Shareholders’ influence on the decision to pay cash dividend
Evidence from the S&P 500
Abstract
This study investigates the effect large shareholders have on the decision to pay cash dividend. The data used in this paper consists of the firms included in the S&P 500 from the period 2003 to 2009.
The effect large shareholders have on the decision to pay cash dividend has been tested by using summary statistics and logit regressions. As proxies to measure the effect of large shareholders on the decision to pay cash dividend, I use the percentage of shares outstanding and a dividend prone indicator. I find that the largest shareholder decreases cash dividend pay and the second largest shareholder increases cash dividend pay. These results confirm the findings of Gugler and Yurtoglu (2003), who conducted a study on German listed companies. I also find that both the largest and second largest dividend prone shareholder have a positive effect on the decision to pay cash dividend.
JEL classification: G35
Keywords: dividends; payout policy; agency problems; large shareholders
Mohamed Bouazizi 1699474
August 2011
University of Groningen Faculty of Economics and Business
MSC Business Administration
Specialization Finance
1. Introduction
The relation between cash dividend pay and shareholder ownership has generated much interest and controversy in the research literature. Especially if one looks at the apparent decline of cash dividend payers observed by authors as Fama and French (2001) and DeAngelo et al. (2004). Both of them noticed a decline in cash dividend pay in almost the same period. Fama and French (2001), which examined firms listed at the NYSE, AMEX, and NASDAQ, saw that the proportion of listed firms paying cash dividends fell from 66.5% in 1978 to 20.8% in 1999. DeAngelo et al. (2004) researched the firms listed on the same indexes (NYSE, AMEX, and NASDAQ) and found that a smaller proportion of firms with a given level of real earnings paid dividends in 2000 than did so in 1978.
Fama and French (2001) attribute the decline in the proportion of listed firms paying cash dividends mainly to the changing characteristics of publicly traded firms. This is due to the increase in the population of publicly listed firms with low profitability and strong growth opportunities. Publicly listed firms with low profitability and strong opportunities are characteristics of firms that have never paid dividends. The second cause of the decline in the proportion of listed firms paying cash dividends, indicated by Fama and French (2001), is a lower propensity to pay of publicly traded firms.
DeAngelo et al. (2004) agree with the fact that the number of dividend payers decreased, but also see that the aggregate dividends paid by publicly listed firms increased over the past two decades. The reason for these two seemingly contradictory observations is according to DeAngelo et al. (2004) the fact that the reduction in payers occurs almost entirely among firms that paid very small dividends.
Secondly, the increased dividends from the top payers overwhelm the modest dividend reduction from the loss of many small payers. The paper by Von Eije and Megginson (2008) show that there is also a decline in the fraction of firms paying cash dividend in Europe, while total real dividends paid have increased in the period 1989 to 2005.
Another reason for the decline in dividend pay could be the increase of share repurchases.
Fama and French (2001) notice an increase in average share repurchases from 3.37% in the period 1973-1977 to 31.42% in the period 1983-1998 of company earnings. Bagwell and Shoven (1989) mention that the increase in share repurchases show that firms have substituted share repurchases for cash dividends in order to lower the taxation of stockholders. Fama and French (2001) see share repurchases as an unimportant factor of explaining the decline in dividend pay because their research shows that share repurchases are primarily conducted by dividend payers. So in their view share repurchases are used more to increase the cash payouts of dividend payers instead of being used as a substitute of cash dividend.
One factor Fama and French (2001) do not look at is the influence of shareholders on the decision to pay cash dividend as a possible explanation to explain the changes in cash dividend pay.
They are more concerned with company characteristics like size, profitability, and investment
opportunities. This paper will investigate if there is also a relation between the decision to pay cash
dividends and large shareholders. The large shareholders this paper will focus on are the largest and second largest shareholder. In several studies the importance of the effect the largest and second largest shareholder have on dividend pay have been researched in the form of a conflict between large and minority shareholders (Gugler and Yurtoglu, 2003; Maury and Pajuste, 2002). Gugler and Yurtoglu (2003) found that larger holdings of the largest shareholder reduce the dividend payout, while larger holdings of the second largest shareholder increase the dividend payout. Maury and Pajuste (2002) also found that larger holdings of the largest shareholder reduce dividend payout, but the second largest shareholder reduces dividend payout too. It is therefore interesting to find out if Gugler and Yurtoglu (2003) are right in concluding that the second largest shareholder supports minority shareholders in increasing dividend pay, or that the opinion of Maury and Pajuste’ (2002) is correct that dominant shareholders in control may collude to generate private benefits. Another innovative way to research the influence large shareholders have on cash dividend pay is to look at the effect dividend prone shareholders have on the decision to pay cash dividend. Dividend prone shareholders are shareholders found to be often present as the largest or second largest shareholder at dividend paying companies. I measure these by the number of times the shareholder was previously present at dividend paying companies minus the number of times they were shareholders of non- dividend paying companies. The dividend prone shareholders are presumed to induce firms to pay dividends. Specifically, the focus will be on the influence dividend prone shareholders have on the firms’ decision to pay dividends and to discover if there is a significant effect between dividend prone shareholders and cash dividend pay.
The research methods to be used are summary statistics and logit regressions. These two methods were also used by Fama and French (2001), but now with the addition of large shareholders as a variable. Fama and French (2001), divided their sample into dividend payers, non-payers, and firms that have never paid dividends. The focus in this paper will only be on payers and non-payers.
Managing the data in this fashion, the publicly listed companies will be divided into two distinct groups: dividend paying companies and non-dividend paying companies. The data used in the research consists of firms included in the S&P 500 from the period 2003 to 2009. Consistency of the shareholders will be examined each year. As measure of large shareholders’ ownership a minimum of 5% of the voting power, or if this is absent 5% beneficial ownership (direct and indirect ownership of common stock), is used. The reason for this minimum percentage is that the US Securities and Exchange Commission (SEC) compels al listed firms in the US to reveal all shareholders holding of more than 5% of the common shares. Also, the S&P 500 consists only of large American companies (minimum capitalization of US$ 4 billion) and the US is a common-law country where ownership of companies is often widely held (La Porta et al, 1999).
When studying the effect of shareholder ownership on dividend pay an important factor to
consider are the so called ‘agency problems’. Mülbert (2009) divides the agency problem into three
categories. The first category (type I agency problem) is the conflict between shareholders and managers/directors resulting from the separation of ownership and control. The reason for this conflict can be that managers show a higher degree of risk-aversion than diversified shareholders since ‘pure’
managers have most of their wealth (income & stock options) tied up in the firm and, hence, are less diversified. Managers can also be seen by shareholders as underperformers with respect to their managerial tasks because of a preference for empire building, shirking, and/or wasting the company’s assets for personal expenses. The second category (type II agency problem) is the conflict between blockowners and dispersed shareholders. Blockowners can prefer to receive payouts from the firms in the form of private benefits instead of dividends since the latter payments would also benefit other (dispersed) shareholders. Blockowners can often show a comparatively higher degree of risk-aversion since, typically, a higher percentage of their wealth is tied up in the firm. The third category (type III agency problem) is the conflict between shareholders and creditors (bondholders/depositors).
Creditors in general, and bondholders/depositors in particular, are more risk-averse than shareholders given that the former are only interested in their claims being paid back in full on time and, hence, in the firm choosing the least-risky strategy possible, whereas shareholders are interested in a riskier business strategy with a higher expected return. This study will mainly focus on type II agency problems whereby the conflict is between blockowners and dispersed shareholders.
When viewing the difference between cash dividend payers and non-payers using summary statistics, I find that the largest shareholder overall decreases cash dividend pay and the second largest increases it. The largest shareholder also holds a higher percentage of the outstanding shares in non- dividend paying companies than in dividend paying companies. This result is reversed when viewing the second largest shareholder. These findings are confirmed when using logit regressions on the relation between the holdings of outstanding shares by both the largest and second largest shareholder and cash dividend pay. The results validate the assumptions of Gugler and Yurtoglu (2003), and disaffirm the view of Maury and Pajuste (2002) who look at the role of the second largest shareholder as having a negative effect on cash dividend pay. The results concerning dividend prone shareholders show that both the largest and second largest dividend prone shareholder have a positive effect on the decision to pay cash dividend.
This paper adds several new insights to the existing literature. Firstly, there are only a few studies that research the difference in effect the largest and second largest shareholders have on cash dividend pay. Secondly, this is one of the few researches which looks at the relationship between shareholder ownership and cash dividend pay in a common-law country like the US. Other researches involve for example Germany (Gugler and Yurtoglu, 2003) and Finland (Maury and Pajuste, 2002).
Thirdly, the ownership data is collected each year and used over multiple years. Therefore, the
assumption of ownership stability does not have to apply in this paper which improves the validity of
the results. Fourthly and lastly, this is the only research that uses dividend proneness to determine the effect these types of shareholders have on cash dividend pay.
The paper is structured as followed. In section 2 the relevant literature for the paper is going to be reviewed. A description of the data will be given in the section 3. In section 4 the methodology of the paper will be explained. Section 5 will reveal the results of the study. And finally, the conclusion will be presented in section 6.
2. Literature review
Different studies have been conducted to study possible factors influencing dividend pay. The main parts of a couple of these studies will be mentioned in this section. The study will be divided into four parts. The first part will concern articles about changes in dividend payers, seen this has changed in the last 20 years. The second part unravels the several agency problems companies have to cope with when determining whether or not to pay dividend. The third part explains the effect different shareholders have on dividend payout. The fourth and final part will discuss the influence of share repurchases on dividend payout.
2.1 Changes in dividend payers
Fama and French (2001) document a large decline over the period 1978 to 1998 in the number and percent of non-financial and non-utility firms that pay dividends. Their research indicates a change in dividend pay due to changes in the population of firms that are publicly held and due to a reduced propensity to pay dividends by firms. If one looks at the fact that the number of dividend paying industrials has declined by more than 1,000 firms (over 50%) over the last 20-25 years, they suggest that dividends themselves are disappearing.
DeAngelo et al. (2004) notice an increase in aggregate real dividends in the same period. The
explanation they give is that the large reduction in payers occurred almost entirely among firms that
paid very small dividends and that dividends simultaneously increased substantially among the largest
payers. Therefore, the increase in real dividends paid by firms at the top of the dividend distribution
swamps the dividend reduction associated with the loss of many small payers at the bottom. DeAngelo
et al. (2004) call this a ‘two-tier structure’ wherein publicly traded companies are divided in two
groups based on earnings. The first tier contains a few high earners, most of which pay dividends, and
these firms’ dividends collectively dominate the aggregate supply. The second tier contains many
firms which have modest earnings and which collectively contribute little to the aggregate dividend
supply. DeAngelo et al. (2004) confirm in their study the ‘two-tier structure’ because besides large
firms paying most of the aggregate dividends, they too generate a large amount of the earnings, which
increases the dividend and earnings concentration. An example of the importance of earnings is that
DeAngelo et al. (2004) found that 100% of the firms with at least $1 billion in real earnings paid
dividends in 1978, whereas 85.7% paid dividends in 2000. DeAngelo et al. (2004) also discovered that a ‘reduced propensity to pay’ does not automatically mean that dividend paying firms now distribute a lower proportion of their earnings than they used to. They found a flat payout ratio over the period 1978 to 2000, which means that there has not been a reduced propensity to pay in this sense. The flat payout ratio is also according to DeAngelo et al. (2004) an indication that income tax law changes cannot be a cause of the decline in payers.
Chetty and Saez (2005) analyzed the effects of dividend taxation on corporate behavior using the tax cut on individual dividend income enacted in 2003. They found that firms where top executives held more shares and fewer unexercised stock-options were much more likely to initiate dividend payments. This result reveals the importance of top executives’ self-interests in determining corporate responses to taxation. Chetty and Saez (2005) also state that increases in dividends are likely to have efficiency benefits only if total payout changes. They find that share repurchases have risen since the tax cut was enacted. Besides share repurchases, the total payout also rose significantly among the companies that initiated dividends after the tax reform. The explanation for the increase in dividends and repurchases according to Chetty and Saez (2005) is that in a world without taxes and with perfect information, share repurchases and dividends are equivalent. Under the former US tax law, share repurchases were a more efficient way of distributing profits because realized capital gains are taxed less than dividend income. But despite the rise in share repurchases, dividends have remained an important way of distributing profits. The question why dividends have not been replaced by share repurchases is termed the ‘dividend puzzle’ (Black, 1976).
Denis and Osobov (2008) conducted a study where they looked at the dividend policies of companies settled in the US, Canada, UK, Germany, France, and Japan using data from 1989 to 2002.
The conclusion of their study was that the propensity to pay dividends is higher among larger, more profitable firms, and those for which retained earnings comprise a large fraction of total equity. The fact that larger and more profitable firms have a higher propensity to pay coincides with the characteristics of dividend payers mentioned by Fama and French (2001). Denis and Osobov (2008) also discovered that the propensity to pay dividends, in most of the studied countries, declined over the period 1994 to 2002 and that this is driven by a failure of newly listed firms to initiate dividends. The increase in newly listed firms and their characteristics of firms that have never paid (low profitability and strong growth opportunities) have also been observed by Fama and French (2001).
Twu (2010) examines the importance of prior payment status in determining the likelihood to
pay dividends using the signaling and life-cycle theory. DeAngelo et al (2006) explain the life-cycle
theory as a theory proposing that dividends are driven by the need to distribute the firms’ free cash
flow. The theory predicts that firms pay little dividends in the early years to secure investment
opportunities and pay excess cash in the later years to mitigate the possibility that free cash flows
could be wasted. So firms initiate dividend payments when they have the concern of wasting earned
equity. The signaling model states that firms are against reversing their dividend pay decision to avoid unintended information conveyed by dividend changes. Twu (2010) categorizes the sample into firms that paid dividends previously and firms that did not pay dividends. The sample consists of firms from 34 countries during 1990 to 2005 and excludes financial institutions and state-owned institutions. Twu (2010) uses a logit model studying the influence of the key characteristics of the firm, such as growth, profitability, size, ownership structure, risk, and legal protection, on prior payers to continue to pay and prior nonpayers to initiate payment. The results show that strong dividend stickiness exists.
According to Lintner (1956) dividend stickiness means that managers are reluctant to make dividend changes that might have to be reversed. Implying that firms currently paying dividends prefer to continue paying dividends and firms that do not currently pay dividends prefer not to pay dividends.
Twu (2010) also find that when comparing the number of new entrants who pay dividend with those that do not pay dividend over the sample period, the proportion of payers decline from 84% in 1990 to 16% in 2004. This is an indication that there is an increase of firms that do not pay dividends. Another result is that the factors of dividend pay differ for the two groups. Factors influencing dividend pay are for example high growth and low insider ownership. These two factors make priors payers more likely to pay dividend but prior non-payers less likely to pay dividends. Other factors influencing dividend pay found by Twu (2010) are profitability and earned/contributed equity mix. Prior payers are more sensitive to these factors, while nonpayers are more sensitive to the factors risk and dividend premiums. Summing up, prior payers show higher dividend payouts, are larger in size, have more profitability, have larger earned equity, higher insider holdings, and have a greater increase in total assets, but with lower growth. The results of the research conducted by Twu (2010) mentioned above indicate that both the signaling model and the life-cycle theory influence firms’ likelihood to pay dividends. Another interesting result from the study conducted by Twu (2010) is that when taking prior payment status into account, this eliminates overestimating the portion of payers.
2.2 Agency problems
Jensen and Meckling (1976) introduced agency problems in firms. As the ownership of the firm
changes from having one owner to multiple owners, in the form of individual shareholders, Jensen and
Meckling (1976) suggest that managers then increase the consumption of company funds for private
benefits. Thus, managers of a firm tend to extract liquid funds from the firm in the case of increased
separation between ownership and control. Another factor, besides separation between ownership and
control, influencing the tendency of managers to extract liquid funds from the company is the
monitoring of management. Jensen and Meckling (1976) therefore suggest that diverging interests
between managers and shareholders is the result of the separation between ownership and control.
As Easterbrook (1984) and Rozeff (1982) point out, payouts to shareholders reduce the resources under managers’ control, thereby reducing managers’ power. Managers will then be subject to the monitoring of the capital markets, which occurs when the firm must obtain new capital. The capital market monitoring reduces the possibility that funds are invested below the cost of capital.
Monitoring of the capital market therefore helps to ensure that managers act in the best interest of the shareholders. A manager can avoid monitoring of the capital market when projects are financed internally.
Another important study on agency problems is the study done by Jensen (1986). In this study the conflict between the interests and incentives of managers and shareholders are researched. One of these conflicts is whether or not to pay dividends to the shareholders. Jensen (1986) mentions that debt can reduce agency costs of free cash flows. Jensen (1986) calls this effect ‘control hypothesis’ for debt creation. If managers are not hindered in their spending of free cash flow their firms would grow beyond the optimal size. The managers’ power will increase due to the fact that the manager will have more resources under their control. Managers’ compensation is also often linked to the growth of the company (Murphy, 1985). Not to mention that the investments made would be investments in low- return projects or wasted if the company has low growth prospects. To sum up, Jensen (1986) concludes that agency problems can lead to inefficient investment of retained earnings and that distrusting shareholders will often demand dividends. Dividend policy can therefore be used by large shareholders as a corporate governance mechanism to monitor or discipline management. Ownership structure is also an important factor in determining the dividend policy and reducing the expected agency costs.
Faccio et al. (2001) investigate the type II agency problem, whereby expropriation of outside shareholders takes place by the controlling shareholders. Faccio et al. (2001) found that group- affiliated corporations in Europe pay higher dividends than in Asia, and thus dampen insider expropriation. Dividend rates are also higher in Europe, compared to Asia, when there are multiple large shareholders. This suggests that multiple large shareholders decrease expropriation in Europe, but increase it in Asia. Faccio et al. (2001) find that the systematic expropriation of the outside shareholders is done at the base of extensive corporate pyramids. The controlling shareholders can therefore extract returns from projects that yield negative returns to the corporation. Faccio et al.
(2001) consider paying out dividends as a means of limiting insider expropriation because it removes
corporate wealth from insider control. This view is shared by La Porta et al. (2000), who report that
higher dividends are paid by corporations in countries with strong legal protection of minority
shareholders. Countries with strong legal protection have codes based on common-law rather than
civil-law. Dittmar et al. (2003) find that firms in countries with poor legal protection have twice as
much cash in their firms than the firms in countries with good legal protection. The results therefore
imply that there is a positive relationship between dividends and corporate governance in terms of
country-based mechanisms. Faccio et al. (2001) also find that significantly higher dividends are paid by corporations that are ‘tightly affiliated’ to a business group via a chain of control that comprises at least 20% of the control rights at each link. La Porta et al (1998) find evidence that ownership concentration is due to poor legal protection, whereby shareholders are well protected in common-law countries and concentrated ownership is therefore less urgent. In the opposite case where shareholders are less well protected, in civil-law countries, ownership is more concentrated to increase the monitoring of managers.
2.3 Institutional holdings and dividend payout
Bagwell and Shoven (1989) mention in their research the influence of asymmetric information on dividend pay and share repurchases. A payout in the form of dividends or share repurchases may contain information about the status of the company. The investors can see an unexpected dividend pay or repurchase of shares as a signal of an increase in firm value because the investor sees management as being better informed than the market place. The good news revealed by dividends or share repurchases can therefore include unobserved firm value, unobserved current cash flow, or unobserved future cash flow. Bagwell and Shoven (1989) mention that research has confirmed that unexpected changes in dividends provide information about unexpected increases in the level of earnings in addition to the information found in earnings announcements.
Short et al. (2002) used a UK panel data set to examine the role of institutional ownership in association to dividend payout ratios by analyzing this relationship within the context of several dividend models (Lintner, 1956, Waud, 1996, and Fama and Babiak, 1986). Their results produce strong support for the hypothesis that there exists a positive association between dividend payout policy and institutional ownership. Short et al (2002) also find a positive earnings trend component to the association between institutional ownership and the dividend payout ratio. Additionally, Short et al. (2002) find evidence supporting the hypothesis that a negative association exists between dividend payout policy and managerial ownership.
Maury and Pajuste (2002) research the effect ownership and control structures have on dividend policy in Finnish listed firms. They find that the dividend payout ratio is negatively related to the control stake of the controlling shareholder. Surprisingly, they also find that the presence of another large shareholder also influences the payout ratio negatively. This finding contradicts the assumption of a positive monitoring role by another large shareholder, as proposed by Faccio et al.
(2001) for European firms and Gugler and Yurtogly (2003) for German companies. Maury and Pajuste
(2002) conclude that the dominant shareholders in control collude to generate private benefits of
control which are not shared with minority shareholders. An indication of wealth extraction of
minority shareholders is a lower dividend payout level. This is an example of agency problems in
dividends whereby a firm’s profits are not paid out as dividends; instead corporate managers may
divert the cash flow for personal benefits or pursue unprofitable investment opportunities. Therefore Lang and Litzenberger (1989) see dividend payouts as a means to reduce the free cash flow managers can use at their own discretion. Maury and Pajuste (2002) divide the agency models on dividends into two groups. The first group considers dividend payouts as an outcome of agency conflict between managers and shareholders, or large shareholders and minority shareholders (La Porta et al., 2000;
Faccio et al., 2000). With ‘outcome of agency conflict’ is meant that dividend payouts are an outcome of legal shareholder protection. La Porta et al. (2000) for example find that companies in common-law countries pay higher dividends than civil-law countries, because minority shareholders can use the legal system of common-law countries to force companies to payout dividends. The second group sees dividend payout policies as substitutes for governance problems in a company by lowering the amount of free cash flow a manager has at its discretion (Lang and Litzenberger, 1989). By lowering the amount of free cash flow a manager has at his or her discretion, the more the manager has to rely on external capital markets to raise funds. For the manager to be able to raise funds with reasonable low costs, the controlling shareholder or manager must establish a reputation for not expropriating outside investors (Gomes, 2000). Maury and Pajuste (2002) find empirical evidence in favor of the outcome agency model of dividends.
Gugler and Yurtoglu (2003) study the large-small shareholder conflict by analyzing dividend
announcements and dividend pay-out ratios in Germany over the period 1992 to 1998. They find
larger holdings of the largest shareholder to reduce dividend payout ratio, while larger holdings of the
second largest shareholder increases the dividend payout ratio. Gugler and Yurtoglu (2003) therefore
conclude that the presence of a second largest shareholder with a considerable large equity stake
influences the governance of a firm in a positive way. This conclusion confirms the theory that due to
the reduction of the free-rider problem of monitoring and/or increased alignment of incentives, large
shareholders can potentially add value. The risk in this case is that the large shareholders may increase
the size of the shared pie, but the private benefits of control that are not shared with minority
shareholders may also increase. Another interesting finding of Gugler and Yurtoglu (2003) is that
deviations from the one-share-one-vote rule, due to for example pyramidal and cross-ownership
structures, are associated with lower pay-out ratios. This is because dividends are received in
proportion to cash flow rights and control is established by voting rights. A separation between the
cash flow rights and voting rights gives rise to the incentive and ability to search for other forms of
compensation than in dividends. Gugler and Yurtoglu (2003) divide their sample into majority-
controlled firms and minority-controlled firms and into ‘unchecked’ firms and ‘checked’ firms. Gugler
and Yurtoglu (2003) define majority-controlled firms as firms whereby the largest shareholder
controls more than 50% of the voting shares and minority-controlled firms as firms whereby the
largest shareholder controls less than 50% of the voting shares. The definition of an ‘unchecked’ firm
is a firm with a second largest shareholder that holds less than 5% of the voting shares and a ‘checked’
firm is a firm with at least another shareholder than the largest shareholder that has more than 5% of the voting shares. Gugler and Yurtoglu (2003) expect that the large-small shareholder conflict will be severe in majority-controlled firms and ‘unchecked’ firms. The conflict will be most severe in both
‘unchecked’ and majority controlled firms. They expect this result because the risk of expropriation is higher in majority-controlled firms. Secondly, the absence of a second large shareholder above 5%
provides no check on the largest shareholder expropriating wealth from the minority shareholders.
This in turn will have negative effects on dividend payouts. Gugler and Yurtoglu (2003) found in their research confirmation in what they expected. They found that majority-controlled and ‘unchecked’
firms have the smallest payout ratio, while majority-controlled and checked firms have the largest target payout ratio.
Grinstein and Michaely (2005) examine the relationship between institutional holdings and payout policy in US public firms. They mention three reasons why ownership structure and payout policy may be related. The first reason is that agency theories suggest that managers are likely to share more of the profits with the investors in the case of improved monitoring of managers. Grinstein and Michaely (2005) support this argument by referring to Jensen (1986), who argues that with enhanced monitoring, firms are more likely to payout their free cash flow. Grinstein and Michaely (2005) hereby assume that institutions are better monitors and that these theories imply that larger institutional holdings will lead to higher payouts. For the second reason, Grinstein and Michaely (2005) refer to Allen et al. (2000) who argue that to increase value, firms want institutions to monitor or to facilitate takeovers. Grinstein and Michaely (2005) mention that institutions prefer dividends because of common institutional charter and prudent-man rule restrictions, and because of the comparative tax advantages that some institutions have for dividends. So, higher dividends will lead to larger institutional holdings. The third reason is the adverse selection problem, whereby uninformed investors prefer dividends over repurchases (Barclay and Smith, 1988; Brennan and Thakor, 1990).
Grinstein and Michaely (2005) state that informed shareholders do not face this problem and so they prefer stock repurchases, which is the least costly payout for them. Better informed shareholders therefore prefer firms that payout in the form of repurchases rather than in the form of dividends. The research done by Grinstein and Michaely (2005) reveals that institutions prefer dividend-paying over non-dividend-paying stocks. Grinstein and Michaely (2005) also find that institutions prefer low- dividend stocks to high-dividend stocks and that firms increasing their dividends do not attract more institutional investors. Another finding of Grinstein and Michaely (2005) is that institutions prefer firms that repurchase shares as opposed to firms that do not repurchase their shares.
Mancinelli and Ozkan (2006) researched the relationship between dividend policy and
ownership structure for a sample of Italian firms. The sample consists of 139 listed Italian firms and
was conducted with data from 2001. They find that the voting right of the largest shareholder has a
significantly negative impact on the dividend payout ratio. Mancinelli and Ozkan (2006) predicted this
result because they hypothesized that a higher level of concentration of ownership, measured in voting rights of the largest shareholder, leads to a higher probability of expropriation by large shareholders at the cost of minority shareholders. Expropriation by large shareholders can occur by using a method called tunneling. With tunneling, the largest shareholders not only have a majority stake in the firm, but also possess a large part or have complete ownership of the cash flow rights of other businesses that have relationships with the firm (Shleifer and Vishny, 1997).
2.4 Repurchases
Modigliani and Miller (1961) argue that a firm’s financial policy, paying dividend or share repurchases, is irrelevant because it has no effect on the firm’s value. If the firm chooses a share repurchase, each shareholder can sell sufficient shares to match the amount they would have received if dividends had been paid out. Therefore, dividends and share repurchases can be seen as substitutes of one another.
Bagwell and Shoven (1989) mention in their article the possible advantages of repurchases compared to dividends. They indicate that the advantage share repurchases have over dividends is that the taxation of share repurchases at the personal level are lower (at that time) than on dividends.
Investors receiving cash after a repurchase of their shares are taxed on their capital gain. But only increases above the basis value are treated as a realized capital gain and subject to taxation. A second advantage of repurchases is that when outstanding shares are repurchased, the price of the remaining shares will be higher than if an equivalent amount of cash had been distributed as a dividend. The higher price on the remaining shares is accrued capital gains whereby the taxes can be deferred until the gain is realized by selling the shares. A third advantage of using repurchases instead of dividends is that firms have also learned how to use repurchase as part of an anti-takeover strategy, and firms and shareholders have recognized that repurchases can be a less costly signal of favorable news than dividends. Dividends namely creates a higher expectancy of investors to continue the same level of dividends in future years than repurchases, also named dividend stickiness by Lintner (1956). Bagwell and Shoven (1989) support their arguments by showing that in 1977 dividends was the primary mechanism for 80% of the companies. This percentage had fallen to 40% in 1986.
A possible explanation for a lower propensity to pay is share repurchases. Fama and French
(2001) show that repurchases are largely carried out by payers, which leaves the decline in the percent
of payers unexplained. They conclude that repurchases are mainly used to increase earnings payouts of
dividend payers. Skinner (2008) states that, ‘Because managers use repurchases to pay out earnings
increases, this helps to explain why dividend policy becomes increasingly conservative’. That the
importance of repurchases is increasing in Europa can be shown from the article of Von Eije and
Megginson (2008), which state that the fraction of companies repurchasing shares has steadily
increased. They study cash dividend payments and share repurchases over the period 1989 to 2005. A
cause of the increase in share repurchases could be that dividends are being replaced by share repurchases. This is because Von Eije and Megginson (2008) notice that the fraction of European firms paying dividends declines, while total real dividends paid increase. Share repurchases account over half of the total value of cash dividends. Von Eije and Megginson (2008) therefore conclude that the propensity to repurchase shares has steady increased. They also find that certain company characteristics like size, market-to-book, and profitability explain the likelihood to pay dividends or repurchase shares in the US (as was shown by Fama and French (2001)) also influence the likelihood that EU firms will distribute cash to investors. Older EU firms for example are more likely to pay cash dividends than younger firms and older firms also pay higher cash dividends. Whereas larger cash holdings, an important source of agency problems, reduces the probability of paying cash dividends, but increases the probability of share repurchases. Another interesting finding of Von Eije and Megginson (2008) is that companies with a majority shareholder repurchase less. In comparing the amount of repurchases carried out in the US compared to the EU, fewer EU than US firms repurchase shares. But both the EU and the US see dividend payments and share repurchases as complements and also share repurchases are significantly more sensitive to earnings than dividend payments.
Hypotheses
Mainly based on the literature, the following research questions were formulated:
1. What are the characteristics of the sample of cash dividend payers and non-payers?
2. What role do large shareholders play in the decision of firms to pay cash dividend?
3. Do dividend prone shareholders positively influence the decision to pay cash dividend?
The first research question focuses on the difference between cash dividend payers and non-payers. As
Fama and French (2001) found, non-payers often show low profitability and strong growth
opportunities, while payers are more profitable and larger in size. The second research question tries to
find an answer to the question what effect large shareholders have on the decision of firms to pay cash
dividend. As the literature suggests, the largest shareholder often lowers dividend payout, whereas
another (second) large shareholder often increases dividend payout because of its monitoring function
on the largest shareholder (Gugler and Yurtoglu, 2003). There is of course a risk that the large
shareholders collude to extract private benefits and thereby lower dividend payout (Maury and Pajuste,
2002). The third research question looks at the positive effect dividend prone shareholders may have
on the decision to pay cash dividend. Dividend prone shareholders are shareholders found to be
present as the largest or second largest shareholder in dividend paying companies. The more often a
shareholder is found to be present as the largest or second largest shareholder at a dividend paying
company the higher the dividend proneness of that particular shareholder will be.
In order to answer the research questions the following hypotheses are presented:
H0a: There is no relation between the decision to pay cash dividend and the largest shareholder.
H1a: There is a negative relationship between the decision to pay cash dividend and the largest shareholder.
Both researches carried out by Gugler and Yurtoglu (2003) and Maury and Pajuste (2002) find that the largest shareholder reduced the dividend payout ratio. An explanation for their finding is that the largest shareholder extracts private benefits at the cost of minority shareholders, and thereby reduces the dividend payout ratio. Therefore a negative relationship between cash dividend pay and the largest shareholder is presumed.
H0b: There is no relation between the decision to pay cash dividend and the second shareholder.
H1b: There is a positive relation between the decision to pay cash dividend and the second shareholder.
Gugler and Yurtoglu (2003) find in their research that another large (second) shareholder increases the dividend payout ratio due to its monitoring function on the largest shareholder. Faccio et al. (2001) also find in their research that multiple large shareholders decrease expropriation of minority shareholders in Europe. Therefore, a positive relation is presumed between cash dividend pay and the second shareholder is expected.
H0c: There is no relation between the decision to pay cash dividend and dividend prone shareholders.
H1c: There is a relation between the decision to pay cash dividend and dividend prone shareholders.
The dividend prone shareholders were selected on the basis of being often present at dividend paying companies. It is interesting to see if the dividend proneness of these shareholders remains the same after subtracting the number of non-dividend paying firms the dividend prone shareholders are present at as largest or second largest shareholder. By taking into account the non-dividend paying firms, a better measure of dividend proneness is used to predict dividend pay.
3. Data description
The research I conduct uses data from 2003 to 2009. Because I also use one year lagged
variables (as will be explained later on) the regressions will be carried out over the period 2004 to
2009. The reason I started collecting data in 2003 is that from 2003 onward the data is more
comprehensibly collectible from the SEC proxy statement than in previous years. The collection of
data has been halted in 2009 because at the moment of data collection, the data on dividend payments
and shareholder ownership is no longer fully available after this period. The company data for this research were extracted from Thomson Reuters DataStream. The data concerning ownership of company stock was collected from proxy statements filed with the SEC. The SEC compels all listed firms to reveal their shareholders if the shareholder holds at least 5% of the common shares of voting power or beneficial ownership. With beneficial ownership is meant that the shareholder beneficially owns common shares not only by holding the shares directly but also indirectly (such as through a relationship, a position as a director or trustee, or a contract or understanding) and has or shares the power to vote, sell or acquire the shares within 60 days.
The sample consist of dividend and non-dividend paying firms included in the S&P 500. The S&P 500 is made out of large publicly held companies that trade on the New York Stock Exchange or the NASDAQ. To be included in the S&P 500 the publicly held companies have to be: (i) US companies, (ii) have a market capitalization in excess of US$ 4 billion, (iii) have at least four consecutive quarters of positive reported earnings, and (iv) the shares must have a public float of at least 50%.
1Firms without recorded shareholders or with missing variables are dropped from the sample of S&P 500 companies. At least the largest, and preferably also the second largest shareholder, must be known each year for every company over the period 2003 to 2009. Therefore, consistency of the shareholders is examined each year. This is different from Gugler and Yurtoglu (2003) who assume ownership stability and look at the ownership structures of companies every 3 years over the period 1991 trough 1998. They examine ownership structures of companies every 3 years instead of yearly because they presume that the ownership structure of German companies is very stable. By registering the possible ownership changes each year this research does not have to assume that the ownership structure is stable in the US. Shareholder information must be available each year separately from 2003 to 2009 or else the firm in that specific year will be dropped from the sample.
Also, financial firms (SIC code between 6000 and 6999) and utility firms (SIC code between 4900 and 4949) will be excluded from the sample. The reason for this is that these companies face additional regulations and hence might have a different payout behavior. The historical dividend patterns of utilities and financials are also quite different from other companies. The data elements from Thomson Reuters DataStream that are being used are: earnings before interest and tax (EBIT), total assets, total cash dividend paid, total debt, market capitalization, treasury stock, sale of common and preferred stock, and purchase of common and preferred stock. If the data of one of the mentioned variables is missing and therefore makes the calculation of the regressions impossible, the firm is removed from the sample. Also outliers have been removed from the sample to give a better representation of the data. An example of outliers removed is firms with shareholders with more than 60% of the voting power or beneficial ownership. These firms have been removed from the sample because the US is a common-law country whereby ownership of companies is often widely held (La Porta et al, 1999).
1http://www.standardandpoors.com
Having firms in the sample with ownership above 60% does not give a correct view of ownership in the US. The total number of firms excluded from the sample due to outliers is about 2% of the entire sample. Table 1 gives a detailed description of the sample construction for every year.
Table 1
Construction of the entire sample
Listed firms are all the firms included in the S&P 500. No recorded shareholders are firms that do not have any shareholder holding at least 5% of the outstanding shares. Financial and utility firms are the firms excluded from the sample on the basis of their SIC code. The SIC code for financial firms lies between 6000 and 6999. The SIC code for utility firms lies between 4900 and 4949. Removed outliers are firms that have a shareholder owning shares above 60% of the outstanding shares.
2003 2004 2005 2006 2007 2008 2009
Listed firms 500 500 500 500 500 500 500
No recorded shareholders -102 -98 -85 -72 -56 -59 -52
Financial & utility firms -87 -83 -92 -97 -102 -105 -104
Missing control variables -11 -8 -12 -7 -9 -6 -8
Removed outliers -7 -6 -6 -6 -7 -6 -5
Number of firms in the sample 293 305 305 318 326 324 331
Table 2 shows the percentage of payers and non-payers each year. The proportion of dividend paying firms increases with 7% between 2003 and 2004. After 2004, the proportion of dividend payers steadily increases from 67% in 2004 to 71% in 2008 and ends 1% percent lower in 2009. This increase in the percentage of dividend paying companies is contrary to the findings of Fama and French (2001) and DeAngelo et al. (2004), which observed a decline in dividend pay and it may be due to the dividend tax cuts starting in 2004.
Table 2
Number of firms with large shareholders (above 5%) and data on dividend pay
Payers are firms that have paid out cash dividends in year t. Non-payers are firms that have not paid out dividends in year t.
2003 2004 2005 2006 2007 2008 2009 2003 – 2009
All firms 293 305 305 318 326 324 331 2202
Payers 177 203 208 219 231 230 231 1499
Non-payers 116 102 97 99 95 94 100 703
Percent of Payers
0.60 0.67 0.68 0.69 0.71 0.71 0.70 0.68
Perfect of non- Payers
0.40 0.33 0.32 0.31 0.29 0.29 0.30 0.32
The dependent variable for the logit regressions is a dummy representing dividend pay or non-
dividend pay of a company. If the company pays dividend the dependent variable will have a value of
1 and if the company does not pay dividend the dependent variable will have a value of 0. The control
variables for this research have been collected over the period 2002 to 2009. The variables concerning
ownership (independent variables) have been collected over the period 2003 to 2009. All the independent variables have been lagged one year.
Variables concerning ownership first of all consist of the percentage of outstanding shares held by the largest and the second largest shareholder. The percentage of outstanding shares is presented in beneficial ownership or in voting power of common shares. The sample only includes shareholders that hold at least 5% of the shares in beneficial ownership or voting power. Percentages below 5% do not have to be registered by the SEC in the company’s yearly proxy statements. The percentages of ownership of the largest and the second largest shareholder have namely been collected from these yearly proxy statements. I aggregated the individual shareholders as much as possible to be able to find the real controlling shareholder. An example of a large shareholder that has multiple subsidiaries is Capital Group Companies. Capital Group Companies has 7 large subsidiaries, namely: (i) Capital Guardian Trust Company, (ii) Capital International Research and Management, Inc., (iii) Capital International Limited, (iv) Capital International S.A, (v) Capital Research Global Investors, (vi) Capital World Investors, and (vii) Growth Fund of America, Inc. The information concerning the shareholders subsidiaries or majority ownership in other shareholders is often found in the footnotes of the SEC proxy statements. Shares held by the largest shareholder (Share 1e) are presumed to have a negative effect on the decision to pay cash dividend due to private benefits of control that are not shared with minority shareholders. In contrast, shares held by the second largest shareholder (Share 2e) are expected to positively influence the decision to pay cash dividend because of the second largest
shareholders’ monitoring function of the largest shareholder (Gugler and Yurtoglu, 2003). The variables concerning the percentage of outstanding shares held by both the largest and second largest shareholder has been lagged one year to prevent endogeneity.
The second independent variable is ‘dividend prone’. Dividend prone shareholders are
assumed to have a positive effect on the decision to payout cash dividend. Dividend prone
shareholders are shareholders found to be often present as the largest or second largest shareholder in
dividend paying companies. If a shareholder is more often present as the largest or second largest
shareholder at a dividend paying company than as the largest or second largest shareholder of a non-
dividend paying company, the dividend proneness of this particular shareholder will be higher. To
calculate the exact figure of dividend proneness, I subtract the number of non-dividend paying
companies from the dividend paying companies each year where a frequently registered shareholder is
the largest or second largest shareholder. It can therefore also be possible that dividend proneness is
negative if the shareholder is present at more non-dividend paying companies than at dividend paying
companies. Dividend proneness is lagged one year in the sample because the pressure of a dividend
prone shareholder on a company can be better measured using the dividend payout one year after the
dividend prone shareholder has become involved in the company. Dividend prone shareholders that
were present one year ago as the largest shareholder are referred to as ‘Div. Prone 1’ and dividend
prone shareholders that were present one year ago as the second largest shareholder will be referred to as ‘Div. Prone 2’. Both dividend prone shareholders (Div. Prone 1 and Div. Prone 2) are expected to have a positive influence on the decision to payout cash dividend. To view the list of dividend prone shareholders and the calculation, I refer to appendix A. This list shows that the dividend prone shareholders are not constant each year but change, as can be seen in table 3. Table 3 shows that the change in dividend prone shareholders does not decline in percentages each year. The reason could be that the number of dividend prone shareholders is not constant each year but changes from 16 dividend prone shareholders in 2003 to 19 dividend prone shareholders in 2004.
Table 3
Percentage change in dividend prone shareholders each year for the cohorts of 2003-2008
Div. prone shareholders of 2003
Div. prone shareholders of 2004
Div. prone shareholders of 2005
Div. prone shareholders of 2006
Div. prone shareholders of 2007
Div. prone shareholders of 2008
Div. prone shareholders of 2009 Dividend prone shareholders
of 2003 82.35% 70.59% 70.59% 64.71% 76.47% 76.47%
Dividend prone shareholders
of 2004 83.33% 83.33% 77.78% 72.22% 72.22%
Dividend prone shareholders
of 2005 84.21% 78.95% 73.68% 68.42%
Dividend prone shareholders
of 2006 90.00% 80.00% 70.00%
Dividend prone shareholders
of 2007 81.82% 81.82%
Dividend prone shareholders
of 2008 85.71%
I also include five control variables. The five control variables are: (i) earnings before interest
and tax over total assets (E
t/A
t), (ii) change in total assets (dA
t/A
t), (iii) market-to-book ratio (V
t/A
t),
(iv) size, and (v) share repurchases. Four of these control variables (E
t/A
t, dA
t/A
t, V
t/A
t, and size) are
the same control variables used by Fama and French (2001) in calculating logit regressions and
summary statistics. The first control variable earnings before interest and tax, or EBIT, over total
assets is expected to have a positive relationship with dividend pay as found by Fama and French
(2001) and Twu (2010). Fama and French (2001) and Twu (2010) found that dividend payers have
higher measured profitability than non-payers. Also Von Eije and Megginson (2008) found
profitability influencing the likelihood that EU firms will distribute cash to investors. Profitability
(E
t/A
t) will be measured by dividing the earnings before interest and tax (DataStream item WC18191)
over total assets (DataStream item WC02999). Change in total assets (dA
t/A
t) is included as the second
control variable and is an indicator for investment opportunities. The control variable dA
t/A
tshows the
growth rate in assets by dividing the change in total assets (total asset t – total assets t-1) by total
assets t. In the study of Fama and French (2001) growth rates of assets of payers are lower than the
growth rate of non-payers. DeAngelo et al. (2006) also predict that firms pay little dividends in the
early years to secure investment opportunities. Only in later years the firm pays out dividends when
there are less investment opportunities. Twu (2010) additionally found that dividend payers show les growth than non-dividend payers. Based on the results of Fama and French (2001) and Twu (2010) a negative relationship is expected between the growth rate in assets and dividend pay. The third control variable is market-to-book ratio (V
t/A
t) as a measure of investment opportunity. Market-to-book ratio, consists of total debt (DataStream item WC03255) plus market capitalization (DataStream item WC08001) divided by the same years’ total assets (DataStream item WC02999). This additional method of calculating investment opportunities is chosen because firms that never pay dividends have the best growth opportunities, as Fama and French (2001) study points out. As is the case with the control variable change in total assets (dA
t/A
t), I expect a negative relation between the market-to-book ratio (V
t/A
t) and dividend payout. The fourth control variable is size. The control variable size has been calculated by taking the logarithm of the market value, or market capitalization, of the firm. Fama and French (2001) show in their study that dividend payers are often larger in size than non-dividend payers. A reason mentioned by Fama and French (2001) is that more mature firms are often larger in size, and therefore spend less cash on investment opportunities than younger firms, that are consuming their cash to grow the firm. A more mature and larger firm will consequently have more cash on hand to pay dividends. Twu (2010) also found in her research that dividend paying firms are often larger in size than non-dividend paying firms. I therefore expect there to be a positive relationship between the size of a firm and the decision of a firm to pay dividends.
The last control variable is the dummy variable repurchases. To be exact, I use net repurchases
computed in the same manner as Fama and French (2001) and Skinner (2008). Net repurchases
involves removing from the share repurchases the effect of shares issued for employee stock option
programs, to fund acquisitions, and for other corporate purposes. The approach involves the treasury
stock method for repurchases when there is a change in common treasury stock. But if the treasury
stock is zero in the current and prior year, the retirement method is applied. With the retirement
method net repurchases are calculated by the difference between stock repurchases and stock
issuances. The treasury stock method is applied first, rather than the retirement method, because the
change in treasury stock nets out any associated issuances. I apply the variable repurchases as a
dummy variable to smooth out the large differences in share repurchases between the several
companies. The dummy variable share repurchases will have a value of 1 if the outcome of the
calculation is positive, and a value of 0 if the calculation had a negative or zero outcome. Share
repurchases have been lagged to prevent an interaction between share repurchases and cash dividend
pay, as these two factors can influences one another. Fama and French (2001) show that repurchases
are largely carried out by payers and therefore conclude that repurchases are mainly used to increase
earnings payouts of dividend payers. I also expect there to be a positive relationship between dividend
payout and share repurchases.
Table 4 provides some descriptive statistics for the firms in the sample each year over the period 2003 to 2009 and in aggregated figures over the entire period. The first and second variable shown in table 4 are the percentage of outstanding shares held by the largest and the second largest shareholder as Share 1e and Share 2e. Comparing these two variables reveals that the average (median) percentage of shares held by the largest shareholder is about the 13% (10%) and 9% (8%) for the second largest shareholder. Both medians are above the largest shareholder’s median stake of votes of 5% in US firms. A common-law country like the US has a lower concentration of votes than the median control stake of 44% as an average of eight continental European countries
2. The median size of the second largest shareholder in the continental Europe is 9%, which is 1% higher than the median percentage of shares of the second largest shareholder in the sample. These results of a lower ownership concentration in the US compared with continental Europe confirms the findings of La Porta et al. (1999), who suggest that in common-law countries like the US, ownership is often widely held, and therefore lower. Secondly, table 4 displays relative stability of ownership concentration over the period 2003 to 2009. This puts forward the premises of ownership stability mentioned by Gugler and Yurtoglu (2003) in their research of German companies. Thirdly, Gugler and Yurtoglu (2003) distinguished between ‘unchecked’ and ‘checked’ firms. The definition of an ‘unchecked’ firm is a firm with a second largest shareholder holding less than 5% of the voting shares and a ‘checked’ firm is a firm with at least another shareholder than the largest shareholder that has more than 5% of the voting shares. The importance of ‘unchecked’ and ‘checked’ firms is that the conflict between the largest and second largest shareholder is more severe in unchecked firms. Firms marked as
‘unchecked’ will therefore have a lower dividend payout ratio, as found by Gugler and Yurtoglu (2003). This research only has ‘checked’ firms when another large shareholder is present because only shareholders holding at least 5% of the outstanding shares are reported in the proxy statements. In my research 77% of the firms in the sample are ‘checked’ firms, and consequently, a higher dividend payout ratio can be assumed if the assumptions of Gugler and Yurtoglu (2003) are correct. The third and fourth variable shown in table 4 present figures on the variable Div. Prone 1 and Div. Prone 2.
The largest dividend prone shareholder is represented as Div. Prone 1 and the second largest dividend prone shareholder as Div. Prone 2. The average dividend prone of both variables starts low in 2003 with 0.01, but then remains quite stable around 0.05. The reason for a low average dividend prone in 2003 compared to the overall average could be the dividend tax cut in 2004. Also the large shareholder FMR, most often present at dividend paying companies as largest or second largest shareholder (38 times) in 2003, has a negative dividend prone figure of -4.44%, as shown in appendix A. Div. Prone 1 is on average almost higher each year compared with Div. Prone 2, suggesting that the largest shareholder is more often at dividend paying companies than the second largest shareholder. The remaining variables of table 4 consist of the five control variables. The mean of the variable
2Source: European Corporate Governance Network, 2001
repurchases changes quite a bit. It increases each year from 43% of the firms repurchasing shares in 2003 to 77% in 2008, afterwards declining to 73% in 2009. This result is in line with the findings of Fama and French (2001) and Von Eije and Megginson (2008) which found share repurchases increasing in recent years.
Table 4 Descriptive statistics
The table shows the descriptive statistics of all the variables each year and aggregated over the period 2003 to 2009. Share 1e is the percentage of outstanding shares held by the largest shareholder. Share 2eis the percentage of outstanding shares held by the second largest shareholder. Div. Prone 1 is the figure defined as the largest shareholder present at dividend paying companies minus present at non-dividend paying companies. Div. Prone 2 is the figure defined as the second largest shareholder present at dividend paying companies minus present at non-dividend paying companies. E/A is defined as earnings before interest and tax (EBIT) divided by total assets. dA/A is defined as the change in total assets (total asset t – total assets t-1) divided by total assets t. V/A is defined as total debt plus market capitalization divided by total assets. Size is defined as the logarithm of the market value. Repurchases is the dummy variable taking the value 1 if the firm repurchases shares and 0 otherwise.
2003 2004 2005 2006 2007 2008 2009 2003-2009
Share 1e
Mean 13.43% 12.84 12.75 12.93 13.08 12.92 12.35 12.89
Median 10.68 10.54 10.50 10.33 10.75 10.47 10.20 10.49
Maximum 53.80 54.20 54.50 56.40 56.21 58 57.78 58
Minimum 4.87 5.10 5.20 5 4.72 5.02 5 4.72
Std. Dev. 8.90 8.34 8.03 8.41 8.79 8.96 9.01 8.64
Share 2e
Mean 8.71 8.64 9.12 9.37 9.02 8.70 8.46 8.86
Median 7.54 7.66 7.86 7.80 7.58 7.58 7.30 7.60
Maximum 43.3 43.3 43.30 51.52 41.04 42.80 43.14 51.52
Minimum 4.80 4.75 5 4.87 4.68 4.76 4.81 4.68
Std. Dev. 4.80 4.54 4.79 5.75 4.71 4.49 4.79 4.85
Div. Prone 1
Mean 0.01 0.05 0.07 0.07 0.08 0.05 0.05 0.05
Median 0.01 0.03 0.08 0.07 0.06 0.03 0.04 0.03
Maximum 0.06 0.12 0.16 0.17 0.18 0.15 0.18 0.18
Minimum -0.04 -0.01 -0.01 -0.01 -0.01 -0.01 -0.01 -0.04
Std. Dev. 0.04 0.04 0.06 0.06 0.07 0.05 0.06 0.06
Div. Prone 2
Mean 0.01 0.04 0.06 0.06 0.06 0.05 0.05 0.05
Median 0.02 0.03 0.05 0.04 0.03 0.04 0.04 0.03
Maximum 0.06 0.12 0.16 0.17 0.18 0.15 0.18 0.18
Minimum -0.04 -0.01 -0.01 -0.01 -0.01 -0.01 -0.01 -0.04
Std. Dev. 0.03 0.04 0.06 0.06 0.06 0.05 0.07 0.06
E/A
Mean 0.10 0.12 0.13 0.13 0.13 0.11 0.09 0.12
Median 0.10 0.11 0.12 0.12 0.13 0.12 0.09 0.11
Maximum 0.72 0.58 0.66 0.85 0.68 0.51 0.61 0.85
Minimum -0.46 -0.16 -0.08 -0.16 -0.56 -0.92 -0.60 -0.92
Std. Dev. 0.09 0.08 0.08 0.09 0.10 0.14 0.11 0.10
dA/A
Mean 0.09 0.11 0.09 0.14 0.16 0.01 0.04 0.09
Median 0.08 0.09 0.08 0.07 0.10 0.01 0.04 0.07
Maximum 0.85 0.66 0.68 4 6.54 0.89 0.63 6.54
Minimum -0.45 -0.53 -0.35 -0.29 -0.50 -1.17 -0.74 -1.17
Std. Dev. 0.14 0.13 0.15 0.34 0.50 0.19 0.15 0.27
