The effects of family ownership, other controlling
shareholders and diffuse ownership on firm performance:
Evidence from the 2008 crisis in Western-Europe
Mitchel Koopal: 10454276
mitchel.koopal@student.uva.nl University of Amsterdam
Finance Department Master thesis
Business Economics: Finance track Supervisor: Florencio López de-Silanes
Abstract
This paper investigates the relationship between majority ownership and firm performance during the 2008 crisis in Western Europe. In a sample of 1189 listed firms of the five largest European economies for the years 2002-2012, I compare firm performance between family owned firms, other controlling shareholders and diffusely owned firms, by using Return on Assets and Tobin’s Q as performance measures. I find that family firms have a higher Tobin’s Q than other controlling shareholders and that active family ownership leads to increased Return on Assets. However, diffuse ownership outperforms the other types of majority ownership. My findings support that family ownership increases firm performance compared to non-family majority shareholders, but do not support the benefit of large majority shareholders over diffuse ownership. I do not find evidence that during the 2008 crisis family firms consistently outperform other types of shareholders, although diffuse ownership leads to higher performance both before and after the 2008 crisis.
I. Introduction
Berle & Means (1932) laid the foundation of the common belief that ownership in the United States is dispersed. Nowadays, this view has been argued by different articles. La Porta et. al. (1999), Holderness (2009), Faccio & Lang (2002) showed that in modern society, corporate ownership around the world is concentrated in the hands of families.
There is no consensus if majority ownership increases firm performance. In general it is believed that the unification of ownership and control (Fama & Jensen, 1983) will result in higher performance, but expropriation of minority shareholders and using the firm for private benefits will reduce firm value (Shleifer & Vishny, 1997).
Family ownership has often been investigated separately, but yielding different results. Morck et. al. (1988) state that family ownership in the US leads to poor performance due to managerial entrenchment, while Maury (2006) studies family-firm performance in Western Europe and states that firm’s which are majority held by families outperform non-family majority firms.
Family ownership is commonly associated with retaining family members in management positions of the firm, while not necessarily being qualified for the position. Despite this notion, different articles (Maury 2006, Anderson & Reeb 2003 & Barontini & Caprio 2006) find that family firms tend to perform better when the firm is actively managed by the family. The reason for this increase in performance, is the fact that families have long term investment horizons and are associated with the firm’s heritage. Families govern their firm in a different way than non-family firms, resulting in differences in performance.
Economists often claim that the 2008 crisis was caused by bad governance practices, while Gupta et al. (2013) contradicts this by stating that different types of governance did not have any effect on stock prices claiming that governance during the crisis does not matter. Every type of owner has its own type of governance. Different types of governance could yield different firm performance. In a high demanding time like the 2008 crisis, better types of ownership and governance could mean higher performance.
This research will investigate the relationship between family ownership and firm performance. Next to that, this research will investigate if there are differences in performance by ownership type before and after the 2008 financial crisis. In previous research like Maury (2006), family performance is measured against other types of non-family controlling shareholders. This research will also look if non-family ownership outperforms diffuse ownership. This analysis will also investigate if active family management increases firm performance.
Due to differences in regulations and the ratio of family owned firms to non-family owned firms, this research will focus on the five largest economies in Western Europe, i.e. the United Kingdom, Germany, France, Italy and Spain.
Using ownership and financial panel data of 1189 individual firms for the period 2002-2012 in the five largest European countries, I find that family ownership leads to an increase in firm performance. In particular, family ownership results in a higher Tobin’s Q than non-family majority owned firms. However, this research shows that diffuse ownership significantly performs better than majority ownership. This result is true before and after the 2008 crisis, saying that the best type of ownership during a crisis would be diffuse ownership. I find that families have large cash flow stakes in the company, while cash flow compared to control stakes are larger in non-family owned firms. This means that in non-family firms, the control enhancing devices are larger than for family firms.
Consistent with Maury (2006) & Anderson & Reeb (2003), active family ownership results in higher Return on Assets. Meaning that if a controlling family member holds a top two position in the firm’s management, this increases firm performance. There are several reasons why active ownership can increase firm performance. The most important reason is that active ownership can result in a decrease in the agency problem due to a better alignment of incentives between shareholders and the management. The second reason could be that families have long term investment horizons, therefore reducing myopic investment behavior, resulting in better performance. I find that when a firm is owned and actively managed by a family, this results in an increase in firm performance. However, due to expropriation of minority shareholders and using the firm for private benefits, family ownership underperforms compared to diffuse ownership.
In 2008 the crisis had a large impact on the firm’s performance. This research does not significantly support that family firms performed better during the crisis than non-family firms due to different types of corporate governance. However, during the crisis, diffuse ownership leads to higher performance. In countries with both high and low developed investor protection rights, diffuse ownership performs better.
This paper is organized as follows. Section II discusses prior literature about the effects of ownership on firm performance. Section III describes the sample, and provides the summary statistics. Section IV provides the empirical results. Section V provides the robustness tests of the results. Section VI concludes the paper and provides its limitations.
II. Different types of ownership on performance 1. The effects of a majority shareholder
La Porta et. al. (1999) studied the ultimate ownership of firms and conclude that in modern society, corporate ownership is very concentrated, and not as dispersed as the original model by Berle & Means (1932), who believed that ownership is too small in large firms for managers to be interested to act in the shareholders benefit. When there is a majority shareholder present, this results in different outcomes for shareholders benefits. In past research there are different views about having a majority shareholder.
A. The upside of having a majority shareholder
The agency theory, introduced by Jensen & Meckling (1976), and the separation of ownership and control by Fama & Jensen (1983), showed that conflicts arise in firm’s due to utility maximization of different stakeholders, and the negative effects of these conflicts have on firm performance. According to Holderness (2003), much of this agency theory is based on diffuse ownership, and is therefore less applicable to majority ownership. The obvious upside of having a majority shareholder is that there is a better alignment between incentives of principals and agents. This alignment reduces the separation of ownership and control, which increases firm performance.
Shleifer & Vishny (1997) built a foundation for corporate governance and investor protection. They state that having a large shareholder reduces the agency costs since the voting rights will be higher, therefore the large shareholder is able to put more pressure on the management. When the large shareholder puts more pressure on management, this could result in less agency conflict due to more monitoring. Another reason for majority shareholders to monitor, is that they own large stakes in one company. Since large stakes in the same company involve substantial risk, this risk can be mitigated by monitoring and putting pressure on the management. In a widely held firm, it can be costly for one minority shareholder to monitor. Therefore if all minority shareholders share this view, nobody is monitoring the management. Shleifer & Vishny (1986) create a model to show that when a firm has a majority shareholder, it gives an incentive to minority shareholders to monitor. This comes due to the fact that the majority shareholder can influence the activities of the management to make sub-optimal decisions for minority shareholders. Therefore it is less costly for minority shareholders to monitor the large shareholder, than to avoid monitoring. Holderness (2009) hand collects United States ownership data, to compare the ownership structure in the United States with the rest of the world. He shows that ownership structure in
the US is very comparable to the Western-European ownership structure, meaning that even in the US, ownership is less dispersed than originally believed. As a result of this ownership, large shareholders want to play an active role in the firms business, because they can actually make ‘a difference’ compared to the passive minority shareholders. This can be explained by the fact that majority shareholders can set large long term goals, instead of aiming for quick returns on their investment, therefore resulting in higher performance.
Franks & Mayer (1994) conduct a study of ownership and control in Germany and find that in Germany firms with large shareholders have a higher turnover of managers. High management turnover means less managerial entrenchment, and less underperforming managers. As stated by Jensen and Ruback (1983): “there is nothing more costly than an
underperforming manager”.
B. The downside of having a majority shareholder
A majority shareholder does not always result in benefits for the firms’ performance. As stated by Demsetz & Lehn (1985), the greatest disadvantage of having a large shareholder is shirking. When the largest shareholder is neglecting the duties for a firm, the benefits accrue to him, but the costs are spread out over all the shareholders, resulting in lower firm performance and lower return for minority shareholders. When the majority shareholder holds a larger stake in the company, this means that the costs will relatively accrue more to himself. Therefore if the shareholder becomes larger, he will try to avoid shirking since he is relatively cutting more into his own benefits.
Shleifer & Vishny (1997) create a framework for different costs of having a large investor. In their paper they state that there are three main problems with having large shareholders; expropriating other investors, managers and employees, value diminishing due to personal utility maximization, and expropriation of other stakeholders. Large investors treat themselves better than minority shareholders. The majority shareholder can distribute wealth according to his own benefit. Therefore it could be less attractive for minority shareholders to acquire shares in a firm with a majority shareholder. This happens especially when control rights are larger than cash flow rights. When control rights are larger than cash flow rights, this means that the large shareholders would encounter less financial risk compared to the amount of control they can exert. According to Barontini & Caprio (2006), families are often associated with these control enhancing structures.
Barclay & Holderness (1989) analyzed 63 block trades in the US, and show that firm value is reduced because the premiums in takeovers accrue to private benefits of large
blockholders. Next to that the majority shareholder can be involved in tunneling or empire building (Johnson et. al 2000), which means that the owner uses corporate benefits for his own value, which results in a decrease of firm value. Shleifer & Vishny (1997) also state that because large investors are undiversified they bear more risk. Due to this extra risk, large shareholders could make decisions that are sub-optimal for the firm.
Lehmann & Weigand (2000) conducted a study for 361 German corporations between 1991 and 1996, to find a relationship between performance and corporate governance. They say that the existence of a large shareholder in Germany does not necessarily increase firm value due to over-monitoring. Over-monitoring can result in bad decision making by the management, therefore reducing firm value. This is confirmed by Burkart et. al. (1997), which state that there is a trade-off between control and initiative. They build a model in which it is proved that managerial incentives, to make good investment decisions, are reduced by monitoring from large shareholders.
C. The effects of diffuse ownership
Diffuse ownership was originally examined when the foundation for modern day corporate governance was built by Berle & Means (1932). Their research created the basis for the common belief that ownership around the world is diffuse. They state that the separation of ownership and control is the trade-off between profit maximization and resource allocation. Nowadays it is acknowledged that in Europe the firms are largely owned by controlling shareholders (Faccio & Lang, 2002), and Holderness (2009) shows that also in the United States ownership is not dispersed.
Holderness & Sheehan (1988) pair majority controlled firms with diffusely owned firms, and find that majority ownership does not statistically outperform diffuse ownership. But due to the fact that majority ownership often pay out large compensation fees to executives, firm value is reduced by a small amount. Demsetz & Lehn (1985) try to explain why there are diffusely held firms since blockholdership unifies ownership and control. Since every company wants to maximize profit and value, they state that there have to be other reasons why widely held firms exist. The choice for widely held ownership is often the trade-off between monitoring management, and efficiency. They find no significant relationship between ownership concentration and firm performance. In other words, majority owners do not necessarily increase firm performance.
2. The effects of family ownership
The effect of family ownership on governance and performance often have been studied in past research, all yielding different results. It is a type of majority ownership that is considered to be different than others in terms of governance and performance.
A. The upside of family ownership
Anderson & Reeb (2003) conduct a study for the S&P 500 firms to find a positive relationship between founder-family ownership and firm performance due to the economic incentives to maximize firm value and the power that controlling families have to obtain this value maximization. They state that this can be explained by the fact that family wealth is linked to the welfare of the company they own. In theory because they have “all their eggs in one basket”, they want to maximize their own utility since it is linked to the company. As a result, families have strong incentives to minimize free riding costs and to diminish the agency problem by monitoring managers. Morck et.al. (1988), conduct a study to find the characteristics of a target firm in a takeover. In their sample of 1980 Fortune 500 firms, they find that family ownership is less likely to be involved in a hostile takeover. Hostile takeovers generally reduce firm value due to the possible use of takeover mechanisms. (poison pills, golden parachutes etc.), therefore families could increase firm value.
Maury (2006) examines the relationship between family ownership and firm performance in Western European countries. One of the most important reasons why family firms perform better is due to active control. This entails that a member of the owning family has a top executive position. Active control can reduce the separation of ownership and control, and can result in better firm performance. This results in long-term investment decisions that can benefit the firm (Stein 1988), so that they can carry on the family legacy. According to different research (Villalonga & Amit 2006, and Barontini & Caprio 2006), founder CEO’s are the ones that cause the increase in firm performance. They find no results that when descendants take the CEO position that firm performance is still higher than its non-family peers. Daily & Dalton (1992) study the relationship between corporate governance and corporate performance for entrepreneurial firms. They state that “founder- CEO’s apply good governance codes, while common belief is that managerial executives should perform better than family executives, who are not instantly qualified for the position. He (2008) studies the relationship of a founder-CEO on firm performance. He finds that when the founder CEO is also the chairperson, this leads to higher performance than professionally managed firms. This is due to their intrinsic motivation and commitment to the
firm they created themselves. He (2008) also states that founder-CEO’s are associated with less compensation than professional managers because they have less job insecurity. This research will not investigate if founder-CEO’s have a higher performance. However, in this research I will examine if active family management, i.e. a family member holds a top two management position in the firm, could increase firm performance.
Next to the quantitative ways in explaining why family ownership could increase performance, Habbershon & Williams (1999) built a Resource Based View (RBV) to explain in a qualitative matter, why family firms generally perform better. They state that family firms have a competitive advantage due to the so called “familiness”, which is the bundle of resources in a way the firm interacts between the family, the members and the business. This can be seen as an idiosyncratic resource for firms that could be turned to their advantage. They state that firms should not just be compared by performance such as ROA, but also at a behavioral level. Since family firms tend to behave differently and have a different social impact on society this means that family firms can perform better in some ways.
Carney (2005) uses this Resource Based View to explain the competitive advantage that family-owned firms have over non-family owned firms in a qualitative analysis. According to Carney (2005), this competitive advantage that family-owned firms have, comes from three different components; parsimony, personalism and particularism. The parsimony component in family-owned firms says that there is a unification of ownership and control, and reduces agency costs since the management and shareholders are often the same persons. People tend to be more careful investing when it is their own money that is being used. Personalism entails the fact that a family firm has fewer constraints on the type of managing, due to the high concentration of ownership they have better incorporation of the ultimate control. Particularism means that families see their firm as their “own business”, and prefer specific types of doing business as is usual in the family.
B. The downside of family ownership
Hermalin & Weisbach (1991) study the effect of monitoring and board composition on managerial incentives and firm performance. They state that family controlled firms are notorious for putting their own value above the value of other shareholders, by reserving top management seats for family members, instead of professionally qualified managers, and using company resources for their own value. Morck et. al. (1988) state that founding family ownership reduces Tobin’s Q for older firms because the managers are entrenched. In contrast, in young firms when the founding family takes position among the top two officers,
this has a positive effect on Tobin’s Q. The reason for this is that in younger firms the founder has a lot of entrepreneurial influence on the business. When a majority shareholder becomes too large, this could result in a decrease in performance due to less entrepreneurial input, and more costs for the firm. This is shown by Morck et. al (1988) who state that the benefits of family ownership starts to taper off around at the 25% level.
Morck et. al. (1998) conduct a study in which they find the so called “Canadian disease”, which means that in firms owned by wealthy families, growth is blocked due to inherited control. This organizational form entails that old family capital makes managers entrenched, causing a lack of creativity and entrepreneurship due to protecting the inheritance and family name. Morck & Yeung (2003) conduct a study to reveal the fact that family firms have their own agency conflicts, and they find that family firms have different agency costs than widely held firms because managers will act in the interest of the family, i.e. protecting the inheritance and family name, instead of the other shareholders.
Furthermore family firms are often associated with control enhancing structures (Barontini & Caprio 2006), for instance by having more control rights than cash flow rights, pyramidal structures and dual-classes shares. By creating a construction in which the family has more control rights than cash flow rights, they can exert more control over the company without bearing risk. La Porta et. al. (1999) state that listed family owned firms have to pay a premium to minority shareholders in order to compensate them from having the risk that large shareholders will use the firm for private benefits.
3. The effects of other types of ownership
Next to family ownership there are other types of majority ownership. The other types of majority ownership can be divided according to La Porta et. al. (1999) into financial firms, non-financial firms and the State1.
A. The effects of (non) financial firms as a controlling shareholder.
Corporation ownership is another type of blockholdership and can be divided into financial firm, and non-financial firm ownership. Since they own a majority of the stock, they should be able to reduce the separation of ownership and control, and increase firm performance, but due to different types of governance, this could give different results than family ownership. In this research financial controlling shareholders are similar to
1 Miscellaneous ownership is not considered a separate type in this part, since it is the remaining majority
financial controlling shareholders. Both types are companies that have majority ownership in another firm, the only difference between the two is the industry that they are in. The main effects of having a non-financial controlling shareholder are already explained under the effects of having a majority shareholder earlier in this research, therefore in this part, I will focus on financial institutions as majority shareholders, since they could have specific governance practices.
Jiambalvo et. al (2002) empirically examine if stock prices of stocks owned by institutional investors, better reflect future earnings. They state that institutions are sophisticated shareholders, and that the current stock price of institutionally owned firms reflect future earnings better than other types of control. Therefore institutional investors create a higher stock return. Lehmann & Weigand (2000) find that having large institutional shareholders perform better than family shareholders due to the fact that they are better monitors. Chaganti & Damanpour (1991) conduct a pairwise regression on 40 manufacturing firms, to find the relationship between institutional ownership, capital structure and performance. They find that the amount of shares held by institutions (both inside and outside the firm) affect capital structure and firm performance. Firms held largely by outside financial institutions tend to have a higher firm performance. Gompers et. al. (2003) examine the relationship between institutional shareholders and stock prices for the period 1980 to 1996. They show that institutional owners like to invest in large firms, and this leads to higher stock prices. The reason for this is that institutions can spot the high momentum stocks, and need to invest in stocks that outperform the market. Barber & Odean (2008), conduct a study which shows that individual investors specifically invest in eye-catching stocks. According to them, institutional investors have an advantage when trading in stocks compared to individuals, because they have the power to search for stocks that fit their investment strategy, and are therefore outperforming individuals.
As stated by Porter (1992) institutional investors are often characterized as owners who are overly focused on current earnings. This short-termism results in a decrease in long-term performance. The reason for this is that institutional investors want to have quick investments and short term profits before they cash-out. Bushee (2001) empirically examines that institutional investors are focused on near-term earnings, which leads to near-term earnings inflation that managers want to avoid. When managers are trying to avoid this, they can fall in sub-optimal decision making for the firm, which decreases firm value.
B. The effects of State ownership
“State control is a separate category because it is a form of concentrated ownership in which the State uses firms to pursue political objectives, while the public pays for the losses”
(Shleifer and Vishny 1994). This view already shows the major effect of state ownership. State ownership is often associated with a lack of innovativeness, but to keep regulations, and competitiveness under control while attaining social goals. Different research has focused on State ownership. Shleifer (1998) states that there has to be a certain tradeoff between capitalism and socialism, where private ownership is preferred over public ownership. Ding et. al. (2007) conduct a study of State ownership on earnings management, and state that when the firm is largely State owned, this has an extra benefit in the principal-agent relationship, since the separation of ownership and control is reduced by government officials in the firm’s management. They state that the downsides of State ownership is that the State uses the companies’ resources for their own purpose, and that State officials in the firms are entrenched. Shleifer & Vishny (1997) state that State ownership is a very inefficient type of ownership since politicians have an indirect concern about profits, while pursuing their own political goals and potentially socially harmful objectives. Therefore they have a large proportion of control rights, with zero cash flow rights, since these will flow back to the tax-payer. Since this type of ownership is supposed to be different, it is taken as a separate group in the analysis part of this research.
4. Governance during crises
In times of crises, governance shifts are inevitable. Sometimes, the entire organizational structure is re-invented, and risk management becomes top priority. During the recent financial crisis, regulators and economists suggested that practicing good corporate governance would have prevented stock price drops. This research will try to see if decision making and governance of controlling shareholders have an impact on performance.
Aebi et al. (2012), conducted a study on risk management, governance and performance of financial firms during the 2007/2008 financial crisis. They find that in the United States, active risk management increased performance during the crisis. Gupta et al. (2013), argue the common view that good governance could have prevented the 2008 crisis, by researching stock prices and governance practices of a large global dataset during the crisis. They argue that investors rapidly changed their risky investments into safe investments which would mean that governance would not affect stock prices, and that well governed firms didn’t outperform poorly governed firms. This could mean that family ownership perform equal to
other controlling shareholders during the crisis. Lemmon & Lins (2003) empirically examine the relationship between ownership structure, governance and firm performance during the East Asian Financial crisis. They conclude that firms with pyramidal structures and more control than cash flow rights on average performed worse than firms who did not have these structures. Family firms are known for applying these structures (Barontini & Caprio 2006), and therefore I can investigate if this had any impact during the crisis. They state that ownership structure plays an important role in determining expropriation of minority shareholders. Johnson et al (2000) state: “Corporate governance can be of first-order
importance in determining the extent of macroeconomic problems in crisis situations”. They
try to explain exchange rate depreciation and stock market decline, by looking at governance variables during the Asian financial crisis. They state that it is possible that in countries with weak corporate governance, the chance of expropriating by managers is larger when there are bad economic prospects, reducing asset prices. Mitton (2002) conducted a study to determine the relationship between corporate governance and firm performance during the Asian financial crisis. He showed that during the Asian financial crisis, companies with good governance and concentrated ownership performed better than those with bad governance.
5. Research focus
In sum, majority ownership is a reason for differences in firm performance. Different ownership types yield different types of governance. During times of crisis, governance is said to be of major importance for a firms performance. This research will investigate if family owned firms outperform other types of ownership, and if there are differences during the 2008 mortgage crisis in Europe. In the next section the data and methodology used will be explained.
III. Data and Methodology
Firstly, this part begins by determining majority ownership. Majority ownership in this research is defined by Faccio & Lang (2002) and La Porta et. al. (1999). They state that a firm is considered majority owned when there is at least one single owner that controls more than 10 percent of the voting shares in another company. The reasons for this threshold is that it provides a significant amount of votes, and most countries require disclosure at the 10 percent cutoff point.
Second, the definition for family ownership is determined. In past literature, different definitions of family ownership are presented. Colli et al. (2003) describes a family firm
when “a family member is chief executive, there are at least two generations of family
control (and) a minimum of 5 percent of voting stock is held by the family or trust interest associated with it.” Miller and Le Breton-Miller (2003) define the family firm as “one in which a family has enough ownership to determine the composition of the board, where the CEO and at least one other executive is a family member, and where the intent is to pass the firm on to the next generation.” These are quite qualitative descriptions. To quantify family
ownership, I use the Faccio & Lang definition, which is defined as when one family owns more than 10 percent of the voting shares in another company. Later in this paper, the 20 percent threshold will be added to determine differences in ownership structure. This means that a single owner controls at least 20 percent of the voting shares in another company
Active family management will be taken as a separate variable among family firms, to determine if this increases firm performance. A firm is considered actively managed by a family when a founding family member holds the CEO, Honorary Chairman, Chairman, or Vice-Chairman position according to Faccio & Lang (2002).
When a company is owned by an unlisted company this is considered to be family owned and is in line with Faccio & Lang (2002). In their dataset they try to find the ultimate owner of unlisted firms, which is difficult to determine since unlisted firms are not required to disclose their owners. They explain that the ultimate owners of unlisted firms are not likely to be widely held, a corporation or institution, or the State, so most likely the owners of an unlisted firm are families. This is also in line with Claessens et. al. (2000), and therefore unlisted firms are considered family firms. In this research, ultimate family ownership and ultimate unlisted ownership are considered as one category; family ownership.
Sample and data
The ownership data in this sample is taken from Faccio and Lang’s (2002) sample of Western European firms with ultimate ownership data that is hand collected from different data sources, which contains 5.232 corporations. From this dataset, I distinguish the five largest European economies; the United Kingdom, Germany, France, Italy and Spain. The financial data of these firms is collected and updated from Worldscope and Datastream for the years 2002-2012. The base year is 2002 because the ownership data of Faccio & Lang (2002) is gathered before that year. This timespan gives the opportunity to investigate differences in performance during the financial crisis, which started in 2008. Most large shareholdings will remain equal over time according to La Porta et. al. (1999), therefore this timespan should not cause a bias. Financial entities (SIC 6000-6999) will be included later
on in the robustness part of this research. Regulated utilities (SIC 49) will be excluded from this research since these are difficult to compare with other industries due to regulations.
Definition of variables: Ownership variables
This research will use similar variables as in Maury (2006) and Barontini & Caprio (2006). The dummy variable Family_own will equal one if the largest controlling shareholder that holds at least 10% of the voting rights is a family member and zero otherwise, and is an indicator for family ownership. The second family variable will be Family_active. This variable will only be determined if the ultimate shareholder is a family and will equal one if a family member holds a top two management position, i.e. CEO, Honorary Chairman, Chairman, or Vice-Chairman, and will be zero otherwise. This variable measures active family control.
In former research, family ownership is related to non-family majority ownership. This research will create separate dummies for all types of ownership to see if these groups yield different results. These groups are according to La Porta et. al. (1999) and are: State, Financial institutions (Widely held financial), corporations (Widely held corporation) and Miscellaneous to measure the other forms of controlling shareholders. The Miscellaneous2 contains of cooperative groups, voting trusts, private equity funds etc., where there is no single controlling investor and equals one if all other groups are zero. Widely held financial and widely held corporations are taken together later in this research, since the separate groups are quite small, and the only difference of these owners is the type of industry in which they operate. These dummy variables equal one when the firm has a shareholder exceeding the 10% ownership threshold, and belongs to that controlling shareholder category, and zero otherwise. The last ownership dummy variable used will be widely_held, which will equal one if the firm does not have a shareholder that exceeds the 10% threshold. This variable is included to see if majority owned firms outperform diffusely held firms.
Governance variables
The variable Cashflow rights measures the proportion of cash flow rights held by the largest shareholder. This variable captures the effect of cash flow incentives on firm performance. To measure the entrenchment effect of excess control rights, the variable
Control minus cashflow will be constructed. This is the difference between cash flow and
control rights held by the largest shareholder.
Multipleblock will be a dummy variable that will equal one if there is another large
shareholder that exceeds the cut-off point of 10%. This variable is included to find differences in performance when there are multiple large shareholders controlling the firm. The Antidirector will be added to determine the legal classification of antidirector rights, according to La Porta et. al (1998), and can range from 0 (lowest), to 5 (highest), to determine the minority shareholder protection per country. The classification per country is as follows: the United Kingdom has Antidirector level 5, Germany 1, France 3, Italy 1 and Spain 4, with Germany and Italy having the lowest and the UK having the highest.
Control variables
There will be some additional control variables to avoid bias and multicollinearity. There are four firm specific control variables used in the model. Three year average growth in net sales will be used as a proxy for the firm’s growth opportunities. Capital expenditure over sales is a proxy for the firm’s investment intensity. There needs to be controlled for size, and leverage. These control variables will be calculated respectively by the natural logarithm of total assets and the amount of debt over total capital. Besides the firm specific control variables, country-, time- and industry fixed effects are included in the model.
Performance variables
As by rule, the performance indicator used will be the return on assets (ROA). In this research two different definitions of ROA will be used to show a clear picture of the effect of ownership on performance. The first ROA is similar to the one in Maury (2006), based on EBIT and is defined by:
( ) )))
The second one is according to Anderson & Reeb (2003), where Net Income is used and is computed as:
To indicate the valuation and future investment opportunities Tobin’s Q will be used. This is the market value of total assets divided by the replacement cost of total assets, and is defined as:
The regression formula looks as follows:
( ) ( ) )
) ) )
) ) )
) )
o Where Performance variables are ROA(EBIT), ROA (Net Income) or Tobin’s Q. o Family_own a dummy variable equaling one if the controlling shareholder is a family
with at least 10% of the outstanding voting shares
o Family_active a dummy variable equaling one when the firm is family owned, and a founding family member holds a top two management position
o Othercontrollingshareholders is a vector of dummy variables equaling one if the controlling shareholder owns more than 10% of the outstanding shares and is either a financial corporation, non-financial corporation, the State or miscellaneous.
o Widelyheld is a dummy variable equaling one when there is no majority shareholder that owns more than 10% of the voting rights, and zero otherwise.
o Cashflowcontrol variables is a vector composed by the share of ultimate cash-flow rights and the difference between cash-flow and control rights.
o Multipleblock and Antidirector determine the remaining governance aspects; Multipleblock equals one if there is another blockholder present, and Antidirector represents the Antidirector rights according to La Porta et al. (1998).
o Control variables is a vector composed by size (ln Total Assets), leverage (debt/total capital), sales growth (3-year average sales growth) and investment intensity (Capital Expenditures/Sales).
o The SIC code dummy captures industry fixed effects, the year dummy captures time fixed effects and the country dummy variable will capture the effects in each country.
Descriptive statistics:
In table I the ownership distribution per country is shown. The ownership type is measured at the 10% and 20% cutoff point, meaning that if the largest shareholder owns at least 10% (20%) of the outstanding shares, he is considered to belong to that particular type of ownership.
The first column is the weighted average of the ownership type of all the countries combined. In total there are 1189 firms individual . Family is with approximately 35% at the 10 percent threshold the largest category in the sample, while at the 20 percent threshold, widely held is the largest category with approximately 38,60%. Out of the 1189 firms, 35% is family owned, and 31% are unlisted companies. This means that there are 416 family firms in the sample and 374 unlisted companies. In this research unlisted companies and family ownership are considered to be one category of family ownership, therefore creating the largest ownership category. This is in line with former research like Maury (2006) and Faccio & Lang (2002) who state that unlisted ownership is considered to be family ownership.
France, Germany and Italy have large proportions of family ownership, which support the finding of La Porta et. al. (1999) that ownership is largely concentrated in the hands of families. There is a difference between these three countries and Spain, in which 14% of the total public firms are owned by families, and 53% unlisted firms. In the UK approximately 30% of the firms is family owned, and 20% owned by an unlisted firm. The difference between the Anglo-American type of firms (UK), and the Continental European type firms is shown in the results. These numbers support the findings of Faccio & Lang (2002) that ownership in Continental Europe is more family focused than in the UK. In Spain and the UK widely held and financial institution ownership is more common. In Italy a large fraction of the firms is owned by the State, while in the UK this is not very common.
When the threshold for majority ownership is raised to 20%, there are more widely held firms. The increase in widely held firms is caused by a decrease in family and widely held financial firms. This shows that families and non-family controlling shareholders, in general have approximately 10% of the voting rights. In the Continental European countries miscellaneous ownership is less common than in the UK.
Since the widely held corporation type is very small in the sample, and because the industry they are in is the only difference with widely held financial institutions, this group will be combined with the widely held financial institution shareholder group into the non-family shareholder group to get more significant results.
Table I Ownership per country
This table shows the ownership distribution per country in the sample . Panel A shows the 10% cutoff point, meaning that in the Faccio & Lang (2002) dataset the largest shareholder owns at least 10% of the voting rights, and belongs to that type of ownership. In panel B the 20% cutoff point is shown. Family; the largest shareholder having at least 10% (20%) of the voting rights is a family. Unlisted firms; the ultimate owner of a company is an unlisted firm. Widely held financial/corporation; the largest shareholder is a financial institution/non-financial corporation holding 10% (20%) of the voting rights, which is considered widely held at the threshold. Widely held; there is no shareholder that holds more than 10% (20%) of the outstanding voting rights. State; the largest shareholder that holds at least 10% (20%) of the voting rights in the firm is the State. Miscellaneous; different types i.e. venture capitalists, non-profit organizations etc. owning more than 10% (20%) of the firm. Crossholdings; the firm is held majority held through multiple chains.
Panel A: 10% cutoff
Ownership type
All
countries UK Spain Italy Germany France
Family 34.99% 28.70% 13.56% 59.72% 44.31% 38.16%
Unlisted firms 31.46% 21.57% 52.54% 22.22% 39.22% 45.18% Widely held financial 11.10% 16.87% 10.17% 2.78% 4.71% 6.58% Widely held corporation 1.09% 1.04% 3.39% 0,00% 0.39% 1.75%
Widely held 14.47% 24.52% 13.56% 1.39% 3.53% 5.70% State 2.61% 0.35% 5.08% 13.89% 4.71% 1.75% Miscellaneous 4.12% 6.96% 1.69% 0,00% 2.35% 0.88% Crossholdings 0.17% 0.00% 0.00% 0.00% 0.78% 0.00% Total 100% 100% 100% 100% 100% 100% Panel B: 20% cutoff Ownership type All
countries UK Spain Italy Germany France
Family 28.09% 17.04% 5.08% 59.72% 41.57% 36.84%
Unlisted firms 22.71% 10.09% 40.68% 19.44% 34.12% 38.16% Widely held financial 4.79% 5.57% 1.69% 2.78% 3.92% 5.26% Widely held corporation 1.51% 0.70% 1.69% 0,00% 2.35% 3.07%
Widely held 38.60% 64.17% 44.07% 4.17% 11.37% 14.04%
State 2.27% 0.17% 5.08% 13.89% 3.53% 1.75%
Miscellaneous 1.85% 2.26% 1.69% 0,00% 2.35% 0.88%
Crossholdings 0.17% 0.00% 0.00% 0.00% 0.78% 0.00%
In table II, Panel A the number of firms per country are shown for the sample used in this research. The largest amount of firms is in 2002, which consists out of 1172 firms. In total there are 1189 individual firms over the timespan. Every year the amount of firms listed is fluctuating due to going company death, privates, mergers & acquisitions, bankruptcy or delisting’s. The overall tendency is that every year the amount of firms in the sample decrease. The reason for this is that the companies used are from Faccio & Lang (2002), combined with fundamentals from Worldscope. The Faccio & Lang dataset was created before 2002, while there are no new companies entering the sample. This phenomenon happens because I would not have immediate access to new ownership data, and therefore the ownership data would be missing. Therefore the amount of firms decreases every year. This company “Death-Rate” will be investigated in the robustness section of this research, to see if there is a bias in the sample.
Furthermore the numbers show that approximately half of the firms in the sample are located in the United Kingdom. This is due to the fact that the UK is a large market with many listed firms. Germany is the second largest group, followed by France. Spain and Italy are the smallest country groups in the sample. Because of the fact that the sample consists for almost fifty percent out of UK companies, and the other fifty percent the Continental Europe countries, Germany, France, Spain and Italy, differences between the Anglo-American model, and the Continental European model can be analyzed.
In Table II, Panel B the Campbell (1996) SIC industry classification is used to determine the industry category per ownership type for the complete sample. In the first column the total amount of firms per sector is shown for all countries. The dataset is divided over the different industries with Consumer Durables as the largest industry in the dataset. Petroleum and Unregulated Utilities are the smallest industries. The table shows that the sectors are divided quite equally across the ownership types. Families are more often the final controlling shareholder in Construction, and less often the final controlling shareholder in Consumer Durables than other types of ownership. Non-family shareholder shows the numbers for ownership by other corporations, either widely held financial or widely held non-financial according to the La Porta (1999) classification of major shareholders. Non-family shareholders and the State have large stakes in the Consumer Durables industry. The table shows that Miscellaneous ownership often occurs in the Construction and Capital Goods industry.
Table II
Firms per country and industry
In Panel A the number of firms for the five largest European economies for the years 2002- 2012 are presented. In Panel B the industry classification per ownership category is presented. The industry classification is based on Campbell (1996), and is constructed as: Petroleum (SIC 13,29), Consumer durables (SIC 25, 30, 36, 37, 50, 55, 57), Basic industry (SIC 10, 12, 14, 24, 26, 28, 33), Food and tobacco (Sic 1, 2, 9, 20, 21, 54), Construction (SIC 15, 16, 17, 32, 52), Capital goods (SIC 34, 35, 38), Transportation (SIC 40, 41, 42, 44, 45, 47), Unregulated utilities (SIC 46, 48), Textiles and trade (SIC 22, 23, 31, 51, 53, 56, 59), Services (SIC 72, 73, 75, 76, 80, 82, 87, 89) and Leisure (SIC 27, 58, 70, 78, 79). This excludes financial utilities (SIC 60 until 69), and regulated utilities SIC (49). The Other category consists out of companies that do not fall under the other SIC industry classifications. The ownership categories are Family owned; the controlling family owns at least 10% of the voting rights; Non-family shareholder; a financial or non-financial firm owns at least 10% of the voting rights; Widely held; no controlling shareholder owning more than 10% of the voting shares; State; the State owns at least 10% of the voting shares and Miscellaneous; other types of controlling ownership like voting trusts or private equity funds own at least 10% of the voting shares.
Panel A
Year
All
countries France Germany Italy Spain UK
2002 1172 226 253 70 58 565 2003 1090 212 237 70 54 517 2004 1022 201 225 69 52 475 2005 947 187 213 64 49 434 2006 894 180 201 63 48 402 2007 827 168 194 57 44 364 2008 792 162 183 53 45 349 2009 759 158 176 51 45 329 2010 729 156 168 51 44 310 2011 698 153 159 48 43 295 2012 457 89 77 25 26 240 Panel B Industry Total number of firms Family owned Non-family owned Widely
held State Miscellaneous
Petroleum 29 2% 3% 4% 3% 2%
Consumer Durables 193 15% 18% 16% 29% 20%
Basic Industry 145 12% 10% 15% 10% 8%
Food and Tobacco 96 9% 8% 7% 3% 2%
Construction 112 10% 8% 6% 6% 14%
Capital goods 143 12% 10% 12% 6% 18%
Transportation 48 4% 3% 5% 10% 2%
Unregulated utilities 23 1% 3% 3% 13% 0%
Textiles and trade 124 10% 12% 10% 3% 10%
Services 147 13% 11% 11% 13% 14%
Leisure 91 8% 10% 8% 0% 6%
Other 38 3% 3% 4% 3% 2%
In table III the descriptive statistics are presented for the different variables. The different performance and control variables used in the analyses are winsorized at the 5% level to control for outliers. The average ROA (EBIT) in the sample is 4,21, the average ROA (Net Income) is 5,15 and the average Tobin’s Q is 1,37. As mentioned in table I, family represents approximately 66% of the sample, and out of the total family controlled firms, on average 37% is actively managed, which means one of the family members holds a top two management position in the firm. In the total sample 12% of firms is controlled by other non-family firms, while 14% of the firms is considered widely held. The average debt to equity ratio is 58%, and there are approximately three times as much control than cashflow rights in the sample. In 40% of the sample there are other large shareholders present at the 10% cutoff point. The average Antidirector level is 3,47, and is upward skewed due to the fact that the UK has Antidirector level 5, and is a large fraction of the sample.
Table IV reports the comparison of means using an independent t-test with equal variance. In Panel A, the average ROA (EBIT) for family firms in the sample is 2,83, the ROA (Net Income) is 3,46 and Tobin’s Q 1,26. Non-family firms have an average of 2,80 ROA (EBIT), 4,81 ROA (Net Income) and 1,20 (Tobin’s Q). Widely held firms have the highest ROA (EBIT), ROA (Net Income) and Tobin’s Q with respectively 3,38, 6,11 and 1,38. Miscellaneous ownership, has the lowest ROA (EBIT) and Tobin’s Q, while the State has the lowest ROA (Net Income). The table shows that widely held firms significantly have higher Tobin’s Q than all other types of ownership except for the State, which gives no significant result. Return on Assets based on EBIT does not give significant results in differences of means. Return on Assets based on Net Income shows that active family control is higher than passive family control, but that family ownership is lower than non-family ownership. The results show that widely held firms have higher ROA (Net Income) means than all other forms of control.
Family active control reduces the amount of other large blockholders in a firm. This could be explained by the fact that when a family actively manages the firm, they could expropriate other shareholders, and therefore other investors are less likely to invest in a firm where the owning family is also in the management. It is shown that family owned firms have more cash flow rights than non-family owned firms, but less cash flow compared to control rights. Firms where a family member holds a top two position is in general smaller than a firm which is passively managed by the family. Companies where the State is the controlling shareholder, are in general larger than with other types of shareholders.
Table III
Descriptive statistics of the variables
This table shows the summary statistics per variable used in this paper for the total sample over all the years at the 10% cutoff point. The variables shown are ROA:EBIT or Net Income divided by Total assets; Tobin’s Q; Family ownership, a dummy variable equaling one when at least 10% of the firms voting rights belong to one family; Family active, equals one if a family member holds a top two management position in a firm; Non-family ownership, another widely held (non)-financial company holds at least 10% in a firm; Widely held, equals one when there is no shareholder that owns more than 10% of the voting shares in the company; State; the State owns at least 10% of the voting shares and Miscellaneous; other types of controlling ownership like voting trusts or private equity funds own at least 10% of the voting shares, Cashflow rights, the percentage of cash flow rights the largest controlling shareholder possesses; Control minus cashflow rights, the difference between the cashflow and control rights held by the largest shareholder; Multipleblockholder, a dummy variable that equals one if there is more than one shareholders that has at least 10% of the shares; Size, the natural logarithm of total assets; Investment intensity, Capital expenditures over sales; Sales growth, three year average growth in net sales; Leverage, debt over equity in a firm; Antidirector rights, the countries legal classification which can range from 0 to 5, and are according to La Porta et. al (1998).
Variable Mean Median S.D. Min Max
ROA (EBIT) 4.21 4.70 6.54 -12.22 15.81 ROA(Net Income) 5.15 5.64 7.42 -12.13 18.29 Tobin's Q 1.37 1.20 0.56 0.74 2.84 Family 0.66 1.00 0.47 0.00 1.00 Family active 0.37 1.00 0.44 0.00 1.00 Non-family 0.12 0.00 0.33 0.00 1.00 Widely held 0.14 0.00 0.35 0.00 1.00 State 0.03 0.00 0.16 0.00 1.00 Miscellaneous 0.04 0.00 0.20 0.00 1.00 Cashflow rights 30.70 24.37 25.82 0.00 100.00
Control minus cashflow rights 3.07 0.00 7.51 0.00 50.00
Multipleblockholder 0.40 0.00 0.49 0.00 1.00
Size (LN total assets) 12.76 12.55 2.05 9.44 16.90
Leverage 0.58 0.59 0.19 0.23 0.92
Investment intensity 4.88 3.30 4.80 0.36 19.08
Sales growth 4.11 3.79 10.89 17.64 27.08
Table IV
Test of means
The table presents summary statistics for the 1189 firms in the five largest Western European countries by ownership. Panel A shows the summary statistics for the variables, Panel B shows the t-statistics for the comparison of means based on a two sample t-test with equal means, also before the crisis for the years 2002-2007 and after the crisis, the years 2008-2012; ROA:EBIT or Net Income divided by Total assets; Tobin’s Q; Family ownership, a dummy variable equaling one when at least 10% of the firms voting rights belong to one family; Family active, equals one if a family member holds a top two position in a firm when it is zero it is family non-active; Non-family ownership, another widely held (non)-financial company holds at least 10% in a firm; Widely held, equals one when there is no shareholder that owns more than 10% of the voting shares in the company; State; the State owns at least 10% of the voting shares and Miscellaneous; other types of controlling ownership like voting trusts or private equity funds own at least 10% of the voting shares, Cashflow rights, the percentage of cash flow rights the largest controlling shareholder possesses; Control minus cashflow rights, the difference between the cashflow and control rights held by the largest shareholder; Multipleblockholder, a dummy variable that equals one if there is more than one shareholders that has at least 10% of the shares; Size, the natural logarithm of total assets; Investment intensity, Capital expenditures over sales; Sales growth, three year average growth in net sales; Leverage, debt over equity in a firm; Antidirector rights, the countries legal classification which can range from 0 to 5, and are according to La Porta et. al (1998); N-is the number of separate observations of firms over the years. The multiple blockholder variable could not be identified for widely held firms and therefore those t-statistics are missing in the table.
*,**,***Significant at the 10% , 5% and 1% levels, respectively. Panel A Variable Family Family Active Family Passive Non-family Widely held State Miscel-laneous Total before crisis Total after crisis
Mean Mean Mean Mean Mean Mean Mean Mean Mean
ROA (EBIT) 2.83 3.25 2.59 2.80 3.38 1.58 1.46 4.23 4.18
ROA (Net Income) 3.46 4.14 3.05 4.81 6.11 2.42 4.19 5.02 5.37
Tobin's Q 1.26 1.22 1.28 1.20 1.38 1.32 1.18 1.43 1.28
Cashflowrights 39.34 42.05 37.74 21.52 0.00 40.09 20.64 30.55 29.78
Ctrl min cshflw 3.08 3.24 2.98 4.77 0.00 7.97 5.41 3.21 3.27
Mult. Blockh. 0.38 0.33 0.41 0.45 - 0.39 0.47 0.39 0.38
Size (Ln total assets) 12.23 11.89 12.44 12.28 12.81 14.39 12.19 12.57 13.08
Invest. Intens. 5.10 4.86 5.24 5.76 5.64 7.65 5.28 5.02 4.62
Sales growth 6.38 6.73 6.17 4.67 6.89 7.22 3.99 4.35 3.67
Leverage 0.59 0.56 0.60 0.56 0.56 0.68 0.57 0.58 0.57
Antidirector 3.09 3.05 3.12 4.26 4.57 1.81 4.27 3.40 3.37
Table IV (continued) Test of means Panel B Variable Family vs Non-family Active vs Passive Family vs
widely held Family vs state
Family vs Miscellaneous
t-stat t-stat t-stat t-stat t-stat
ROA (EBIT) 0.06 1.31 0.94 0.97 1.38
ROA (Net Income) 2.00** 1.93* 4.17*** 0.73 0.66
Tobin's Q 1.14 1.57 2.68*** 0.63 0.95
Cashflowrights 8.48*** 2.46** 21.59*** 0.17 5.44***
Control minus cashflow 2.40*** 0.47 5.27*** 3.37*** 2.12**
Multiple blockholders 1.5 2.39** - 0.05 1.25
Size (Ln total assets) 0.28 4.08*** 3.61*** 6.25*** 0.16
Investment intensity 1.44 1.05 1.30 2.77*** 0.26 Sales growth 1.64 0.66 0.51 0.38 1.43 Leverage 1.20 2.41** 1.67* 2.60*** 0.71 Antidirector rights 7.92*** 0.49 10.96*** 4.16*** 4.85*** Variable Non-family vs widely held State vs widely held Miscellaneous vs widely held Total: Before vs after crisis
t-stat t-stat t-stat t-stat
ROA (EBIT) 0.74 1.24 1.62 0.30
ROA (Net Income) 1.67* 2.57** 1.66* 2.21**
Tobin's Q 2.97*** 0.49 2.12** 12.06***
Cashflowrights 14.55*** 19.90*** 12.14*** 1.41
Control minus cashflow 7.35*** 8.04*** 11.31*** 0.36
Multiple blockholders - - - 0.45
Size (Ln total assets) 2.31** 3.73*** 1.92* 11.91***
Investment intensity 0.20 1.98** 0.43 3.81***
Sales growth 1.60 0.13 1.42 0.677***
Leverage 0.29 3.18*** 0.25 3.54***
Antidirector rights 2.27** 12.88*** 1.59 0.89
Before the crisis, years 2002-2007, the mean for ROA (EBIT) and Tobin’s Q is higher than after the crisis, while ROA (Net Income) is higher after 2008. This can be explained by the fact that the sample size decreases every year. Tobin’s Q is significantly higher before the crisis, while size is significantly larger after the crisis. The reason for this could be that Tobin’s Q is a measure of a firm’s investment opportunities. After the 2008 crisis, many firms were affected by lower profits, therefore resulting in lower investment opportunities. The size increases because this is the normal growth cycle of the firms, because the variable measured is the natural logarithm of the total assets.
Table V
Correlation matrix
The table presents correlation matrix for the different variable used. Variables shown are: ROA:EBIT or Net Income divided by Total assets; Tobin’s Q; Cashflow rights, the percentage of cash flow rights the largest controlling shareholder possesses; Control minus cashflow rights, the difference between the cashflow and control rights held by the largest shareholder; Multipleblockholder, a dummy variable that equals one if there is more than one shareholders that has at least 10% of the voting rights; Size, the natural logarithm of total assets; Investment intensity, Capital expenditures over sales; Sales growth, three year average growth in net sales; Leverage, debt over equity in a firm; Antidirector rights, the countries legal classification which can range from 0 to 5, and are according to La Porta et. al (1998).
ROA (EBIT) ROA (Net Income) Tobin's Q Cashflow rights Ctrl minus cf Multiple block. Size (Ln
assets) Inv. Intens.
Sales
growth Leverage Antidirector ROA (EBIT) 1.00
ROA (Net Income) 0.77 1.00
Tobin's Q 0.35 0.37 1.00 Cashflow rights -0.01 -0.10 0.02 1.00 Ctrl minus cf 0.01 -0.01 -0.02 -0.25 1.00 Multiple block. -0.04 -0.02 -0.02 -0.19 0.04 1.00 Size (ln Assets) 0.15 0.20 0.01 -0.07 0.14 -0.06 1.00 Inv. Intens. 0.03 -0.02 0.02 0.06 0.06 0.01 0.10 1.00 Sales growth 0.30 0.36 0.15 -0.02 -0.02 0.02 0.18 0.10 1.00 Leverage -0.26 -0.23 -0.01 -0.02 0.04 0.02 0.20 -0.10 -0.03 1.00 Antidirector 0.05 0.20 0.06 -0.35 -0.25 0.04 -0.22 -0.06 0.11 -0.15 1.00
Table V presents the correlation between the different variables. The table shows that most variables are not largely correlated with each other. Tobin’s Q and Return on Assets are positively correlated, but these variables will be used as dependent variables in the regression. Leverage is negatively correlated with Return on Assets. Which could be explained by the fact that when there is more leverage, the firms profitability becomes less. Sales growth is largely positively correlated with Tobin’s Q, which is true since Tobin’s Q is a measure of valuation and investment opportunities. The remaining variables have small correlation coefficients.
IV. Empirical results
To move on to the core of the paper, I examine the relationship between ownership and performance. In table V the first regression results are reported at the 10% cutoff point using a country fixed-effects specification with Huber-White sandwich estimators to control for heteroskedasticity, with N being the number of separate observations. The results show that active ownership has a higher relative profitability of approximately 20% for ROA (EBIT), and 16% for ROA (Net Income), calculated by Active ownership coefficient divided by the ROA (EBIT) or ROA (Net Income) mean for non-family ownership. The results also show that family ownership has an increasing effect on Tobin’s Q of approximately 5,25% over non-family ownership. (family ownership coefficient/ mean Tobin’s Q for non-family firms). Active ownership does not have a significant effect on Tobin’s Q. These results support the hypothesis that family ownership increases firm performance compared to non-family ownership, and that active ownership has a positive impact on profitability. Widely held also has a significant increasing effect on both profitability and valuation, with a 23,6% increase on ROA (EBIT), 22,6% ROA (Net Income), and 10,4% (Tobin’s Q) over family ownership. (Widely held coefficient for the performance variables/mean of family ownership for the performance variables). Therefore resulting in the fact that widely held firms have the highest effect on firm performance.
In table VI the regression results for the 20% cutoff point are shown. It shows that active ownership has a positive impact on ROA (EBIT) of 12,6%, and 11% on ROA (Net Income), but no significant effect on Tobin’s Q. It affects profitability less than when the large shareholder is determined at the 10% threshold. This means that when in actively managed firms the majority family shareholder becomes larger, this has a decreasing effect on Return on Assets. Non-family and Miscellaneous ownership, do not have a significant effect on profitability or Tobin’s Q while Widely held firms at the 20% cut-off point have a
positive effect on both ROA’s, but no significant effect on Tobin’s Q. The State has a significant negative effect on both ROA (Net Income) and Tobin’s Q, at both thresholds, since State ownership remains approximately equal over the different thresholds. This is in line with Shleifer & Vishny (1997), who say that State ownership is an inefficient type of ownership.
At both cutoff points it is shown that family ownership, and in particular active family ownership has a positive impact on firm performance compared to non-family ownership. The results also show that the amount of cash flow rights has little effect on the firms profitability and valuation. The control enhancing structures, the wedge between cash flow and control rights, do not have a significant effect on firm performance. More than one large blockholder, has a significant decreasing effect on ROA (EBIT) at the 10% threshold. Meaning that when there is more than one owner holding more than 10% of the voting rights, this harms firm performance. The control variables influence the dependent variables in different ways. Leverage and size have a large positive influence on Return on Assets. Return on Assets is a formula that includes components of the size and leverage ratio of the firm. Countries with higher Antidirector rights have higher performance than in countries with low Antidirector rights. Investment intensity has a significant negative effect on performance, while sales growth has a significant positive effect.
Table VI
Regression results
This table presents the results of the fixed effects regression with Huber-White sandwich estimators for the total sample of all the firms over all the years for the five largest Western European economies. Panel A reports the regression at the 10% cutoff point and Panel B reports the regression at the 20% cutoff point. Variables reported: ROA:EBIT or Net Income divided by Total assets; Tobin’s Q; Family_own, a dummy variable equaling one when at the threshold, the firm belongs to a family; Family_active, equals one if a family member holds a top two position in a firm when it is zero it is family non-active; Non-family ownership, another widely held (non)-financial company is the ultimate owner at the threshold; Widely held, equals one when there is no shareholder that owns more at the threshold of the voting shares in another company; State; the State is the ultimate owner at the threshold and Miscellaneous; other types of controlling ownership like voting trusts or private equity funds are the ultimate owner at the threshold, Cashflow rights, the percentage of cash flow rights the largest controlling shareholder possesses; Control minus cashflow rights, the difference between the cashflow and control rights held by the largest shareholder; Multipleblockholder, a dummy variable that equals one if there is more than one shareholders at the threshold; Size, the natural logarithm of total assets; Investment intensity, Capital expenditures over sales; Sales growth, three year average growth in net sales; Leverage, debt over equity in a firm; Antidirector rights, the countries legal classification which can range from 0 to 5, according to La Porta et. al (1998); Two-digit SIC codes included.
