RuG Master Thesis Title:

(1)

Faculty of Economics and Business

University of Groningen

RuG

Master Thesis Title:

The Commercial Network as a Determinant of

Family Ownership

in Italian Textile Industry

Author: Chun Ling Zhang Student Number: s1624563 Email Address: C.L.Zhang@student.rug.nl

(2)

Abstract

The purpose of this paper is trying to address the gap by exploring the relationship between commercial network and family ownership in Italian textile industry. The analysis is based upon the data of 271 companies in Italian textile industry from Amadeus. Ordinary Least Square regression model and Logistic regression model are used depending on whether the dependent variable is continuous or binary variable. The results from these two models are in consistent indicating that commercial network has positive effects on family ownership in Italian textile industry.

(3)

1. Introduction

Anthropologists have often proposed that the family in a traditional society forms the primary economic unit. The economic roles of the family have gradually diminished in some cases. However, in some other cases, the family’s socio-economic mode of production remains important1.

Many firms that are now public firms actually were family firms in the past2. Family ownership is always a prevalent topic and a stream of empirical researches has already been done with the intention of understanding the existing of family firms well.

Under the agency theory, James (1999) and Greenwood (2003) argue that the agency costs are lower within family firms compared to non-family firms because of the fusion of ownership and control. Gedajlovic and Carney (2009) applied transaction cost theory to their study of family firm and argue that family governance provides advantages in developing, sustaining and appropriating value from generic non-tradeables together with other assets. Under institutional theory, Leaptrott (2005) sees a family firm as an organization that has an identifiable structure, because variation in family structure is not always considered as a predictor or control variable in descriptive research models. Naldi and Nordqvist (2008) argue that under the resource dependency theory family firms with external ownership, an external Chief Executive Officer (CEO) and larger top management team are more likely to attract the resources needed and therefore increase the scale of a family firm’s international operations. Under stewardship theory, altruism, participative strategy process and control concentration are important family specific resources that may help easing relationship conflicts and therefore maintaining the family firms’ success (Eddleston and Kellermanns, 2007). The application of network theory in the governance of firms focuses on vertical and horizontal relationships, where the different types of ties play important roles in different business environments (Mesquita and Lazzarini, 2008; Wallevik, 2009a).

The theories applied in these researches have different focuses on different elements, for example, agency theory focuses on effects of the separation of ownership and management in family firms; Transaction cost theory focuses on firm boundaries, ownership and financing structures in family firms; Institutional theory focuses on effects of regarding family firms as an institution; Resource dependency theory focuses on power and resource aspects of relationships in family firms; Stewardship theory focuses on effects of no alignment problems between owners and managers; Network theory focuses on effects of relationships in vertical and horizontal networks.

In this paper, the social capital theory is chosen to be used to study family firms. This theory focuses on resources embedded within networks of relationships and argues that networks of relationships comprise valuable resources for the management of social affairs by offering their

1_{http://en.wikipedia.org/wiki/Family}

(5)

members with the backing of collectively-owned capital, a “credential” which entitles them to credit (Bourdieu, 1986). The resources within networks give important and great contributions to family firms’ success however this is neglected by the other theories. Social capital theory filled the gap by contributing to examine different resources to different levels within networks.

There are a few other researches examining the situation of family ownership in textile industry. For example, Thomsen and Pedersen (1998) examined six corporate ownership structures between main industries based on a dataset of 100 largest companies in 12 European countries and found out that family ownership was significant in textile, shipping, retailing industries compared to other industries. Further, Franks, Mayer, Volpin and Wagner (2009) compared the evolution of family ownership in the top 1000 companies in the largest four European economies, France, Italy, Germany and United Kingdom, and the data showed that the Italian textile industry had the highest degree of family ownership among all of these four economies. And in Italy, textile and financials were the industries with highest concentration of family ownership. In conclusion, there are two interesting facts about Italian textile industry. The first one is that the case of concentrated family ownership in textile industry is not only in Italy but also in some other countries. The second one is that the concentration of family ownership in textile industry is higher than other industries in Italy.

The textile industry is characterized as capital intensive, highly automated and less flexible (Nordas, 2004) because of the integrated manufacturing processes, the long-term lead time and the difficulties to adjust to the customers’ tastes. Traditional theory argue that ownership is supposed to be dispersed in industries of being highly risky, capital intensive and limited access to capital (Thomsen and Pedersen, 1998). However, the case of Italian textile industry is a contradiction of this which is the third interesting fact.

The fourth interesting fact is that some researches argue about the disadvantages against family ownership arising within families, between family members (Klein, 2010), such as complexity, inefficient information, lack of discipline, limited access to capital sources and risk aversion3. And many of the family firms failed to survive for a long life and around 95% of them can not survive the third generation of family ownership4_{. Some family firms evolve} into widely held companies as they age, while some family firms such as family firms in Italian textile industry choose to stay family owned in stead of widely-held for several generations. Why is this? However, no research has been done before explaining these interesting facts.

Motivated by these interesting facts, this paper tries to address this gap by focusing on Italian textile industry as a case study. Based on Wallevik (2009c), where commercial network is studied as a determinant to family ownership in the shipping industry in Norway, this paper aims at analyzing the relationship between family ownership and network structure. The main research question is “what is the impact of network structure on ownership”? The

(6)

sub-questions are:

1) What is family ownership and how can we measure this (This question will be answered in section 2 and section 4).

2) What is network structure and how can we measure this (This question will be answered in section 3 and section 4).

3) What is the impact of network structure on family ownership (This question will be answered in section 3)?

4) Can we identify the impact of networks on family ownership in reality (This question will be answered in section 4)?

Commercial network can be expressed in the way of investments in other industrial firms, based on partnerships, co-investments, or other business cooperation (Wallevik, 2009d; Curto and Molho, 2002; Sugiyama and Grove, 2001), and offers ways to overcome shortages of capital sources, to access to other markets and social organizations, to reduce the risks, to enhance the competitive capability and to facilitate the expansion of business (Rose, 2000; Wallevik, 2009d) through the social capital embedded within. Thus I argue that the commercial network can be the determinant of family ownership in Italian textile industry due to the social capital embedded within. The argument is examined using data of 271 companies in Italian textile industry from the database of Amadeus. The results are in consistent with Wallevik (2009c) showing that commercial network has a significant and positive effect on family ownership in Italian textile industry.

This paper makes three main contributions.

First, this paper fills the research gap by examining the prevalence of family ownership with a focus on firms in Italian textile industry about which no researches has been done yet. Second, this paper applies a very important theory, social capital theory, to study the determinant of family ownership in Italian textile industry. Under this theory, this paper provides a deepened perspective to understand the existing of family ownership.

Third, one purpose of this paper is to compare the result with that of Wallevik (2009c). Based on the same definitions of variables as Wallevik (2009c) used, the result from this paper shows a significant and positive effect of commercial network on family ownership which is in line with Wallevik (2009c) just as I expected. This result may be useful for future researches in two directions. First, commercial network may be also a determinant of family ownership in some other industries of a certain country or some countries when other industries share common characteristics with Norwegian shipping industry and Italian textile industry. Second, commercial network may be also a determinant of family ownership in textile industry in some other countries.

(7)

hypotheses and part four consists of method. The last part consists of conclusion.

2. Family Characteristics

One concern about this paper is how to define family firm and there is no authoritative common definition. Generally speaking, a family firm is a firm in which one or more members of a family or one or more families have a certain level of ownership interests, management or control over the firm5. This perspective emphasizes three elements: ownership interests, management and control. And family firms can be owned at different levels by persons who are not family members. Family firms can also be managed or controlled by someone who is not members of the family. However, mostly family members will be involved in the operations of the firm to some extent. Thus, the specific definition of family firm depends on how these three elements are included in the definition (Litz, 1995).

For example, Randoy and Goel (2003) focus on founder influence on family control (CEO or chair) because they believe this kind of firms have a unique governance structure. Ehrhardt and Ehrardt, Nowak and Weber (2006) also focus on family control documented by a voting rights concentration of over 50 percent. Under this kind of perspective, participation of families in the management or ownership interests is no long important for defining a family firm.

Some papers focus on both ownership interests and family control (Sraer and Thesmar, 2006; Amit and Villalonga, 2006; Villalonga and Amit, 2004). They report a family firm when the family members have both shares and voting rights.

While some other papers focus on ownership and management (Anderson and Reeb, 2003). They report a family firm when the family members have both shares and board positions. The definition of family firm in this paper is in line with Wallevik, (2009c) with a focus on the family’s involvement of ownership because one purpose of this paper is to compare the result with that of Wallevik, (2009c). More specifically, when the percentage of family ownership owned by one or more known individuals or families is at least 20% (20% is included), then this company is considered as a family company, otherwise, not. The terms “family-owned firms”, “family-controlled firms” will be used in this paper that all refer to family firms.

2.1 Advantages of Family Firms in a Social Capital Theory Perspective

Family firms are prevalent and substantial in many industries and countries in the world details of which are listed in Table 1 in the Appendix.

--- Insert Table 1 about here

(8)

---

Many researches focus on investigating this prevalence and conclude that there are some distinctive advantages of family firms from which they can derive significant competitiveness (Cadbury, 2000). Social capital theory focuses on different valuable resources embedded within networks of relationships (Bourdieu, 1986) and it responds to these researches. Under social capital theory, family researches argue that some valuable resources bring advantages to family firms. Thus some of these advantages are as follows6:

Commitment7

The family members have more incentives to make the family firms successful and they would be very proud to pass the firms to next generations. In order to achieve this purpose, family members have to work harder than owners from non-family firms.

Altruism

Altruism is selfless concern of the welfare of others8. Parents are often selfless on behalf of their children and the same other way around. Families tend to be deeply altruistic (Becker, 1981; Schulze and Lubatkin et al., 2001; Zick, 1992). And it was argued that family traits, such as trust, altruism and paternalism can encourage an atmosphere of love and commitment towards the business (James, 1999; Danco, 1975; Poza, 1989; Steier, 2003), also a unique organizational setting and approach to decision-making (Zick, 1992 ). Altruism is treated as one of the most important factors that make the performance different between family firms and nonfamily firms.

Trust

Trust generated by family networks and close-knit communities has the advantage of ensuring a combination of incentives, effective monitoring, and loyalty to protect against the danger of poor governance, poor performance and external uncertainty (Colli, Perez and Rose, 2003; Rose, 1999).

2.2 Disadvantages of Family Firms

However, it is also true that many of the family firms failed to survive for a long life and around 95% of them can not survive the third generation of family ownership9. This is because there are also risks and disadvantages associated with family firms arising within

6_{IFC, IFC family business governance handbook. Available at: http://www.ifc.org/ifcext/corporate}

governance.nsf/AttachmentsByTitle/Family+Business_Second_Edition_English+/$FILE/English_Family_Business_Final_2008.pdf

governance.nsf/AttachmentsByTitle/Family+Business_Second_Edition_English+/$FILE/English_Family_Business_Final_2008.pd.

8_{http://en.wikipedia.org/wiki/Altruism}

(9)

families, between family members who are actively working in the business (Klein, 2010). Some of these disadvantages are the same ones that could also explain the failure of any other business, i.e., bad management, insufficient cash to fund growth, inadequate control of costs, firm, industry, national and other macro conditions10. However, some other disadvantages are family specific and they are shown as follows:

Complexity

Usually, family businesses are more complex than non-family businesses in terms of governance because of the additional element: the family. This additional element will result in some other consequent influences, such as, different management and ownership style compared to non-family businesses, and non-aligned incentives11. Klein (2010) defines three characteristics, which comprises of a family business and the relevant elements that will increase the complexity of a family business. According to Klein (2010), the first characteristic is family. And the number of family members, the range of these members, the geographic distance, the difference between the individual goals and the subjective relevance of those goals can increase the complexity of family and consequently increase the complexity of family business. The second characteristic is management. And the number of directors, the difference between the particular qualifications, the difference of personal origin and life experience, the difference of the respective goals and their subjective relevance for the individuals will increase the complexity of management and consequently increase the complexity of family business. The third characteristic is ownership. And the number of owners, the differences of their perceived roles as well as the differences of the different share sizes will increase the complexity of ownership and consequently increase the complexity of family business.

Informality12

Most of the time, the informality situation within family firms is less efficient than non-family firms. This is because in the earlier stages of family firms when the firms are managed by founders or second generation, they seldom pay attention to the informality situation of their firms. However, when the firms become bigger in size or non-family- member managers are involved, problems show up.

Lack of discipline13:

The families ignore disciplines about the succession of CEO and important managers and how to get good non-family managers. This would cause problems for the firm afterwards.

10_{IFC, IFC family business governance handbook. Available at: http://www.ifc.org/ifcext/corporate} 11_{IFC, IFC family business governance handbook. Available at: http://www.ifc.org/ifcext/corporate} 12_{IFC, IFC family business governance handbook. Available at: http://www.ifc.org/ifcext/corporate}

(10)

Limited access to capital sources:

When the size of the company grows, the need for capital in public market place increases, and the ownership is expected to be more dispersed (Berle and Means, 1932). However, the family ownership is very concentrated. Thus family ownership has this limit in access to capital sources (Shleifer and Vishny, 1997).

Risk aversion

Risk aversion is the reluctance of a person to accept a bargain with uncertain but higher payoffs rather than a bargain with certain, but maybe lower payoffs14. One reason is that concentrated ownership’s capital source is always more limited compared to collective capital sources. Another reason is that with concentrated ownership, the risks can not be shared by investing in more firms (Wallevik, 2009a). Third reason is that larger stakes maybe difficult to be sold in the secondary markets than smaller stakes (Becht, Bolton and Roell, 2002). So in the end, family ownership is expected to be negative in this way.

2.3 Family Governance15

Corporate governance refers to the structures and processes for the direction and control of firms. More specifically, it deals with the relationships among managers, board of directors, controlling shareholders, minority shareholders and other shareholders. A good corporate governance structure is expected to ensure a sustainable accountability by enhancing the performance of firms, reducing or eliminating the conflicts and increasing the firms’ access to outside capital. Most of the corporate governance experts focus on balancing the interests among managers, shareholders and boards in non-family firms.

However, the governance in family firms is more complex than non-family firms because of the additional element: the family.

In a typical non-family firm, any individual can be an employee, a manager, a director, a shareholder or a combination of these roles. However, when an individual in a family firm, carries multiple roles or responsibilities, the situation within the family firms will become more complex. This is because these multiple roles have different incentives and will therefore create challenges and conflicts for family firms (Neubauer and Lank, 1998).

Owners (Shareholders)

If owners in a family firm carrying multiple roles, conflicts and non-aligned incentives will show up. For example, there is an opportunity to reinvest the profits in the family firm and

14_{http://en.wikipedia.org/wiki/Risk_aversion}

(11)

otherwise these profits would be distributed as dividends. An owner who also works in the family firm would probably decide to reinvest because the owner already gets salary from the firm. But an owner, who does not work in the firm, would probably reject this reinvestment because the owner relies on the dividends as the main source of income. This kind of situation would be more complex with more conflicted incentives if the owner carries more roles.

Managers (Senior management)

Different roles of managers will also create conflicts in the family firm. For majority of the family firms, it is common to save the senior manager’s position for family members. This will create difficulty to attract and retain non-family talented managers because they always know that no matter how hard they work, they almost can never get equal treatment compared to family managers. So a clear and fair governance structure is needed to align the non-family managers’ incentives and their performances.

Directors (Board of directors)

Firstly, family firms usually prefer to reserve the directors’ positions for family members in order to keep the directions of business under family control. Only under some rare occasions trusted non-family directors will be accepted. Secondly, when there is an opportunity to reinvest the profits as stated earlier, family directors who is also the manager would probably choose to agree in order to get better growth, while family directors who does not work in the firm would probably choose to reject the investing opportunity in order to maintain the dividends. And this will cause conflicts to the firm.

Family members (The family and its institutions)

The access to information about the firm’s operations is not balanced between family members who work in the firm and those who do not work in the firm. Usually, the family members who work in the firm can get more information within a shorter time period than family members who do not work in the firm. This will cause conflicts and therefore a governance structure with transparent, timely information channels are needed to keep all the family members well informed of all the relevant information about the firm.

Considering of these particular potential conflicts resulting from multiple roles and the family specific disadvantages which were discussed in the earlier part, a well-functioning governance structure is needed.

(12)

with the business by making clear of the family vision, core values, mission and policies. The actual content or components of constitution differ from one family firm to another because of the differences of the firms in their size, age, development and the evolvement of families, but the core contents usually include:

Family Values, mission statement, and vision;

Family institutions, including the family assembly, the family council, the education committee, the family office and so on;

Board of directors (board of advisors if there is one); Senior management;

Authority, responsibility, and relationship among the family, the board and the senior management;

Policies regarding important family issues such as family members’ employment transfer of shares, CEO succession, and so on.

The family institutions make the relationships within family firms harmonious by allowing family members to get together discussing aspects related to the firm under one or more organized structures. These organized structures will provide mechanisms to help understanding and achieving agreements among family members. And they can have different forms, such as, family assembly, family council, family office and other family institutions.

3. Theory and Hypotheses

3.1 Theory

There are a few theories with different focuses on different elements that researchers used to study family firms. Which theory to choose to use is determined by the research problem and the contextual environment of the research. Some of these theories are as follows (See also Table 2):

--- Insert Table 2 about here ---

Agency theory

(13)

interests, because the principal just hires the agent to pursue the interests of principal16. Different governance mechanisms are means to deal with this problem to align the interests of principals and agents in different ways with minimum costs. And agency theory predicts a lower incidence of agency costs within family firms compared to non-family firms because of the fusion of ownership and control (James, 1999; Greenwood, 2003).

Transaction cost theory

Transaction cost theory is concerning about the costs of writing good contracts, contractual incompleteness as well as the cost of renegotiating contracts (Coase, 1937; Williamson, 1975; Williamson, 1985; Hart, 1995; Wallevik, 2009a). In transaction cost economics (Coase, 1937; Williamson, 1975; Williamson, 1985), the focus is on whether the firm decides to outsource goods or services through an arm’s length market transaction or to produce the goods or services on its own (Mustakallio, 2002; Wallevik, 2009a). Market price is not the only factor and there are also some other factors, such as, transaction costs, search costs, contracting costs and coordination costs17_{. These costs can be determinants for a firm to decide to outsource or} produce within the firm which is the essence of the market-vs-buy decision18.

Gedajlovic and Carney (2009) applied transaction cost theory to their study of family firm and they identified a group of assets termed as generic non-tradables (GNTs) which are firm-specific, but generic in application. They further reasoned that family firm’s governance provides advantages in developing, sustaining and appropriating value from GNTs through combinations with other types of assets. And these advantages together with other combinations explain the success of family firms.

Institutional theory

Institutions can be defined as governance structures based on rules, norms, understandings and routines (March and Olsen, 1989), or as social patterns characterized by standard sequences of interaction (Jepperson, 1991; Wallevik, 2009a). Institutional theory focuses on the deeper and more resilient aspects of social structure and it concerns about the processes by which structures, including schemas, rules, norms and routines, become established as authoritative guidelines for social behavior19. Different components of institutional theory explain how these elements are created, diffused, adopted and adapted over space and time; and how they fall into decline and disuse20. Family firms can be seen as an organization that has an identifiable structure, because variation in family structure is not always considered as a predictor or control variable in descriptive research models (Leaptrott, 2005).

16_{http://www.investopedia.com and http://en.wikipedia.org/wiki/Agency_theory.} 17_{http://www.sourcingmag.com/dictionary/Transaction_cost_theory-203.htm} 18_{http://www.sourcingmag.com/dictionary/Transaction_cost_theory-203.htm} 19_{http://en.wikipedia.org/wiki/Institutional_theory}

(14)

Resource Dependency Theory

Resource dependency theory is concerning about strategic behaviors of organizations in the process of managing and controlling scarce resources within organizations through interdependencies with other organizations in their environments (Mustakallio, 2002; Wallevik, 2009a). Organizations are not self-sufficient and therefore become dependent on outside suppliers who have different resources (Pfeffer, 1982). Pfeffer (1982) reasoned that there are two elements under resource dependency theory: the first one is the external constrains, where the organization needs to respond to the demands of the organization that controls the critical resources. Second element argues that managers try to manage their external dependencies to ensure success of the organization and also to get more autonomy and freedom from external constrains.

Naldi and Nordqvist (2008) argue that under the resource dependency theory family firms with external ownership, an external CEO and larger Top Management Team are more likely to attract the resources needed to increase the scale of a family firm’s international operations. Stewardship theory

Different from agency theory which focuses on the monitoring role of management in order to deal with the conflicts between principals and agents, stewardship theory offers an alternative view which focuses on the managerial motivation, which has applications to family firms (Donaldson and Davis, 1991; Wallevik, 2009a). More specifically, stewardship theory indicates that mutually trusting relationships, involvement-oriented environments and empowering organizational structures will increase pro-organizational behaviors and firm performance (Corbetta and Salvata, 2004; Davis, Schooman and Donaldson, 1997; Eddleston and Kellermanns, 2007). Altruism, participative strategy process and control concentration are important family specific resources that may help easing relationship conflicts and therefore maintain the family firms’ success (Eddleston and Kellermanns, 2007).

Network theory

Usually social networks are used to investigate how organizations interact with each other through informal connections that combine executives, associations, and connections between individual employees at different organizations (Wallevik, 2009a). The use of network theory in the governance of firms focuses on vertical and horizontal relationships, where decision makers can build cooperative governance structures and where the different types of ties play important roles in different business environments (Mesquita and Lazzarini, 2008; Wallevik, 2009a).

(15)

give important and great contributions to family firms’ success however they are neglected by other theories.

Thus, social capital theory filled the gap by contributing to examine different resources to different levels within networks. Under this theory, I argue that social capital as a core resource embedded within commercial network can explain the success of family firms.

Social capital theory

Family researches use social capital theory to understand family firms and they argue that higher level of social capital creates more advantages for family ownership (Arregle, Hitt, Simon and Very, 2007).

The central proposition of this theory is that networks of relationships comprise valuable resources for the management of social affairs by offering their members with the backing of collectively-owned capital, a “credential” which entitles them to credit (Bourdieu, 1986). And most of the valuable resources are social capital which is embedded within networks of mutual acquaintance and recognition (Bourdieu, 1986). In order to clarify the applications of social capital theory to family firms’ researches, the term of “social capital” is needed to be defined first.

The term of social capital initially showed up among community studies, highlighting the central importance ( for the survival and functioning of city neighborhoods) of the networks of strong, crosscutting personal relationships developed over time that provide the foundation for trust, cooperation, and collective action in such communities (Jacobs, 1965; Nahapiet and Ghoshal, 1998). Early usage also showed the characteristic of social capital for the individual: the set of resources inherent in family relations and in community social organizations useful for the development of the young child (Loury, 1977; Nahapiet and Ghoshal, 1998). However, there is not a clear, undisputed definition of social capital (Dolfsma and Dannreuther, 2003; Foley and Edwards, 1997). Some of the definitions focus on substance, sources or the effects of social capital. Other definitions focus on the relations between actors, the structure of these relations between actors, or both of them (Adler and Kwon, 2002).

However, there seems to be a consensus among all these different definitions of social capital. That is, social capital constitutes both the network and the assets through different networks (Burt, 1992; Bourdieu, 1986; Mustakallio, 2002).

Thus the definition of social capital can be seen as assets embedded within different types of networks (Bourdieu, 1986; Leana and Van Buren, 1999).

There are two important elements in this definition: assets and networks which indicate the content and source of social capital.

(16)

As discusses earlier, there is not a clear, undisputed definition of social capital. This results from the fact that the content of social capital is multi-dimensional rather than one-dimensional21. And each one dimension alone can not fully capture the picture (Hean, Cowley and Forbes, 2003). Putnam (1995) has argued that a high research priority is to make the dimensions of social capital clearly. Asset from the definition of social capital means valuable resources, belongs, quality, or skills which is a multi-dimensional definition as well22_. Many researches have been done trying to clarify the dimensions and some of them are as follows:

Trust (Putnam, Leonardi and Nanetti, 1993; Leana and Van Buren 1999; Coleman, 1988; Collier, 1998; Snijders, 1999)

Rules and norms governing social action (Coleman, 1988; Collier, 1998) Types of social interaction (Collier, 1998; Snijders, 1999)

Network resources (Snijders, 1999)

Other network characteristics (Putnam, 1995) Network

Nahapiet and Ghoshal (1998) argue that much of the social capital is embedded within networks of mutual acquaintance and recognition. For example, the durable obligations arising from feelings of gratitude, respect, and friendship or from the institutionally guaranteed rights derived from membership in a family, a class, or a school (Bourdieu, 1986). Nahapiet and Ghoshal (1998) further argue that social capital can be also derived from contacts or connections from networks, for example, from “weak ties” (Granovetter, 1973) and “friends of friends” (Boissevain, 1974). In conclusion, no matter what form the network takes, much of the social capital is always derived from it.

The Relationship between Social Capital and Family Firm

Arregle, Hitt, Simon and Very (2007) argue that social capital brings a leveraged, powerful network to a firm unconsciously. And this will benefit the business in the end. For family firms, one source of success is social capital which is comprised of internal and external relationships (Arregle, Hitt, Simon and Very, 2007). Internal social capital can be a relationship based on trust, a culture based on confidence between the executives and workers, or close relationships that facilitates information flow and effective monitoring (Arregle, Hitt, Simon and Very, 2007). What is more, Newton (1997) argues that the trust generated by family networks and close-knit communities comprises of the social capital in family firms. External social capital can be close relationships between suppliers and other participants in the business (Rosenfeld, 1997). Nahapiet and Ghoshal (1998) and Arregle, Hitt, Simon and Very (2007) argue that social capital can not be traded, imitated or transferred easily. They also argue that this social capital

(17)

is from family firms’ social capital and it can create competitive advantages to family firms (Arregle, Hitt, Simon and Very, 2007).

The Relationship between Social Capital and Network

From the definition of social capital we can see that social capital is treated as assets from different networks which can facilitate the efficient actions of individuals (Burt, 1992; Bourdieu, 1986; Mustakallio, 2002; Granovetter, 1973).

Social capital as a resource, also acts as a constraint in enforcing norms of behaviors among partners (Walker, Kogut and Shan, 1997). Fewer constraints will lead to fewer difficulties in finding information about current and potential partners (Walker, Kogut and Shan, 1997). They argue that network is an important factor in the development of constraints directing information flows in the developing of social capital (Walker, Kogut and Shan, 1997). And they identify two types of network structures, named “closed” and “open” network. “Closed” network means all the firms in an industry have relationships with each other and therefore information flows quickly. While “open” network means there is no relationship among all the firms in an industry (Coleman, 1992). However in real life, the network in an industry is neither “open” nor “closed” network. It is between “open” networks and “closed” networks (Knoke and Rogers, 1978; Van de Ven, Walker and Liston, 1979; Nohria and Garcia-Pont, 1991). If there are more relationships within the firms in an industry, there will be more information flows and it will be easier to have successful cooperation. So the degree of social capital depends on the type of network where social capital is embedded within (Walker, Kogut and Shan, 1997).

Social capital can also help building a leveraged, powerful network to a firm unconsciously (Arregle, Hitt, Simon and Very, 2007). The formation of the network is associated with the establishment of new relationships based on existing inter-firm ties (Walker, Kogut and Shan, 1997). In order to achieve successful formations, a lot of resources are needed of which the organizations are lack (Walker, Kogut and Shan, 1997). And social capital is just a valuable resource to help managing the relationships and strengthen the network (Walker, Kogut and Shan, 1997).

In conclusion, social capital comes from the network and the degree of social capital depends on the type of network where social capital embedded within. At the same time social capital reinforces the network’s development.

3.2 Hypothesis Development

(18)

(La Porta et al., 1999; Roe 2000). These researches focus on examining the determinants in different ways. However, they indicate that no matter how different the environments where firms operate are, the firms will always seek an optimal ownership structure to suit the environment in order to maximize the profits. Motivated by this, this paper will investigate family ownership in Italian textile industry where there is a preponderance of family firms (Owen, 2003).

Most of the literatures are talking about family ownership as a determinant to firm performance (Himmelberg et al., 1999; Kole, 1994; Wallevik, 2009d). And most of these literatures agree that the main difference between family ownership and other ownership types is the highest dedication for family members to see their business grow, prosper and pass on to the next generation23_{. In order to achieve this purpose, family firms have to overcome their} shortcomings, keeping competitive and innovative, informative. This paper argues that commercial network can be a determinant of family ownership due to the social capital embedded in, family ties, altruism, trust and partnerships. However, there may be a potential endogeneity problem with simultaneity which supposes that family ownership and commercial network are codetermined with each other24_{. This is because there is a potential question that} while stronger commercial networks enhance family ownership through social capital sources, commercial network may be also determined by family ownership through social capital. Details of this will be discussed in the later part of this paper.

First, the relationship between social capital and network is complicated. Social capital is embedded within network and the degree of social capital depends on the position at which it lies in the network (Walker, Kogut and Shan, 1997). However, the other way around, social capital can also strengthen the network by offering resources during the managing process. Commercial network, as one type of networks, can be expressed in the way of investments in other industrial firms, or investments in suppliers or customers based on partnerships, co-investments, or other business cooperation (Wallevik, 2009d; Curto and Molho, 2002; Sugiyama and Grove, 2001). Commercial network offers ways to overcome shortages of capital sources and opportunities, to access to other markets and social organizations, to reduce the risks, to enhance the competitive capability and to facilitate expansion of business (Rose, 2000; Wallevik, 2009d) due to the social capital and other ties embedded within. Social capital can be defined as an investment in social relations with an expected return (Lin, 1999). Because social capital can enhance commercial network by adding more investments, family firms’ shortcomings, limited access to capital sources, risk aversion can be greatly mitigated and thus maintaining family ownership.

Second, Arregle, Hitt, Simon and Very (2007) and Putnam (1995) state that the assets of social capital can be norms, information, technology, knowledge etc and these sources may have effects on a firm directly or indirectly. Commercial networks with these assets from social capital increase the diffusion of information and norms mitigating family firms’ shortages of

(19)

efficient information and disciplines (Burt, 1992)25. In this way, family ownership is maintained.

Based on Wallevik (2009c) where she defined commercial network as investments in family firms from other firms in the same industry and found that commercial network has a positive and significant effect on family ownership. However, the conclusion from Wallevik (2009c) was limited to the shipping industry in Norway. Will commercial network also be a determinant for family ownership in Italian textile industry? This paper’s focus fills this gap. And I expect a positive result.

What is more, the size of commercial network indicates how strong the network is and consequently indicates the degree of the social capital embedded within. So the hypotheses are as follows:

Hypothesis 1: There is a positive relationship between the size of commercial network and family ownership percentage.

Hypothesis 2: There is a positive relationship between the size of commercial network and the likelihood of being a family firm.

4. Method

4.1 Sample

The data is mainly based on an up-to-date sample of 271 firms listed in Amadeus26 in Italian textile industry. The criterion for textile industry is the industry classification code NACR Rev 2 from Amadeus. Cross-sectional data are used in this paper in stead of panel data. In both of the two regression equations, all the dependent variables and independent variables are not time-variant variables. But the control variables except firm age are time-variant. So the averaged value based on three years, 2006, 2007 and 2008 for the control variable except firm age is used in order to be in line with the dependent and independent variables.

Amadeus is a comprehensive, pan-European databank that contains accurate, comprehensive and timely financial information about more than 11 million public and private companies in 41 European countries27. Amadeus integrates the best source of information through in-depth market research in each country and applies strict inclusion criteria to prevent any bias in coverage28. And the data cover firms with different sizes, including, small, medium, large or international firms; different aspects of textile, such as, manufacture of machinery for apparel

(20)

and leather production, washing an dry cleaning of, and fur products, weaving, wholesale and retail sale etc. It turns out that there are 788 firms reaching the criteria. Then firms with incomplete ownership or financial data, incomplete data from 2006 to 2008 are excluded from the sample to ensure the accuracy of the sample. In the end, this dataset contain 271 firms. What is more, there are also some data from the companies’ website because some companies’ data from Amadeus were missing. In conclusion, I believe this dataset gives a balanced trade-off between sample and population.

Some general characteristics of textile industry:

Textile industry is considered as a classic small-scaled industry where the advantages of flexibility are more important relative to the economies of scale and scope (Thomsen and Pedersen, 1998). In all, the textile industry may be less complicated compared to those industries with more technology or research and developments. So it would be possible to be more efficient to allocate the ownership rights to the individual owners or families.

The textile industry is also characterized as capital intensive, highly automated and less flexible (Nordas, 2004). The textile manufacture consists of spinning, weaving and finishing which are done in integrated factories (Nordas, 2004). Most of the time, the lead time of in the textile industry is very long which leads to a relatively large minimum orders (Nordas, 2004). Thus it is difficult to adjust flexibly to suit the customers’ taste which makes the industry less flexible (Nordas, 2004).

Some characteristics of textile industry in Italy

The textile industry has always been one of the most important industries in Italy, and this is evidenced not only by its national position among all the industries but also the position that Italian textile industry has gained in the world through export.

Firstly, the textile industry is of paramount importance to the national economy of Italy. Until end of 2008, in textile industry, there were 56608 firms in total; 0.5 million employees which accounted for 11.6% of the whole country’s employment; 54.1 billion Euros’ output which accounted for 7.3% of the total manufacturing output of Italy29. The textile industry in Italy is also specialized in exports (Graziani, 1998). In 2008, the total export value of textiles was 27.8 billion Euros which accounted for 51.3% of the whole industry’s output in Italy and accounted for 7.8% of the output of the whole manufacturing industry in Italy30.

Secondly, textile industry is a very traditional industry which has a very long history, operating in highly regulated economies facing competition from other countries with lower labor costs and levels of workplace regulation (Owen, 2003). However, Italy is still one of the top three main countries around the world which has the competitive technological production capability of textiles (Bianchi and Bottacin et al., 2001). A survey from Stengg, (2001) based on the

(21)

combined data of “turn over”, “value added”, and “employment”, indicates that Italy turns out to be the most important country in the textile industry in Europe which holds shares of 31% of the EU’s total. The second important country is United Kingdom which holds 15% share and Germany with 14%, France with 13%, Spain with 9% and Portugal with 6%31.

For firms in Italian textile industry, there are generally four characteristics about them: 1. Some medium sized companies with very famous brand names32.

2. Some medium-to-large companies have no brand names, but they mainly produce for large distribution compartment stores or as providers of the first group33.

3. Plenty of small firms specialized in only one production part, whose customers are one or more of the other companies of the sector34_.

4. A preponderance of family firms35.

4.2 Methodology

One of the most important steps in carrying out a successful research is to choose a right methodology. There are so many factors needed to be considered, the research questions and all the other considerations in the design. But firstly, if the methodology is suited to the research questions, the whole process will be much easier.

Each methodology has a different approach to evaluate. The fact is that no single one would be the best methodology. Which methodology would be the most appropriate for the research firstly depends on the research questions and then the design of the paper.

Overviews of the methodologies in the relevant literatures

In the appendix, Table 3, there are summarized overviews of the methodologies in relevant literatures. And the conclusions from overviews of methodologies are:

1. These studies range from sample, definitions of variables, but they do not range much with methodologies.

2. The main criterion for choosing the methodology is to better answer the research questions. One main purpose of this paper is to compare the result with Wallevik (2009c), also considering of the research questions and sample specification, so the methodology for this paper is close to Wallevik (2009c).

31_{Source: Euratex (“Memorandum on Preferential Rules of Origin”, February 2000), based on Eurostat and Stengg, (2001).} 32_{(Bianchi and Bottacin et al., 2001)}

(22)

3. There were some advantages and disadvantages in the overviews when they dealt with methodology problems which offer clues about potential problems that I should pay attention to.

4. In the overviews, some checks for the assumptions (for example, normality and heteroskedasticity etc) of the methodologies are neglected. But this paper will provide more details of check for assumption.

The methodology of this paper

Through out this paper, a big challenge is to deal with the potential problem of endogeneity with simultaneity. Simultaneity is also called reverse causality. It supposes that two variables are codetermined, with each other affecting the other36. And the variables that encounter endogeneity problem are called endogenous variables which are determined by other variables within the system; contrast with exogenous variables which are variables that are considered as outside of the system. In this paper there is a potential question that while stronger commercial networks enhance family ownership through social capital sources, commercial network may be also determined by family ownership through social capital. Demsetz and Lehn (1985) argue that managerial ownership is an endogenous variable influenced by various firm specific variables representing the business environment, different managerial contracting environment and performance. What is more, Himmelberg et al. (1999) argue that managerial ownership and performance are endogenously determined by the firm’s exogenous contracting environmental changes. Therefore, ownership can be an endogenous variable that will be determined by exogenous changes in the firm’s environment. But the endogeneity problem in this paper is reverse causality because social capital is not exogenous but within the system as an independent variable of the model. So now the question is that will social capital be determined by family ownership. Literatures about determinant of social capital are various and they do not reach one agreement. Some argue that social capital is mainly determined by cultural evolution (Fukuyama, 1995; Putnam, Leonardi and Nanetti, 1993), while some others argue that political and economic development are also important (Brown and Ashman, 1996; Tendler and Freedheim, 1994). Further, Halpern et al (2002) and Pantoja (1999) suggested a list of main determinants of social capital including family and kinship connections, history and culture, economic inequalities, wider social networks and so on. Winter (2000) sees family as a bed-rock of social capital, because it can create good relationships with trust, cooperation, activeness and civic virtues, therefore enhances social capital (Putzel, 1997; Cox, 1995; Stone and Hughes, 2001). In conclusion, commercial network and family ownership can be determinants of each other through social capital.

This reverse causality problem may lead to biased ordinary least square (OLS) estimates. And it can be solved if good instrument variables can be found, but this will beyond the scope of this paper and future researches may focus on this.

So, in the end, according to the research questions and design, also in line with Wallevik (2009c), the linear regression Ordinary Least Square (OLS) and Logistic Regression models

(23)

are used depending on whether the dependent variable is continuous variable or dummy variable and controlling for some firm-specific characteristics.

4.3 Measurements of Variables (Table 4)

Before going deep into the detailed explanations about Yes-or-No (the dependent variable in model 1), family ownership (the dependent variable in model 2) and commercial network (the independent variable in both of the two models), one possible question may come up. That is, who else hold shares in a firm except families and industrial companies (shares held by industrial companies is used as the measurement for commercial network and details you will find in the following part of this section)? Because when there are only these two types of shareholders in a firm, the relationship between commercial network and family ownership would be obviously negative. Then regressions are not really necessary. In order to ensure this point, the following details are needed.

There are actually 14 types of shareholders in the sample which is from the database Amadeus37_{. This means it can happen to a firm that there may be only these two types of} shareholders. It could also be possible that there are 14 types of shareholders. Of course there are also many other possibilities. So it is not always that there are only these two types of shareholders. Therefore, in this respect, it would be necessary to use regressions.

Dependent variable 1: YN

YN (Yes or No) represents whether the firm is a family firm or not, which is a dummy variable with 1 =yes, 0 otherwise.

The definition of family firm in this paper focuses on the family’s involvement of ownership because one purpose of this paper is to compare the result with that of Wallevik, (2009c). More specifically, when the percentage of family ownership owned by one or more known individuals or families is at least 20% (20% is included), then this company is considered as a family firm, otherwise, not. I use dummy variable for this, 1=yes, 0 otherwise. And the definition of “one or more individuals or families” is from Amadeus as follows:

Besides single private individuals or families38, shareholders are designated by more than one

37_{The 14 types of shareholders contains bank; financial company; insurance company; industrial company; mutual and pension}

fund/Nominee/trust/trustee; foundation/research institute; public authorities, states, governments; one or more known individuals or families; employees/managers/directors; self ownership; private equity firms; publicly quoted company; unnamed private

shareholders, aggregated; other unnamed shareholders, aggregated.

38_{Family is a group of people affiliated by consanguinity, affinity or co-residence. And the definition of consanguinity is not only}

(24)

named individual or families in this category. The idea behind this is that they would probably exert their voting power together. This latter case corresponds to entries like:

“Mr. Gregory Edward Bailey & Mrs. Margaret Ethel Bailey”; “Mme Bringaud et son fils”;

“Mme Sotto et M Cohen”

“Families Courault and Andrivon”. Dependent variable 2: FO

FO refers to family ownership. It measures the percentage of family ownership in a firm. And the definition of family ownership is the same as what was stated above.

Independent variable 1: CN

The definition of commercial network is based on the theory of social capital theory. Commercial network can be expressed in the way of investments in other industrial firms, co-investments, partnerships or investments in suppliers or customers based on trust or particular business cooperation (Wallevik, 2009d).

Industrial ownership may happen through direct investments, co-investments or partnerships within industrial companies locally, nationally or internationally (Harlaftis and Theotokas, 2004). Therefore this paper follows after Wallevik (2009c) by using industrial ownership as a proxy for commercial network. Industrial ownership is measured by the invested shares from other industrial firms. Different levels of invested shares from other industrial firms indicate how strong the commercial network of this firm is.

The definition of industrial companies (Amadeus) is39: 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.). They include also companies active in B2B40 or B2C non-financial services.

This variable is divided into four levels in order to get a clearer conclusion about the effect of commercial network: 1= the percentage of invested shares from other industrial firms is above 50% (the existence of commercial network in general); 2= the percentage of invested shares from other industrial firms is between 50% and 20% (both 20% and 50% are included in this category); 3= the percentage of invested shares from other industrial firms is below 20%; 4= the percentage of invested shares from other industrial firms is zero.

39 Pedersen and Thomsen (2003) argued that owners from this definitions are typical large and they may have better access to capital sources.

40 Business-to-business (B2B) involves the transactions between businesses, for instance, between a manufacturer and a wholesaler or between a wholesaler and a retailer. Its opposite is business-to-customer or business-to-consumer (B2C).

(25)

Control variables

There are also some variables not included in the hypotheses. However they may also have influences on family ownership. So this paper controls for these variables, including debt ratio, debt to equity ratio, equity ratio, firm age, firm size, profit before tax, return on assets, return on equity, sales to assets. The averaged value based on three years, 2006, 2007 and 2008 for the control variable except firm age is used.

Financial Leverage Measurements: Debt Ratio %, Debt to Equity Ratio % and Equity Ratio % Debt ratio is a ratio that indicates what proportion of debt a company has relative to its assets41. Debt to Equity Ratio is a measure of a company’s financial leverage calculated by dividing its total liabilities by stockholders’ equity42. The equity ratio indicates the relative proportion of equity used to finance a company’s assets43_.

These three variables are financial leverage measurements which are commonly used as control variables in past research papers about financial leverage. This paper argues that all of them are needed to be added as important control variables into the model, because they indicate different aspects of the financial leverage of a firm. Debt ratio focuses on the potential risks that the firm faces in terms of its debt-load. Equity ratio is used to measure solidness and survival of a firm and the higher this ratio is the more benefits that the equity holders can get from the firm (Wallevik, 2009c). Debt to Equity Ratios focuses on the proportion of equity and debt that the firm is using to finance its assets. However, there may be a potential problem of collinearity between DR and ER. This is because they are related to each other by definition (i.e. DR + ER =1, because there are only two types of capital sources, debt and equity.). So even though they indicate different aspects about financial leverage, one of them will be dropped from the model.

For any firm, in order to maintain competitive, it is necessary to access to external financial resources. And there are many types of financial resources, mainly categorized into two main types, debt and equity. Debt is about borrowing money to be paid back with interests, while equity is about getting money by selling interests in the company. The main differences between debt and equity are: 1) Debt will not dilute the firm’s ownership because the lender does not have any claims to equity in the firm. The lender can only get the loan back together with the interests, no matter the firm is successful with management or not. 2) But for equity, the firm’s ownership will be diluted; meanwhile, if the firm is successful, then stockholder can get a bigger portion of benefits. When a company decides on how to raise additional capital, the advantages and disadvantages of each type of funding will need to be considered.

Family firms maybe can not get enough access to outside capital sources because of the poor investor protection within its ownership (La Porta et al., 1997). Therefore they may prefer

41 http://en.wikipedia.org/wiki/Debt_ratio

(26)

raising debt to issuing equity. Raising debt can help the family firms to overcome the shortcoming of limited capital accesses, keep competitive, by getting external funds without diluting control power which is the most concerned thing for a family firm (Franks, Mayer, Volpin and Wagner, 2008; Wallevik, 2009d). Additionally, Wallevik (2009c) found that debt ratio, debt-to-equity ratio have significant and positive effects on family ownership in the shipping industry in Norway.

Firm Age

It is measured by natural log form of the number of years since the founding of the firm. When the size of the firm grows with age, the needs for capital and further development of the firm increase, (Berle and Means, 1932), therefore the company founders have to sell their stakes gradually over time for diversification reason or sell the equity for acquisitions and therefore diluting the ownership of the firm (Holderness and Clifford et al., 2008). This is also the case for family firms. So age of a firm is an important determinant of family ownership.

Additionally, Franks, Mayer, Volpin and Wagner (2009) found the development of family firms follows a life circle by evolving into widely held firms as they age and the result also showed a strong negative correlation between family ownership and firm age in U.K. Wallevik (2009c) also found that firm age has a significant and negative effect on family ownership in the shipping industry in Norway.

Firm Size

This paper measures firm size as the natural log form of total assets of a firm. When the size of a firm is bigger, firms can possess economies of scale (Carney and Gedajovic, 2002; Wallevik, 2009b). The capital resources would be more (Hall and Weiss, 1967) and the market value of a given fraction of ownership would be greater. When the market value of a given ownership is greater, the ownership concentration would be reduced. This is because the risk-neutral effects of size on ownership (Demsetz and Lehn, 1985). Risk neutral is in between risk aversion and risk seeking. For example, if offered either $50 or a 50% chance of $100, a risk adverse person would choose to take the $50, the risk seeking person would take the 50% chance of $100, and the risk neutral person would have no preference between these two choices44. Risk aversion suggests that the owners of the firms would purchase additional shares only at lower, risk-compensating prices. And they have to dilute the ownership because of the increased costs of capital. As the size of a firm grows, both the risk-neutral and risk-aversion weigh more than shirking cost, therefore the concentrated ownership would be expected to be reduced (Demsetz and Lehn, 1985; Pedersen and Thomsen, 2003). Some previous literatures found out that this applies to family ownership as well. That is the size of a firm has a negative effect on family ownership (Carney and Gedaljovic, 2002; Wallevik, 2009c; Schulze and Lubatkin et al., 1999; Demsetz and Lehn, 1985).

(27)

Performance Measurements: Return on Equity %, Profit before Tax %, Return on Assets % and Sales to Assets %

Return on equity measures the rate of return on the ownership interest of the common stock owners. Profit before tax is a profitability measure that looks at a company’s profits before the company has to pay corporate income tax. Return on assets shows how profitable a company’s assets are in generating revenue. Sales- to- assets is a measurement about the revenues generating from total assets.

These four variables are variables about the financial performance situation of a firm which are commonly used as control variables in previous research papers. This paper argues that all of them are needed to be added as important control variables into the model, because they indicate different aspects of the financial performance. Return on assets focuses on indicating how efficiently resources are allocated and return on equity focuses on indicating how effectively management is in generating profits through shareholders’ money. Profit before tax is included in order to control for different taxation rules among firms and therefore it is easy to compare profits. Sales-to-asset indicates a low capital intensity level because sales are higher compared to assets (capital intensity is computed as a ratio of total assets to sales). Several studies have analyzed the relationship between family ownership and financial performance (Demsetz , 1983; Demsetz and Lehn, 1985; Demsetz and Villalonga, 2001; Himmelberg et al., 1999; Kole, 1994; Wallevik, 2009c; Wallevik, 2009d) with measurements of financial performance by return on assets, return on equity, profits, sales to asset, Tobin’q etc, but most of these literatures talk about the effect of family ownership on financial performance (Himmelberg et al., 1999; Kole, 1994; Wallevik, 2009d). Few literatures talk about the effect of financial performance on family ownership (Demsetz, 1983; Demsetz and Lehn, 1985; Demsetz and Villalonga, 2001; Wallevik, 2009c). For the former literatures, almost every author mentioned the endogeneity problem from reverse causality because financial performance and family ownership are determined simultaneously.

The first arguments about the effects of financial performance on family ownership were from Demsetz (1983) and Demsetz and Lehn (1985). They argued that ownership should be more concentrated in firms for which amenity potential is greater. Amenity potential is a financial performance indicator refers to a characteristic of the good produced by the firm that allows for the creation of non-profit related utility for owners of the firm (Demsetz and Villalonga, 2001). Further, Demsetz and Lehn (1985) showed that the ownership structure is more concentrated when amenity potential is greater which measured by return on equity, stock price volatility etc in the companies in America.

(28)

4.4 Model

To examine the influence from commercial network on family ownership in Italian textile market, this paper applies different regression models depending on whether the dependent variable is continuous or binary variable. So this paper uses logistic Regression for the first model and Ordinary Least Squares (OLS) regression for the second model.

Regression models:

Model one : YNi =β0 +β1 (CN)i + β2 (DR) i + β3 (DTER) i + β4 (ER) i + β5 (FA) i + β6 (FS) i + β7 (PBT) i + β8 (ROA) i + β9 (ROE) i +β10 (STA) i +ei

Model two: FOi =β0 +β1 (CN)i + β2 (DR) i + β3 (DTER) i + β4 (ER) i + β5 (FA) i + β6 (FS) i + β7 (PBT) i + β8 (ROA) i + β9 (ROE) i + β10 (STA) i +ei

The βs are the regression coefficients, denoting the amount the dependent variable Y changes when the corresponding independent changes 1 unit. And “i” means firm from Italian textile industry. e is the error coefficient.

Representations of variables: Dependent variables:

YN (Yes or No) represents whether the firm is a family firm or not, which is a dummy variable with 1 =yes, 0 otherwise.

FO (Family Ownership) is measured by the percentage of family ownership in a firm. Independent variable:

CN (Commercial Network) is measured by the invested shares from other industrial firms. Control variables (averaged value based on three years, 2006, 2007 and 2008 is used except firm age):

DR (Debt Ratio %) is measured by Interest bearing debt as a ratio of total assets in percent. DTER (Debt to Equity Ratio %) is measured by dividing its total liabilities by stockholders’ equity in percent.

ER (Equity Ratio %) is measured by equity as a ratio of total assets in percent.

FA (Firm Age) is measured by natural log form of the number of years since the founding of the firm.

FS (Firm Size) is measured by the book value of total assets in balance sheet.

(29)

ROE (Return on Equity %) is measured by profit before tax as a ratio of the total equity in percent.

STA (Sales to Assets %) is measured by total sales as a ratio of total assets in percent. Model estimation

First Model

This paper estimates the first model with logistic regression. Logistic regression, sometimes called logistic model or logit model, is often used for prediction of the probability of occurrence of an event (The “event” is a particular value of the dependent variable) by fitting data to logistic curve when the dependent variable is a dichotomy and the independent variables are of any type. Note that in logistic regression the dependent variable can not be continuous variables. Logistic regression uses maximum likelihood estimation and transforms the dependent variable into a logistic variable that is the natural log of the odds of the dependent variable. The natural log of the odds of an event equals to the natural log of the probability of the event occurring (often means the dependent variable takes the value of “1”) divided by the probability of the event not occurring (often means the dependent variable takes the value of “0”): Ln (odds (event)) = Ln (prob (event)) / prob (nonevent)).

The reasons why this paper chose to use logistic regression for model 1:

1) The dependent variable from the model 1, YN, is a dummy variable. And it takes the value of 1 when the firm is a family firm, otherwise 0. By using Logistic regression, I can predict the probability of being a family firm what is exactly the purpose and specialty of logistic regression.

2) In statistics and econometrics, there is also another similar estimation to logistic regression which is called probit regression. Probit regression is a popular estimation for a dichotomous response model which often employs maximum likelihood estimation45.

The similarities and differences of logistic regression and probit regression:

Both of these two regressions are used for dichotomous dependent variables, use maximum likelihood, and are not linear. In order to change into linear relationship, transformations are needed to be used on dependent variables in both of the two regressions. But they use different transformations. Logistic regression uses the logistic function which is the natural log of odds, while probit regression uses the inverse of the standard normal cumulative distribution.

Making a choice between logistic regression and probit regression often depends on personal preference. But there is an advantage that logistic regression has over probit regression. That is, the transformation of logistic regression can be easily interpreted as log-odds; however the probit regression does not have a direct interpretation. What is more, because this paper is