Corporate ownership and firm performance: family vs. non-family involvement in Western Europe.

(1)

Corporate

ownership

and

firm

performance: family vs. non-family

involvement in Western Europe.

J.C.G. van den Enk S2211424

Msc International Financial Management

Faculty of Economics and Business, Rijksuniversiteit Groningen Supervisor: Dr. M. (Mario) Hernandez Tinoco

Co-assessor: Prof. Dr. C.L.M (Niels) Hermes August 16th 2016

Abstract

This study examines the corporate ownership structure with firm value and performance for 395 listed firms in 5 different West European countries during the years 2011 until 2014. The selected firms are divided into 3 different categories: family firms, non-family firms and widely held firms. This research is mainly focused on the performance of family firms. There are many different sorts of family firms with each different firm performance and value. The results showed that family firms that are active managed by the family have higher results than non-active managed family firms. There are statistically significant results regarding the relation between the legal environment and firm valuation. Active family firms have statistically significant higher firm values and profitability compared to non-family firms in countries with high level of legal shareholder protection.

Keywords: Ownership structures, firm performance, family firms, agency theories, Corporate Governance

(2)

1. Introduction

The aim of this paper is to examine the relationship between corporate ownership and firm performance. Especially, the performance of family controlled firms will be evaluated. In Western European countries, family control is common for public listed firms (Faccio and Lang, 2002). Therefore I want to examine if corporate ownership affects firm performance. Especially, if family ownership affects firm performance and whether it makes a difference in which way a family is involved in a business. Researchers stated, that family control and management could minimize the agency costs which resulted in higher firm valuation (Maury, 2006). Although family control is widely analyzed, empirical evidence on the performance still remains inconsistent nowadays. To analyze this relationship, I will make use of panel-data from public listed companies in Germany, the Netherlands, Belgium, France and Denmark.

The purpose of this research is to provide new knowledge on the literature on family controlled listed firms. The goal of this research is to describe, explain and predict different variables which differentiate family listed firms from non-family listed firms in Western European countries. These variables will be compared and reflected on the firm valuation and profitability of the specific firms. The main research question for this study is: Will family controlled listed firms in Western Europe outperform non family listed firms? This research question is divided into multiple research questions that would help us to determine the main research question. First of all, the study focused on the effect of different corporate ownership type on corporate performance. This will be conducted by implementing the principal-agent and principal-principal theories, the resource-based view and stewardship theory. Secondly, does active family management affect the performance of family firms compared to non-family firms. For the second research question, some sub-hypotheses are developed and examined to provide empirical evidence regarding this question. Due the widely categorization of family firms in this study, we can investigate which type of family firms perform the best compared to non-family firms.

(4)

profitability results compared to non-family firms. In line with the results of Maury (2006), the results based on this sample concluded that there is not a linear relationship between increasing family control and firm valuation and profitability. Families with less than the half of the control rights perform statistically significant higher than non-family firms in terms of firm value and profitability. When the control rights exceed the 45% it will perform lower but insignificant compared to non-family firms. For the relationship between family participation and firm performance follows the same pattern. Active family controlled firms with a maximum of 25% representation in the management board perform statistically significant higher in comparison with non-family firms and other family firms. When family presence exceeds 25% the firm, valuation becomes lower than non-family firms due to the chance of expropriation of minority shareholders. A rise in family participation does not lead to higher firm performance compared to non-family listed firms in this sample. In line with Villalonga and Amit (2004), this study finds statistically significant higher returns on firm valuation and profitability when the founder of the family firm is the CEO. When a family firm becomes older and a descendant takes the position of the CEO, it has negatively but nog significantly lower firm valuation in comparison with non-family firms. When a family firm has an outsider as CEO, they increases statistically significant the performance of the firm compared to non-family firms. An independent CEO added value because they could act as a guardian over the resources of the firm and minimize the family expropriation (Claessens et al, 2002). In line with previous research of Maury (2006), this research showed that the legal environment affect the performance of family listed firms compared to non-family firms. The profitability and firm valuation are higher for active family firms compared to non-family firms in countries where the investor protection is high. In countries where the investor protection is low, the family firms do not statistically significant outperform non-family firms. This could be explained because, the agency problems between the controlling shareholder and minority shareholders. In the countries where the investor protection is high, the minority shareholders have more influence in the decisions that minimize the risk of expropriation of assets (Maury, 2006). These results complement previous work that presented that country factors and institutions influence corporate governance and corporate ownership (Pederson & Thomsen, 1997).

(5)

(6)

2. Theoretical background

The theory section will briefly review existing theories around corporate ownership and performance, and discusses expectations on the relation between family control and the performance of a company.

The aim of this section is to explain in general the corporate governance and ownership structures of a typical listed company to provide some theoretical background information for this research. Then, multiple theories will be used to explain the link between (concentrated) ownership and the performance of a company. The widely known principal-agent theory will be explained as well as the principal-principal problems in publicly traded firms. The principal-agent problem is the result of contradicting interests between shareholders and managers of a firm. The second agency problem, the ‘principal-principal’ problem will also play an important role in this research. This problem is also known as a form of expropriation of non-controlling shareholders by the controlling (family) shareholders. The resource based view and ‘stewardship’ theories will be discussed and those theories suggest that family involvement within companies could have advantages due to their unique corporate governance characteristics compared to non-family controlled firms. Finally, the legal environment will be discussed. The current literature shows us that the legal environment and shareholder protection influences the corporate governance and ultimately firm performance.

2.1 Corporate Governance

(7)

Figure 1: Corporate governance structure in listed European companies.

This figure shows the general corporate governance structure in a typical European company. It presents the influences of multiple stakeholders on a company whereby the management board and shareholders play an crucial role within the firm.

Figure 1 shows that the management is the spin in the web of all stakeholders. The management act on behalf of the shareholders, but this relation is affected by many more stakeholders than solely the shareholders as presented in figure 1. All these stakeholders have their own interests and incentives to participate in a certain company. The shareholders have the incentive to monitor the management to see if they pursue the same goal: maximizing firm value. By monitoring, the shareholders want not only to observe the managers but they also try to influence the decisions of the managers through restrictions, policies and rules (Jensen & Meckling, 1976). It is hard for shareholders to monitor the management because it is time-consuming and expensive, and there is information asymmetry.

(8)

Enron and Lehman brothers related to misleading financial reporting by the management. This led to the introduction of regulations in financial reporting of listed firms and management structures (Sarbanes-Oxley act in the US and Code Tabaksblat in the Netherlands).

2.2 Scope and definition of family firms

Family firms are a specific form of ownership structure and are examined for decades. Family firms play a prominent role in the economy, so it is interesting to examine the nature of such family companies in an attempt to explain why family-controlled businesses actually outperform other types of businesses. In the current literature, there is still no consensus about the exact definition of a family firm and the relation with performance. Different outcomes are more or less depending on the used definitions of ‘family firms’ and family involvement.

Maury (2006) for example, use the threshold of 10% ultimate voting rights that would be considered as enough to regard a company as a family firm. La Porta, Lopez-de-Silanes & Shleifer (1997) and Miller and Miller (2006) use the threshold of 20% of the total voting rights of the firm to regard a company as a family firm and Ang, Cole and Lin (2000) use a threshold of 50%. There are also researchers that not only define a family firm based on the number of votes they control but also on the level of active family involvement. For example, Anderson and Reeb (2003) define a family as a family firm when a ‘founding’ family has control rights and has at least one family member serves on the executive board, orwhen a descendant of the founding family serves as CEO. Claessens and Fan (2002) regard a firm as a family firm when the founder or a family member with the same surname as the founder, serves as CEO of the company.

For this research a threshold of 20% from the voting rights is used to determine a firm as a family controlled firm. The reason for this 20% threshold will be explained in the method section. To determine the different sorts of active family management, this study examine a wide variety of definitions compared to the research of Villalonga and Amit (2004).

2.3 Agency problem; Principal-agent

(9)

management of the firms. This is in line with the analyses of Jensen and Meckling (1976) who concluded that the largest shareholders are highly interested in keeping the management and company under control because of their large stake of financial input. This will lead to maximized firm performance and thus firm value. This will always be an ongoing conflict because managers will always want to maximize their own goals and wealth under the control of the shareholders. Our goal for this research is to investigate in which type of ownership structure could minimize these principal-agent problems.

(10)

So if the company is controlled by a non-family blockholder, principal-agent problems occur and that could lead to value decreasing performances. Nevertheless, if a company is managed by an outsider as CEO with active participation of the family in the board, it could lead to enhanced value. However, it could also lead to decreased value, when family participation becomes too high in the management board composition.

In line with the model of Burkart et al. (1997), Anderson and Reeb (2003) has found that US publicly traded family-controlled firms have higher market values (Tobin’s Q) and higher return on assets than comparable non-family firms. Maury (2006) found out in his research that European family firms outperform non-family firms. This is in line with the study of Barontini and Caprio (2006) that reported the same results for West-European listed companies. These studies were conducted in the year 1999. These studies support the view that family ownership reduces the classical agency problem between managers and shareholders and therefore outperform non-family firms. Based on these results, the following hypothesis has been drawn for the Western European listed firms in our sample:

H1: A family controlled firm has a higher firm value than a non-family controlled firm due less agency problems between shareholders and the management.

(11)

a company could make a difference between voting rights and cash flow rights. This structure provide the controlling shareholder/family the power to control the voting rights with a small amount of equity in the firm (Bebchuk, Kraakman & Triantis, 2000). In this situation, the controlling family could use their voting power to use other family companies for cooperation in projects against unfavorable prices. In this situation the controlling family could extract profits from a certain firm to another. Burkart et al. (1997) even observe that controlling families who act on their own behalf can negatively affect employee productivity and effort. This loss in productivity could also lead to lower profitability and firm value. The legal environment could mitigate these problems by giving minority shareholders more rights by law to give them more power when they could vote for different projects or investments.

There are also studies that investigates the relation between corporate performance and family control. For example, Anderson and Reeb (2003) found out that there is a U-shape in the performance of US listed family firms based on the amount of shares a family owns. Anderson and Reeb (2003) have shown that the benefits from family control starts to decline when the amount of shares exceed 30%. This is in line with the results of Maury (2006) that showed that there is a nonlinear relationship between the rise of family control and the increase of firm value compared to non-family firms. Their results suggest that family opportunism may increase at high levels of control that influence the firm performance. This study investigated 1672 listed-firms in 13 Western European countries based on the dataset of Faccio and Lang’s (2002) research. This nonlinear relationship between ownership concentration and performance is also recognized by the study of Thomsen and Pedersen (200).Thomsen and Pedersen (2000) stated that beyond a certain point of ownership leads to entrenchment and has adverse effects on performance of a firm. Therefore the following is hypothesized:

H2: There will be a family ownership range in which Western European family controlled firms will perform better compared to non-family controlled firms.

2.4 Family Involvement & Stewardship

(12)

involvement is considered as a unique asset because companies can align the interests of shareholders and management. Dyer (2006) mentioned in his research that human capital is a unique asset, certainly when the interest of the management is fully aligned with the interest of the controlling shareholders (family). Based on the research-based view of Schumpeter (1934), an entrepreneur has an unique set of competencies (like entrepreneurial motivation and unique knowledge) that help a company to perform better. This research-based view is strongly connected to the stewardship theory. Stewardship is another perspective from which to view the advantages (and disadvantages) of a family firm. Miller and Miller (2006) explained the stewardship theory in their research as the fact that family members are not only self-serving economic individuals but often work altruistic for the benefits of the family firm. They will be motivated by higher-level intrinsically motivation like family pride in favor of the collective good of the firm. Davis, Schoorman, Mayer and Tan (2000) stated that the active family members within a family firm identify themselves with the firm and their objectives, even at personal sacrifice. Miller et al (2006) and Villalonga and Amit (2004) concluded in their papers that only active family management has a positive effect on family performances and should outperform non-family controlled firms. Villalonga and Amit (2004) conclude that active family management could reduce the principal-agent problems and therefore perform better than non-active family firms. Non-active family firms still have the principal-agent problems that reduces the firm value. These findings are based on US listed firms and therefore it would be interesting to see if this is also the case in our sample: Western European listed firms.

Therefore the following hypothesis is formulated:

H3: Western European listed firms with management presence will be higher valuated than non-family and non-active family controlled firms.

Family management can reduce agency costs and increase the power of stewardship that will result in better firm performance but family management can also be dangerous. When the family has a majority in the management board, they could overrule the proposals of non-family board managers that could be beneficial for the company (Buckart et al., 2002).

(13)

H4: There will be a family board member presence range in which Western European family controlled firms will perform the best compared to non-family firms.

Morck et al. (1988) reported that firm value increase if founding family members are present are one of the two top officers (CEO and CFO functions). In this case, the CEO is often the founder of the company. This is based on the resource based view that a founder of a firm has unique knowledge and an entrepreneurial drive that results in better firm performance.

Therefore the following hypothesis is formulated:

H5: A firm with a founder as a CEO will be more positively valued compared to another family and non-family controlled firms.

When a company is managed by a founder’s descendant this can have both positive and negative implications for the value of the company. At first, Rosenblatt et al (1985) argue that family members will be more motivated and committed to the business, because they are more naturally part of the company. However, the restricted nature of the human resource pool supplied by the family means that the family may not have enough qualified personnel to operate a business successfully. Some researchers argue that the presence of a family descendant can work destructive. Therefore the following hypothesis is formulated:

H6: A family firm with a descendant in a key position may be negatively valued due the restricted nature of the human resource pool supplied by the family.

2.5 Legal environment and shareholder protection

(14)

expropriate assets. This could be for example overcompensation of (family)-managers, excessive consumption of perquisites and tunneling by ‘non-arms lengths’ transactions with affiliated (family)-companies (Claessens et al., 2002). When these problems arise and are recognized by the market, the attractiveness of the shares of the companies will become lower. The shares need to be sold at a lower price and that leads to lower firm value. Firm performance could be affected as well due the costs associated with for example over paying family management, tunneling and installing family board management that is not the most qualified applicant for the job (Claessens et al., 2002). In the end, the controlling family bear some of the costs of expropriation of the assets because this will be recognized by the market and the attractiveness of the shares will become lower.

Therefore it will be interesting to investigate if there are any differences between performance in our sample of Western-European firms. Although the countries are known as well-regulated and transparent markets, the shareholder protection regulations are different. To study this relation, the sample is divided in two panels. Panel A consisted of firms from Western European countries with relatively high shareholder protection (e.g. The Netherlands, France and Denmark), compared to panel B, countries with relatively low shareholder protection (e.g. Germany and Belgium) based on the minority shareholder index of La Porta et al. (1997). Therefore the following hypothesis is formulated:

(15)

2.6 Review of empirical literature

In the past decades, the effect of corporate ownership on firm performance is widely discussed in numerous research with mixed results. Corporate ownership is important because (major) shareholders could have other interests than minority shareholders. Jensen and

Meckling (1976) stated that the largest shareholders are highly interested in keeping the management and company under control to maximize firm performance and so on value. Previous research generally mentioned the cause of principal-principal agency problems between controlling and non-controlling shareholders in firms and especially in family enterprises (Memili & Misra, 2015). Ali, Chen and Radhakrishnan (2007) stated that families are often involved as stock owners and are involved in controlling functions as a manager and/or board members which might increases altruistic behavior towards family-centered goals that may not be beneficial for non-controlling owners. Anderson and Reeb (2003) investigated the performance between family and non-family owned firms and find out that family firms perform better than non-family firms. Miller et al. (2007) on the other hand, find no evidence that family firms outperform non-family owned companies. There is also

evidence that there is a U-shaped relationship between family ownership and firm

performance (Memili & Misra, 2015). The discrepancy is mostly derived from different forms of family ownership, involvement and performance definitions and different geographical areas. Therefore I want to make an overview of articles with their definitions of a family firm, performance indicators, sample period and size and their main findings.

This study contributes to the current literature on different aspects. First of all, based on the information given in the literature review there is little consistent evidence about the performance of European family firms. This study is compared to Maury (2006) but has a different threshold for family control (25% instead of 10%). Thereby, this study will have no threshold regarding to the size of the companies to prevent a selection bias. Furthermore this study will focused on a recent time span (2011-2014) whereby Maury (2006) and Barontini and Caprio (2006) focusses on the end of the 20th century. This study is also unique because it will comprehends not only family control but also active family management regarding to performance in Western-European countries. Table 1 presents the main highlights and used variables of the current literature about the relation between family firms and firm

(16)

Table 1: Review of empirical literature in the field of family firms.

Table 1 reviews the empirical literature in the field of family firms. The main trends that we observe is that family firms perform better in western developed markets and less in Asian countries. Thereby influences the threshold and the definition of the family firms the analyses that gives us mixed findings. Active firm participation is positively influences the firm performance.

Authors Definition of family firm Performance

measures

Sample size & geographical location

Main findings and results

Anderson & Reeb (2003)

- The founder and/or his family have an equity stake in the firm.

OR:

- Family possesses board seats. OR: - Founder/descendant is CEO. - Tobin’s Q - Return on Assets - Return on Equity

The sample consisted about 403 firms from the S&P 500 in the period 1992-1999. The sample has 141 family firms. Banks were excluded from the sample.

The main finding was that family firms have a higher Tobin’s Q and a higher return on assets compared to non-family firms.

Barontini & Caprio (2006)

- More than 10% of the capital rights. AND: - More than 50% of the ultimate voting rights possessed by the family. - Tobin’s Q - Return on Assets

The sample consisted of 675 listed firms form 11 West-European countries. In the time frame of 1999 till 2001. There were 375 family firms in the sample. Banks were excluded from the sample.

Tobin’s Q and Return on Assets is significantly higher when a founder has an active role as CEO or non-executive director. Family effect continues to be positive at the

descendants stage, when descendants limit

themselves to non-executives roles. Family effect is neutral when descendants assume the role of CEO.

Faccio and Lang (2002)

- Family or individual or unlisted firm as the ultimate owner by more than 20% of

- Not applicable

Coverage of 13 western Europe countries from 1996 till the end of 1999. All listed firms

(17)

Maury (2006)

- More than 20% of the voting rights are owned by one single person or family. - Return on assets - Return on Equity - Tobin’s Q - Sales Growth

Coverage of performance and control variables is found for 1672 non-financial firms from 13 countries. Financial data come from the fiscal year-end closet to end of 1998.

Active family ownership, in which the family holds at least one of the top two officer positions, improves profitability, whereas active ownership does not change the

value premium of family firms. Passive family ownership does not affect the profitability

of family firms compared with non-family firms. Memili &

Misra (2015)

- The founder and/or his family have an equity stake in the firm. - Tobin’s Q - Return on Assets - Return on Equity - Return on Investment

386 S&P firms from a period of 2001 till 2007.

Findings support the hypotheses suggesting the moderation effects of the use of provisions

(a) protecting controlling owners in terms of their sustainability of controlling status, and

(b) protecting management legally on the inverted U-shaped relationship between family

ownership and firm

performance. We also found support for the moderation effects of

(18)

protecting

non-controlling owners, and (e) protecting management monetarily on the inverted U-shaped relationship between family management and firm performance.

Miller et al. (2007)

- More than 5% of the equity is held by a family or founder, the founder or a descendants that holds a management position

- Tobin’s Q More than 850 firms in the years

1996 until 2000 from the USA. 404 firms where family/founder firms.

U.S. family firms do not outperform in their market valuation compared to non-family firms. Founder controlled companies do significantly better than other firms in Tobin’s Q. Morck et al.

(1988)

- It is a family firm when a member of the founding family is the CEO or the CFO.

- Tobin’s Q Based on the Fortune 500 in

1980 in the US.

There is a U-shape relation founded between Tobin’s Q and ownership. The

performance rises when the ownership of the board also rises. The performance is going down when there is a descendant of the family founder as CEO than by an CEO unrelated to the family.

Pedersen & Thomsen (1997)

- Controlling shareholder is one owner (person, family,

company) that owns a sizeable voting share (20% - 50%) of the company.

- Market value Based on the hundred largest independent ownership units among non-financial companies in twelve European nations.

(19)

person or a family owns a (voting) majority of the company. We include in this category foundation (trust) ownership because it reflects the will of a personal founder and often gives the family some degree of control.

a large stake in a company are likely to oppose investment projects that stretch beyond their ownership period.

Villalonga and Amit (2004)

- Family owns at least 5% of firms equity.

- CEO is a family member of the founder’s family

- 1 or more has a key position.

- Tobin’s Q Fortune 500 companies in the

US from 1994-2000.

(20)

2.7 Hypotheses

Based on these theories stated in the theory section the following overview of the hypotheses is given.

H1: A family controlled firm has a higher firm value than a non-family controlled firm due less agency problems between shareholders and the management.(+)

H2: There will be a family ownership range in which Western European family controlled firms will perform better compared to non-family controlled firms. (+)

H3: Western European listed firms with management presence will be higher valuated than non-family and non-active family controlled firms. (+)

H4: There will be a family board member presence range in which Western European family controlled firms will perform the best compared to non-family firms. (+)

H5: A firm with a founder as a CEO will be more positively valued compared to another family and non-family controlled firms. (+)

H6: A family firm with a descendant in a key position may be negatively valued due the restricted nature of the human resource pool supplied by the family. (-)

H7: The firm value of a family firm will be higher in countries with a legal environment that protects the minority shareholders against controlling (family)-shareholders opportunism. (+)

(21)

3. Research methods & Data

In this section the data and method are presented. In the first section the used databases will be explained. In the second section the procedure of the data collection and construction will be provided. The third section will explain the definitions of the used variables. The fourth section will explain the methodology and robustness checks.

3.1. Databases

The data is from firms listed in west-European countries with ultimate ownership data that is provided in Orbis. Orbis is a database containing information about more than 170 million companies around the world. The database consists of legal, financial, sector and M&A activities information. Orbis gives us accurate information about the ultimate shareholders of the companies and to which group they belong (Family, non-family or widely held shareholders). Orbis also provides the board members of the companies. Therefore it is a useful database for our research. The financial information is subtracted by CapitalIQ. Capital IQ is a database that is comparable to Orbis and is a division of McGraw Hill Financial. By using two databases I also could compare and check the board members. If there was discrepancy between the two databases I verify the results by the annual reports of the companies. When it remains unclear after the results of the annual reports the firms are removed from the sample.

3.2 Data collection & construction

The data will consist only publicly traded companies due the accessibility of the information. The countries that will be used are the Netherlands, Germany, Belgium, France and Denmark. These countries are the best representative for West-European countries. I excluded Ireland and the UK due their more Anglo-Saxon governing style (Maury, 2006). The timeframe will be from 2011 up to and including 2014. The sample starts with 716 firms. Financials companies (SIC 6000-6999) are excluded from the sample due to their characteristic nature (Maury, 2006). Financial firms are excluded because their high leverage ratio is normal for these firms and that does not have the same meaning as for non-financial firms (Fama & French, 1992). In case that ownership data is not available by the different databases and annual reports, the concerning firm will be left out of the research.

(22)

origin are all based in general on common law. The data contains observations of the identity and shares of the largest owner, main industry classification and four annual observations (2011-2014) of performance measures about the return on assets, Tobins Q, growth rates, capital expenditures, leverage and age of the firms. This leaves a maximum total of 445 X 4 = 1780 observations. In the results the observations are generally lower because of missing values.

To classify the largest shareholders I follow the approach that is used by Faccio and Lang (2002). Faccio and Lang (2002) classify the largest shareholders into groups as Family, State, Financial, Widely Held and Other. In this sample the family firms are firms who are hold by an individual or a family that has more than 20% of the control rights at the end of 2014. From this point the family is assumed to have the effective control over the company. Non-family firms are firms with an ultimate shareholder with more than 20% of the control rights that are controlled by the State, Financials or another named institution. If there is no shareholder that has more than 20% of the control rights, then the company is classified as a widely held company. This assumption is also used by Pedersen & Thomsen (1997). The available ownership data for the firms in this sample are from the end of 2014. The fact that this is only from one year is not likely to be a significant problem. La Porta et al. (1999) stated that the amount of shares held by the largest shareholder are relatively stable over time. This assumption is also used by the study of Maury (2006) and Thomsen & Pedersen (2000). Thomsen & Pedersen (2000) supports this assumption by indicating that there is reason to assume that ownership structures are quite stable and do not adjust quickly to changing economic circumstances. Thomsen & Pedersen (2000) found this assumption to be supported by high structural stability in their research, which makes it reasonable to regard ownership structure as a fixed variable over a 4-year period.

Data about the management of the family firms were retrieved from Orbis, CapitalIQ, annual reports and gathered through the Google Search engine. Active family involvement is measured by the presence of people with the same family name as the ultimate family shareholder. Founder CEO family information is always verified and checked by Google.

3.3 Performance & Control variables

(23)

of the years in the timeframe. Tobin’s Q was defined as the book value of total assets minus the book value of shareholders equity plus market value of shareholders equity divided by the book value of the total assets. Return on assets was measured by net income divided by the book value of the total assets. This measure will be used in this analysis. All these variables will be taken from CapitalIQ database.

Active family control was measured by a dummy variable. There is active control when there is a CEO or another board member that have the same family name as the family shareholder. The variable Family CEO was subtracted in the same way. The variable Governance measure was the amount of control rights a certain family has of a company.

This analysis control for variables that are expected to influence firm performance. The most important ones are legal shareholder protection, firm size, firm age, firm industry, company growth and debt/equity ratio. Firm size was considered to be important because large companies should have performance advantages over smaller firms because of economies of scale and scope (Hansen & Wernerfeld, 1986). Legal shareholder protection was measured by the shareholder protection index of La Porta et al. (1997). This measured how strongly the country’s law favor outside investors against managers and dominant shareholders. This could varies between 0 (lowest) and 5 (Highest) depending on the amount of ‘pro-investors’ rules are written in the law of a country. Firm size will be controlled and measures via the natural logarithm of the average total assets. Firm age could also affect companies performance because of their established name and successes (Hansen & Wernerfeld, 1986). Firm age will be controlled and measured by the natural logarithm of the numbers of years incorporated until 2014. Firm industry variable was used to capture fixed industry effects. Debt to equity ratio will also be measured as a control variable on firm performance. Berger and di Patti (2006) find evidence that highly leveraged companies reduces agency costs and increases firm value by better monitoring by external funds providers. This lead to better monitoring of managers and therefore act more in the interest of the owners and shareholders. Table 2 gives an overview of the descriptions of the variables used in this study.

3.4 Methodology & Robustness checks

(24)

performance are measured by the different ownership types (Family ownership, Non-family ownership and widely hold ownership), governance variables and control variables.

Table 2: Descriptions of the variables

This table shows the variables used in the analysis to evaluates the effect of corporate ownership on firm performance.

Variables Descriptive

Tobin’sQ The Tobin’s Q of the company over the available years from 2011 till

2014. It is calculated as the Market Capitalization divided by the total assets.

ROA The total net income divided by the total assets of a company over the years 2011 till 2014 will measure the performance.

Countryit Country dummy variables that are equal to one if a company is settled

in the corresponding country. These country variables will deal with country specific fixed effects.

Industryit Industry dummy variables. These variables equal to one if a company

operates in corresponding industry. These variables capture fixed industry effects.

Family_Firmit Dummy variable that equals to one when a named individual or a

family holds more than 25% of the control rights of a firm.

Non_Family_Firmit Dummy variable that equals to one when the controlling shareholder

(>25%) is a state, (financial) institution or another non-family groups.

Widely_heldit Dummy variable that equals to one when there is no controlling

shareholder that has more than 15% of the control rights of a firm.

Ownershipit This is the amount of voting shares a controlling shareholder or a

family has of a company.

Legal_environmentit Legal environment is measured by the shareholder Rights index of La

Porta et al. (1998). It varies between zero and six depending on the degree of legal shareholder protection in a country. In this sample it varies from 0 to 3.

Sales_growthit The average growth over the 4-year period 2011-2014, or available

years.

Leverageit

CAPEXit

The average of the debt to equity ratio of a company over the years 2011 till 2014

The ratio of capital expenditures divided by the total sales over the 4-year period 2011-2014, or available 4-years.

LN(Size)it The natural logarithm of the total assets over the years 2011-2014.

LN(Age)it The natural logarithm of the age of a company over the years 2011

till 2014.

(25)

This gives the following regression:

Firm_Performance i,t = α0 + β1 * Family_firmi,t + β2 * Non_Family_firmi,t + β3 * Widely_heldi,t + β4 * Ownership I,t + β5 * Legal_environment I,t + β6 * Sales_growth I,t + β7 * Leveragei,t + β8 *

Capexi,t + β9 * Sizei,t + β10 * Agei,t + i,t (1)

The regressions were performed using panel data with a country fixed-effects specification. All the analyses and results are based on the difference in performance of family firms against the non-family controlling firm group. . This cross-country data is also very helpful because it accounts for unobserved country specific variables that are not measurable within county level data (Sraer and Thresmar, 2004).

In this study there is a possibility for an endogeneity problem. Maury (2006) also stated in his paper that there is a reason to believe that ownership of a particular company is affected by the performance of the company due the fact that families would increase their stake when the company is outperforming the market. Nevertheless, based on the theory of Thomsen & Pedersen (2000) that ownership structures are quite stable and do not adjust quickly to changing economic circumstances. By using data on insider shareholdings we regard ownership structures as an exogenous variable in this sample. To address this self-selection or reverse causality problem statistically, we should using the Heckman (1979) two-step treatment effects model. The Heckman correction approach offers a means of correcting for non-randomly selected data samples what should lead to better conclusions and policy implications. Nevertheless, this Heckman (1979) test is out of scope for this research, for further research it would strongly improve the strength of the conclusions of this research.

4. Descriptive statistics

(26)

there is active family involvement in the firm. Family involvement is in this case family membership in the highest management board of the firm. In 51% (0.18/0.35) of the family-controlled firms, the CEO is a family member. In 20% (0.07/0.35) of the family-family-controlled firms the active family CEO is also the founder of the firm.

Table 3: Summary statistics of variables

The table presents the summary statistics for the 445 non-financial firms from Germany, Belgium, the Netherlands, France and Denmark. The variables used are: Tobin’s Q; Return on assets; Family dummy, equals one if the controlling shareholder with more than 20% of the votes is a named individual or a family, and zero otherwise; Family_active dummy, equals one if a member of the controlling family is also active in the highest management board of the firm, and zero otherwise; Family_ceo dummy, equals one if a member of the controlling family is also the CEO of the firm, and zero otherwise; Family_Founder dummy, equals one if the CEO of the family company is also the founder of the firm, and zero otherwise; Non-family dummy, equals one if the controlling shareholder (like the state, (financial) institution or others) has more than 20% of the votes, and zero otherwise; Firm size as the log of the total assets in the year 2014; Firm age as the log of 2014 minus the year of incorporation; CAPEX as the ratio of capital expenditures divided by the total sales; Sales growth (4 years); Leverage as the total debt divided by the total equity of the firm; and Legal_environment as the antidirector rights index in a country.

Variable Mean Median S.D. Min Max

Tobin's Q 0.94 0.66 1.04 0.00 9.97 Return on Assets 0.03 0.03 0.11 -0.89 1.43 Return on Equity -0.01 0.08 0.70 -11.40 10.77 Family 0.35 0.00 0.48 0.00 1.00 Family_active 0.65 0.00 0.48 0.00 1.00 Family_CEO 0.18 0.00 0.39 0.00 1.00 Family_Founder 0.07 0.00 0.25 0.00 1.00 Non-Family 0.33 0.00 0.47 0.00 1.00 Widely-held 0.32 0.00 0.47 0.00 1.00 Firm size 2.42 2.33 1.09 0.02 5.28 CAPEX 0.12 0.03 0.79 -0.27 24.85 Sales growth 0.12 0.04 0.74 -1.00 14.05 Leverage 0.83 0.45 1.66 0.00 22.25 Antidirector rights 1.58 2.00 0.93 0.00 3.00

In 33% of the firms in this study, the controlling shareholder with at least 20% of the control rights is another group than a family like government (19), institution (33), other company that is not held by a family/person with more than 20% (82) or a bank (15). The group is mentioned with the variable ‘Non-Family’. For 32%, this sample consisted of firms without any controlling shareholders with more than 20% of the total control rights. This group of firms is called ‘Widely-held’. The average sales growth, is about 12% and the ratio ‘capital expenditures of total sales’ is 12%.

(27)

(28)

Table 4: Descriptive statistics by different ultimate shareholders

This table represents summary ownership statistics of the 445 non-financial firms coming from the Netherlands, Belgium, Germany, France or Denmark. Family firms are the firms with an individual or more members of a family that holds more than 20% of the control rights. The Family active is a family controlled owned company that has at least one family member in the highest management board. The family non-active category consisted of family controlled firms that does not have a family representative in the highest management board. The Non-Family firms are the firms with a controlling shareholder (PE firms, institutions, governments e.g.) that holds more than 20% of the control rights but is not classified as a family. The widely held firms are firms without an ultimate shareholder that has more than 20% of the control rights of a firm. The variables used are: Tobin’s Q; Return on assets; Ownership as the % of direct control rights; Firm size as the log of the total assets in the year 2014; Firm age as the log of 2014 minus the year of incorporation; CAPEX as the ratio of capital expenditures divided by the total sales; Sales growth (4 years); Leverage as the total debt divided by the total equity of the firm; and Legal environment as the antidirector rights index in a country.

Variable Family Family

active Family nonactive Non-family Widely held Fam active vs. Non active Family vs nonfamily Family vs. Widely held Nonfamily vs. Widely held

mean mean mean mean mean t-stat t-stat t-stat t-stat

Tobin's Q 0.94 1.00 0.88 0.86 0.98 1.11 1.27 0.65 1.82*

Return on assets 0.05 0.05 0.04 0.02 0.03 1.19** 4.59*** 3.14*** 1.12

Ownership 0.49 0.50 0.50 0.49 0.10 1.46 1.09 45.66*** 44.07***

Firm size (log of total assets) 2.14 2.10 2.20 2.33 2.83 1.19 3.14*** 11.63*** 7.94*** Firm age (log of 2014-incorporation) 1.59 1.56 1.63 1.52 1.53 1.83* 2.70*** 2.12** 0.52

Sales growth 0.11 0.15 0.05 0.07 0.19 1.95* 1.09 1.60 2.67***

Leverage 0.79 0.81 0.76 0.98 0.70 0.45 1.78* 1.16 2.60***

CAPEX 0.06 0.07 0.04 0.19 0.10 2.02** 2.37** 1.89* 1.51

Shareholder_protection 1.52 1.46 1.63 1.65 1.56 2.10** 2.29** 0.88 1.71*

Observations 620 404 216 596 564

(29)

5. Regression results

In this chapter, the relationship between firm ownership and performance is analyzed. The regressions were performed using a country fixed-effects specification. All the analyses and results are based on the relative performance against the non-family controlling firm group. In the first section of the results, the analyses of the different ownership types and firm control will be discussed. In the second part of the results, the different types of family involvement will be compared to firm performance.

5.1 Ownership types & firm performance

Table 5 till 9 shows the results of the regression on the relationship between different ownership structures and firm performance. The regressions are performed by panel data using a country-fixed effects model. The country fixed-effects specification is supported by the Hausman Test. These results are given in Appendix A.

Table 5 shows the results of the regression on the relationship between ownership types and firm performance. The performance were measured by the Tobin’s Q (TQ) for firm value and Return on Assets (ROA) for firm profitability. In columns 1 and 4, companies controlled by a family or widely held firms are compared with non-family controlled firms based on firm value (TQ) and profitability (ROA). In columns 2 and 5 we analyze the relationship between active and active family firms and widely held firms versus non-family firms as control group based on firm value (TQ) and profitability (ROA). In columns 3 and 6 the relationship between active and non-active family firms is analyzed on a sub sample of only family firms.

(30)

Table 5: Relationship ownership type and firm performance

The table reported the results of the country fixed-effects regression for a sample of 445 non-financial Western-European listed firms. The dependent variable in the tables are Tobin’s Q in columns 1-3, and return on assets in columns 4-6. The independent variables are: Family, a dummy variable that equals one if there is an individual or more members of a family that holds more than 20% of the voting rights and zero otherwise; Fam_active, a dummy variable that equals one if there is at least one family member of the ultimate shareholder is participating in the highest management board of a firm and zero otherwise; Fam_nonactive, a dummy variable that equals one if there is no active participation in the board from an ultimate family shareholder; Widely Held, a dummy variable that equals one if the company hasn’t a controlling shareholder that has more than 20% of the voting rights and zero otherwise; Ownership, a variable that measures the total shares that a controlling shareholder holds (or a widely held firm); Firm size as the log of the total assets in the year 2011-2014; CAPEX as the ratio of capital expenditures divided by the total sales; Sales growth (4 years); Leverage as the total debt divided by the total equity of the firm; Firm age as the log of the 2014 minus the year of incorporation; and dummy variables for two-digit sic codes. Values between parentheses represent T-statistics.

Variable Full sample Full sample Family firms Full sample Full sample Family firms

TQ TQ TQ ROA ROA ROA

(1) (2) (3) (4) (5) (6) Constant 1.098*** (8.45) 1.231*** (8.45) 1.304*** (5.10) -0.048*** (-3.6) -0.044*** (-3.61) 0.044** (1.79) Family 0.119* (1.87) 0.015*** (2.65) Fam_active 0.140** (1.96) 0.224** (2.34) 0.015** (2.50) 0.003 (0.35) Fam_nonactive 0.079 (0.91) 0.013* (1.72) Widely Held 0.228** (2.55) 0.228** (2.55) 0.000 (0.02) 0.000 (0.02) Ownership 0.296* (1.92) 0.297* (1.92) -0.129 (-0.60) 0.032** (2.41) 0.032** (2.41) 0.043** (2.04) Firm size -0.024 (-0.80) -0.023 (-0.81) 0.097** (1.91) 0.018*** (7.8) 0.0187*** (7.79) 0.020*** (4.14) CAPEX 0.316*** (4.87) 0.314*** (4.86) -0.191 (-0.77) -0.020*** (-6.66) -0.020*** (-6.66) -0.066*** (-2.66) Sales growth 0.09** (2.35) 0.0924** (2.32) 0.0856 (1.42) -0.004 (-1.3) -0.004 (-1.31) 0.011** (1.82) Leverage -0.076*** (-4.37) -0.067*** (-3.59) -0.077** (-2.24) -0.001 (-0.97) -0.001 (-0.97) -0.012*** (-3.78) Firm age -0.286*** (-4.05) -0.286*** (-4.04) -0.360*** (-2.75) 0.011* (1.88) 0.011* (1.89) -0.034*** (-2.75)

SIC-code incl. YES YES YES YES YES YES

R-squared adj 18% 17% 22% 17% 14% 11%

Observations 1457 1457 526 1561 1561 555

(31)

When family firms are compared to non-family firms (in column 1 and 4 in table 5), it can be seen that family control has significant higher firm value (TQ) (β =0.119, p < 0.1) and significant higher profitability (ROA) than non-family firms (β =0.015, p < 0.01). Thus, the results in table 5 provide evidence on benefits with family control compared to non-family controlled firms, plausibly due to lower agency costs between shareholders and managers, in Western European firms. Further, based on table 5 we can conclude that widely held firms perform statistically significant higher on firm value (TQ) compared to non-family firms but not significantly different based on profitability (ROA).

In columns 2-3 and 5-6 in table 5, the effects of active versus non-active family control on firm value and profitability are also examined. Active family control defined as holding at least one family member participating in the highest management board of a firm. Passive family control occurs when a controlling family has no family member within the highest board of the firm. Table 5 column 2 shows us that active family control improves the firm valuation positive significantly. The coefficient from family firm in column 1 from 0.119 (p < 0.1) rises to 0.140 (p < 0.05) in column 2. Thus, active family firms deliver significant higher firm valuation and profitability. When active family control is analyzed on a sub sample of only family firms in table 5 (columns 3 and 6), the same pattern is found for firm valuation (β =0.224, p < 0.05) but there is no statistically significant difference in profitability (β =0.003, ns) between active and non-active family firms. Overall, the findings on (active)-family ownership support the hypothesis that a (active)-family controlled firm has a higher firm value than a non-family controlled firm due less agency problems between shareholders and the management. Table 5 also supports the hypothesis that Western European listed firms with management presence will be higher valuated than non-family and non-active family controlled firms.

(32)

Table 6: Relationship between family control and firm performance

This table presents the results of the regression for a sample of 445 listed firms. The dependent variables are Tobin’s Q in column 1 and return on assets in column 2. Family_control (%) [20-35], [35-45], [45-55] and [55-+] are dummy variables that equals one if the controlling shareholder is a family within a corresponding voting rights range, and zero otherwise. Widely Held, a dummy variable that equals one if the company hasn’t a controlling shareholder that has more than 20% of the control rights and zero otherwise; Ownership, a variable that measures the total shares that a controlling shareholder holds (or a widely held firm); Firm age as the log of the 2014 minus the year of incorporation; CAPEX as the ratio of capital expenditures divided by the total sales; Sales growth (4 years); Leverage as the total debt divided by the total equity of the firm; Firm size as the log of the total assets in the year 2014. Values between parentheses represent T-statistics.

Variable TQ ROA (1) (2) Constant 1.054*** (6.67) -0.049*** (-3.65) Family_control (%) [20-35] 0.275*** (2.70) 0.016** (1.80) Family_control (%) [35-45] 0.3844*** (2.95) 0.034*** (3.01) Family_control (%) [45-55] -0.013 (-0.11) 0.006 (0.55) Family_control (%) [55- +] -0.060 (-0.62) 0.010 (1.23) Widely Held 0.359*** (3.51) 0.003 (0.35) Ownership 0.670*** (3.26) 0.041** (2.33) Firm size -0.014 (-0.49) 0.019*** (8.00) CAPEX 0.309*** (4.80) -0.020*** (-6.67) Sales growth 0.092** (2.33) -0.004 (-1.29) Leverage -0.067*** (-3.62) -0.001 (-0.93) Firm age -0.297*** (-4.20) 0.011* (1.81)

SIC-code incl. YES YES

R-squared adj 18% 15%

Observations 1457 1561

*, **, *** Significant at the 10%, 5% and 1% levels respectively.

Table 6 investigates the relation between different family control ranges and firm performance compared to non-family controlled listed firms. Table 6 give some interesting results regarding the value and profitability of the family firms. Again there is a difference between firm valuation and firm profitability within the defined family control ranges. Family control for the first two ranges have a significant positive relation on the firm value. The firm value coefficient rises from the first range [20-35] from 0.275 to 0.384 in the [35-45] range and then starts to decline when the level of control exceeds the range of 45% voting rights. Both ranges are performing positive and significant on a 1% level compared to non-family firms. When a family firm controls more than 45% of the voting rights they underperform related to non-family firms. Nevertheless, this relation is insignificant. These results are in line with existing literature of Maury (2006). The results of Maury (2006) indicates that family firms performance in non-majority family controlled firms are significantly higher compared to non-family firms, while no significant relation is found when the family has a majority of the voting rights.

(33)

In conclusion, we can conclude that family firms with ownership rights till 45% outperform non-family firms and that family firms that exceed that amount does not

outperform non-family firms due the increasingly chance of family opportunism and private benefit extraction. The results of table 6 supports the assumption that there is an ownership control range in which Western European family controlled firms outperform non-family controlled firms in terms of firm valuation and profitability. Therefore, hypothesis 2 is supported.

5.2 Family involvement & firm performance

To investigate the hypotheses that there will be a family board member presence range in which west-European family controlled firms will perform the best, we use the same procedure as presented in table 7 columns 1 and 3. In the analysis the family firms are divided in dummy variables that equals 1 if it’s within the certain range and zero otherwise. These variables will then be compared to non-family firms as control group and widely held firms as another dummy variable. The results in table 7 columns 2 and 4 shows the relation between the different sorts of CEO’s and firm value and profitability related to the non-family control group.

Table 7: Relationship between family involvement and performance

This table presents the results of the regression for a sample of 445 listed firms. The dependent variables are Tobin’s Q in columns 1 and 2 and return on assets in column 3 and 4. Family_presence (%) [0], [1-25], [25-50] and [50-+] are dummy variables that equals one if the controlling shareholder is a family and has active firm participation within a corresponding board participation range, and zero otherwise. Widely Held, a dummy variable that equals one if the company hasn’t a controlling shareholder that has more than 15% of the control rights and zero otherwise; Ownership, a variable that measures the total shares that a controlling shareholder holds (or a widely held firm); Firm age as the log of the 2014 minus the year of incorporation; CAPEX as the ratio of capital expenditures divided by the total sales; Sales growth (4 years); Leverage as the total debt divided by the total equity of the firm; Firm size as the log of the total assets in the year 2014. Values between parentheses represent T-statistics.

Variable TQ TQ ROA ROA

(1) (2) (3) (4)

Constant 1.23*** (8.43) 1.139*** (7.78) -0.045*** (-3.64) -0.057*** (-4.63)

Board participation range:

Family_presence (%) [0] 0.08 (0.88) 0.013* (1.78) Family_presence (%) [1-25] 0.209*** (2.65) 0.017** (2.53) Family_presence (%) [25-50] -0.002 (-0.02) 0.013 (1.21) Family_presence (%) [50- +] -0.510 (-1.48) -0.021 (-0.72) Founder CEO 0.446*** (3.89) 0.068*** (6.87) Descendant CEO -0.127 (-1.36) -0.007 (-0.83) Non-Family CEO 0.010 (1.31) 0.004 (0.66) Widely Held 0.233*** (2.60) 0.225** (2.55) 0.000 (0.07) 0.000 (0.05) Ownership 0.313** (2.03) 0.308** (2.00) 0.033** (2.48) 0.033** (2.54) Firm size -0.026 (-0.92) -0.031 (-1.12) 0.019*** (7.75) 0.0182*** (7.65) CAPEX 0.311*** (4.83) 0.317*** (4.93) -0.020*** (-6.68) -0.019*** (-6.70) Sales growth 0.091** (2.30) 0.097** (2.45) -0.004 (-1.32) -0.004 (-1.24) Leverage -0.068*** (-3.66) -0.067*** (-3.64) -0.001 (-0.99) -0.001 (-0.94) Firm age -0.284*** (-4.02) -0.210*** (-2.89) 0.011* (1.92) 0.020*** (3.38)

SIC-code incl. YES YES YES YES

R-squared adj 17% 18% _14% _16%

Observations 1457 1457 ₁₅₆₁ 1561

(34)

Table 9 column 1 shows us that family presence in the board till 25% is significant value enhancing (β =0.209, p < 0.01) and over the 25% value destructive but not significant compared to non-family controlled firms. Family firms without active family participation does not significantly outperform non-family firms. The family firms that have board presentation from at least 1% till 25% scores the best compared to non-family firms. The results show that the beta coefficient of this group scores 0.219 higher than others on Tobin’s Q at a 1% significance level. When the active board participation then rises it will result in lower but not significant firm valuation compared to non-family firms. We can concluded based on table 7 that the firm valuation doesn’t increase when firms are installing more family members in the board. Based on the profitability of firms (column 3) we can conclude that non-active family firms and family firms with board representation until 25% outperform non-family firms but only marginally (β = 0.013 and β = 0.017) at a significance level of 10% and 5%. In terms of profitability we can conclude that active family participation until 25% of the board has the best performance compared to non-family listed firms. Therefore we can’t reject the hypotheses 3 that there is an active family board participation range that increases firm valuation. Based on these results that is the range of [1%-25%].

Related to the existing literature we can conclude that based on the resource based view that family firms only have a limited pool of talented family member that could add value within a management board (Schumpeter, 1934). This could also be from the fact that when the family involvement becomes higher but their controlling shares doesn’t exceeds, the proportion of presence of the ultimate shareholder and cause an unfavorable situation for minority shareholders that also want active participation in the board. Burkart et al. 1997 also stated that when minority shareholders are also active managers, they could add value because they have an incentive to act as guardians over the resources of the firm. In that case, the expropriation of minority shareholders will be a problem that cause firm devaluation (Miller et al. 2006). The results are also in line with Anderson & Reeb (2003) that argued that independent board membership from outside is value enhancing because such parties are independent and contribute expertise, objectivity and provide alternative perspectives.

(35)

founder as CEO. The third dummy CEO variable is a non-family CEO. The results show that a founder CEO has significant higher performance than all the other forms of firms. The firm valuation increases by 0.446 at a 1% significant level. The return on assets is also positive related at 0.068 at a 1% significant level. For the firms with a descendant as CEO the performance are lower compared to non-family firms but not significant. Family firms with an outsider as CEO is also positive correlated with firm valuation and profitability but not significant. In this case, the widely held firms also perform significant better on firm valuation compared to non-family firms. This is in line with the previous results.

(36)

5.2 Legal environment and firm performance

In table 8 we split our dataset into firms that are based in countries with equal to and above median scores on shareholder protection (Panel A) and those below the median score (Panel B). The countries in panel A are the Netherlands, France and Denmark and in Panel B Belgium and Germany are selected. In Germany firms are not regulated by law that protects investors and in Belgium there are barely any rules in terms of minority shareholder protection. In columns 1 and 3 the firm valuation and profitability data is based on family firms and widely held firms in comparison with non-family firms. In columns 2 and 4 the firm value and profitability is measured in relation with active, non-active and widely held firms compared to the non-family firms as control group.

Based on table 8 column 1 family firms do not have significant higher firm valuation (β =0.065, ns) in panel A. Nevertheless, table 8 column 3 shows that family firms do have marginally, but significant higher results for profitability (β =0.040, p < 0.01) compared to non-family firms. The widely held firms have significant higher firm valuation (β =0.367, p < 0.01) in countries where investor protection is high but that doesn’t count for return on assets (β =0.007, ns).

Based on table 8 panel B, family firms have a positive coefficient but insignificant in terms of firm valuation (β =0.476, ns) compared to non-family firms. The same pattern holds for the widely held firms (β =1.162, ns). We see no significant differences in firm valuation when the family firms are divided in active and non-active family firms. In terms of profitability (table 8 columns 3 and 4) family firms have a positive coefficient but again insignificant (β =0.005, ns). compared to non-family firms in the countries where the investor protection is low. This pattern holds when we divided the family firms in active and non-active family firms. Widely held firms do statistically significant outperform widely held firms in terms of profitability (β =0.018, p < 0.05).

(37)

high investor protection countries. In our panel of high investor protection countries, we find significant positive results for active family firms in terms of profitability and valuation in comparison with non-family firms. The family firms that operates in countries with low investor protection from our sample do not perform better in terms of value regarding non-family firms. This is in line with the results of Maury (2006). We can conclude from our sample that firm valuation does differ in panel A and B. In terms of profitability, the family firms in legal environments that favors shareholder protection do perform better than family firms in countries with lower investor protection.

Regarding our hypotheses 6, we cannot reject the hypotheses. Family firms in panel A perform better than firms in panel B. This is in line with the theory provided in the theory section. Overall, family firms have better firm performance than nonfamily firms and especially in active family firms in countries with high shareholder protection. These results are in line with the theory in the theory section that the level of legal shareholder protection has a different impact on the firm performances in Western European listed firms. This difference is driven by the agency problems between the controlling family and the minority shareholders. Even though increased globalization, that would causes similarities in firms around the world, different legal regimes in different countries can result in differences in corporate governances and performance (Memili & Misra, 2015).