), we find that the largest shareholding is a function of dividend policy.
We also find that the decision to pay dividends is linked to the identity of the largest shareholder. Firms tend to pay dividends when the largest shareholder is a financial institution, but this is not the case for firms where the largest shareholder is an insider. There is also evidence that dividend payout ratios are lower when the largest shareholder is either an insider or a financial institution. Finally, consistent with La Porta et al. (2000b), we find that that legal origin (common law or civil law) does matter in dividend policy decisions. While common law firms are less likely to pay dividends, they tend to pay higher dividends compared to those in civil law countries.
This paper is structured as follows. Section 2 reviews the literature dealing with the linkage that exists between the largest shareholder and dividend policy. Section 3 describes the data, while Section 4 outlines the empirical models used in the analysis. Section 5 reports and discusses the empirical results. Section 6 concludes.
This study examines the interaction between the largest shareholder and dividend policy across 8,279 firms from 37 countries around the world. Unlike prior studies, this study focuses both on the decision whether or not to pay dividends and on the decision concerning how much to pay. We find that firms in our sample choose to pay dividends and are inclined to pay more dividends when they have high levels of profitability and low levels of investment opportunities, consistent with previous studies such as Fama and French, 2001 and Fama and French, 2002 based on US firms. This is also found to be the case where firms have lower levels of debt.
It is evident that the largest shareholder, a distinctive class of large shareholders, has some influence on dividend policy that the firm adopts. We observe a convex relation between the largest shareholding and dividend payout, and this result is robust even after controlling for endogeneity in ownership variables.26 This reflects the classical agency view that the largest shareholder may act as a substitute for dividends in mitigating agency costs.
Dividend policy is also linked to the identity of the largest shareholder. Firms are fewer prone to pay dividends when the largest shareholder is not an insider, however they tend to pay fewer dividends when the largest shareholder is either an insider or a financial institution. We further note that the interaction between the largest shareholding and dividend payout is a two-way relation. It appears that shareholding can be partly explained by dividend payout. Finally, consistent with La Porta et al. (2000b), we find evidence that dividend policy is also an outcome of legal origin (common law vs. civil law). While common law firms are less likely to pay dividends, they tend to pay higher dividends compared to those dividend paying firms in civil law countries.
One may ask why firms pay dividends rather than repurchases shares since the latter may constitute a cheaper way to return cash to shareholders. Brav et al. (2005) note that widely-divergent views exist among firm executives on whether to distribute firm profits via dividends or repurchases or both. Our analysis would have been more complete if the effect of repurchases was considered but the lack of data ensures that we leave consideration of this question for future research.
