The importance of ownership identity in the relationship between ownership concentration and firm performance

(1)

The importance of ownership identity in the

relationship between ownership concentration and

firm performance

Master thesis

By

Dirk-Jan Petersen University of Groningen Faculty of Economics and Business

(2)

(3)

Preface 3

Preface

In September of 2008 I started the bachelor study Business Administration at the University of Groningen. After having completed this bachelor I continued with the master Organizational Management and Control in September 2011. Now, May 2014, it is finally time to close this chapter of my life and finish my studies.

During the writing of this master thesis I encountered many difficult moments when I was lost and did not know how to continue. Completing this thesis was an enormous hurdle that had to be conquered. My supervisor dr. Teye Marra has been a great source of inspiration that provided me guidance from the start. I would like to thank dr. Marra for remaining patient and understanding with me and providing me with a lot of good advice and feedback as this project slowly took shape.

I would also like to thank my parents for facilitating my studies and supporting me during this period with words of love and advice. My sister, for getting angry with me when I needed to speed up and telling me to work harder. And finally my girlfriend Carola who has kept supporting me during this final phase of my study and during other difficult periods that my family has gone through this and last year.

New challenges lie ahead now. An academic career comes to an end, a professional career will begin.

(4)

Abstract 4

Abstract

This master thesis consists of a study on the importance of ownership identity in the relationship between ownership concentration and firm performance. This thesis includes qualitative and quantitative research into this relationship. Based on relevant literature described in the theory section, different ownership identities are expected to have a positive or negative relationship with performance of listed firms. However, using a sample of 900 firms and controlling for firm size, capital structure, firm value, industry and nation, my empirical analysis finds no significant relationship between ownership concentration and firm performance or ownership identity and firm performance.

(5)

Table of contents 5

1. Introduction

1.1 Introduction

The relationship between firm performance and ownership concentration has received increasing attention in corporate governance literature during the past decades. Several authors have studied the manner by which large shareholders in a firm can exercise power over the firm to guide it in a certain direction, and so influencing the performance of a firm. However, the literature that is available to me does not agree on whether or not firm performance is influenced by ownership concentration (Demsetz and Villalonga, 2001; Sánchez-Ballesta and García-Meca, 2007; Heugens, Van Essen and Van Oosterhout, 2008; Weiss and Hilger, 2012).

Berle and Means’ (1932) study was the first to pioneer the existence of a positive relationship between ownership concentration and firm performance. The first most notable study to challenge that idea was that of Demsetz and Lehn (1985). They argue that a firm’s ownership structure is determined from within with the aim of optimizing the costs and benefits at different levels of ownership concentration. They find that there is no significant relationship between ownership concentration and firm performance. Other studies have found empirical evidence for a curvilinear relationship where firm performance first increases when ownership concentration increases, but when ownership concentration level gets beyond a certain point firm performance decreases (Morck, Shleifer and Vishny, 1988; Claessens, Djankov, Fan and Lang, 2002).

(8)

Introduction 8

1.2 Goal and relevance

The goal of this study is to conduct an empirical analysis of the relationship between ownership concentration and firm performance, in which the analysis of ownership identity plays a key role. Not just the influence of family ownership is examined but several types of ownership identities are taken into account. By using a recent dataset from the year 2012, this study can provide an updated view on ownership concentration, firm performance and the importance of ownership identity. By combining data of two different performance measures and controls for nation and industry effects together, I try to find new insights regarding the relationship between ownership concentration, ownership identity and firm performance. Earlier studies that focused on Western European firms included between 100 and 200 firms spread over 12 countries. This study uses a larger sample spread over just three countries, Great Britain, Germany and France. By taking these three countries, that host the three largest economies of Europe, an additional analysis can be done to see if differences exist between firms active in the Anglo-Saxon model which is present in Great Britain, or firms active in the Continental model which is present in Germany and France.

1.3 Structure

(9)

Theory 9

2. Theory

2.1 Introduction

This chapter gives an overview of the relevant theories regarding ownership concentration and owner identity related to firm performance. First, the origins of studies regarding ownership structure related to firm performance are presented. This is followed by a short introduction to ownership concentration and a further description of the influence that ownership concentration can have on firm performance. In the fourth section of this chapter varying classifications of owner identities are discussed. Paragraph five continues to link different ownership identities to firm performance. The final paragraph provides a brief summary of the chapter.

2.2 Ownership structure

The ownership structure of a firm exerts great influence on the corporate governance and performance of a firm (Short, 1994; Shleifer and Vishny, 1997). An important link between ownership structure and firm performance is the agency theory. This link was pioneered by Jensen and Meckling (1976) and initiated many successive studies on the relation between ownership structure of a firm and firm performance. Jensen and Meckling argue that ownership structure is an important corporate governance mechanism that influences a firm’s agency costs. The separation of ownership and control can lead to the pursuit of different goals by managers which can reduce shareholder value. However, large shareholders can influence managers to align their goals with those of the shareholders in order to increase the economic performance of the firm. The shirking of managers can be reduced by providing optimal contracts and control. Separation of ownership and control can also have benefits. It promotes democratic and unbiased decision making.

2.3 Ownership concentration

(10)

Theory 10 2.3.1 Positive effects of concentrated ownership on firm performance

In firms with dispersed ownership, it is more difficult for owners to exercise control over the management of the firm. In such a situation, dispersed owners lack the power and often the will to handle possible agency problems. When information asymmetry and a difference in interests between owners and managers of a firm occur, problems with managerial opportunism can easily spread (Fama and Jensen, 1983). When owners possess a higher amount of shares of a firm, their stake in the firm increases. By increasing their stake in the firm, the incentive for owners to exercise their control rights and to influence performance maximizing decisions increases (Jensen and Meckling, 1976). Concentrated ownership can thus stimulate or even force firm management to act in the interest of the owners. It is also more likely that concentrated shareholders will use their personal resources and knowledge to invest in managerial and organizational capabilities (Carney and Gedajlovic, 2001). When a firm is underperforming, a large shareholder with a high stake in the firm can decide to use private recourses to boost performance in order to secure a share of future profits (Friedman, Johnson and Mitton, 2003).

2.3.2 Negative effects of concentrated ownership on firm performance

(11)

Theory 11 2.3.3 Curvilinear effects of concentrated ownership on firm performance

Some studies provide evidence that the effect of ownership structure on firm performance varies with the level of ownership concentration (Morck et al., 1988; Claessens et al., 2002). In a situation where the concentrated owner's stake is relatively low, such an owner has little power to let tunneling take place or implement other strategies that can be harmful to minority owners. In this case the concentrated owner is served best by influencing management in a way that increases general firm performance the most. Thus relatively low levels of concentrated ownership have a positive influence on firm performance (Heugens et al., 2008). The effect of concentrated ownership on firm performance will cease to be positive once the concentrated owner reaches a level where he can exercise more power over the firm while there still are minority shareholders. At this point tunneling becomes attractive to the concentrated shareholder. When the concentrated owner's stake increases further, it will reach a level where tunneling becomes less prudent to increase personal wealth. This is due to the fact that there will be less minority shareholders to expropriate. The best strategy for high concentration shareholders to increase their personal wealth then becomes directing the firm to higher performance. Thus higher levels concentrated ownership again have a positive influence on firm performance. Such a positive/neutral/positive relationship between ownership concentration and firm profitability was found by Morck et al. (1988).

In summary, the effects of ownership concentration on firm performance can have different results. From the previously described literature it becomes clear that different studies have achieved different outcomes. I propose the first hypothesis for empirical testing in this study:

Hypothesis 10: The level of ownership concentration of a firm has no effect on firm performance of

listed firms.

Hypothesis 1a: The level of ownership concentration of a firm has a significant effect on firm

performance of listed firms.

(12)

Theory 12

2.4 Ownership identity

A firm can have different goals or a combination of goals. Such goals can be to maximize firm value, maximize growth, enter new markets, develop new products etc. Goals are usually set top-down by firm management, which acts in service of the firm’s owners. Firm owners are often not the final firm owners because banks, institutional investors, other firms or governments can act as intermediate agents (Thomsen and Pedersen, 2000). Varying owner types may prefer to pursue different corporate strategies due to difference in preferences and backgrounds. Thomsen and Pedersen (2000) group firm owners in a set of ownership categories which cover the largest owners in the largest European firms. Their categories are: institutional investors, banks, (nonfinancial) companies, single individuals/families, and governments. These categories will first be explained briefly.

Institutional ownership is characterized by an arm’s length relationship with the firm and portfolio investments. Institutional investors normally have low risk aversion and a long term focus. They are quite specialized owners compared to the other ownership categories. Financial success is their key performance indicator and therefore their main objective is to maximize shareholder value. The belief that institutional investors have a positive impact on shareholder value, despite their inability to influence firm management due to their low ownership shares, is supported by findings by Nickell, Nicolitsas and Dryden (1997) and McConnell and Servaes (1990).

Bank ownership is not allowed in the United States and highly uncommon in the United Kingdom. However in what Charkham (1994) calls 'the German model', a bank can act as the universal provider of financial services to industrial firms. Because of the close relationship between the bank and the firm, the firm can enjoy special perks provided by the bank.

(13)

Theory 13 In a family ownership the family often acts as owners and managers of the firm at the same time. Families prefer to keep control over the firm because they make firm-specific investments in human capital. Combined with the fact that in most cases the founding family owners are quite wealthy, family ownership usually results in a long term commitment to the well-being of a firm. Family owned firms are risk averse since a large amount of the owner’s personal wealth is invested in the firm. They are also hesitant about letting in outside investors from stock markets because the families do not want to lose control of the firm. Within family owned firms it is not uncommon that families direct benefits towards themselves at the cost of minority shareholders (Thomsen and Pedersen, 2000).

A government owned firm is likely to implement a strategy in the firm aimed at political goals. Such goals can be to maintain low output prices, employment or external effects relative to profitability (Hart, Shleifer and Vishny, 1997). Government owned firms often pursue a nonprofit strategy in order to best service the public benefit. Compared to similar firms in their market, government-owned firms normally perform less than their privately owned peers based on 'regular' performance measures such as net profits and return ratios. Being a government owned firm does have its advantages, as governments normally have a large amounts of capital and other resources available that can be used to aid the firm.

Another way of classifying ownership identity is by looking at preferences and priorities regarding corporate risk, stability, growth, and performance (Douma, George and Kabir, 2006; Gedajlovic, Yoshikawa and Hashimoto, 2005). These differences can influence the relationship between ownership concentration and firm performance. There are many different classifications of concentrated firm owners. Heugens et al. (2008), who focus their study on the Asian market, define two broad influential categorizations:

 Foreign versus Domestic owners

 Stable, market and inside owners

(14)

Theory 14 Domestic ownership is often linked to lower performance (Heugens et al., 2008). Domestic owners normally involve business groups that are legally separate but bound together by formal or informal ways. The effects of group ownership can differ between countries (Khanna and Rivkin, 2001). There are, however, two characteristics of group ownership that are likely to have a negative impact on firm performance in Asian countries. First, groups have a strong common identity through the brands they sell or through dominant family ownership. When a firm in the group is not performing well, others in the group can transfer resources and capital in order to aid that firm. By this inter-firm loyalty the performance of all firms in the group is affected in a negative way. Second, the effects of tunneling, which were explained earlier, can also occur in ownership groups.

Stable, market, and inside owners This categorization of firm owners is based on motivations and objectives that apply to their ownership share. Each of these types will now be explained shortly. Stable investors usually have multiple connections to the firm in which they own shares. Besides being shareholders they are also often a supplier, buyer, creditor, debtor or partner. Typical stable investors are parties such as affiliated firms, banks or insurance companies.

Inside investors own a significant share of the organization and also fulfill a managerial role within the organization (Fama and Jensen, 1983). This category is very similar to family ownership as classified by Thomsen and Pedersen (2000). There are three reasons why inside owners pursue strategies that are not aimed at obtaining the highest possible profits for the organization. First, inside owners are likely to be averse which puts pressure on the firm's potential profitability. Their risk-aversion is the result of an undiversified investment portfolio, which places a large amount of the investor's wealth in a single firm. Second, inside owners often appoint trusted relatives in high positions within the firm. In most cases these relatives cannot live up to their expectations due to their lack of managerial experience. Third, inside investors may partake in tunneling activities, exploiting minority shareholders within the firm.

(15)

Theory 15

2.5 The relationship between ownership identity and firm performance

A study by Hamadi (2010) among listed Belgian firms between the period 1991 and 1996, which used Tobin’s Q as a measure of firm performance, finds that ownership concentration has a negative effect on firm performance. Hamadi refers to tunneling as the main cause for this result, but does not further define any forms of tunneling. In her study, firm size also has a negative effect on firm performance which she links to the argument that larger firms have a lower Tobin's Q. Hamadi (2010) also concludes that, in contrast to her other results, large shareholders in family owned firms have a positive influence on firm performance. However when family shareholders are organized in voting blocks, this results in a lower firm performance.

Andres (2008) performed a similar study amongst listed firms in Germany. He finds that family owned firms are more profitable than firms with dispersed shareholder ownership and non-family owned firms. However Andres (2008) also provides a requirement for his results to be true. The founding family needs to play an active role in the firm on the executive or supervisory board. If the founding family is just a large shareholder, there is no positive relation to firm performance.

A study by Maury (2006) among Western European firms also finds that active family ownership, in which a family member holds one of the top two positions within the firm, is required to improve firm profitability. Passive family ownership does not increase firm profitability compared to other firms. Silva and Maljuf (2008) find that family ownership is not always positively related to performance when linked to voting rights. Their results are based on data from publicly traded Chilean firms. Silva and Maljuf (2008) argue that a lower concentration family ownership has a positive influence on firm performance, which becomes even stronger when the family is actively involved in the management of the firm. However, when the concentration becomes too high, the family gains enough control over the firm to let tunneling, as described in section 2.3.2, occur, which has a negative effect on firm performance.

The conflicting results of these earlier studies show that there is no clear conclusion to be taken regarding the influence of concentrated family ownership on firm performance. As the most recent studies point in the direction of a positive link between family ownership and firm performance, I propose the second hypothesis for empirical testing:

Hypothesis 20: Family ownership has no effect on firm performance of listed firms.

(16)

Theory 16 Thomsen and Pedersen (2000) investigated how ownership share and identity are related to the market-to-book value, asset returns and sales growth of large European firms during the period between 1990 and 1995. Their results show a first increasing then decreasing (bell-shaped) curve effect of ownership share on market-to-book value and return on assets, but no effect on sales growth. The type of ownership identity (from their classification, which was explained in section 2.4) is just as important as the concentration of ownership. Family, corporate and government ownership generally result in a lower market-to-book value while institutional ownership is of a positive influence. Government ownership is also found to have a significant negative effect on return on assets. From their findings Thomsen and Pedersen (2000) conclude that the type of ownership structure and identity of a firm affects the level of priority a firm gives to profit versus growth. This prioritization further influences the economic performance of the firm. In addition to the previously defined hypothesis concerning a specific ownership identity, I propose the following hypotheses for empirical testing:

Hypothesis 30: Corporate ownership has no effect on firm performance of listed firms.

Hypothesis 3a: Corporate ownership has a negative effect on firm performance of listed firms.

Hypothesis 40: Institutional ownership has no effect on firm performance of listed firms.

Hypothesis 4a: Institutional ownership has a positive effect on firm performance of listed firms.

Government ownership and bank ownership only occurs 8 and 4 times respectively in the data sample of 912 firms that I collected for empirical testing. Therefore no hypothesis is formulated for these two ownership identities as there is insufficient data for a regression analysis.

(17)

Theory 17 on firm value and vice versa. Corporate ownership promotes the transfer of knowledge and resources between firms which can increase firm value. The final category, government ownership, has a negative effect on firm value. The level of government ownership also increases with lower firm values. This could indicate that government's goals are not always firm value maximization.

2.6 Summary

From the previous sections it becomes clear that different levels of ownership concentration can have varying effects on firm performance. Concentrated ownership can stimulate closer monitoring and control of firm management, enabling shareholders to steer management towards a performance increasing strategy. A higher level of concentrated ownership can however result in large stakeholders using their power to assign firm benefits to themselves. The expropriation of minority shareholders by larger stakeholders is defined as tunneling earlier in this chapter.

(18)

Methodology 18

3. Methodology

3.1 Introduction

The aim of this study is to perform an analysis of the obtained data. The data sample consists of 900 listed firms spread over three European countries (Great Britain, Germany, France). The results of this analysis will either support or reject the hypotheses that are defined in the previous chapter. This chapter explains the methodology of the empirical part of this study. Section 3.2 explains how the data for this study is obtained and analyzed. The next two sections define the variables that are used for statistical analysis. The end of this chapter includes a table that gives an overview of all the discussed variables.

3.2 Methodology

This study focuses on the importance of ownership identity in the relationship between ownership concentration and firm performance of listed firms. This sentence holds the three central themes or variables that are important in this study:

- Ownership concentration - Ownership identity - Firm performance

I briefly explain the meaning of these variables in the context of this study first, and discuss them in more detail in the next section of this chapter. The ownership concentration level for each firm is based on the amount of shares owned by the combined largest type of investor present among the top 5 largest investors in that firm. The ownership identity of each firm is assigned based on the identity of the highest combined ownership concentration level present in that firm. In the previous chapter the five different ownership identities were defined as: family, corporate, institutional, government or bank ownership. Firm performance is determined by two performance measures: return on assets and return on equity.

(19)

Methodology 19 Earlier studies that were discussed in the theory section focused on firms from the US, Asia, Chile and a broad selection of Western European countries with between 4 and 33 firms per country. This study focuses on listed firms based in the three countries that hold the largest economies of Europe; Great Britain, Germany and France. I zoom in on just these three countries with a larger sample of firms with more recent data in order to get more up to date knowledge. All the data for all the variables is collected for the year 2012. This is because information about ownership concentration is only available for the most recently finished year. All listed firms with annual sales in 2012 of over $ 50 million are selected. This threshold is selected because larger firms proved to have the most complete data available. Firms that still have incomplete data are eliminated. Firms that are active in the finance, insurance, and real estate industry according to their SIC industry category are also removed from the sample. After elimination, a sample of 912 firms remains of which 475 are from Great Britain, 246 from Germany and 191 from France. Table 1 shows the steps that were made that resulted in the final data sample which is used for empirical testing.

Table 1: Elimination table

# of firms in sample

All listed firms in Great Britain, Germany and France with annual sales of over $ 50 million 1464

Sample minus firms that are active in finance, insurance and real estate 1226

Sample minus firms that have incomplete data for any variable that is used for testing 912

Sample minus bank ownership and government ownership (used for empirical analysis) 900

All the data obtained is analyzed using SPSS. Several variables are used for multiple linear regression in order to test for a relationship between firm ownership structure and firm performance in the sample. The variables are discussed in further detail in the next section.

3.3 Variables

(20)

Methodology 20

Table 2: Firms by share and identity of the largest single investor

Ownership

share (%) Bank Corporate Government

Institutional

investor Person/family Total (%)

0-10 0 16 0 198 19 233 (25.5) 10-25 2 50 1 178 68 299 (32.8) 25-50 1 72 3 41 95 212 (23.2) 50-75 1 46 3 5 67 122 (13.4) 75-100 0 33 1 3 9 46 (5.0) Total (%) 4 (0.4) 217 (23.8) 8 (0.9) 425 (46.6) 258 (28.3) 912 (100)

The single largest investor is not always the largest represented identity among the investors of a firm. For example, firm A can have a person/family investor holding 20 percent of the shares and two institutional investors holding 17 and 18 percent of the shares. The two institutional investors together own more shares than the single largest investor. In such a case, two investors with the same identity can pursue similar goals and together they can exercise more influence than the single largest investor. In order to take this into account I take the identity that is represented by the highest combined ownership concentration of the same identities among the top 5 largest investors. The identity of each firm is assigned in this way. The ownership concentration level for each firm also consists of the combined shares of the largest represented identity. To illustrate this, firm A from the earlier example would have an ownership concentration of 35 percent with the identity of institutional investor. Table 3 provides an overview of the firms in the sample by share and identity of the largest combined identity represented among the top 5 investors.

Table 3: Firms by share and identity of the largest combined identity represented among the top 5 investors

Ownership

Institutional

(21)

Methodology 21 Table 3 shows that by combining investors with the same identity there is a shift upwards in ownership concentration level compared to table 2. Table 3 also shows an increase in the amount of institutional investors compared to table 1 by 47, of which 23 were corporate investors and 24 were person/family investors. Not all the changes in the assigned identities are visible through the differences between table 2 and 3 because a few institutional investors changed into corporate or person/family investors. The total amount of changed identities as a result of combining top 5 investors as opposed to taking the single largest investor is 64.

Tables 4, 5 and 6 show the ownership concentration levels of the combined investors per country. By presenting the data from table 3 individually per country some interesting things become visible. In Great Britain the identity of institutional investor is present in a large majority of firms (344; 72.4%). These numbers far exceed those of Germany (80; 32.5%) and France (48; 25.1%). A possible cause for this difference could be the different economic model by which Great Britain is governed. The investor identities for German firms are quite evenly divided among the corporate, institutional investor and person/family identity. Also a relatively large part (77; 40.3%) of the French firms is owned by a person or family, which occurs in relatively few firms in Great Britain (84; 17.7%). Firm ownership in Germany is quite evenly divided amongst corporate (88; 35.8%), institutional investor (80; 32.5%) and person/family (73; 29.7%) ownership. Higher levels of ownership concentration (50%+) occur relatively the most in Germany and France. In Great Britain, in 15.8% of the firms from the sample the largest combined identity of investors hold 50% or more of the shares compared to 41% of the firms in the sample that are from Germany and 44.5% of the firms in the sample that are from France.

Table 4: Firms from Great Britain by share and identity of the largest combined identity represented among the top 5 investors

Ownership

Institutional

(22)

Methodology 22

Table 5: Firms from Germany by share and identity of the largest combined identity represented among the top 5 investors

Ownership

Institutional

0-10 0 1 0 16 3 20 (8.1) 10-25 2 6 0 42 7 57 (23.2) 25-50 1 23 1 17 26 68 (27.6) 50-75 1 28 0 2 32 63 (25.6) 75-100 0 30 0 3 5 38 (15.4) Total (%) 4 (1.6) 88 (35.8) 1 (0.4) 80 (32.5) 73 (29.7) 246 (100)

Table 6: Firms from France by share and identity of the largest combined identity represented among the top 5 investors

Ownership

Institutional

0-10 0 3 0 9 1 13 (6.8) 10-25 0 8 1 21 3 33 (17.3) 25-50 0 21 2 12 25 60 (31.4) 50-75 0 17 1 6 42 66 (34.6) 75-100 0 10 3 0 6 19 (9.9) Total (%) 0 (0) 59 (30.9) 7 (3.7) 48 (25.1) 77 (40.3) 191 (100)

Table 7 shows the average ownership concentration by identity for the single largest investor and the combined largest investor among the top 5 investors. It shows that the combined largest investor identity that is assigned to each firm consists mostly of the single largest investor. The others among the top 5 investors contribute only a small amount of shares.

Table 7: Average ownership concentration level by identity

Bank Corporate Government

(23)

Methodology 23 From table 2 and 3 it becomes clear that bank ownership and government ownership only occur 4 and 8 times respectively in the chosen sample of 912 firms. This low amount of observations is not sufficient to perform a meaningful multiple linear regression analysis and therefore these 12 firms are removed from the sample. The statistical analysis will continue with a sample of 900 firms and focus on the ownership identities corporate ownership, institutional ownership and person/family ownership, as was already mentioned in section 3.2. Ownership identity is a categorical variable and therefore a dummy variable is created for two of the three ownership identities. The first dummy variable is linked to corporate ownership. With this dummy all firms that have the corporate ownership identity get assigned a 1, all other firms get assigned a 0. The second dummy variable is for person/family ownership. With this dummy all firms that have the person/family ownership identity get assigned a 1, all other firms get assigned a 0. The institutional ownership identity has the most observations in the sample and will therefore act as the reference category and remain without dummy.

The variable ownership concentration is set between fixed boundaries (0 and 100 percent), which makes it unsuitable for linear regression. To overcome this problem I use a log transformation of the variable for the regression analysis.

Firm performance is measured by two different variables: return on assets (ROA) and return on equity (ROE). These two measures are fundamental indicators of firm performance (Edmonds, Tsay and Olds, 2009; Hillier, Ross, Westerfield, Jaffe and Jordan, 2010). Numerous other studies that investigated the relationship between ownership structure and firm performance have used ROA and ROE as measures firm performance as well (Oswald and Jahera, 1991; Lehmann and Weigand, 2000; Heugens et al., 2008; Weiss and Hilger, 2012).

3.3.4 Control variables

(24)

Methodology 24 classification consists of just 12 categories that are based on SIC codes. Table 8 presents the 12 categories defined by Fama and French. Because the category Money does not occur in the sample, 11-1 dummy categories are created and assigned accordingly to the firms in the sample. The category Other acts as the reference category.

Table 8: Fama and French industry classification

Number Name Description

# Of firms in sample

1 NoDur Consumer non-durables – Food, tobacco, textiles, apparel, leather,

toys

89

2 Durbl Consumer durables – Cars, TV’s, furniture, household appliances 31

3 Manuf Manufacturing – Machinery, trucks, planes, off furn, paper, com

printing

133

4 Energy Oil, gas, and coal extraction and products 25

5 Chems Chemicals and allied products 28

6 BusEq Business equipment – Computers, software, and electronic equipment 144

7 Telcm Telephone and television transmission 30

8 Utils Utilities 18

9 Shops Wholesale, retail, and some services (laundries, repair shops) 117

10 Hlth Healthcare, medical equipment, and drugs 48

11 Money Finance 0

12 Other Other – Mines, constr, bldmt, trans, hotels, bus serv, entertainment 237

For larger firms investors require a larger amount of capital to own a given share of the firm. Firm size can therefore be associated with firm specific risk. A larger firm requires more investment from an owner to own a certain percentage of the firm. Larger firms are expected to have lower levels of ownership concentration. The variable for firm size I use is based on the total assets of the firm (Demsetz and Villalonga, 2001; Pedersen and Thomsen, 2003). For this variable I also use a log transformation, which is common in previous literature.

(25)

Methodology 25 Finally, I take market-to-book ratio as a variable to control for firm value. Pedersen and Thomsen (2003) found that firm value is related to ownership concentration and firm performance for certain types of owner identities. I expect firm value to have a strong positive relation with the performance measures employed in this study.

Table 9 presents a quick overview of all the variables that are used for this empirical study. It lists the variable, the abbreviation of the variable that is used while describing the statistical analysis in this study, a description of what the variable means, and a definition of how the variable is computed.

Table 9: List of variables

Variable Abbreviation Description Definition

Ownership concentration

CON The combined percentage of

shares owned by the identity that is most represented among the top 5 investors in a firm

Log (ownership concentration)

Return on assets ROA Measure of firm performance (first)

Return on equity ROE Measure of firm performance (second)

Owner identity IDENTITY The identity that is most represented among the top 5 investors in a firm

Categorical variable (3-1 dummies)

C = Corporate ownership IC = Institutional investor ownership

FA = Person/family ownership

Nation NATION Country where the firm is listed

Categorical variable (3-1 dummies)

GBR = Great Britain GER = Germany FRA = France

Industry INDUSTRY The industry, based on the Fama and French classification, in which the firm is active

Categorical variable (11-1 dummies)

Size TA Total assets of the firm Log (TA)

Debt-equity ratio DEBT EQ The leverage ratio of the firm Total debt/common equity

(26)

Results 26

4. Results

4.1 Introduction

This chapter presents the results of my empirical study. Section 4.2 starts with giving the descriptive statistics of the sample. The next section discusses outliers in the sample, how they are treated and what the effects of treating them are. The fourth section provides a correlation table. In the fifth section the results of the regression analysis are presented. The final section of this chapter will give a short summary of the most important findings of the empirical analysis.

4.2 Descriptive statistics

The descriptive statistics that are presented are based on the original data sample. This means that the data has not yet been transformed in preparation for the regression analysis and the outliers have not yet been treated.

Table 10 shows descriptive statistics for all the numerical variables that are used in the regression analysis of the complete sample. The table reports the spread of the variables by listing range, minimum and maximum value, median, mean, standard deviation, variance, skewness and standard error of skewness for each variable.

(27)

Results 27

Table 10: Descriptive statistics total sample (s=900) before transformation and winsorizing

Range Minimum Maximum Median Mean Standard deviation Variance Skewness Std. Error of Skewness

CON 98.84 0.07 98.91 35.6950 39.6174 20.48827 419.769 0.713 0.082 ROA 163.77 -94.64 69.13 5.4850 5.1919 10.23971 104.852 -2.171 0.082 ROE 677.15 463.90 213.25 10.9250 6.4632 43.99050 1935.164 -4.800 0.082 TA 397806.90 5.60 397812.50 654.3800 7802.9877 29044.09709 843559575.6 8.052 0.082 DEBT EQ 15.13 0.00 15.13 0.4050 0.7379 1.34482 1.809 5.820 0.082 MTB 76.39 0.12 76.51 1.4850 2.2826 3.44275 11.853 12.529 0.082

Table 11: Descriptive statistics Great Britain (s=475) before transformation and winsorizing

(28)

Results 28

Table 12: Descriptive statistics Germany (s=241) before transformation and winsorizing

CON 98.01 0.07 98.08 39.5300 42.6568 25.96458 674.160 0.383 0.157 ROA 121.13 -69.12 52.01 5.5100 4.7180 11.97510 143.403 -1.984 0.157 ROE 535.22 -463.90 71.32 10.9500 0.3759 57.22766 3275.005 -5.394 0.157 TA 397784.41 28.09 397812.50 619.9100 9257.2747 35827.29006 1283594713 7.302 0.157 DEBT EQ 14.62 0.00 14.62 0.4400 0.7792 1.39187 1.937 6.023 0.157 MTB 76.23 0.28 76.51 1.5300 2.3129 5.08460 25.853 13.184 0.157

Table 13: Descriptive statistics France (s=184) before transformation and winsorizing

CON 97.84 0.07 97.91 44.4000 44.4342 23.34672 545.069 0.145 0.179 ROA 49.14 -25.41 23.73 4.3350 4.1124 5.65504 31.980 -0.837 0.179 ROE 212.21 -160.78 51.43 9.2300 5.9198 20.15190 406.099 -4.413 0.179 TA 224095.74 35.62 224131.36 1586.8350 9691.1752 24084.01069 580039570.9 5.235 0.179 DEBT EQ 8.24 0.00 8.24 0.5200 0.7334 0.92220 0.965 4.452 0.179 MTB 11.05 0.15 11.20 1.1950 1.6360 1.52570 2.328 3.593 0.179

(29)

Results 29

Table 14: Descriptive statistics by industry before transformation and winsorizing

Range Minimum Maximum Median Mean

(30)

Results 30

4.2 Data transformation and outlier detection

Table 10 shows varying levels of skewness for ROA, ROE, TA, DEBT EQ and MTB, which is problematic for performing a regression analysis. To reduce the level of skewness for these variables in order to make them fit for regression analysis, I detect outliers and apply Winsorizing. Winsorizing is the modification of one or more data points at the end of the tails of the distribution to the next highest/lowest values within the distribution that are not suspected to be outliers. By Winsorizing the tails of the distributions of the variables with outliers are drawn inwards. In addition to outlier detection and Winsorizing, I do a log transformation for the variables DEBT EQ and MTB in order to normalize the data.

Outlier detection is performed by multiplying the interquartile range by a factor of 2.2. The result of this is then subtracted from the first quartile to determine the lower boundary, and added to the third quartile to determine the upper boundary. All observations that fall outside the lower and upper boundaries are considered outliers. This rule was suggested by Hoaglin and Iglewicz (1987) after showing that using a 1.5 multiplier for the interquartile range to detect outliers was inaccurate approximately 50% of the time. Table 15 shows the results of outlier detection using the 2.2 multiplier rule by Hoaglin and Iglewicz (1987).

Table 15: Outliers Q1 Q3 Lower boundary Upper boundary # Of observations outside lower boundary # Of observations outside upper boundary Percentage of observations outside boundaries CON 25.03 51.86 -33.99 110.87 16 0 1.78% ROA 2.63 8.95 -11.26 22.84 33 20 5.89% ROE 4.42 17.55 -24.45 46.41 55 33 9.78% TA 2.28 3.49 -0.38 6.15 0 0 0% DEBT-EQ 11.00 83.18 -147.78 241.96 0 44 4.89% MTB 0.99 2.56 -2.46 6.01 0 41 4.56%

(31)

Results 31 rule. DEBT EQ and MTB both have no outliers outside the lower boundary and 44 (4,89%) and 41 (4,56%) observations respectively falling outside the upper boundary. ROE is the only variable with a relatively high amount of outliers. The other variables, with the exception of ROA, all have less than 5% outliers. All the observations in the sample are valid points as they are exported directly from the Thomson One Banker database.

After outliers have been detected and Winsorized, the log transformations are performed. Because DEBT EQ has values of zero, one is added to each observation before doing a log transformation. Table 16 gives an overview of which transformation is done to which variable (including the transformations already described in table 9) and the resulting skewness levels after Winsorizing and data transformation.

Table 16: Data transformation and skewness levels, after Winsorizing

Transformation Skewness CON Log(CON) -0.717 ROA - -0.076 ROE - -0.092 TA Log(TA) 0.58 DEBT EQ Log(DEBT EQ + 1) 0.79 MTB Log(MTB) -0.125

4.4 Correlation

(32)

Results 32 TA has a negative correlation with CON (-0.182**). This confirms the expectation from section 3.3.4, which stated that larger firms require a greater investment to own a higher percentage of shares which carries more risk for the investor. TA also has a weak positive correlation with ROE (0.097**). DEBT EQ has negative correlations with ROA (-0.268**) and ROE (-0.100**) and a positive correlation with TA (0.320**). As described earlier MTB has a strong positive correlation with ROA (0.532**) and ROE (0.578**), and a weaker positive correlation with TA (0.120**). The dummy variables show no signs of multcollinearity with the other predictor variables, with the exception of institutional investor ownership which shows a correlation of -0.428** with CON. Because of the strong correlations of MTB with both the dependent variables I will perform a regression analysis which includes MTB and one that excludes MTB. This will result into four regression models, which will be explained in the next section.

Table 17: Correlation matrix

CON ROA ROE TA DEBT EQ MTB

(33)

Results 33

4.5 Regression results

Because I have two separate performance measures I start with formulating two regression formulas that are tested in this study. These regression models will test the hypotheses defined in the theory section. The two regression models are:

- Model I: ROA = α + β1 CON + β2 IDENTITY + β3 TA + β4 DEBT EQ + β5 MTB + β6 NATION + β7 INDUSTRY

+ ε

- Model II: ROE = α + β1 CON + β2 IDENTITY + β3 TA + β4 DEBT EQ + β5 MTB + β6 NATION + β7 INDUSTRY

+ ε

Because of the strong correlation between MTB and both of the dependent variables ROA and ROE, I perform two additional regression analysis that are similar to models I and II but exclude MTB as an independent variable. Any changes in the outcomes of the models will show the level of influence the variable MTB has on the model.

- Model III: ROA = α + β1 CON + β2 IDENTITY + β3 TA + β4 DEBT EQ + β5 NATION + β6 INDUSTRY + ε

- Model IV: ROE = α + β1 CON + β2 IDENTITY + β3 TA + β4 DEBT EQ + β5 NATION + β6 INDUSTRY + ε

Table 18 provides an overview of the regression results of all four models. It shows for each model the unstandardized coefficients (B) with the corresponding p-value behind it in brackets and its significance level. Additional outcomes for each regression model can be found in Appendix A.

I start with the analysis of model I, which uses ROA as a performance measure. Table 18 shows an R square of 0.358 which means that the model accounts for 35.8% of the variance in ROA scores. A little over 1/3rd_{of the total variance that is observed between ROA scores of listed firms in the sample}

(34)

Results 34 The findings of model I show no support for a relationship between either ownership concentration and firm performance or any of the in chapter two defined ownership identities and firm performance. Based on the results so far, there is not enough evidence to reject any of the null hypotheses of the four hypotheses that are defined in chapter two.

In model II, the dependent variable that contains the performance measure ROA is substituted by the other performance measure ROE. All the other predictor variables in model IIremain the same. Table 18 reports a slightly higher R square of 0.359 which means that 35.9% of the variance in ROE can be explained by the predictor variables used in the model. Just as in model I, DEBT EQ (-11.071**) has a negative relationship with the independent variable. MTB (26.260**) again has a positive relationship with the independent variable. In model II, TA (1.061*) also has a positive relationship with the independent variable. Again there are no NATION effects. There is one INDUSTRY effect visible, namely that of business equipment (-2.770*). Ownership concentration and ownership identity again have no significant influence in this second model.

Models III and IV are similar to models I and II respectively, but exclude MTB as an independent variable. The R square values of models III (0.108) and IV (0.061) are lower than those of models I (0.358) and II (0.359) respectively. This means that by excluding MTB from the equation, the models lose some of their predictive power. In both models III and IV TA and DEBT EQ are significant variables. They also show that when MTB is excluded from the regression analysis a negative NATION effect occurs for French firms for both model III (-2.024) and model IV (-4.672). Model III also shows positive INDUSTRY effects for consumer non-durables (1.861) and telecom (2.704). In model IV the only significant INDUSTRY effect is for telecom (6.123).

Similar to model I, the results from model II, III and IV show no evidence for any significant relationship between ownership concentration and firm performance or any of the ownership identities and firm performance.

(35)

Results 35

Table 18: Regression results models I, II, III and IV

Independent variables Dependent variables

Including MTB Excluding MTB

(36)

Results 36 4.5.1 Interaction effects

Table 19 presents a further analysis of factors that influence ROA and ROE. I examine to what extent the impact of ownership concentration is conditioned by ownership identity. To test if such an interaction effect is present I compute two new variables and add them to the regression equations of model I and II. The two new variables are CON * IDENTITY Corporate andCON * IDENTITY Family. When these variables are added to models V and VI their regression equation becomes as follows:

- Model V: ROA = α + β1 CON + β2 IDENTITY + β3 TA + β4 DEBT EQ + β5 MTB + β6 NATION + β7

INDUSTRY + β8 CON * IDENTITY Corporate + β9 CON * IDENTITY Family + ε

- Model VI: ROE = α + β1 CON + β2 IDENTITY + β3 TA + β4 DEBT EQ + β5 MTB + β6 NATION + β7

INDUSTRY + β8 CON * IDENTITY Corporate + β9 CON * IDENTITY Family + ε

(37)

Results 37

Table 19: Interaction effects, models V and VI

Including MTB

Model V: ROA Model VI: ROE

(38)

Results 38 4.5.2 Other effects

Table 20 shows the regression results when comparing firms from Great Britain with firms from continental Europe. For this analysis model I and II are used, but in both models the separate NATION dummies are replaced by one dummy for firms from Great Britain. The tables show that for both ROA and ROE the influence of a different governance system, the Anglo-Saxon or the Continental model, does not have a significant relationship with performance measures ROA and ROE. More detailed regression results are shown in the appendix as models VII and VIII.

Table 20: Regression results Great Britain vs Continental Europe, models VII and VIII

Including MTB

Model VII: ROA Model VIII: ROE

(Constant) 5.480* 5.803 CON -0.410 -1.373 IDENTITY Corporate -0.166 0.105 IDENTITY Family 0.362 0.957 TA 0.104 1.071 DEBT EQ -11.995** -11.086** MTB 11.443** 26.241**

NATION Great Britain 0.502 1.467

(39)

Results 39 Finally I test if firms being owned by institutional investors behave differently than other firms in the sample. For this I replace the two IDENTITY dummies being used in model I and II with a single IDENTITY dummy for firms with the institutional investor identity. The results for this are shown in table 21 and in the appendix as models IX and X. Table 21 shows that firms being owned by institutional investors do not have a significantly different relationship with performance measures ROA and ROE than firms not owned by institutional investors.

Table 21: Regression results institutional identity vs other identities combined, models IX and X

Including MTB

Model IX: ROA Model X: ROE

(Constant) 6.250** 8.136*

CON -.445 -1.448

IDENTITY Institutional investor -0.143 -0.591

(40)

Results 40

4.6 Summary

The results in tables 18 throughout 21 show that in regression models I throughout X only a few of the control variables are relevant predictors of the dependent variables ROA and ROE. In all of the models MTB (when included) and DEBT EQ are relevant variables. In none of the models the independent variables CON and IDENTITY have a significant relationship with the dependent variables ROA and ROE. Only in models III and IV, when MTB is excluded, a significant NATION effect is visible for French firms. In some of the models one or two INDUSTRY types are significant. The significant industries are either Consumer non-durables, Business Equipment or Telecom.

I have also tested for differences between firms from continental Europe, active in the Continental model, and firms from Great Britain, active in the Anglo-Saxon model. Table 20 shows no significant relationship regarding performance measures ROA and ROE respectively. Furthermore, I tested if being owned by institutional investors or not has a significant effect on performance measures ROA or ROE. The results in table 21 show no sign that this is the case.

(41)

Discussion and conclusion 41

5. Discussion and conclusion

This thesis focuses on the importance of ownership identity in the relationship between ownership concentration and firm performance. The relationship between ownership concentration and firm performance has received increasing attention from scholars since a positive link between the two subjects was first pioneered by Berle and Means in 1932. Especially during the last two decades, the available literature regarding ownership concentration and firm performance has increased steadily. Different authors have taken various approaches in their research towards this subject. The outcomes of these different studies are not always in line with one another. The theory chapter of this thesis describes in further detail the results of the most relevant studies regarding ownership concentration and firm performance. There are proponents that argue for a negative relationship between ownership concentration and firm performance (Fama and Jensen, 1983; Shleifer and Vishny, 1997; Hamadi, 2010). Others argue for a positive link between ownership concentration and firm performance (Maury, 2006; Andres, 2008). There are also studies that point to a curvilinear relationship between ownership concentration and firm performance (Morck et al., 1988; Claessens et al., 2002) where different levels of ownership concentration are either positively or negatively related to firm performance. Finally, some studies do not find a significant relationship between ownership concentration and firm performance at all (Cho, 1998; Himmelberg, Hubbard and Palia, 1999; Demsetz and Villalonga, 2001).

All these studies take different approaches in their empirical analysis. Some focus on firms from different countries around the world, others use varying performance measures, and only a few specifically look into the influence of one or more ownership identities. Because all the different and contradicting outcomes of all these studies there is no general consensus about how ownership concentration and ownership identity are related to firm performance. The goal of this study is to contribute to the existing body of literature by conducting an empirical analysis of the relationship between ownership concentration and firm performance, in which the analysis of ownership identity plays a key role.

(42)

Discussion and conclusion 42 is also unique because all the firms in the sample are from either Great Britain, Germany or France, the three countries that host the largest economies in Europe. The performance measures that are used for this study are ROA and ROE. The available data is statistically analyzed by multiple regression using SPSS.

A salient feature of my dataset is that Bank ownership (0.4%) and government ownership (0.9%) rarely occurs. Compared to the dataset from 1990 of Thomsen and Pedersen (2000), which have a total sample of 435 firms spread of 12 European countries (including Germany, France and Great Britain), of which 6.4% show bank ownership and 17.0% government ownership, this is a large change. The lower amount of government ownership can be attributed to the ongoing trend of privatization that was launched in the 1980s. The decrease in bank ownership suggests that banks have diversified their investment portfolios and lowered the amount of shares they own of large listed firms. Because the number of firms with bank or government ownership is not sufficient for a regression analysis they are removed from the sample.

In the results section I analyze the impact of ownership concentration and the ownership identities (family, corporate and institutional) on firm performance (ROA and ROE). Control variables are used for capital structure (DEBT EQ), size (TA), firm value (MTB), NATION, and INDUSTRY. The two performance measures are used as dependent variables in two separate regression models. Results show that ownership concentration is not a significant predictor in any of the regression models. Based on the data used in this study, I cannot conclude that ownership concentration is related to firm performance. This is consistent with studies by Cho (1998), Himmelberg et al. (1999), and Demsetz and Villalonga (2001). None of the ownership identities that I use is found to have a significant effect on firm performance. This is in contrast with results by Thomsen and Pedersen (2000), Pedersen and Thomsen (2003), Maury (2006), Andres (2008), Silva and Maljuf (2008), and Hamadi (2010) who all find evidence for either a positive or a negative relationship between ownership identities and firm performance.

(43)

Discussion and conclusion 43

5.1 Limitations and recommendations

While the data used in this study is more recent than the data used in earlier studies, it does only include information about the year 2012. Although ownership concentration levels are quite stable in most firms (Thomsen and Pedersen, 2000), their performance levels may well be influenced by the widespread impact of the financial crisis that has struck all of Europe since 2008. Furthermore, because of the use of different performance measures and control variables, the outcomes of this study are not fit for a direct comparison with previous studies. Moreover, this study focuses on just three countries in Europe. Even though the three countries that are included in this study contain the three largest economies in Europe, the generalizability of this study is limited. There is no guarantee that the findings can be applied to firms in countries elsewhere in the world.

(44)

References 44

References

Andres, C., 2008. Large shareholders and firm performance - An empirical examination of founding-family ownership. Journal of Corporate Finance, 14(4): 431-445

Berle, A.A., Means, G.C., 1932. The modern corporation and private property. Macmillan, New York Carney, M., Gedajlovic, E., 2001. Corporate governance and firm capabilities: a comparison of

managerial, alliance, and personal capitalisms. Asia Pacific Journal of Management, 18(3): 335-355

Charkham, J.P., Keeping good company: a study of corporate governance in five countries. Clarendon Press, Oxford

Cho, M., 1998. Ownership structure, investment, and the corporate value: an empirical analysis. Journal of Financial Economics, 47(1): 103-121

Claessens, S., Djankov, S., Fan, J.P.H., Lang, L.H.P., 2002. Disentangling the incentive and entrenchment effects of large shareholdings. Journal of Finance, 57(6): 2741-2771

Cramer, D., Howitt, D.L., 2004. The sage dictionary of statistics: a practical resource for students in the social sciences. Sage publications ltd, Thousand Oaks

Demsetz, H., Lehn, K., 1985. The structure of corporate ownership: causes and consequences. Journal of Political Economy, 93(6): 1155-1177

Demsetz, H., Villalonga, B., 2001. Ownership structure and corporate performance. Journal of Corporate Finance, 7(3): 209-233

Douma, S., George, R., Kabir, R., 2006. Foreign and domestic ownership, business groups, and firm performance: evidence from a large emerging market. Strategic Management Journal, 27(7): 637-657

Edmonds, T.P., Tsay, B.Y., Olds, P.R., 2009. Fundamental managerial accounting. McGraw-Hill Irwin, USA

Fama, E.F., French, K.R., 1997. Industry costs of equity. Journal of Financial Economics, 43(2): 153-193 Fama, E.F., Jensen, M.C., 1983. Agency problems and residual claims. Journal of Law and Economics,

26(2): 327-350

Friedman, E., Johnson, S., Mitton, T., 2003. Propping and tunneling. Journal of Comparative Economics, 31(4): 732-750

Gedajlovic, E., Shapiro, D.M., 2002. Ownership structure and firm profitability in Japan. Academy of Management Journal, 45(3): 565-575

(45)

References 45 Hart, O., Shleifer, A., Vishny, R.W., 1997. The proper scope of government: theory and an application

to prisons. Quarterly Journal of Economics, 112(4): 1127-1161

Hamadi, M., 2010. Ownership concentration, family control and performance of firms. European management Review, 7: 113-131

Heugens, P.P.M.A.R., Van Essen, M., Van Oosterhout, J.H., 2009. Meta-analyzing ownership concentration and firm performance in Asia: towards a more fine-grained understanding. Asia Pacific Journal of Management, 26(3): 481-512

Hillier, D., Ross, S., Westerfield, R., Jaffe, J., Jordan, B., 2010. Corporate finance. McGraw-Hill Higher Education, UK

Himmelberg, C.P., Hubbard, R.G., Palia, D., 1999. Understanding the determinants of managerial ownership and the link between ownership and performance. Journal of Financial Economics, 53(3): 353-584

Hoaglin, D.C., Iglewicz, B., 1987. Fine tuning some resistant rules for outlier labeling. Journal of American Statistical Association, 82(400): 1147-1149

Jensen, M.C., Meckling, W.H., 1976. Theory of the firm: managerial behaviour, agency costs, and ownership structure. Journal of Financial Economics, 3:305-360

Jensen, M.C., 1989. Eclipse of the public corporation. Harvard Business Review, 67(5): 61-75

Johnson, S., Lopez-de-Silanes, F., La Porta, R., Shleifer, A., 2000. Tunneling. Yale School of Management's Economics Research Network (working paper), 1-15

Kester, W.C., 1992. Industrial groups as systems of contractual governance. Oxford Review of Economic Policy, 8(3): 24-44

King, M.R., Santor, E., 2008. Family values: ownership structure, performance and capital structure of Canadian firms. Journal of Banking & Finance, 32(11): 2423-2432

Khanna, T., Rivkin, J.W., 2001. Estimating the performance effects of business groups in emerging markets. Strategic Management Journal, 22(1): 45-74

Lehmann, E., Weigand, J., 2000. Does the governed corporation perform better? Governance structures and corporate performance in Germany. European Finance Review, 4(2): 157-195 Maury, B., 2006. Family ownership and firm performance: empirical evidence from Western

European corporations. Journal of Corporate Finance, 12(2): 321-341

McConnell, J.J., Servaes, H., 1990. Additional evidence on equity ownership and corporate value. Journal of Financial Economics, 27(2): 595-612

Morck, R., Shleifer, A., Vishny, R.W., 1988. Management ownership and market valuation: an empirical analysis. Journal of Financial Economics, 20(1/2): 293-315

(46)

References 46 Oswald, S. L., Jahera Jr., J.S., 1991. The influence of ownership on performance: an empirical study.

Strategic Management Journal, 12(4): 321-326

Pathak, R., Ranaejee, Pradhan, S., 2012. The role of ownership structure in firm performance: a study of Indian manufacturing firms. The IUP Journal of Corporate Governance, 11(3): 36-47

Pedersen, T., Thomsen, S., 1997. European patterns of corporate ownership: a twelve-country study. Journal of International Business Studies, 28(4): 759-778

Pedersen, T., Thomsen, S., 1999. Economic and systemic explanations of ownership concentration among Europe’s largest companies. International Journal of the Economics of Business, 6(3): 367-381

Pedersen, T., Thomsen, S., 2003. Ownership structure and value of the largest European firms: the importance of owner identity. Journal of Management & Governance, 7(1): 27-55

Sánchez-Ballesta, J.P., García-Meca, E., 2007. A meta-analytic vision of the effect of ownership structure on firm performance. Corporate Governance: An International Review, 15(5):879-892

Shleifer, A., Vishny, R.W., 1997. A survey of corporate governance. Journal of Finance, 52(2): 737-783 Short, H., 1994. Ownership, control, financial structure and the performance of firms. Journal of

Economics Surveys, 8(3): 203-249

Silva, F., Majluf, N., 2008. Does Family ownership shape performance outcomes? Journal of Business Research, 61(6): 609-614

Thomsen, S., Pedersen, T., 2000. Ownership structure and economics performance in the largest European companies. Strategic Management Journal, 21(6): 689-705

Thomsen, S., Pedersen, T., Kvist, H.K., 2006. Blockholder ownership: effects on firm value in market and control based governance systems. Journal of Corporate Finance, 12(2): 246-269

(47)

Appendix 47

Appendix A

Model I

Model summary

R R Square Adjusted R square Std. Error of the Estimate

0.599 0.358 0.345 5.49779

Anova

Sum of Squares df Mean square F Sig. Regression 14881.420 18 826.746 27.352 0.000

Residual 26628.867 881 30.226

Total 41510.287 899

Coefficients

N = 900 Unstandardized coefficients Standardized coefficients

B Std. Error Beta T Sig.

(48)

Appendix 48

Model II

Model summary

0.600 0.359 0.346 11.57239

Anova

Residual 117983.740 881 133.920

Total 184187.302 899

Coefficients

B Std. Error Beta t Sig.

(49)

Appendix 49

Model III

Model summary

0.328 0.108 0.091 6.48001

Anova

Residual 37035.604 882 41.990

Total 41510.287 899

Coefficients

B Std. Error Beta t Sig.