The effect of ownershipstructure and
internationalization of ownership on nationality
diversity in European Top Management Teams
Master Thesis IB&M 2007-2008
University of Groningen Faculty of Economics and Business
Landleven 5 Zernikecomplex Paddepoel 9747 AD Groningen Joost Rusius Stud nr: 1362208 Email: j.h.rusius@student.rug.nl
Abstract
This thesis empirically examines the effect of ownershipstructure and internationalization of ownership on nationality diversity in Top Management Teams (TMTs). The research has been conducted on 307 European companies in 14 EU countries from the year 2005. The thesis examines the effect of four independent variables (ultimate ownership, percentage shares held by the five largest shareholders, percentage foreigners among the five largest shareholders and the percentage foreigners on all recorded shareholders) on the dependent variable, nationality diversity in TMTs. The main result of this thesis is that nationality diversity among the five largest shareholders is positively related to nationality diversity in TMTs. For the other three independent variables no significant relation with nationality diversity in TMTs could be found.
Key words: ownershipstructure, internationalization of ownership, shareholders, nationality
diversity and TMTs.
Table of contents Section Page Abstract 1. Introduction 4-7 1.1 Problem statement 5 1.2 Outline 7 2. Literature review 7-9 3. Corporate Governance 10-14
3.1 Fundamental issues and examples 10
3.2 Differences in European Corporate Governance 12
3.3 Recent developments 13
4. Indicators and hypotheses 14-18
5. Data and descriptives 18-32
5.1 Data and sources 18
5.2 Description of general indicators 21
5.3 Distribution of ownership 23 5.4 Internationalization of ownership 28 6. Methodology 32-44 6.1 Dependent variable 32 6.2 Independent variables 33 6.3 Overview of variables 34 6.4 Description of variables 35 6.5 Empirical results 38
6.5.1 Regression results control variables 39 6.5.2 Regression results without country dummy variables 40 6.5.3 Regression results including country dummy variables 41
7. Discussion and conclusion 44-47
References 48-50
1. Introduction
This thesis examines the effect of ownershipstructure and internationalization of ownership on nationality diversity in Top Management Teams (TMTs) in 14 European countries. As Heijltjes et al. (2003) argue, the internationalization of ownership is a potential determinant of internationalization of TMTs. Linck et al. (2007) and Krivogorsky (2006) argue that there is a relatively limited amount of research on the determinants of board structure and that the factors that affect board composition are not well understood. However, it is known to be related to a firm’s ownership structure (Bhagat and Black, 2001). Also, board composition is supposed to be affected by the national context, as argued by Heijltjes et al. (2003). While growing across borders, companies are confronted with the question how to fill executive positions (Van Veen and Marsman 2007). Over the 1990s, some larger European companies made first steps to recruit foreign TMT members (Alexander and Esser, 1999; Heijltjes, Olie, and Glunk, 2003). Several factors might have contributed to the international mobility of executives, e.g. the increasing use of English as a common language in large multinationals, a convergence of educational formats, and increasingly international capital markets as well as markets for goods and services (Ruigrok and Greve 2008). Selection of TMT members is subject to different corporate governance laws, regulations and codes in the countries of the EU. Although, there are differences in corporate governance systems, there are also company determinants known which explain nationality diversity in TMTs. For instance, Staples (2007), Kaufman (2007) and Van Veen and Marsman (2007) show that TMTs become more globalised through cross border M&As.
TMTs have the highest nationality diversity, followed by the UK and Germany. This is mainly caused by different selection procedures and ownershipstructure differences of TMT members in The Netherlands, UK and Germany to which TMT members are subject to. Van Veen and Elbertsen (2007) primarily rely on most dominant and general ownershipstructures on country level, such as showed by Van Oijen (2000). As well Van Veen and Elbertsen (2007) as Van Veen and Marsman (2007) argue that more sophisticated research into the effect of ownershipstructure and internationalization of ownership on nationality diversity in TMTs is needed. Therefore, the aim of this thesis is to describe how differences and similarities in ownershipstructure and internationalization of ownership and might affect nationality diversity in TMTs.
1.1 Problem statement
The relationship between ownershipstructure and internationalization of ownership on nationality diversity in TMTs has hardly been researched before. Ruigrok and Greve (2008) make an attempt by measuring how foreign ownership might affect nationality diversity in TMTs. For the relationship between foreign ownership levels and nationality diversity in TMTs they use data from 2000, 2002, 2004 and 2005 from Denmark, Finland, the Netherlands, Norway and Sweden. They find that there is an overall increase in the international composition of TMTs and in the degree of nationality diversity in TMTs. Ruigrok and Greve (2008) also find that foreign share ownership has a strong positive relation (beta coefficient =0.75) with nationality diversity in TMTs. They measure foreign share ownership by looking at the reported number of shares held by owners (and in a few cases voting rights if share data were ambiguous or unavailable) who are domiciled outside the home country. In their research they control for TMT size. Further they use year dummy variables for temporal effects and use country dummy variables for geographical effects. Their dataset for the examined relationship includes 117 companies.
Another difference is that this thesis also focuses on ownershipstructure as a determinant of nationality diversity in TMTs. This will be done by looking at ownership in terms of voting power (=control). Ruigrok and Greve (2008) focus on share data, instead of voting rights. To operationalize ownershipstructure, this thesis identifies the most powerful shareholders. These are the major shareholders and ultimate owners, which might have implications for nationality diversity in TMTs. Overall, this thesis might add new insights on how ownershipstructure and new measurement methods for internationalization of ownership are related to nationality diversity in TMTs.
In the field of comparative corporate governance in a European context not much is known about different ownership structures, the composition of the shareholder population and TMTs. La Porta (1999) and Faccio and Lang (2001) do give a description of ultimate ownership structures in different countries in the world, however, less is known about the nationality, identity, and the type of shareholder. Further, very little research deals quantitatively with a relationship within a corporate governance system, namely the effect of ownershipstructure and the internationalization of ownership on nationality diversity in TMTs. Given all the above, the objective of this thesis is to examine how ownershipstructure and internationalization of ownership might affect nationality diversity in TMTs. To meet this objective, this thesis will have the following research question:
What is the effect of ownershipstructure and internationalization of ownership on nationality diversity in European top management teams?
1.2 Outline
The thesis’ outline will be as follows; in the next section a relevant literature will be discussed. Section 3 will discuss corporate governance mechanisms, this is to show through which mechanism ownershipstructure and internationalization of ownership affect nationality diversity in TMTs. Then, indicators will be discussed and hypotheses will be formulated. Section 5 discusses the data and findings about different ownershipstructures and degree of internationalization of ownership in Europe. In section 6 the methodology with the (in) dependent and control variables will be shown and the empirical results will be shown. Finally, there will be dealt with the discussion, limitations of this research, opportunities for future research and this thesis will end with the conclusion.
2. Literature review
Globalizing markets, technological developments, economic integration of national states and increased international competition have been the most important reasons for the unprecedented internationalization of business (Heijltjes et al 2003). Determining a firm’s “internationalization” is one metric by which stakeholders will assess a firm’s overall performance and future competitiveness (Ramaswamy et al., 1996). This internationalization can be measured by nationality diversity in a corporate board (Caligiuri, 2004). Nationality diversity is important given that individuals’ values, attitudes, and behavior tend to vary depending on nationality (Francesco and Gold, (1998), Hambrick et al. (1998), Hofstede (1980); Ronen and Shenkar, (1985)). Managers from a variety of countries are likely to possess a diversity of cultural values, attitudes, preferences, etc. Their diversity is likely to provide them with broader information resources, skill sets, and “cultural capital” (Caligiuri, 2004). Further, Athanassiou and Nigh (2000) argue intricate knowledge of foreign markets in TMT is seen beneficial for firms. Consequently, Nicholson & Kiel (2004) stress that good performance of the board of directors will contribute to the good performance of the company and, thus, can be said that national diversity in a TMT will contribute to this performance, which is also shown by i.e. Caligiuri (2004). Research where diversity issues in TMTs are considered as an independent variable is embedded in the ‘upper echelon perspective’, put forward by Hamrick and Mason (1984). However, this thesis considers nationality diversity as a dependent variable. This is namely, the effect of ownershipstructure and internationalization of ownership on nationality diversity in TMTs.
of foreign involvement in terms of foreign sales, geographic diversification or cultural diversification, the internationalization of ownership, an international orientation of the company, the type of industry and country of origin (Heijltjes et al. 2003). Several of the mentioned determinants have recently been examined by Van Veen and Elbertsen (2007), Van Veen and Marsman (2007), Kaufman (2007), Staples (2007) and Ruigrok and Greve (2008). Also, other potential determinants might explain nationality diversity in TMTs. These are on firm level, the type of board structure (1 vs 2-tier), whether a company is involved in cross border Mergers and Acquisitions (M&A’s). On country level these are proxies for country size, namely GDP and population, and the length of a countries’ EU-membership. Given the above mentioned determinants of nationality diversity, in-depth and extended research of the effect of ownershipstructure and internationalization of ownership and on nationality diversity in TMTs appears to be lacking. Following Berle and Means (1932), the conception of control in a corporation is focused on the ability to determine the composition of its TMT. There are several reasons why ownershipstructure and internationalization of ownership might affect nationality diversity in TMTs. Ownershipstructure might be a determinant because the degree of concentration of ownership might influence nationality diversity in TMTs. It might enhance nationality diversity because when there a more different shareholders (dispersed ownership), it is likely that these owners have another nationality. The presence of shareholders with different nationalities will increase the chance of selecting TMT members of foreign nationalities. It might diminish nationality diversity because when there a few shareholders (closely held ownership) it is likely that these owners have the same nationality. The chance that these owners will be have another nationality is therefore smaller than where ownership is dispersed. The presence of shareholders with the same nationality will increase the chance of selecting TMT members of the domestic nationality.
Further, Holderness and Sheehan (1988) showed that large blockholders (shareholders with more than 5% of the shares) in the USA often serve as members of TMTs which puts them in the position to influence strategic decisions directly. The reason for blockholders to influence strategic decisions directly is that the blockholders have a strong incentive to increase firm-value. Given the ongoing globalization of ownership, foreign investors are able to buy large share portfolios and it is likely that this might increase nationality diversity in TMTs. The next section will discuss in detail (with examples) how ownershipstructure and internationalization of ownership may affect nationality diversity in TMTs through the corporate governance mechanism.
Ruigrok and Greve (2008) examine the relationship between internationalization of ownership by focusing on share data. However, to measure internationalization of ownership there are several different and more reliable indicators (see section 4.1). Especially when you take into account that there exist deviations from the proportionate (“one share – one vote”) allocation of ownership and control across listed EU companies. This is not taken into account by Ruigrok and Greve (2008). Examples of deviations from the proportionate allocation between ownership and control are multiple-voting rights, voting-rights ceilings, ownership-ceilings and non-voting shares. These examples might prevent shareholders with ‘ordinary’ shares to exert influence on selection procedures of TMT members. Also, these kinds of constructions might protect the original owners of company from a takeover. Since this thesis examines the relationship between ownershipstructure and internationalization of ownership and nationality diversity in TMTs, this thesis uses the degree of control (=voting rights) as a proxy to exert influence on TMT selection procedures and strategic decision making1.
1
3. Corporate Governance
In this section corporate governance will be discussed. This is to show in detail how ownershipstructure and internationalization of ownership will affect nationality diversity in TMTs through the corporate governance mechanism. First, the fundamental issues of corporate governance will be discussed and recent examples will be given about how shareholder influence might affect TMT composition. Then, main differences of corporate governance among EU countries will be discussed. Finally, a brief summary will be given about recent developments which might contribute to increased nationality diversity in TMTs. These differences might affect nationality diversity in TMTs in the countries examined, as will be explained.
3.1 Fundamental issues and examples
Agency theory assumes that there is one party (the principal) that delegates work to another (the agent), who performs that work. Agency theory is concerned with resolving two problems that can occur in agency relationships. The first is the agency problem that arises when (a) the desires or goals of the principal and agent conflict and (b) it is difficult or expensive for the principle to verify what the agent is actually doing. The problem here is that the principal cannot verify that the agent has behaved appropriately. The second is the problem of risk sharing that arises when the principal and agent have different attitudes towards risk. The problem here is that the principle and the agent may prefer different actions because of the different risk preferences (Eisenhardt 1989). A major shareholder may be able to influence the composition of a TMT; the two problems that arise from the relationship between shareholder and TMT could be largely avoided then.
ways how ownershipstructure and internationalization of ownership might affect nationality diversity in TMTs. As indicated in section 2, the shareholder has an incentive to influence strategic decisions. Three ways how this can be accomplished will be shown, supported by practical examples.
First of all, to maximize shareholder value a shareholder might choose an executive with international experience or from another nationality than the firm’s country of origin since several studies have shown that nationality diversity seems to be an important driver for corporate performance (i.e. Caligiuri (2004)), also intricate knowledge of an executive seems to enhance performance (i.e. Athanassiou and Nigh (2000) and Tan and Mahoney (2006)). Such a situation might have happened in the Dutch company Vendex KBB in 2004. Shortly, after the American investment company KKR bought the majority of the shares the Dutch CEO was replaced by an Englishman. The reason given by the company to appoint the new CEO was that this foreigner had much international experience in restructuring organizations. Second, to determine and control strategic decision making of an organization a major shareholder might request for a board seat. Scott (1997) argued that it may well be that a bank insists that one of its executives be placed on the board of an enterprise that it seeks to dominate. This board seat might be executive or non executive in a one-tier board. For a two-tier board, a major shareholder might request an executive board or supervisory board seat. This way of taking a seat in a board (one or two-tier) is not uncommon, such happened at Fortis in the Belgium/Netherlands and Motorola in the USA. In 2007, a Chinese insurance company became the largest shareholder (4.2% of the shares) in the Dutch-Belgium bank Fortis. The CEO of this insurance company entered the TMT of Fortis because of his knowledge of the Chinese market, but it is likely that his shareholder would like to influence strategic decision making in order to increase firm value. In 2008, an American investor in Motorola increased his share to 6.4% and two representatives of this investment company entered the board to split up the company in two profitable divisions. Motorola further announced that the CEO of the investment company would be involved in the appointment of the new CEO of one of the two divisions.
All three examples show that through the corporate governance mechanism the influence of shareholders might affect nationality diversity in TMT. The influence of shareholders is measured in terms of voting power. In all examples there was a major shareholder who was able to influence the TMT composition, as argued by Scott (1997). Only in the case of Motorola there was no foreign party involved. Corporate governance practices such as the influence of shareholders on TMT’ composition differs per country; therefore in the next section the main differences in corporate governance between European countries will be discussed. Section 3.3 will discuss general developments of corporate governance in Europe.
3.2 Differences of European corporate governance
National differences in corporate governance result from variations in finance, law and culture. For an overview of relevant differences in corporate governance among EU countries see table 1. A large difference in corporate governance practice among EU Member States relates to the role of employees in corporate governance. In Austria, Denmark, Germany, Luxembourg and Sweden, employees of companies of a certain size have the right to elect some members of the supervisory body. France has three characteristics of employee influence. First of all, all companies can provide employees with the right to elect some members of the supervisory board. Second, when employee shareholding reaches three percent (3%) employees are given the right to nominate one or more directors. Finally, employee representatives may have the right to attend board meetings, but not vote. Employee representatives in the The Netherlands also have the right to attend board attend meetings, but are not able to vote on TMT selection procedures. In Finland, all companies provide employees the right to nominate one or more directors (which is the same as in France). In all other EU Member States, it are the shareholders alone who elect all the members of the supervisory body. This results in a fundamental difference among EU Member States in the strength of shareholder influence in the corporation (EU 2002). The employees’ role in supervisory boards result in less seats left for board members which are elected by the shareholders. Given this, it is also likely there are less foreigners in the board; see also Van Veen and Elbertsen (2007)
the size of the supervisory board. Both the one-tier board of directors and the supervisory board (two-tier) are elected by shareholders, but, as explained above, in some countries employees may elect some supervisory body members as well. A discussion of benefits of each system is beyond the scope of this thesis. Relevant is the way how shareholder might exert influence on TMT composition. In companies which have a one-tier structure shareholders might exert direct influence on TMT composition. In companies which are two-tier structured the shareholder does not have direct influence on TMT composition. The shareholder in these companies has an indirect influence on TMT composition, because the shareholder elects the supervisory board and they elect the executive board. However, it may be that the nationality diversity in an executive board may be a reflection of the nationality diversity in a supervisory board, since an overlap might exist between the supervisory and managerial board (see table 1). See also Van Veen and Elbertsen (2007) for the differences and similarities in the nationality diversity in one and tier board in Germany, the Netherlands and the UK.
Table 1 Characteristics of corporate governance in EU countries (EU 2002)
Country One/two tier board Employee
Role in Supervisory
body
Separate Supervisory & Managerial Leadership
Finland one no yes
Sweden one yes yes
Denmark two yes yes
Germany two yes not required
Austria two yes yes
The
Netherlands
two no yes
Belgium one no not required
France one no not required
Spain one no not required
Portugal one no not required
Italy one no not required
Greece one no not required
Ireland one no not required
UK one no not required
Luxembourg one yes not required
3.3 Recent developments
people across EU borders, the competitive pressures of globalization, the realization of new technologies, privatization of state-owned enterprises, the growth and diffusion of shareholding, and increased merger activity among large European corporations all stimulate interest by European issuers and investors, Member States and the European Commission in the similarities and differences between national corporate governance practices. Differences in corporate governance practices among EU Member States appear to be the result of differences in law rather than from differences in recommendations that emanate from the types of codes analyzed in this Study” (EU 2002).
The growing interest in corporate governance codes among EU Member States may reflect an understanding that equity investors, whether foreign or domestic, are considering the quality of corporate governance along with financial performance and other factors when deciding whether to invest in a company. Investors are willing to pay more for a company that is well governed, all other things being equal (McKinsey Investor Opinion Survey, June 2000).
4 Indicators and hypotheses
As indicated in section 1.1, this thesis uses self developed indicators for ownershipstructure and internationalization of ownership which might affect nationality diversity in TMTs. First a discussion will be held about the indicators used by Ruigrok and Greve (2008). After this the indicators and hypotheses of this thesis will be discussed.
The data of Ruigrok and Greve (2008) involves the geographical distribution of ownership origin and is expressed in percentages per country and has been retrieved from companies’ annual reports. A good aspect of their research is that they use panel data which shows an ongoing internationalization of TMTs in seven European countries. Further, their research is one the first which link internationalisation of ownership with nationality diversity in TMTs. Their data for ownership also has some limitations. First of all, their data are partly based on estimations since not all shareholders are identifiable. Therefore, the measure they use is in some ways not reliable.
Finally, the measure used by Ruigrok and Greve (2008) is computed as follows, shareholders with nationality X / the total number of all shareholders. This is not representative because, for example, when a shareholder from the UK holds 5% of the shares, it is not notified that 5% of all shareholders come from the UK, but just one shareholder from the UK. This means that the percentage held by a shareholder (=voting power) is not taken into account, while this makes a major difference (as shown in the literature review and corporate governance section). How this thesis simultaneously makes use of voting power and internationalization of the most powerful shareholders, see below. Although Ruigrok and Greve (2008) made a valuable contribution to research on the determinants of nationality diversity in TMTs, there is room left to improve the measurement method for internationalization of ownership. To improve the measurement method this thesis uses 4 indicators which might affect nationality diversity in TMTs. Two indicators for ownershipstructure and two for internationalization of ownership.
1. Ultimate ownership (UO)
independent TMT members) in the TMT. It is therefore likely that the size of the TMT is smaller. A smaller size of the TMT will decrease the chance of the presence foreign TMT members. Therefore, it has been hypothesized that:
H1: Nationality diversity in TMTs will be significantly lower in companies that are ultimately owned than those which are not ultimately owned.
2. The percentage of shares held by the 5 largest shareholders (A5) This number is computed as follows:
A5= sum of the percentages of shares held by the 5 largest shareholders.
H2: Higher ownership levels held by the A5 is negatively associated with nationality diversity in TMTs.
3. The percentage of foreigners among the 5 largest shareholders. (fA5) This number is computed as follows:
fA5 = the number of different nationalities of (A5 / 5 )*100.
This measure could directly be retrieved from the AMADEUS company reports, because the nationalities of the five largest shareholders are explicitly mentioned. This measure has been chosen to see the internationalization among the five largest shareholders of European listed companies. This measure includes the five largest shareholders with the most voting power and also nationality diversity of the five most powerful shareholders is taken into account. As shown in section 3.1 there are several ways how ownershipstructure and internationalization of the five largest shareholders might affect nationality diversity in TMTs. The main difference between hypothesis 2 and this hypothesis, is that this hypothesis focuses on the internationalization of the five largest shareholders. Therefore, the argument to construct this hypothesis will be related to internationalization of ownership, however, take into account that this only includes the 5 most powerful shareholders.
Following Ruigrok and Greve (2008), foreign (incl. minority) ownership is likely to lead the appointment of foreign executives, e.g. from the owners' home country. Firms may appoint one or more foreign TMT members in order to signal to markets and stakeholders that due attention will be paid to the interests of foreign as well as domestic interests. Therefore, it has been hypothesized that:
H3: Higher foreign ownership levels among A5 are positively associated with nationality diversity in TMTs.
4. The percentage foreigners among the recorded shareholder. (fRs) This number is computed as follows:
fRs = Foreign nationalities / The total number of recorded shareholders *100
the nationality where the country is listed. In general, the bias is there because there are more employees who have the nationality of the country-of-origin of the company; however, a larger number of recorded shareholders might also imply more foreign shareholders. When the number of recorded shareholders is higher, it is likely that the number of foreign nationalities is also higher. Also, a reason might be that more foreign shareholders among all recorded shareholders are more globally oriented and this might lead to faster approval of a proposed foreign TMT member.
H4: Foreign ownership levels among all recorded shareholders are positively associated with nationality diversity in TMTs.
To conclude an overview has been given of four testable hypotheses. Two hypotheses deal with ownershipstructure and two deal with the internationalization of ownership. The next section will give an overview of ownership structure and internationalization of ownership in listed European companies.
5. Data and descriptives
In section 5.1 an overview will be given about how the data on ownershipstructure and internationalization of ownership was retrieved, which decisions had to be made and from which sources the data come from. In section 5.2 general characteristics of the data of ownershipstructure and internationalization of ownership will be discussed. Section 5.3 describes the distribution of ownership among different type of shareholders. Section 5.4 describes the distribution of internationalization of ownership among different type of shareholders.
5.1 Data and sources
seemed to be inadequate and inaccurate. Several ownership structure studies use AMADEUS this database seemed to be suited for this thesis. AMADEUS shows the nationality of the shareholders and the percentage held by these shareholders in terms of voting power. So, the data on ownershipstructure and internationalization of ownership used in this thesis is retrieved from the database AMADEUS of Bureau van Dijk (2005). This company is specialized in providing company information to its customers in various forms. Their database AMADEUS is an information source for about 9 million public and private companies in 38 European countries. Bureau van Dijk retrieves its information from over 30 specialist regional information providers, annual reports, self executed surveys by phone, email and mail and from national registers (In The Netherlands this is ‘De Kamer van Koophandel’) which limits the possibility of bias. AMADEUS is only for non-financial companies. For financial companies this thesis relied on BANKSCOPE of Bureau van Dijk (2005). The information from BANKSCOPE is exactly the same as for AMADEUS.
The companies in this thesis are based on the dataset created by Van Veen & Marsman (2007) and Van Veen & Elbertsen (2007). Their dataset contains information regarding TMT composition, as well as firm- and country specific variables for 363 large listed multinational firms from the earliest 15 Members of the EU. The reason for selecting these firms is two-fold. First of all, these companies are often blue-chip companies, which implies that they have a relatively long corporate history. Furthermore, normally they are financially sound, and can be considered to be stable by definition and are considered to be the most important companies in these countries, when looking at market capitalization. Additionally, these companies are required by law to publish company data in their annual report, which means that the required data will be accessible for these sample firms. The selection of the 363 listed European multinational firms leads to a bias towards large, enlisted, and internationally operating firms. As a consequence, the descriptive results probably show an overestimation of the expected effects (Van Veen and Marsman 2007).
not usable because there was nothing disclosed about the nationality and the percentage held by these shareholders in terms of voting power. In some cases several shareholders held more than 100% of shares together. These companies were excluded because this was not reliable. In general, the indicators on ownershipstructure and internationalization of ownership are reliable. This is because legal restrictions for listed companies in Europe. In Germany and the UK a shareholder’ identity need to be disclosed when this shareholder holds at least 3% of the shares. In the other countries involved in this research this disclosure threshold is 5%. These disclosure rules are arranged by European and national law2.
Four companies were from Luxembourg and these have been excluded because these are not representative for listed companies in Luxembourg. This resulted in a minimum of 15 listed companies in Austria to a maximum of 35 listed companies in France, however with no dominance of companies in a certain industry (for which is anyway controlled for, see section 6.3). This resulted in 307 companies suitable for this research.
In some cases the nationality of the recorded shareholders was missing but in all these cases the nationality of the shareholder could be found For companies as shareholders, search engine Google was used and looking to the legal abbreviation for a recorded shareholder (such as the Dutch legal abbreviation NV, then it was clear what the nationality was). For Germany and Austria (both Aktiengesellschaft, abbreviated by AG), UK and Ireland (both Public Limited Company, abbreviated by PLC) and Portugal, Spain and Greece (S.A.) it was difficult, however, each time it was possible to assign a nationality to a shareholder. Determining the nationality of a company remains a difficult issue. In this thesis there has been focused on the legal nationality since this was explicitely mentioned in the AMADEUS reports. However, there are more ways of determining the nationality of a firm. More on this, see section 7. For individuals as recorded shareholders, the nationality was mentioned in the AMADEUS reports. In some cases, the search engine google was used to find which nationality an individual had.
With the retrieved indicators of ownershipstructure and internationalization of ownership from the company reports of AMADEUS and BANKSCOPE, tables with descriptive results are made. The next sections will discuss these indicators.
5.2 Description of general indicators of the dataset
The total number of companies involved in this research is 307 and is distributed among 14 European countries. An overview of the indicators of ownership and internationalization of ownership can be found in table 2.
Table 2 General characteristics data
Companies UO A5 fA5 fRs Finland 19 6 49,79 23,16 19,79 Sweden 22 7 48,22 16,82 18,77 Denmark 19 7 47,38 21,63 28,07 Germany 25 4 55,36 27,36 31,13 Austria 15 12 78,81 24,13 34,69 Netherlands 22 3 33,12 52,00 35,08 Belgium 16 7 61,27 24,73 32,77 France 35 8 78,48 30,43 28,35 SPAIN 31 11 53,06 34,81 8,53 Portugal 18 7 50,83 22,17 27,07 Italy 24 10 44,60 26,67 24,26 Greece 17 9 77,67 36,88 42,11 Ireland 16 3 30,32 31,25 21,88 UK 28 5 20,39 35,11 11,02 Luxemburg 3 0 43,63 60,00 35,71 SUM 307 99 Average 52,09 29,08 25,97
UO = Ultimate Ownership: a company with a minimum of 24.9% controlled by a
shareholder, with no other single shareholder owning a greater percentage (AMADEUS 2005).
A5 = Percentage of shares owned by the 5 largest shareholders.
fA5 = Foreigners among the 5 largest shareholders: the percentage of foreigners among
the 5 largest shareholders.
fRs = Foreign ownership levels among all recorded shareholders: the percentage foreign
nationalities of the total number of recorded shareholders.
In 51 companies out of 307 there were less than 5 shareholders, so the A5 was determined by summing the percentage held by less than 5 direct shareholders and resulted in A 1, 2, 3 or 4.
complete picture of control, in most cases there is mentioned which percentage is controlled by a shareholder. For better understanding see definitions above.
In total 32.24% (99 out of 307) of the listed companies in this dataset do have an ultimate owner. In Austria, 12 out of 15 have an ultimate owner. Of these 12 companies, 2 companies are ultimately owned by the State, namely Telekom Austria AG and OMV Aktiengesellschaft. It is not uncommon in Europe that a national telecom company is controlled by the State. At KPN in the Netherlands the State is a major shareholder with 14.25% of the shares and in France, France Telecom, the State even controls a total share of 34.91%. In Germany the state controls 23% of the shares of Deutsche Telekom and in Portugal the State controls shares of Portugal Telekom, however the percentage is not disclosed. In Belgium, even the State of France controls shares in Mobistar, but again the percentage is not disclosed. In Greece, Cosmote SA is also controlled by the State, but the percentage is not disclosed. In Italy, the State doesn’t control shares of Telecom Italia. In Sweden Teliasonera AB is controlled by the Swedish and Finnish Sate with a share of respectively 45.32 and 13.73%. In Spain, Portugal and Ireland State ownership couldn’t be found. Telecom companies from Denmark, Ireland and Finland were not included in the sample.
Since 12 out 15 companies are ultimately owned in Austria, it is not surprising that the percentage held by the A5 is the highest among all countries, namely 78.81%. This might indicate that ownership is closely held in Austria, especially when you take into account that there are only 13 recorded shareholders on average. Also in Belgium there is a high number of ultimate owners, namely 7 out of 16 companies are ultimately owned. In Belgium, the A5 is 61.27% and this much higher than the average A5 in Europe. The average recorded shareholders in Belgium is 13.82, and this might indicate that ownership is closely held in Belgium. Greece has the highest number of ultimate owners, namely 9 out of 17 and has the lowest number of different nationalities among the recorded shareholder. In Greece, the number of recorded shareholders is also the lowest and the A5 is among the highest. Just like in Austria and Belgium, this might indicate that companies are closely held in Greece.
ultimately owned. Overall can be said that shareholdersin French companies control high percentage of shares on average.
The section above has given an overview of the indicators used for this research. In section 5.3 will discuss which types of shareholders are present in 307 listed companies in 14 European countries.
5.3 Distribution of ownership
In figure 1 an overview is given of the distribution of ownership. This has been done by different types of shareholders which control shares in listed European companies.
Figure 1 Distribution of ownership
Ownership distribution 0,00 10,00 20,00 30,00 40,00 50,00 60,00 70,00 80,00 90,00 Fin land Sw ede n Denm ark Ge rman y Aust ria Net her lan ds Belg ium Franc e Spai n Portugal It aly Greec e Irel and UK Aver age Countries P e rcen ta ge Unnamed State Employees Foundations Individuals Pensionfunds Industrial Banks + Financials Insurance
To get a better understanding of the distribution of ownership among European companies, depicted in figure 1, it might be useful to read this together with table 3. Table 3 is about the the distribution of different types of major shareholders among European companies. This is to see on which numbers averages of ownership held shareholders are based and to see which types of shareholders are present in the companies used in the sample. It has to be read a follows, in Finland there are 19 listed companies which have an insurance company as major shareholder.
The first part of this overview deals with figure 1. The second part of this section deals with table 3.
Industrial companies: this category includes all companies that are not banks or financial companies nor insurance companies. They can be involved in manufacturing activities but also in trading activities (wholesalers, retailers, brokers, etc.). The last column, unnamed, are three types of shareholders combined, namely Public, unnamed private shareholders, and other unnamed shareholders. These are considered as unable to exert, as such, control over a company. These types of shareholders are put together because these shareholders cannot participate in any election procedures on the appointment of a TMT member and strategic decision making.
From figure 1 can be concluded that industrial companies control the highest percentage of shares of listed European companies. Especially in Austria, Spain, Portugal and Greece these are present. In Austria, 26.19% of the shares are controlled by industrial companies. In Austria there is great variation on the shares controlled by an industrial company. Agrana Beteiligungs AG for example, is controlled for at least 95.5% by industrial companies but there are also 5 out of 15 companies in Austria that do not even record industrial companies as shareholders.
In Spain, only Telefonica S.A. is not controlled by an industrial company. In contrast, Telefonica Moviles S.A. is held for 92.46% by an industrial company. This company is Telefonica S.A. There are three Telefonica companies in the dataset, these companies form a kind of holding where they control shares of each other. So, this influenced the percentage controlled by industrial companies in Spain. In Portugal, 31.81% of the shares are controlled by industrial companies; however there are also large variations in the percentage held. Finally, in Greece 24.98% of the shares is controlled by industrial companies. Further, 9 out of 16 don’t record any industrial companies. An interesting aspect is that in Spain and Portugal industrial companies control a large percentage of shares in banks, where in Austria and Greece this is very small.
ultimate owner is L'arche Holding SA and this is controlled by members of the Heineken family.
Interesting is that Austria, Spain, Portugal and Greece have the highest sum of percentages controlled by different types of shareholders. Due to regulations on ownership information disclosure companies have to report their major shareholders and since this percentage is high, this might indicate that ownership is closely held in these 4 countries. The Netherlands, Ireland and the UK report the lowest sum of shares controlled by different type of shareholders. This might indicate that ownership is widely held in these countries, since these countries don’t record a high percentage and therefore have a large number of shareholders who hold less than the minimum required percentage obliged to disclose.
Banks and financial institutions control 8.51% of the shares of European listed companies on average. As Levine and Kunt (1999) argue there are market- and bank based economies. In bank-based systems banks play a leading role in mobilizing savings, allocating capital, overseeing the investment decisions of corporate managers, and markets all grow and become more active and efficient as providing risk management vehicles. In market-based systems securities markets share center stage with banks in getting society's savings to firms, exerting corporate control, and easing financial management. The market-based economies in this research are Sweden, Denmark, The Netherlands, Ireland and the UK and from figure 1 it can be confirmed that this is indeed the case. These countries are among the countries where the lowest percentage of shares is controlled by banks and financial institutions. Surprisingly, Germany is among the lowest percentage controlled by banks and financial institutions. In general, Germany is considered as a country where banks and financial institutions are major shareholders. However, this can not be confirmed by this thesis. A reason for this might be that the 25 largest MNE’s from Germany are chosen and these are not representative for a typical German company. It is also possible that German companies have found other ways to attract capital in the recent years.
These shares are both controlled by one US pension funds, namely Capital Research and Management Company. In Italy, 12 out of 24 companies are controlled by pension funds. Sweden, Ireland, UK and show the highest percentage controlled by pension funds. Lucas (2004) argues that pension funds can be more powerful in listed Swedish companies and indicate a growing influence of pension funds on listed European companies. In Sweden only 1 company is not controlled by a pension fund. In Ireland, 2 out of 16 don’t record any pension fund shareholder. In the UK, this is 5 out 28 companies.
Individuals control 3.31% of the shares of listed European companies. Greece shows the highest percentage controlled by individuals. Actually, only 5 out of 17 of the companies record ownership by an individual; however, the percentages controlled by these individuals are high, varying from 30% to 80% of the total shares.
State governments control 2.4% in European listed companies. Especially in Finland 9 out of 19 companies have the State as shareholder. This all involved the Finnish State. States are also involved in cross border ownership, see section 5.4. Overall, these are mainly companies active in raw materials and natural resources. Spain, Portugal and Ireland show no State involvement at all.
Finally, the column ‘Unnamed’ combines Public, unnamed private shareholders, and other unnamed shareholders. The percentage controlled by ‘Unnamed’ are the highest in Austria, France and Greece. As shown in table 2 these companies also record the highest percentage controlled by the five largest shareholders. This might be due to national regulations. It might be the case that companies in these countries need to disclose low thresholds or they need to distribute all shareholders into a type without disclosing the identity of the shareholder. UK and Ireland do not report any ‘Unnamed’ shareholders, a reason for this might be that a market based economy is characterized by strong shareholder protection and good accounting standards and therefore each major shareholder need to be identified and distributed among the different shareholder types.
Industrial companies seem to be the shareholders that control the most shares in listed companies in Europe. This number is most dominant compared to other shareholder types within all the different countries. In Ireland, 5 listed companies are controlled by industrial companies. Kerry Group Plc is ultimately owned by an industrial company, namely its parent Kerry Co-op creameries Ltd. In Greece there are not many listed companies which are controlled by industrial companies. However, industrial companies as shareholder are still dominant among the shareholder types in Greece. Further, 7 out of 8 listed companies which are controlled by industrial companies are also ultimately owned by industrial companies. Among all the European companies which are controlled by industrial companies, Spain is the country which records the most industrial companies as shareholders.
Banks and financial institutions control shares of 185 listed European countries. In The Netherlands, Spain, Italy and the UK this is most dominant shareholder type. These countries are among the countries where the lowest percentage of shares is controlled by banks and financial institutions. Compared to figure 1, except for Italy, these are not the shareholder types that control the largest percentages of shares. In the Netherlands and the UK listed companies are not ultimately owned by Banks and financial institutions. In Italy there are 4 companies are ultimately owned by Banks and financial institutions, these are Banca Fideuram, Fiat, Mediaset and Terna. In Spain, Gas Natural is ultimately owned by a bank.
infrastructure. These companies were probably totally controlled by the State but since these companies were privatized the State has sold shares to private shareholders.
To conclude, there are great variances in the distribution among the different types of shareholder. However, groups of countries do also show great similarities such as The Netherlands, Ireland and the UK. In these countries control seems to be dispersed and more equally distributed over the different types of control. Also, Austria, Italy, France, Spain and Greece do have similarities. Ownership in these countries seems to be closely held.
5.4 Internationalization of ownership
In table 4 an overview is given of the internationalization of ownership. This has been done by different type of shareholders.
---insert table 4 about here---
To get a better understanding of the internationalization of ownership among European companies, depicted in table 4, it is useful to read this together with table 5. This is to see on which numbers averages of internationalization of ownership held shareholders are based.
---insert table 5 about here---
Table 4 and 5 are discussed together, so this section is not divided in different parts (as in section 5.3). Each mentioned percentage of internationalization in this section belongs to table 4. Each mentioned number in this section belongs to table 5.
The percentages of table 4 have been computed as follows:
The average percentage of foreign ownership in shareholder type A in country X / the average percentage of total ownership in shareholder type A in country X. * 100%
This has been done to see how internationalized ownership is distributed among listed European companies.
Sweden, Denmark and Portugal report the lowest fA5. Interesting is that the Scandinavian countries in this sample report in some cases similar descriptive results. Portugal also reports a low fA5, this was also expected since the number of different nationalities among the recorded shareholders was also among the lowest. The Netherlands, UK, Ireland, Greece and Spain report the highest fA5. The Netherlands, UK and Ireland also show the most dispersed ownership (A5 direct is the lowest among Europe, table 2).
The second column of table 4 shows the percentage foreign insurance companies in relation to total ownership held by insurance companies. Foreign insurance companies who hold shares of listed companies seem to be most active in Germany (2 companies record a foreign insurance company as shareholder), Belgium (1), France (3), Spain (4), Portugal (1), Italy (3) and the UK (3). Continental AG (Germany) reports the highest percentage held by a foreign insurance company (AXA, France) namely 11%.
Foreign banks and financial institutions in relation to total ownership held by banks and financial institutions is the second highest in Europe. In Portugal, 11.17% of the total shares held by banks and financial institutions is owned by foreigners. Still, there are 8 listed Portugese companies that record foreign banks and financial institutions as shareholders. The Portugese Banco BPI shows the highest percentage held by foreign banks, namely 20.72%. This 20.72 is distributed among the Brazilian Bance Itau SA (16%), Chase Bank USA (4%) and the Dutch bank ING (0.72%). The high percentage held by Bance Itau SA might be assigned to the historical and linguistic reasons.
Finland and Ireland show the highest percentage foreign banks and financial institutions in relation to total ownership held by banks and financial institutions, respectively 5 and 4 listed companies have foreign bank and financial institutions as shareholders. The banks and financial institutions who hold shares in listed companies in these countries are almost completely foreign. The Finnish Metso OYJ shows the highest percentage held by foreign banks, namely 9.48%. This is distributed among the American companies JP Morgan Chase (5.33%) and Franklin Resources (4.15%). A final remark on Finland and Ireland is that although a high percentage (from the total) of shares is held by foreign banks and financial institutions, the direct percentage held by these foreign banks and financial institutions on average is low.
The percentage foreign industrial companies in relation to total ownership held by industrial companies is the lowest in Belgium. In Belgium, 2 listed companies are held by foreign industrial companies. Bekaert NV shows the highest percentage held by foreign industrial companies, namely 8.94%. This 8.94% is distributed among companies from Panama (Beuval Enterprises Corp 4.47%) and the British Virgin Islands (Tirholding Inc 4.74%). France and Spain record shows the most companies who are held by foreign industrial companies, respectively 13 and 18. In the percentage shares (of the total shares) held by foreign industrial is in France 35.83% and in Spain 15.65%. The average percentage of shares held foreign industrial companies is in Fance 6.33% and in Spain 6.04%. The largest percentage held by a foreign industrial company has been found in Stmicroelectronics SA in France, namely this for 100% held by its Dutch parent. In Spain, Gestevision Telecino SA is for 54.15% held by foreign industrial companies, namely by the Italian Mediaset S.P.A. (52%) and by the American Fidelity International LTD (2.15%). In The Netherlands, 77.03% of the shares held by industrial companies is held by foreign industrial companies. Only 5 listed Dutch companies record a foreign industrial company. However, the 77.03% is strongly influenced by the shares of a foreign industrial company in Heineken NV; this is Lárche Holding SA with 50.01% of the shares. Also, Numico NV is majority owned by Biltrain AG, this German industrial company holds at least 50.01% of the shares. For the Netherlands, 5 listed companies are held by foreign industrial companies and these foreign industrial companies hold large percentages of shares.
to total ownership held by pension funds varies from 18.57% in Sweden to 100% in Austria. However, in Austria 2 companies record a foreign pension fund. In both cases this foreign pension fund is the American Capital Research & Management Company and holds shares in Erste Bank der Oesterreichischen Sparkassen AG (4.12%) and in Telekom Austria AG (11.3%). Sweden has 8 listed companies which have a foreign pension fund as shareholder. The percentage shares (of the total shares) held by foreign pension funds is 18.57%, this is the lowest in Europe. Spain is among the highest in number of listed companies which a foreign pension fund as shareholder and is also among the highest percentage shares (of the total shares) held by foreign pension funds, namely 65.08%. UK and Ireland report among the highest percentages held by pension funds, however, the percentage foreign pension funds in relation to total ownership held by pension funds is among the lowest in the UK and Ireland. A reasons for this might be that these countries, especially the UK, have strong developed financial markets with wide availability of capital and therefore less need to attract foreign capital. A remark on this is that the number of listed companies which record a foreign pension fund is in Ireland 11 and in the UK 17. So this implies that the average percentage held by foreign pension funds is low in the UK and Ireland.
Foreign individuals who hold shares in listed European companies are very few. In Portugal and the UK there are foreign individuals who hold shares in listed companies. These listed companies are O2 in the UK and EDP Energias in Portugal.
Listed companies which have foreign foundations, employees and unnamed shareholders as shareholders are not many among Europe. Belgium has 3 listed companies which has 3 foreign foundations as shareholders. These are Bekaert NV, Colruyt NV and Inbev, the foundations all come The Netherlands. A similar language and tax reasons might be reasons for this distribution.
Investment Authority and has 7.2% of the shares of Daimler Chrysler. In the UK Old Mutual Plc is held by the Public Investment Commission of South Africa and has 7.13% of the shares. In Belgium, Mobistar is held by the France State, as earlier discussed.
To conclude, there are great variances in the number of listed companies where foreign types of shareholder are active. Banks and Financial institutions, Pension funds and Industrial companies seem to be most international shareholders in listed companies in Europe in the number of listed companies where they are present. Also there are great variances in the percentage (of total shares) held by these shareholders.
6. Methodology
In this section the methodology and variables will be presented. To recall, the objective of this research is to examine the relationship between ownershipstructure and internationalization of ownership on nationality diversity in TMTs. Nationality diversity in TMTs as dependent variable and the indicators from section 4.1 are independent variables in an Ordinary Least Squares (OLS) regression to be runned by SPSS.
6.1 Dependent variable
The data of nationality diversity in European TMTs is retrieved from annual reports from the year 2005. In case of missing data, the electronic databases AMADEUS, Reach and websites Zoom, GoogleFinance, Zoominfo and Topmanagement were consulted (Van Veen and Marsman 2007). The steps taken by Van Veen and Elbertsen (2007) and Van Veen and Marsman (2007) in finding top management team data has led to minimization of the possible wrong nationality assignments. The database is constructed in May 2006 and only involved annual reports of the year 2005.
Nationality diversity in TMTs will be measured by the percentage of foreigners on the board. This is calculated as follows:
(Number of foreigners in the board / all members in the board) * 100%
Another measure, such as the number of foreign board members is not useful since this doesn’t say anything about the size of the board.
For the nationality of the company there were several options. The first option is to choose the country where the company is listed, however some companies examined are cross listed and that makes this option unusable because some companies then has to be excluded for no reason. The second option is to choose the nationality of largest shareholder, which holds at least 50.1%, however, following Heijltjes et al. (2003) board composition is affected by the national context and in a environment of free flow of capital the largest shareholder can be subject to change in short time periods. Finally, the country where the headquarters is located seems to be best measure of the nationality of the company.
6.2 Independent variables
The indicators of ownershipstructure and internationalization of ownership mentioned in section 4.1 will be used as independent variables.
1. Ultimate ownership (UO)
H1: Nationality diversity in TMTs will be significantly lower in companies that are ultimately owned than the nationality diversity in TMTs of those which are not ultimately owned.
To find statistical significance an UO-dummy will be included. There are 99 companies which are ultimately owned and 208 which are not ultimately owned. The purpose of the OLS regression analysis to find out whether the UO-dummy has a significant negative effect on nationality diversity in TMTs.
2. The percentage of shares held by the 5 largest shareholders (A5)
H2: Higher ownership levels held by the five largest shareholders is negatively associated with nationality diversity in TMTs.
The purpose of the OLS regression analysis to find out whether the percentage shares held by the five largest shareholders has a significant negative effect on nationality diversity in TMTs.
3. The percentage of foreigners among the 5 largest shareholders. (fA5)
The purpose of the OLS regression is to find out whether foreign ownership levels among A5 have a significant positive effect on nationality diversity in TMTs.
4. Foreign Nationalities / The total number of Recorded Shareholders *100. ((Fn/Rs)*100) H4: Foreign ownership levels among all recorded shareholders are positively associated with nationality diversity in TMTs.
The purpose of the OLS regression is to find out whether foreign ownership levels among all recorded shareholders have a significant positive effect on nationality diversity in TMTs.
6.3 Overview of independent, control variables and dummies
As indicated above, this thesis is part of a larger set of studies initiated by Mr. K. van Veen. The research of the European Top Managers project resulted in several determinants of nationality diversity in European TMTs, such as Kaufman (2007), Van Veen and Elbertsen (2007) and Van Veen and Marsman (2007). Therefore, several control variables are included:
Firm-Specific:
Country dummy variables: created manually based on Incorporation Code-Foreign
(Country Incorporation Code). Sample firm is given a 1 if it belongs to the relevant country and a 0 if this is not the case. Dummy variable.
Cross-border M&A dummy: created manually to indicate whether a sample firm has been
involved in a cross-border M&A deal involving $ 1 billion or more during the sample period; the sample firms is assigned a value of 1 if such an M&A deal has taken place, and a value of 0 if this is not the case. Dummy variable.
D1 – D11 Industry dummies: created manually by making use of the first digit of the
Primary SIC code. Hence, d1 captures all the firms that have a SIC code starting with 1, etc. Definition of Primary SIC code: The Industry Classification Code is a four-digit system of classification that identifies a company’s primary operations. Standard & Poor’s Compustat assigns these codes by analyzing the sales breakdown from a company’s 10K (annual report required by the Securities and Exchange Commission each year) and annual report. The assigned classification is reviewed each year when the company is updated by analyzing the product-line breakout in the 10K or annual report. Dummy variable.
Employees: represent the number of both full and part time employees of the company.
Number of countries active: represents the number of countries the sample firm is active in
for the year 2005. This information is based on annual reports and corporate websites.
Sales: represents the net sales or revenues of the company converted to U.S. dollars using the
fiscal year end exchange rate (in millions $). Control variable.
Ultimate Ownership dummy: created manually to indicate whether a sample firm has an
ultimate owner. This is based on AMADEUS and BANKSCOPE company reports. Sample firm is given a 1 if it has an ultimate owner, and a 0 if this is not the case. Ownershipstructure indicator.
Ownership concentration: created manually and is the percentage held by the 5 largest
shareholders. This is based on AMADEUS and BANKSCOPE company reports. Ownershipstructure indicator.
Foreigners among the 5 largest shareholders: created manually and is the percentage of
foreigners among the 5 largest shareholders. This is based on AMADEUS and BANKSCOPE company reports. Internationalization of ownership indicator.
Foreign ownership levels among all recorded shareholders: created manually and is the
percentage foreign nationalities of the total number of recorded shareholders. This is based on AMADEUS and BANKSCOPE company reports. Internationalization of ownership indicator.
6.4 Descriptives of the variables
In this section the descriptives of the (in) dependent variable(s) will be discussed. For an overview, see table 6.
Table 6 Characteristics of the (in) dependent variable(s)
Variable N Mean Median Std.
% foreigners among all recorded shareholders independent ((Fn/Rs)*100) 307 20 19 13 0 0,71
Table 6 shows that the average percentage of foreigners in the board is 18% in the 307 firms in this sample. Out of 2149 (7*307) observations, none were missing. There is a great variance in international orientation in terms of number of countries active, this range from 1 to 220 countries. Countries in the EU do have different regulations on the disclosure of shareholders and the shares they hold, (also see section 5.1), in some countries in the EU, such as Spain and Greece, all shareholders need to be identified. In other countries, such as the UK and The Netherlands this is not the case. This influences the reliability of one measure of internationalization of ownership, Fn/Rs*100. The other internationalization of ownership measure, fA5, can be considered as a very reliable because the nationality of the five largest shareholders was explicitly mentioned in the AMADEUS and BANKSCOPE company reports. The ownershipstructure measure, A5, can be considered as a very reliable measure because the percentage of each of the five largest shareholders was explicitly mentioned in the AMADEUS and BANKSCOPE company reports. In 49 companies there were less than 5 largest shareholders. The other ownershipstructure measure (dummy), UO, is also a very reliable measure since this could directly be retrieved from the AMADEUS and BANKSCOPE company reports.
Since this thesis takes one of the first steps in explaining nationality diversity in TMTs by ownershipstructure and internationalization of ownership, different measures of these determinants might be useful to gain better understanding of its effects and to see which measures are useful.
In this thesis there is a dummy variable used for ultimate ownership (H1). The frequency of this independent variable is as follows, see table 7.
Table 7 Frequencies of dummy variables Ultimately owned 99 (32.25%)
Not Ultimately owned 208 (67.75%)
Industry dummy variables are also included in this thesis. The distribution of can be found in table 8.
Table 8 Frequencies of industry dummies
Mining (D1) 11 3.58% Utilities (D2) 21 6.84% Construction (D3) 15 4.89% Manufacturing (D4) 107 34.85% Wholesale (D5) 13 4.23% Retail (D6) 16 5.21% Transportation & Warehousing (D7) 10 3.26% Information (D8) 36 11.73%
Finance, Insurance & Real Estate (D9)
60 19.54%
Services (D10) 13 4.23%
Miscellaneous (D11) 5 1.63%
Total 307 100%
From table 8 can be concluded that the majority of the companies in this sample are from the industries Manufacturing and Finance, Insurance and Real Estate. The sample set may be biased towards a certain industry, therefore, there is controlled for in the multiple regression analysis. To see in which industries ultimate ownership is common, see table 9.
Table 9 industry - ultimate ownership
Industry Non ultimately owned Ultimately owned Total Mining 7 2.28% 4 1,30% 11 Utilities 13 4,23% 8 2,61% 21 Construction 8 2,61% 7 2,28% 15 Manufacturing 68 22,15% 39 12,70% 107 Wholesale 8 2,61% 5 1,63% 13 Retail 11 3,58% 5 1,63% 16 Transportation & Warehousing 7 2,28% 3 0,98% 10 Information 23 7,49% 13 4,23% 36
Finance, Insurance & Real Estate
49 15,96% 11 3,58% 60
Services 11 3,58% 2 0,65% 13
Total 208 67,75% 99 32,25% 307
From table 9 can be concluded that the industries Manufacturing and Finance, Insurance & Real Estate have the most ultimate owners in the sample. This might have implications for nationality diversity in the TMTs of these companies, as hypothesized. This will be found out at the results of the regression analysis.
To see how the industries are distributed among the 14 countries, see table 10 in the appendix.
--insert table 10 about here--
6.5 Empirical results
This section is divided in three parts. As indicated above, this research is part of a larger research project on determinants of nationality diversity in TMTs. Therefore, there has been chosen for a gradual approach for presenting the empirical results. First, the control variables (the variables of the original model) will be tested in a multiple regression analysis. Second, the control variables and indicators of ownershipstructure and internationalization of ownership will be tested. This is to see what the indicators used in this thesis add to the original model. Finally, the control variables, the indicators and country dummy variables will be tested. There has been controlled for separate countries because differences between countries seem to be an important determinant for nationality diversity in TMTs (as shown by Van Veen and Marsman 2007 and Kaufman 2007).
The multiple regression analysis has been used to control for the different relations between the independent variables. To control for factors which can influence nationality diversity in TMTs, there has been made use of number of employees, total sales, number of countries where the country is active, M&A dummy and industry type as control variables.
In order to avoid multicollinearity between the independent variables a correlation test has been employed. For an overview of the correlations between the independent variables, see table 11 in the appendix.
