OWNERSHIP STRUCTURES AND DIVIDEND PAYOUT POLICY: EVIDENCE FROM SWEDEN

OWNERSHIP STRUCTURES AND DIVIDEND

PAYOUT POLICY: EVIDENCE FROM

SWEDEN

Hessel T. de Groot s2211149

Faculty of Economics & Business University of Groningen Master Thesis MSc. Finance

Supervisor: Prof. Dr. C.L.M. (Niels) Hermes January 2017

TABLE OF CONTENT

ABSTRACT 3

1. INTRODUCTION 4

2. THEORETICAL FRAMEWORK 6

2.1 Controlling shareholder and minority shareholder protection 6

2.2 Type of the controlling shareholder 8

2.3 The second largest shareholder 11

2.4 Type of both the controlling and second largest shareholder 11

3. DATA AND METHODOLOGY 12

3.1 Sample 12 3.2 Ownership database 13 3.3 Variables 14 3.4 Descriptive statistics 16 3.5 Correlation matrix 17 3.6 Research model 19

4. RESULTS AND DISCUSSION 19

4.1 Regression results 19

4.1.1 The voting rights of the controlling shareholder 20 4.1.2 The voting rights of the second largest shareholder 21 4.1.3 The voting rights of the controlling and the second largest shareholder 22

4.1.4 Types of shareholder 22

4.2 Robustness checks 29

5. CONCLUSION 33

OWNERSHIP STRUCTURES AND DIVIDEND

PAYOUT POLICY: EVIDENCE FROM

SWEDEN

ABSTRACT

This paper reports the results from an empirical investigation into the relationship between dividends payout policy and ownership structures of companies, using a sample of 167 listed Swedish companies. Results of the empirical analysis reveal that higher voting rights of the controlling shareholder are associated with a higher likelihood to pay dividend and with higher dividend payments, this also holds for the second largest shareholder. Family controlled companies are related to a higher likelihood to pay dividends and pay higher dividends than non-family controlled companies. Institutional controlled companies do also have a higher likelihood to pay dividends compared to non-institutional controlled companies. Controlling owners part of one of the 20 largest business spheres in Sweden have a higher likelihood to pay a dividend and pay higher dividends than controlling owners not part of a business sphere. If both the controlling and the second largest shareholder are institutions the company does have a less likelihood of paying dividends and pay lower dividends than companies controlled by other types of ownership structures. When both the controlling and second largest shareholder are a family, there is a high likelihood that the company pays a dividend.

1.

INTRODUCTION

Why do some firms pay dividends while others do not? This question has raised the attention of financial economists since the publication of Miller and Modigliani (1958, 1961),

stating that in a frictionless world dividend policy has no consequences for shareholders’ wealth when the investment policies of the firm remain constant over time. Higher dividend payouts lead to lower retained earnings and capital gains, resulting in an unchanged

shareholder wealth.

One of the most studied explanations why firms pay dividends is the agency cost theory, which derives from problems between ownership and management and the differences in managerial and shareholder priorities, also called the principal-agent conflict (Jensen and Meckling, 1976). This theory states that cash dividends can be used to mitigate agency problems in a company by reducing the cash flow and forcing management to enter the capital market for financing, which leads to monitoring by the market (Easterbrook, 1984; Jensen, 1986). Prior literature has paid attention to the principal-agent conflict and has focused mostly on developed countries, with publicly held firms and dispersed ownership structures where the managers of the company had the control.

However, apart from developed countries, Faccio and Lang (2002) indicate that family controlled companies are widespread around the world and that they have a growing

importance in current economies. La Porta, Lopez-de-Silanes and Shleifer (1999) examined the ownership structures of large firms in 27 different countries and suggest that these companies are heavily concentrated and are mostly controlled by families or states. Faccio, Lang and Young (2001) found that families have the main control in East Asia and Western Europe, which also holds for Swedish companies (Faccio and Lang, 2002).

generate private benefits not shared with minority shareholders. (Bodnaruk and Ostberg, 2013; Brav, Graham, Harvey and Michaely, 2005; Maury and Pajuste, 2002). These private benefits held by the controlling shareholder reduce or sometimes even remove the dividend payout to the other (minority) shareholders. Following the principal-principal theory, controlling shareholders want to use the companies’ free cash flow for own discretion, meaning that dividend payouts will be low as this is a way to reduce the free cash flow and reduce managerial control (Jensen, 1986). As mentioned in these studies, and many others, it is important to consider the ownership structures of companies to understand dividend policies related to the agency problems in the market.

The purpose of this study is to investigate how ownership structures are related to dividend payout; more precisely to dividend levels and the propensity to pay dividend in Swedish listed firms, leading to the following research question:

What is the relationship between ownership structures and the dividend payout policy of Swedish listed companies?

This study will focus on the controlling shareholder of the company, i.e. the shareholder that holds the largest amount of voting rights, and the influence of the second largest shareholder of the company. Different types of ownership will be discussed in this study, such as family ownership, institutional ownership and ownership part of a business sphere. This study is based on data from Sweden, because of the good data availability for listed firms.

Although the relationship between ownership structures and dividend payout is an intensively studied topic, there is a large variety in outcomes. In the past, different countries have been studied, such as: Italy (Mancinelli & Ozkan, 2010), Germany (Gugler and

Yurtoglu, 2003), United Kingdom (Kahn, 2006) and Finland (Maury and Pajuste, 2002) and several US studies (Bodnaruk and Ostberg, 2013; Brav et al., 2005). The findings of these studies did show different outcomes. Therefore, this could suggest that the relationship between several ownership structures and dividend payout could be country specific. It is interesting to study the relationship between ownership structures and dividends payout policies in Sweden because Sweden is characterized by relatively concentrated ownership structures on the capital market. Whereas the majority of listed companies in the US and UK markets have a highly dispersed ownership structure, the ownership of Swedish listed

Corporate Governance, 2009). These controlling shareholders are expected, even in rough times, to take a long-term responsibility for the company by holding on to their shares. Usually, these major shareholders do play an active role, for example by sitting on the board of directors. Balancing their strong ownership powers, the Swedish Companies Act provides a far-reaching protection of minority shareholders. Altogether, the differences in the Swedish corporate governance compared to other countries make it interesting to study the relationship between ownership structures and dividend payout policies in Swedish companies.

Sweden has its own Corporate Governance code, rules and regulations concerning the power of the controlling shareholders and the protection of the minority shareholders. These rules are different from other countries. Adding to this, Swedish companies distinguish themselves from other European companies by frequently having a family in the controlling ownership position, which is often related to lower dividend payouts (Faccio et al., 2001; Kahn, 2006). Because of their high amounts of voting rights these (family) controlling shareholders have a lot of power in the decision-making of the company. The question arises if these Swedish controlling shareholders take advantage of the minority shareholders and gain benefits of control, or does the Swedish Companies Act provide enough protection of minority shareholders to discourage this? Adding to this, Swedish society takes a positive view of major shareholders taking particular responsibility for the company (for example by sitting on boards) and the Swedish capital market is regarded with considerable skepticism towards widely dispersed ownership structures (The Handbook of International Corporate Governance, 2009). These above raised differences with other countries and questions make Sweden an interesting country to study the relationship between ownership structures and dividend payout, which has to my knowledge not been studied so far.

The remainder of this paper is structured as follows: the next chapter contains a literature review on controlling shareholders, protection of minority shareholders, the second largest shareholder and the type of the controlling shareholder from which hypotheses will be derived. The third chapter discusses the data and research methodologies. Chapter four presents the results and discussion. The final chapter provides a conclusion.

2.

THEORETICAL FRAMEWORK

ownership structures do vary per country. Countries as the UK, US and Canada have

dispersed ownership and therefore the manager of the company is the controller. However, in most European countries a company is controlled by the largest shareholder (La Porta et al., 1999). Large shareholders, who have the power to influence the decisions of corporate

managers, are the most common shareholders in Sweden. The controlling shareholder, i.e. the shareholder holding the largest proportion of the votes, is therefore the relevant shareholder to focus on in this study. The larger the proportion of voting rights held by the controlling shareholder, the larger his influence over corporate decision-making and payout policies (Brav et al., 2005; Maury and Pajuste, 2002; Gugler and Yurtoglu, 2003). These influences can lead to an agency conflict between the controlling shareholders and the minority shareholders, the so-called principal-principal conflict. The principal-principal conflict focuses on the expropriation problem of the wealth of minority shareholders by the controlling shareholders (Young, Peng, Ahlstrom, Bruton and Jiang, 2008). A controlling position gives shareholders certain benefits compared to minority shareholders, the so-called private benefits of control (Dyck and Zingales, 2004). Insiders that control the assets of the company can use these funds for their own purpose without benefiting minority shareholders, for example by paying excessive salaries, choosing specific investment policies in line with their private preferences or allowing other companies under their control to trade with the company at favorable terms (Jensen, 1986; Shleifer and Vishny, 1997). Since the amount of resources under the control of the controlling shareholder increases their ability to gain private benefits, shareholders have the incentive to invest in projects that expand the (resources of the) company even when these projects are not the most profitable ones available (Jensen, 1986).

Dividend payments reduce cash flows and financial resources and therefore lessen the ability of controlling shareholders to gain benefits over other shareholders. High dividend payouts increase the need to turn to the capital market to finance future investments, which will results in increasing power for external investors (Easterbrook, 1984). According to the view of Easterbrook (1984), a controlling shareholder wants to keep as much cash flows as possible to his control to prevent the need for new shareholders. Thus, controlling

shareholders have the incentive to minimize dividend payouts and retain all financial resources to gain private benefits of control.

have laws and regulations to restrict the power of the controlling shareholders, aiming to protect minority shareholders (De Silanes, La Porta, Shleifer and Vishny, 1998). This makes it important to consider the levels of legal protection when studying the relationship between ownership structures and dividends. According to De Silanes et al. (1998) protection of minority shareholders is stronger in common law countries, such as the UK and the US, and the worst in French civil law countries. German and Scandinavian civil law countries are located in the middle. A following study of La Porta, De Silanes, Shleifer and Vishny (2000) reported that dividend policies could be viewed as a substitute or as an outcome of legal protection of minority shareholders. Firstly, as a substitute, controlling shareholders are assumed to raise dividends when there is low protection to show minority shareholders that they do not take an advantage compared to them. However this is contradictory, if the controlling shareholders does not like this sign of goodwill they can just keep the dividends low if they want. Secondly, as an outcome, minority shareholders put pressure on the company to raise dividends when they have strong protection. This is mostly possible when minority shareholders form groups and put pressure on the controlling shareholder as an entire group.

Taking into account the medium protection of minority shareholders in Scandinavian countries (de Silanes et al., 1998) and the fact that most Swedish companies have one or a few major shareholders, which are frequently families (Faccio and Lang, 2002), the minority shareholders in Sweden have lower protection than countries like the UK, US and Canada (Shleifer and Vishny, 1997). Therefore I assume that the large controlling shareholders in Sweden can exercise private benefits of control, therefore they tend to keep the free cash flow of the company available, which result in a lower likelihood to pay and lower dividend payments (Brav et al., 2005; Easterbrook, 1984; Gugler and Yurtoglu, 2003). Leading to the first hypothesis:

H1: The higher the voting rights of the controlling shareholder the lower the dividends payout.

2.2 TYPE OF THE CONTROLLING SHAREHOLDER Family ownership

reduce the effectiveness of the board by carrying out policies that are beneficial for the controlling shareholders and hence increase the expropriation of the minority shareholders’ wealth (La Porta et al., 1999). Because of the controlling power of family owners the principal-principal conflict may be tightened. This may be caused due to the absence of a good monitoring mechanism, family owners can therefore increase the use of the companies’ funds, which may lead to more principal-principal conflicts and thus increase the agency costs. The absence of good monitoring and the increasing agency costs can lead to benefits of control for the controlling shareholder and thus less dividend payments. Morck and Yeung (2003) introduced the ‘other peoples money problem’, meaning that a family has control over the company but only has a very small investment in the firm compared to other investors with less control. This is most often the case when a family is the founder of a company. The controlling shareholder can easily gain benefits of control because of a separation between cash flow and control rights, for example obtaining managerial positions and paying excessive salaries (Shleifer and Vishny, 1997).

Recent studies have reported whether families exacerbate conflicts between the controlling and minority shareholders and how family ownership is related to dividend policies. Faccio et al. (2001) studied the relationship between ownership structures and dividend policies in East Asia compared to West European companies. They showed that in East Asia as well as in West Europe wealth expropriation of minority shareholders occurs, mostly by family owned controlling shareholders, leading to lower dividend payments. Kahn (2006) investigated the relationship between dividends and ownership structures for large quoted UK firms. Results showed that ownership composition matters, finding a negative relationship between dividends payout and family ownership. According to Kahn (2006) family controlled ownership will increase the agency costs in the company and thus will reduce the dividends paid to shareholders.

Based on the discussed literature above, the following is hypothesized: H2: Family controlled ownership is related to lower dividends payout.

Institutional ownership

Institutions differ from individual investors, among other things because institutions manage more funds and therefore make larger investments. Because they have larger amounts at stake, they should have incentives to devote resources to monitoring (Shleifer and Vishny, 1986). According to Hartzell and Starks (2003) and Grinstein and Michaely (2005)

institutional shareholders are likely to be better informed than other shareholders, but also because of the fact that institutional owners will have more information disclosure (Grinstein and Michaely, 2005; Jankensgard, 2015). More information leads to lower monitoring costs and lower agency costs, therefore managers are likely to pay out more of their free cash flow to the shareholders of the company (Jensen, 1986). Assuming that institutional owners are better monitors, the above-mentioned theories imply that large institutional ownership will lead to higher dividends payout (holding all else constant).

Another view on the relationship between institutional ownership and dividend payments is that institutions generally prefer to push the company to pay high dividends. Because of these high dividends payments the company has to approach the external capital market to finance future investments and other financial requirements. According to Farinha (2003) institutional shareholders might do this when they know that their own monitoring mechanism is

inefficient. Likewise, Zeckhauser and Pound (1990) suggested that institutional shareholders are sometimes not involved in direct monitoring, because of their arm’s length investments. Their research showed that institutions prefer to encourage companies to pay high dividends, in this way they have to approach the external capital market for future investments. Kahn (2006) follows the argumentation that institutional ownership is linked to lower agency costs, therefore he argues that dividend payments will become higher when an institutional

shareholder controls the company.

Based on different studies in different countries the following hypothesis will be: H3: Institutional controlled ownership is related to higher dividends payout. Business sphere ownership

According to Khanna and Palepu (2000) groups of investors exist to overcome market failure, this could be done by creating internal capital markets that substitute for external capital markets. Which can result in excessive cash flows within the company and therefore lead to higher dividends because of the free cash flow paid by the company’ managers. Groups of investors are also linked with reducing information asymmetry between the company and the financial market, because of group visibility (Dewenter and Malatesta, 2001). Business spheres can also obtain better information availability due to diversification across a number of industries, meaning that controlling shareholders that are part of a business sphere are better informed than other shareholders. Better informed owners will result in less need to monitor and thus lower agency costs, therefore the company is likely to pay out higher

profits and therefore use retained earnings to create an internal capital market (Khanna and Palepu, 2000). Thus the larger and more diversified the business group is the less sensitive is the dividend policy decision (Manos, Murinde and Green, 2012). This study focuses on the 20 largest business spheres of Sweden, therefore I assume that the business spheres in general are large and well diversified. Based on the above-mentioned theories the following hypothesis will be:

H4: Business sphere controlled ownership is related to higher dividends payout. 2.3 THE SECOND LARGEST SHAREHOLDER

Looking at the amount of votes as a measure of corporate control, the votes of the second largest shareholder also have a significantly influence on corporate decisions and dividends payout. When legal protection does not give enough control rights to minority shareholders, other large shareholders might mitigate the principal-principal conflict by efficient monitoring of the management (Shleifer and Vishny, 1986). Only shareholders with a sufficient amount of money can monitor effectively because of the high costs attached to monitoring.

Existence of large shareholders can mitigate the free rider problem and therefore reduce agency costs, which can result in higher dividends. Gugler and Yurtoglu (2003) suggests that large (minority) shareholders have strong incentives to monitor companies and are able to put pressure on the management of the company to pay dividends. Additionally according to Faccio et al. (2001), the presence of multiple large shareholders increases dividends payouts in West Europe but decreases dividends in East Asia, meaning that in Europe another large shareholder helps to reduce the expropriation by the controlling shareholder. Therefore I assume that the larger the voting rights of the second largest shareholder the less power the controlling shareholder has and thus the larger the dividend payouts will be. This leads to the following hypothesis:

H5: There is a positive relationship between the voting rights of the second largest shareholder and dividend payout.

2.4 TYPE OF BOTH THE CONTROLLING AND SECOND LARGEST SHAREHOLDER As Maury and Pajuste (2002) suggested, the type of the controlling shareholder significantly influences the dividend policy of a company, therefore the type of the second largest

minority shareholders most often do not have the resources to effectively monitor the management and influence managerial decisions.

As mentioned earlier, family ownership is often related to higher agency costs due to inefficient monitoring and therefore the ability to extract benefits of control at the expense of the minority shareholders, resulting in lower dividend payments within family controlled companies (Faccio et al., 2001; Kahn, 2006). The study of Faccio et al. (2001) proved that if both the controlling and the second largest shareholder are families they collude instead of monitoring each other resulting in lower dividend payments. These assumptions lead to the following hypothesis:

H6: Companies with a family as controlling and second largest shareholder pay less dividends.

As said before when discussing the theory of controlling institutional shareholders,

institutional owners have better monitoring mechanisms than non-institutional owners and therefore have lower costs. Thus institutional shareholders are likely able to pay out more of their free cash flow to the other shareholders of the company (Jensen, 1986; Grinstein and Michaely, 2005). Institutional owners focus on maximizing the value of the company

(Shleifer and Vishny, 1986). Therefore institutional owners as controlling and second largest shareholder can strengthen each other with their monitoring mechanisms to reach their shared goals and thus have the ability to pay out more free cash flows to other shareholders, meaning that dividend payments increase. These assumptions from the agency theory framework lead to the following hypothesis:

H7: Companies with an institutional owner as controlling and second largest shareholder pay more dividends.

3.

DATA AND METHODOLOGY

3.1 SAMPLE

the tech services and goods sector. In the following sections the ownership data and the variables used in this study will be described in more detail.

Table 1: Industry composition based on observations

Industry Number of observations Sample (%)

BASIC 64 5.4

Foresty, mining, industrial metals

INDUSTRIALS 463 39.4

Industrial engineering,

construction, aerospace, defense

GOODS 151 12.8

Food production, household goods, personal goods

HEALTH 109 9.3

Healthcare, pharmaceuticals

SERVICES 167 14.2

Travel/leisure, media, retail

TELE 25 2.1

Mobile and fixed line telecom

TECH 178 15.1

Software, computer services, technology hardware

ENERGY 20 1.7

Oil equipment, oil/gas production

TOTAL 1177 100

3.2 OWNERSHIP DATABASE

SIS Ägarservice, a Swedish company specialized in providing and analyzing ownership information collected the ownership data used in this study. They provide ownership lists for Swedish listed companies, except for ownership smaller than 0.1%, which are not published for integrity reasons. An ownership list contains the 50-200 largest shareholders (list is capped at 200 shareholders, so even when there are more separate owners with stakes exceeding 1% the list shows 200) and show for example the number of A and B shares held by an individual or legal entity; the associated cash flow and voting rights; the owner’s nationality; whether the owner is classified as an insider whether the shareholder is a legal entity or not.

One feature of SIS Ägarservice is that it shows the ownership list organized according to spheres of influence. This sphere option means that individual owners who are likely to act in a synchronized way, for example based on a kinship, are listed as a group. This reveals the ultimate control of the company. With ultimate control you will have a bigger say in

Jankensgard (2015) provided an example considering the ownership of Clash Ohlson, a large cap retailer. The raw ownership list suggests that the control lies with the single largest shareholder (17.3% of the votes). However, the sphere view of SIS Ägarservice reveals that the members of the Haid-family (none of them were individually the largest owner) together control 37% of the votes and thus have ultimate control. In fact, when they combine all their shares they are the largest shareholders and have the ultimate control, however at first sight it does not look like they have. In this study I will use the feature of SIS Ägarservice showing the holdings of the 20 most influential spheres in Sweden. These are built around the holdings of successful industrialists. Examples are the Kinnevik sphere, which originates the ventures of entrepreneur Jan Stenbeck, and the Wallenberg sphere, which originates the holdings of the Wallenberg family. These spheres are interesting to study, because these shareholders use a complex web of controlling stakes in cross-holdings to achieve effective control over a company, though cash flow rights are typically much smaller (Collin, 1998; Jankensgard, 2015).

3.3 VARIABLES

This study uses four different variables for dividends payout, which is the dependent variable. The propensity to pay dividends (DVPTP) is a binary variable, which indicates if a firm did pay any dividend (DVPTP=1) or did not pay dividends (DVPTP=0) during the period 2004-2012. A drawback of the DVPTP measure is that the difference between zero dividends and only a small dividend payout is not significant. An advantage is that it shows the propensity to pay a dividend and therefore highlights the so-called ‘minority freeze outs’ (De Angelo, De Angelo and Rice, 1984), i.e. the tendency to not pay a dividend and extract private benefits of control.

Other measures of the dependent variable are the dividends to sales (DTS), dividends to funds from operations (DTFFO) and the dividends to earnings (DTE) ratio. Measured by the total dividends paid out divided over the total sales, funds from operations and earnings respectively.

(INSTITUTIONAL) is defined as a mutual or pension fund, insurance companies, investment companies or other large entities that manage funds on the behalf of others. It takes the value 1 if the controlling shareholder is an institutional owner and a value of zero otherwise. For family ownership the FAMILY variable takes the value 1 if the controlling shareholder is a family and zero otherwise. Furthermore I look at the business spheres in Sweden, i.e. individual owners that act in a synchronized way. The SPHERE variable takes a value of 1 when the controlling shareholder is on SIS Ägarservice’s list of the 20 most influential spheres in Sweden and zero otherwise.

Besides looking at the controlling shareholder I look at the influence of the second largest shareholder. The second largest shareholder has a significant influence on the corporate decision-making and has therefore for instance the opportunity to demand the controlling shareholder to pay a dividend or even collude with the controlling shareholder to extract benefits of control. The influence of the second largest shareholder is done by looking at the proportion of votes of the second largest shareholder (SIV2) relative to the total shares of the company.

To check whether colluding of the largest and second largest shareholder takes place I look at the family-family dummy variable, which takes a value of 1 if both the largest and the second largest shareholder are a family and zero otherwise.

On the other side since I expect that institutional owners pay a higher dividend and pay a dividend more often I look at the institutional-institutional dummy, as I expect that the institutional shareholders strengthen each other due to their shared goal. The institutional-institutional dummy takes a value of 1 if both the largest and the second largest shareholder are an institution and zero otherwise.

CONTROLS

Independent variables were chosen from an agency framework, but there are also other factors that may influence the dividends payouts of the company. Therefore I use size, leverage and the profitability of the company as control variables. The data of the control variables is obtained from Datastream.

Jensen (1986) argues that shareholders are only able to extract private benefits of control in overcapitalized companies, meaning that the company has more cash available than they actually need to finance their investments. According to Jensen (1986) debts increase the interest payments and thus reduce the cash flows available to the controlling shareholder. In Jensen’ view debts reduce the agency costs and can therefore be seen as a substitute for dividends. Therefore I expect that debt is negatively related to dividends. The debt of a company will be measured as the LEVERAGE ratio, which is the total debt divided by the total assets.

Finally according to De Angelo et al. (2004) profitable firms have better access to the capital market and have more cash flows available due to profits. Therefore I expect that the profitability of the firm positively affect the dividend payout. Profitability (PROF) is

measured as the net earnings divided over the average total assets. 3.4 DESCRIPTIVE STATISTICS

Table II shows the descriptive statistics for the variables used in the multivariate analyses. The dataset contained 167 Swedish listed firms with roughly 1100 year observations over a period of 2004-2012.

As the table demonstrates, the mean of the DVPTP is 0.700, indicating that in 70.0% of the observations Swedish companies paid a dividend. The average dividends to sales ratio is 2.9%, the average dividends to funds from operations ratio is 30.3% and the average dividends to earnings ratio is 70.7%. With regard to the ownership structure, we see that the largest shareholder holds on average 33.30% of the voting rights, ranging between 5.1% and 87.7%. The second largest shareholder holds on average 12.0% of the voting rights, ranging between 3.2% and 36.6%. Table II shows that 9.7% of the largest shareholders are an

institution, 48.4% of the companies have family owners and for 30.6% of the observations the largest shareholder is part of one of the 20 largest business spheres in Sweden. These numbers seem logical in comparison with other studies, Faccio and Lang (2002) found that 46.94% of the controlling shareholders in Sweden were a family and 7.8% was an institution, Maury and Pajuste (2002) found that 36.3% of the Finnish companies had a controlling family and 12.6% had an institution as controlling shareholder. In 5.4% of the observations the largest

Table 2: Descriptive statistics

Variables N Mean Median SD Min Max

Dependent DVPTP 1177 0.700 1 0.460 0 1 DTS 1177 0.029 0.017 0.049 0 0.564 DTFFO 1177 0.303 0.205 0.609 0 13.40 DTE 1177 0.707 0.347 5.234 0 164 Independent SIV1 1010 33.30 28.85 19.56 5.10 87.70 INST1 1007 0.097 0 0.297 0 1 FAM1 1007 0.484 0 0.500 0 1 SPHERE1 1007 0.306 0 0.461 0 1 SIV2 783 12.00 10.60 6.126 3.20 36.60 INST1-INST2 783 0.054 0 0.226 0 1 FAM1-FAM2 783 0.261 0 0.439 0 1 Control SIZE 1120 14.555 14.284 2.035 9.056 19.705 LEVERAGE 1112 0.180 0.149 0.160 0 0.758 PROF 1119 0.091 0.071 0.083 0.001 1.149

Notes: DVPTP is a dummy variable indicating firms paying dividends; DTS is the dividends to sales ratio; DTFFO is the dividends to funds from operations ratio; DTE is the dividends to earnings ratio; SIV1 is the size of voting rights of the largest shareholder; SIV2 is the size of the voting rights of the second largest shareholder; INST1 is a dummy a variable indicating if the largest shareholder is an institution; FAM1 is a dummy variable indicating if the largest shareholder is a family; SPHERE1 is a dummy variable indicating if the largest shareholder is part of one of the 20 largest business spheres in Sweden; INST1-INST2 is a dummy variable indicating that the largest shareholder and the second largest shareholder are both an institution; FAM1-FAM2 is a dummy variable indicating that the largest shareholder and the second largest shareholder are both a family; SIZE is the natural logarithm of the total assets; LEVERAGE is the total debt divided by the total assets; PROF is the profitability of the firm which is the earnings dividend by the assets.

3.5 CORRELATION MATRIX