1
Ownership concentration and firm performance during
pre- and post-crisis periods
s2559447
Daniel Lui
Faculty of Economics and Business, Rijksuniversiteit Groningen
MSc International Financial Management
Master Thesis
Supervisor: Prof. Dr. C.L.M Niels Hermes June 20th , 2014
Abstract
This study investigates the relationship between ownership concentration and firm performance in a four-year timeframe from 2007-2011. The whole sample size consists of 138 quoted non-financial firms operating in Germany. I find no significant effect of ownership concentration on any used performance indicators under investigation (e.g. Tobin’s q, Stock return, ROA, ROE). It is difficult to build any conclusions in respect to the possible emergence of minority expropriation effect by larger shareholders during the financial crisis. ___________________________________________________________________________ JEL classification: G30, G32, G34
2
1. Introduction
The relationship between ownership structure and firm performance has always been a central topic of discussion in the global financial literature. Attention increased after the publication of Berle and Means (1932). Their theory deals with the governance problem in public corporations where ownership (shareholder) and control (management) are not represented in form of one person, but are separated distinctively. This ownership dispersion is associated with a negative effect, namely the difficulty to exercise control over the corporation. Consequently, in cases of weak and ineffective outside supervision, managers may face incentives to misuse their own position in favor of individual interests rather than to maximize shareholder’s wealth. As the linked agency cost has a reducing effect on the firm performance, firm performance is inversely related to ownership dispersion. Since then, several empirical studies were published addressing exactly this relationship in more detail (Shleifer and Vishny, 1986; Mørck, Shleifer, and Vishny, 1988; McConnell and Servaes, 1990; Bebchuck and Weisbach, 2010).
Yet, this specific issue took a new point of view, showing also a negative relation between ownership and firm performance - due to the emergence of potential entrenchment or expropriation effects. This means, large shareholders may exploit the situation of dispersed ownership for their own benefits at the expense of small shareholders at a specific threshold, and thus leading to a decrease of firm performance. These two opposed hypotheses help to explain the bell-shaped relationship between ownership structure and firm performance which was examined in the studies by Mørck, Shleifer, and Vishny (1988) and McConnell and Servaes (1990).
3
According to Pearson and Clair (1998), the arising features of skepticism and ambiguity can have a significant impact on the firm performance as managers (increase of agency cost) and shareholders (increase of expropriation cost) have both the incentive to misuse the situation to their individual advantage. “One potential explanation for this finding is, that during a crisis
period, controlling shareholders’ incentives to expropriate minority shareholders tend to go up as the expected return on investment falls.” (Bae, Baek, Kang, and Liu, 2012, p.413).
Shleifer (2000) suggests in his book about behavioral finance that authorities play an important role in securing and guaranteeing the protection of investor’s rights and interests. They should minimize the probability of any action related to theft and expropriation by corporate insiders. Further, he states that these problems occur rarely in the United Kingdom (UK) and United States (US) and are more often observed in Continental Europe and Asia. The main reason is the relatively low number of strong and effective rules, codes or regulations prohibiting the existence of shareholder’s expropriation in any way. These facts open the research question whether the effect of shareholder’s expropriation over minority shareholders occurred during the financial crisis in 2008 and 2009 within Continental Europe, where Germany is analyzed as an example of a Continental European country. In case of an increase of expropriation effects during the financial crisis, the bell-shaped curve is assumed to shift towards the left direction.
This study makes a contribution to the corporate governance literature based on several reasons. Firstly, it deals with the question of how the financial and economic crisis in 2008/2009 had an impact on the corporate structure in a highly industrial European country. More precisely, it provides a better understanding of large shareholders’ attitude during the pre-crisis, crisis and post-crisis years. Are they eventually motivated to expropriate minority shareholder’s wealth during the crisis? Secondly, it provides empirical evidence in regard to the ongoing corporate governance debate about Continental European firms analyzing those operating in the German business environment. Most of the prior studies analyze firms from Anglo-Saxon countries and comparably less research is done on firms from Europe.
4
The remainder of this study is structured as follows. In the next section, the institutional environment in which this thesis is constructed is discussed in more detail. Section 3 briefly summarizes the recent literature about the relation between ownership structure and firm performance during a stable and a crisis time period including the discussion of the relevant theories. Section 4 describes the data set, the methodology and the descriptive statistics. It is followed by empirical analysis where the outcomes are firstly presented and secondly, further elaborated. The last section concludes this research by providing recommendations and implications for future studies as well as showing the limitations of the study at hand.
2. German Corporate Governance
Before the relevant theories are examined in more detail, an overview of the main aspects of German corporate governance1 is provided in order to enhance the understanding of the issue in discussion.
According to the theory of La Porta, Lopez-de-Silanes, Shleifer, and Vishny (1998), Germany, as a civil law country, is associated with a relatively low investor protection compared to Anglo-Saxon countries such as the UK or the US. In contrast to that, secured creditors are supposed to be well-protected in the context of the German legal system. Historically, bank finance dominates the German business landscape, which leads to the fact that Germany is regarded as a bank-based economy. A high number of outside investors providing corporations with new equity, which is prevalent in many Anglo-Saxon countries, is relatively rare and uncommon in Germany.
Therefore, the market for corporate control in terms of investor’s protection is less matured. The existence of ownership concentration is a so-called “salient feature of corporate
governance” (Lehmann and Weigand, 2000, p. 160). In general, German corporations have
the tendency to hold one large shareholder who exercises the overall control (La Porta et al., 2000). Families are usually the biggest blockholder of German firms (Franks and Mayer, 2001). As already mentioned, large joint stock corporations like in the US or the UK with a characteristically widely-spread ownership structure are rarely found. The numbers of equity-ownership is relatively low as those investors are less secured by the national legal system than in common-law countries (Lehmann and Weigand, 2000).
5
Generally, the German ownership structure is classically built in a complex pyramidal ownership structure2, industrial conglomerates or cross-shareholdings where banks and/or insurance companies are usually the ultimate owner (Franks and Mayer, 2001). The involvement of financial institutions as a direct shareholder favors firms to have a better and simpler access to debt capital, strengthens the lender borrower relation and reduces attempts of hostile takeover. Next to acting as a lender, banks are commonly represented in the Board of Directors of a company and hold so-called proxy votes which enable them to derive more influence on the firm not only through the lending relationship (Fohlin, 2005). In contrast, firms in a market-centered financial business environment more frequently use equity as a new finance tool, thereby putting the founders under pressure to give up control in favor of new capital, as well as increasing the minority shareholders’ protection (La Porta, Lopez-de-Silanes, and Shleifer, 1999).
The main reason for the difference in corporate governance between Germany and Anglo-Saxon countries lies in the diverging interpretation of an economic entity. In contrast to UK- or US-firms, German enterprises are seen as multi-functional institutions whose central target is to support the interest of diverse stakeholders such as shareholders, creditors, clients, employees and the society in which the company operates (Gelter, 2009). Compared to Germany, the declared goal of the Anglo-Saxon model is the maximization of shareholder’s wealth. (Fiss and Zajac, 2004). As a consequence of these different targets, shareholders in the German business context e.g. banks, families, governments or other corporations are not only focusing on wealth maximization as it is common in the UK or US. Nevertheless, the increasing competitive pressure in international capital markets, the change of investor demand in favor of a shareholder-value orientation and the pressure of new legislations forces German corporations to adopt the shareholder-value orientation, and thus the characteristic ownership concentration of German firms of the past will disappear slowly in the following years (Deakin, 2005). This trend is mainly driven by the German government trying to abolish the ‘Deutschland AG’, which stands for the typical close ties between inter-sectorial as well as intra-sectorial companies and financial institutions – the so called policy of “Entflechtung” (divestiture) (Klein and Blondel, 2002; Goergen, Manjon, and Renneboog, 2008; Krempel 2008). They are aiming to eliminate the opaqueness and non-transparency of their firm structure, as it has a negative impact on the firm value (Beyer, 2003, Hansmann and Kraakman, 2000). This trend is supported by the results of a recent empirical study conducted
6
by Frønningen and Van der Wijst (2009) showing the emergence of widely held ownership in German corporations.
This current development fosters the interest in examining the German company landscape in respect to the ongoing change of ownership structure in relation to the firm performance. However, the process is still at the very beginning and will take a long time in order to be changed completely.
3. Review of the relevant theory
The following part provides a theoretical overview of the main theories associated to the relationship between ownership structure and firm performance.
Since the main study by Shleifer and Vishny (1986) it is commonly recognized that ownership structure is an important aspect of corporate governance. This topic evolved to attract great interest in the international corporate finance literature, whereby two opposed theories are applied for the explanation of the relationship between ownership structure and firm performance, namely the agency theory and the theory of expropriating minority shareholders.
3.1. Agency theory
7
theory as empire-building (Jensen, 1986). This can eventually lead to higher manager compensations.
Another closely related topic in the context of the agency theory is the free cash flow hypothesis, first mentioned by Jensen (1986). He states that managers with a high level of free cash flow are more likely to invest this capital in other operations or even in negative net present value projects instead of paying this amount out to shareholders in order to increase the amount of potential perquisites.
By monitoring manager’s actions, above mentioned damaging effects can be reduced efficiently. Hence, shareholders strive to implement corporate governance mechanism in order to lower the risk of agency cost, e.g. by increasing the board size (Adams and Ferreira, 2007) or by raising the ownership concentration (Roe, 2003). Monitoring in this respect does not solely mean to observe the manager’s activities, but it also includes to actively control their operations through the implementation of rules, policies and guidelines (Jensen and Meckling, 1976). While all those points lead to a higher monitoring effect over the firm’s management, it is simultaneously costly and time-consuming for shareholders. In the case of minority shareholders, the advantages of monitoring (a decrease of agency costs, and thus an increase of firm performance) usually do not outweigh the associated monitoring cost. Furthermore, small investors do not have enough power through votes in order to force and discipline the management to act in favor of their interests.
However, large shareholders are better able to trade off the benefits of monitoring against the incurred costs. As a consequence, only a small group of shareholders, namely those with high percentages of votes, have the incentive to monitor management actions, since a large amount of their investment is directly affected by managerial actions.
8
depends on the firm performance process, or respectively on the shareholder’s wealth maximization.
When ownership concentration is increased, an additional side effect can be observed: the reduction of free rider conflicts inside the corporation in respect to the stock market. This argument suggests that firms with a dispersed ownership structure are more likely to be confronted with free rider problem as the likelihood of firm asset exploitation grows the more managers work together (Grossman and Hart, 1980; Shleifer and Vishny, 1986, 1997).
Summarizing, ownership concentration is associated with a stronger governing effect by individual large shareholders through a reduction of information asymmetry and agency costs between owners and managers as well as between the firm and external investors. As monitoring is cost-intensive, only large shareholders have the incentive to monitor. Therefore ownership concentration can be positively related to firm performance.
3.2. The theory of expropriating minority shareholders
With the introduction of the expropriation argument in the study of Mørck, Shleifer, and Vishny (1988), a new perspective on the relation between ownership and firm performance was created. It states that more equity ownership results in a decrease of firm performance because shareholders with large ownership stake might be so powerful to ignore the interests of the minority shareholders. The dominant position of large shareholders results in exploitation of firm resources in favor of private benefits which is to the detriment of the firm’s value.
With a high ownership concentration the control over the organization is accumulated through a small number of large shareholders. This condition facilitates large shareholders to benefit from dealing, since fewer investors prevent them from conducting actions related to self-enrichment. Self-dealing means, for instance, to transfer firm’s profits directly to their own possession instead of sharing it with all other investors. This action is known as tunneling. The pyramidal ownership structure helps larger shareholders to tunnel firm resources from one firm to another closely affiliated firm (Wolfenzon, 1999). Moreover, the expropriation effect occurs more often in countries with a relatively low legal protection of minority shareholders since large shareholders do not fear any consequences under corporate and civil law.
9
expropriation of minority shareholders is through newly implemented corporate policies, e.g., changing the payout policy in favor of larger shareholder’s financial advantage, or transfer the corporate profits to another closely related company, which belongs to the same large shareholder. Another way of expropriation conducted by large shareholders can also be through the pursuit of non-maximizing targets instead of targeting shareholder wealth maximization. This can finally have an influence on the firm’s growth as well as the firm value.
Despite of the fact that most of the existing studies analyze the expropriation effect in Asian countries (Claessens, Djankov, Fan, and Lang, 1999, 2000; Claessens, Djankov, and Lang, 2000; Johnson, Boone, Breach, and Friedman, 2000), examples can also be found in the academic literature examining these effects in well-developed countries like Sweden (Bergstrom and Rydqvist, 1990), the US (Barclay and Holderness, 1989) and Italy (Zingales, 1994). As the majority of German corporations have large pyramidal ownership structures and/or cross-shareholdings, the issue of expropriation can be regarded as a potentially existing feature in the context of German shareholders. In addition the effect of expropriation is further supported by a weaker shareholder protection in Germany compared to the US or UK.
To sum up, due to the self-dealing action of large shareholders, ownership concentration can be negatively related to firm performance.
3.3. Causal relation
While agency theory indicates a positive relationship between firm performance and ownership structure, the expropriation theory provides arguments for a negative relationship. Hence, several scholars suggest that this specific relation is curvilinear. According to Demsetz and Villalonga (2001), ownership concentration at a low level increases firm performance but after exceeding a specific threshold it has adverse effects, meaning at some point the disadvantages of ownership concentration outweigh the benefits, ultimately leading to a decrease of firm value. Or put in the words of Shleifer and Vishny - “large investors
represent their own interests, which need not coincide with the interests of other investors in the firm, or with the interests of employees or managers. In the process of using his control rights to maximize welfare, the large investor can therefore redistribute wealth - in both efficient and inefficient ways - from others” (Shleifer and Vishny, 1997, p. 758).
10
ratio firm performance improves because of shareholder’s action in terms of reducing agency costs by better monitoring the management. However, at a certain point, the more the shareholder holds, the more likely it is that the effect of exploitation in favor of private benefits dominates (Point C). This leads to a decrease of firm performance. The breakpoint (Point B) where the relationship changes from positive to negative specifies the optimal ownership concentration which should be held in order to maximize firm performance.
Graph 1
Graphical representation of the non-linear relation between ownership concentration and firm performance.
3.4. Ownership structure and firm performance in Germany - empirical evidence
The topic of ownership structure in Germany was not investigated as intensively as in Anglo-Saxon countries. Therefore, relatively less empirical evidence can be found for any kind of ownership concentration in Germany. However, the few studies in the context of German corporate governance suggest different results that show either positive, negative or even no relationship at all.
One of the first analyses about the relationship between firm performance and ownership structure is conducted by Gedajlovic and Shapiro (1998). They find a significant negative and non-linear relation, whereby the sample size consists of major firms from the five largest industrialized countries (including 99 German firms) in the time-span from 1986 to 1991. Edwards and Weichenrieder (2004) conclude that in most cases the advantages of ownership concentration (a greater monitoring and bonding effect, and thus a reduction of agency problems) equals or even exceeds the negative aspects of ownership concentration (self-dealing and tunneling). Other studies (Edwards and Nibler, 2000; Kaserer and Moldenhauer,
11
2008) strengthen this argumentation by empirically showing a positive and significant relationship between firm performance (measured in two different forms: Tobin’s q and stock price) and insider ownership. Further, Edwards and Weichenrieder (2009) analyze the relationship between ownership structure and firm performance in 102 listed German firms. Examining how firm performance (Tobin’s q) relates on the one hand to voting and on the other hand to cash flow rights of the largest and second-largest shareholders. They conclude that the largest shareholders mistreat their position at the expense of smaller shareholders, indicating the existence of the expropriation effect within the German business environment. Another academic paper by Lehmann and Weigand (2000) examine this topic on the basis of 361 German corporations between 1991 and 1996, concluding that ownership concentration has a significant negative effect on profits.
Despite of all those studies, it should be considered that the outcomes of the diverse academic works are based on different periods of time, industries, available data, statistical tools and sample sizes. Being more specific in respect to the issue of different time periods, Lehmann and Weigand (2000) propose that the relation between firm performance and ownership structure might have changed over time due to an economic transition process, for instance by the opening of the market, and thus an increase of competition. They argue that a significant positive relationship has been found for German governed firms during the 1970s and early 1980s, as is indicated by the studies of Cable (1985), Schmid (1996) and Gorton and Schmid (2000). Then, this relation might have changed in the beginning of 1980s and continues to maintain until now; shown by the academic work of Becht (1999) and Bebchuk (1999). Concluding, the literature provides an inconclusive picture of the relationship between ownership structure and firm performance in German firms.
3.5. The influence of financial crisis in regard to the relation between ownership structure and firm performance
12
losses in respect to their investments increased continuously. Firms had to struggle with the difficult condition of accessing new external funds in order to continue with their own business. Companies experienced a great financial loss during that time and were challenged with a tough and severe national and international business climate. Germany, and hereby especially the manufacturing industry as well as the financial institutions had to suffer from the repercussions of that crisis (Illing, 2013). The turmoil of the financial crisis created a high level of uncertainty, ambiguity and economic instability, which threatened shareholders to invest in German corporations.
Investigating how an organization adjusts the ownership structure in the context of corporate governance to sudden changes within the national economy received comparably low attention. Most of the recent studies investigate the dynamics of corporate governance in response to macro-economic aspects of deregulation (Lehn, 2002; Ovtchinnikov, 2010) or in regard of the East Asian financial crisis (Johnson et al., 2000; Mitton, 2002; Lemmon and Lins, 2003; Baek, Kang, and Park, 2004).
The general findings of these studies suggest that the economic crisis propels large shareholders to expropriate minority shareholders because of a high level of uncertainty of shareholders, expected losses and the anxiety about the firm’s future. Conversely, that means that large shareholders reduce management monitoring. So, since a crisis leads to a significant decline in firm’s future profits and investment opportunities, shareholders tend to divert, for instance, corporate assets for their own benefits at the expense of minority shareholders (Bae, Baek, Kang, and Liu, 2012). So, the sudden economic shock associated with the crisis triggers shareholders to protect their investments from potential losses. They are confronted with a high level of fear, worry and anxiety about the uncertain future and with the open question whether the firm will easily recover from such a negative situation.
The separation of cash flow and control rights leads to the fact that shareholders can better exercise a stronger controlling power (control rights) over the corporate decision-making relative to their hold share size (cash flow right). This separation typically arises in firms with a pyramidal or cross-holding ownership structure (Bebchuk, Kraakman, and Triantis, 2000). In other words, the concentrated and pyramidal ownership structure allows shareholders to “commit low equity investment while maintaining tight control of the firm, creating a
13 shareholders.” (Lee, 2007, p. 394). Lemmon and Lins (2003) state that firms with a larger
discrepancy between cash flow and voting rights experience a lower equity value during financial crisis suggesting that the incentives of large shareholders to expropriate minority shareholders is the main reason for the weak firm performance during crisis times.
According to Friedman, Johnson, and Mitton (2003) and Johnson et al. (2000), the main explanation for the occurrence of expropriation effects by large shareholders during times of crisis is based on the fact that the return on investment drops. They argue that the sudden shock in regard to the returns reduces the marginal cost to shareholders to divert the firm resources away from possible profitable investment projects. As a consequence, the expected degree of expropriation increases.
Another possible reason deals with the fact that large shareholders were not provided with full information whether or not their investments were deployed correctly. The crisis as an exogenous sudden shock might have discovered these existing weaknesses in the corporate governance structure and triggered shareholder’s awareness about this specific problem. This risen awareness resulted in investor’s tunneling or self-dealing behavior.
Most of these studies investigate East Asian countries, mainly focusing on companies from South Korea. Furthermore, the findings and implications drawn from the analysis of East Asian countries were tested during the Argentinian crisis in 2001 with the same results of shareholder’s expropriation to minority shareholders (Bae, Baek, Kang, and Liu, 2012). However, there exist some similarities between Germany and those countries regarding corporate governance. For instance, the phenomenon of pyramidal ownership structure and cross-shareholdings (known, for instance, in South-Korea as chaebols), large ownership concentration, a bank-based financial system and a weak legal protection of minority shareholders compared to Anglo-Saxon countries (Baek, Kang, and Park, 2004; Rosser, 2005; Shleifer, 2000). Thus, it is not unreasonable to think that the recent crisis may have caused the same expropriation effects in Germany that it has caused in Asian corporations before.
4. Hypothesis development
14 P‘‘ X expropriation effect Firm performance Ownership concentration monitoring effect P X P‘ X A‘ A A‘‘
structure and firm performance is assumed to be non-linear. This means that the convergence of interest effect dominates at a lower level of ownership concentration until a specific threshold. After that point, the expropriation effect dominates at a high level of ownership concentration.
In times of crisis and economic instability, larger shareholders tend to expropriate firm resources sooner than during stable conditions. Graph 2 below illustrates the bell-shaped curve already explained in section 3, but introduces an expected left-movement towards the y-axis. Graph A demonstrates the curve-linear relationship between firm performance and ownership structure in a normal situation - in this study understood as the pre-crisis year 2007. Graph A’ shows the same relationship during the financial crisis 2008 and 2009. Since the action of corporate asset exploitation prevails over monitoring, it is expected to observe a significant shift to the left. The breakpoint P from graph A moves too, and becomes P’ for the years of the financial crisis. Thus, P’ is smaller than P. Graph A’’ displays the relationship, in case of P’’ becoming greater than P, indicating an increased monitoring effect. However, this is not in line with the previous stated theory because during the recent financial crisis, the aspect of expropriation of large shareholders over the minority shareholders is believed to dominate the monitoring effect.
The research hypothesis is formulated as follows:
Hypothesis: The financial crisis has a significant influence on the bell-shaped relation
between ownership structure and firm performance.
Graph 2
15
5. Methodology and Data
In this study, the shareholder structure of German-listed corporations in relation to the firm performance is examined over a four-year time period. The data availability in respect to the ownership structure in Germany is very limited because of historical and legislative reasons3, therefore the possibility to build a complex and complete panel data model over a long period of time is restricted. Since this study focuses on the change of the maximum point of the bell-shaped curve as a consequence of the recent crisis, the data consists of corporate information on a yearly basis from 2007 - 2010.
The first part consists of a comprehensive presentation of the descriptive statistics, supported by a comparison of the results with those originating from previous academic studies. Subsequently, the outcomes of the Ordinary Least Squares (OLS) –regression estimation are analyzed in depth in order to expand the knowledge of the special relation between ownership concentration as a corporate governance instrument in Germany and firm performance.
5.1. Data set – Sample selection
The sample selection is based on the top German corporations (legal form: Aktiengesellschaft - AG) listed at the main German stock Exchange, namely the ‘Frankfurter Wertpapierbörse’ (FWB). In this particular case, “top” is understood as those corporations with the highest market capitalization of common equity in the end of 2010. Due to various incomplete and missing values in regard to the ownership structure and control variables, the observed sample consists in total of 138 different firms operating in diverse industries such as Textiles, wearing apparel, leather (5); Food, beverages, tobacco (5); Machinery, equipment, furniture, recycling (48); Metals & metal products (6), Wholesale & retail trade (19); Publishing, printing (7); Chemicals, rubber, plastics, non-metallic products (24); Construction (3); Transport (3); Post & telecommunications (6); Gas, Water, Electricity (9); Primary sector (1) and Wood, cork, paper (2). Banks and insurance companies were purposely excluded from the sample size because the issue of corporate governance as well as the financial structure of these institutions differs from non-financial corporations (Adams and Mehran, 2003). Another reason is based on the fact of regulations. According to Demsetz and Lehn (1985), the financial sector operates in a more regulated environment than other industries. Regulations can influence the management and shareholder actions concerning the question what can be
16
done with the resources owned by the firm. Because the above-mentioned aspects can have an impact on the ownership structure compared to firms operating in a non-regulated industry, financial organizations have been excluded from the analysis.
The relevant data is collected through three different databases. The reciprocal network of ownership and the coherent interdependency complicates the issue of finding the ultimate owners of each single corporation. Thus, the commercial database called ‘Hoppenstedt Aktienführer’ is taken aside in order to receive the relevant voting rights of each individual firm. The database consists of specific financial and corporate information as well as a detailed overview of the ultimate owners for all listed German firms.
Accounting data for the regression variables are mainly extracted through Orbis which is provided by Bureau van Dijk and contains annual report data from the last ten years of each international listed corporation. In case of missing or inconsistent data, the respective information is investigated manually through the annual report uploaded on the individual firm homepage and added in the data set. As German listed corporations are obliged by law to electronically publish their annual report and financial statements, the Federal Gazette ‘Bundesanzeiger’ is used in specific cases to recheck the figures generated from Orbis. The data for the market-based performance-indicator (stock return) is extracted from Datastream – Worldscope.
To analyze the interrelationship between ownership structure and firm performance on the basis of German listed companies for the years 2007-2010, the following quadratic regression model is used:
Regression model:
where is the constant term, is the coefficient, is the error term, determines the firm and determines the year. The given model permits us to analyze whether the dependent variable (Firm performance) is influenced by the variable ownership concentration (OC). It analyses a curve-linear relationship of both variables.
17
applied in the study of Gugler and Weigand (2003) who have also examined the performance-ownership relation in the German economy.
By including the OC as well as its squared value in the model, it is possible to specifically examine the existence of any monitoring or exploitation effect. In order to determine the breakpoint, where the positive relationship between firm performance and ownership structure turns from positive to negative (indicating the optimum equilibrium of ownership concentration firm performance) the model’s partial derivative is set equal zero. This leads to the following breakpoint formula: (
).
Since, the variable OC is always positive, and must have opposed signs. Further, the breakpoint is expected to be the maximum of the bell-shaped curve. To achieve this, the numerator (in this case ) must be greater than zero, and thus the denominator (in this case ) must be less than zero. In the contrary case, when < 0 and > 0, the relationship would indicate a U-shaped curve, which is not supported by the theory. Summarizing, and must have opposite signs whereby > 0 and < 0. This condition must be fulfilled for a normal bell-shaped curve which is assumed in this research.
5.2. Definition of Variables
5.2.1. Dependent variable
Four different proxies of firm performance are used in this study: ROA (1), ROE (2), Tobin’s q (3) and annual stock returns (4). Those four ratios are the most prominent for evaluating performance in the context of corporate governance.
The return on assets (ROA) and the return on total equity (ROE) both have an accounting background. Compared to the return on equity (ROE), ROA is not influenced by the different cost of capital of each firm. ROA has a greater explanation power regarding the operating performance compared to other accounting figures such as ROE. In contrast to ROE, total assets cannot be lower than zero. However, ROE answers the question of how effectively the corporation uses the shareholder’s equity. Both ratios provide a better understanding of firm performance regarding the past. However, these performance indicators also exhibit some drawbacks, for instance both financial accounting ratios can change over time because of new adapted tax laws and accounting regulations and beyond, have a backward view.
18
Tobin’s q is the market value of a firm’s assets divided by their replacement value. Annual stock return is the annual change of the stock price adjusted for the dividend payments at the end of year t.4
5.2.2. Independent variable
The collection of data is explicitly focused on the specific composition of the ultimate ownership in each individual firm regardless whether that firm is built through pyramids / cross-holdings or not. In case of a pyramidal or cross-holdings ownership structure, the target is to find the ultimate shareholder of the whole network. The search is confined to the three largest shareholders of each corporation who hold at least five percent of the total voting rights. The threshold is fixed by this number because the German Securities Trading Act (Wertpapierhandelsgesetz - WpHG) defines a major shareholder as someone who holds at least five percent of the voting rights. The five percent limit is also in line with the recent study of Frønningen and Van der Wijst (2009). In case of having less than three large shareholders, the sum of the voting rights of the remaining large shareholders is used. For instance, assuming to have only two shareholders holding at least more than five percent of the voting rights, the sum of both investors is taken. If only one shareholder holds more than five percent of the voting rights, this exact percentage is used for the further regression.
The approach of measuring ownership concentration by taking the three largest shareholders is identical to the academic works of Demsetz and Lehn (1985), Mørck, Shleifer, and Vishny (1989) and La Porta et al. (1998). Further, several studies dealing with the issue of ownership structure of German firms conclude that the majority of the single-shareholders hold on average more than 50% of voting rights (Franks and Mayer, 2001; Edwards and Nibler, 2000). So, by taking the three largest shareholders it is assured that the main important investors are considered in the sample.
Another approach in the context of measuring the expropriation incentives of large shareholders deals with the divergence of cash-flow and voting rights. In the case of an owner holding a high number of control rights, the owner’s ability to exploit firm resources for private gains increases. But, this action can simultaneously reduce the amount of dividends. Moreover, if this owner has a small number of cash-flow rights, the costs of transferring the firm resources are lower. Hence, the disparity between control and cash-flow rights provides a better understanding of the ability as well as the incentives to expropriate corporate resources.
4
19
Further, as mentioned above, controlling shareholders are less likely to take benefit from minority shareholders if the ratio of cash flow rights is higher than the percentage of shareholder’s control rights (Claessens et al., 1999; Claessens, Djankov, and Lang, 2000). However, this study focuses only on voting rights and does not include any cash flow rights. Main reason for that is based on the difficulty to generate this specific data from German corporations. Normally cash-flow rights are calculated on the basis of dividends and with the use of a formula derived, for example, by Brioschi, Buzzacchi, and Colombo (1989) and Flath (1992). This estimation requires considerable efforts to come to the exact figure and might be hampered by the large number of interconnections among German corporations in respect to the dividend payments.
Nevertheless, it should be mentioned that all figures regarding ownership concentration are not weighted. In other words, systematic differences of the size of each corporation are not respected. Considering any size effects, smaller businesses will presumably have a higher ownership concentration compared to larger businesses which is confirmed by the studies of Himmelberg, Hubbard, and Palia (1999).
5.2.3. Control variable
Several control variables are included in the regression analysis in order to strengthen the validity of the conducted research. In accordance with recent literature, four important variables are used, namely liquidity, financial leverage, firm size and firm age.
The first introduced control variable is liquidity (LIQUIDITY) reflecting the investment opportunity associated to the firm. It is the ratio of liquid assets to short-term liabilities. Further, according to Lie (2000) it may also play an important role in curbing the potential overinvestment by corporate managers.
Following Jensen (1989) and Kim and Sorensen (1986), the internal debt structure of a firm can provoke the existence of potential agency problems. Moreover, both scholars showed in their respective studies a positive relation between leverage and firm performance, because a high financial leverage status encourages debt holders to monitor the agent, and thus reduces agency costs. Hence, the financial leverage (LEVERAGE) is introduced as the second control variable, dividing total debt by the firm’s total assets.
Firm size (SIZE) is measured as the natural logarithm of total assets and is added in the
20
scale and an increase of market power. In addition, Haniffa and Hudaib (2006) argue that firm size is related to firm performance as large firms have a higher degree of financial resources to weather times of financial and economic crisis, changes and transitions.
Firm age (AGE) is the natural logarithm of the years listed, starting with the IPO until the
year 2007 respectively 2008, 2009 and 2010. It is assumed that age has a positive relation to firm performance, since the longer the firm is listed the more information is available on the market and to the outside investment world. This might attract new investors, leading to a higher cash balance, and thus to a higher firm value and better firm performance. Further, industry dummies are in some cases included in order to control special industry features which might influence the corporate performance.
5.3. Descriptive statistics
Table 1 depicts the descriptive statistics of the whole sample group per variable. Almost all variables show a great divergence between the minimum and maximum value. This is most probably due to the crisis’ effects. In consideration of the existence of several outliers (especially in regard to the performance variables during the crisis) which might significantly influence the regression results, all continuous variables in the entire sample group are winsorized at a 2% - 98% level.
Further, the table shows that the largest three shareholders hold on average 51.7% of the total voting shares. This number is close to the ratio examined by La Porta, Lopez‐de‐Silanes, and Shleifer (1999) in regard to German firms. The mean of their study is 48% and the median 50%. So, the German ownership structure of the corporations under investigation is mostly concentrated. Recent studies have argued that the Anglo-Saxon model of widely held corporations will become more dominant throughout the German business environment in the near future. The study at hand does not confirm this trend.
Moreover, Tobin’s q as one of the dependent variables measuring firm performance is on average 1.489. The other firm performance proxies are around 10.411% (Annual Stock Return), 3.926% (ROA) and 9.758% (ROE). The average liquidity ratio over the four year-timespan is 1.161. The analyzed German corporations have an average leverage ratio of 68.1%. The firm size (calculated as the natural logarithm of total assets) is 21.166 and the average listed years of German corporations is 2.290 (in natural logarithm).5
21
Table 1
Summary statistics.
The sample consists of non-financial German corporations listed on the German Stock Exchange during 2007-2010. The summary statistics are the values at the end of each year. Tobin’s q is the sum of the firm’s market value and the book value of liabilities divided by the book value of total assets. ROA is the Return on Assets ratio. Stock return is the annual change of the stock prices of the present year to the previous year. ROE is the Return on Equity ratio. Ownership concentration (OC) is the sum of direct equity ownership of the three largest controlling shareholders obtained directly and indirectly along the chain of the firm’s ownership structure. OC_2 is the square of OC. All variables are win-sorized at bottom and top 2 percentile to control potential outliers.
Variables N Mean Median Maximum Minimum Std. Dev.
Tobin’s q 552 1.489 1.227 3.820 0.750 0.604
Stock Return (percent) 552 10.411 6.779 153.990 -81.889 46.688
ROA (percent) 552 3.926 3.699 25.070 -24.390 6.887 ROE (percent) 552 9.758 10.597 68.903 -68.879 19.710 OC 552 0.517 0.538 1.000 0.000 0.286 OC_2 552 0.350 0.289 1.000 0.000 0.301 LIQUIDITY 552 1.161 0.981 4.393 0.225 0.701 LEVERAGE 552 0.681 0.630 0.970 0.242 0.152 FIRM SIZE (LN) 552 21.166 20.700 25.629 18.733 1.826 AGE (LN) 552 2.290 2.303 4.431 0.693 0.693
Table 2 shows the distribution of ownership concentration in respect to different industries and years. In general, the ownership concentration declines in almost all industries from 2007 to 2010 by 5.49%. By conducting a paired difference t-test, the change of ownership concentration between 2007 and 2010 becomes even statistically significant at the 10 percent level. It indicates that the mean average of both sample sets differs significantly. So, over the course of time, the biggest German listed companies become relatively less concentrated implying that the number of shares held by the main largest shareholders of each firm reduces significantly. One possible reason is that more minority investors enter the corporate ownership structure and hold just a small number of shares (less than 5 percent).
22
should be mentioned that the number of companies of each these industries is comparably low, 5 and 9. So, any further interpretations related to these data should be treated with caution. Moreover, three different industry sectors overweigh the entire sample, namely Machinery, equipment, furniture, recycling (34.78%), Chemicals, rubber, plastics, non-metallic products (17.39%) and Wholesale & retail trade (13.77%). In consideration of the subsequent analysis, the predominance of those sectors with each respective characteristic industrial features can be seen as a disturbance factor for later interpretative approaches. Thus, the weighted average based on number of companies of given industries is added in Table 2. This number declines during the time-span like the unweighted average.
Table 2
Ownership concentration by years and industries.
This table summarizes the arithmetic mean ownership concentration for 13 different industry sectors in each single year. Last two rows show the total arithmetic mean, both unweighted and weighted based on the amount of companies of each given industry. Last column indicates the percentage change of ownership concentration from 2007 to 2010. A paired sample t-test is used to compare the two related means of 2007 and 2010. ***, ** and * represent the 1%, 5% and 10 % significance levels, respectively.
Year Industry
2007 2008 2009 2010 % change
2007 - 2010
1. Textiles, wearing apparel, leather (5) 0.585 0.567 0.490 0.444 -24.103
2. Food, beverages, tobacco (5) 0.773 0.784 0.693 0.735 -4.916
3. Machinery, equipment, furniture,
recycling (48) 0.461 0.471 0.477 0.442 -4.121
4. Metals & metal products (6) 0.591 0.601 0.558 0.628 6.261
5. Wholesale & retail trade (19) 0.527 0.543 0.542 0.537 1.898
6. Publishing, printing (7) 0.535 0.479 0.464 0.413 -22.804
7. Chemicals, rubber, plastics, non-
metallic products (24) 0.487 0.524 0.489 0.497 2.053
8. Construction (3) 0.637 0.603 0.480 0.444 -30.298
9. Transport (3) 0.510 0.512 0.476 0.459 -10.000
10. Post & telecommunications (6) 0.427 0.426 0.384 0.360 -15.691
11. Gas, Water, Electricity (9) 0.735 0.644 0.641 0.636 -13.469
12. Primary sector (1) 0.669 0.669 0.669 0.669 0.000
13. Wood, cork, paper (2) 0.814 0.814 0.814 0.820 0.737
Total (unweighted average) 0.528 0.531 0.512 0.499 -5.492*
Total (weighted average based on
number of companies of given industries) 0.281 0.277 0.260 0.257 -8.541
23
24
Table 3
Pairwise Correlation Matrix of all used variables.
This table presents the pairwise correlations between all variables of the sample over the period 2007-2010. Industry dummies are added but not reported. Negative values are represented in brackets.Two issues of mulitcollinearity are observed, namely between the variables OC - OC_2 and LEVERAGE – LIQUIDITY. With respect to the first case, the reason for multicollinearity is based on the fact that OC_2 is the simple square of OC. In the latter context, the correlation does not exceed the limit of 0.7 and should not be regarded as a serious multicollinearity problem (Belsley, Kuh, and Welsch, 2005). ***, ** and * represent the 1%, 5% and 10 % significance levels, respectively.
25
6. Empirical Results
6.1. OLS-Results
Before starting the analysis of the empirical results, the problem of endogeneity in regard to the relationship between ownership structure and firm performance, which can cause interpretative difficulties, should be discussed shortly. Being more specific, the endogeneity problem in the case at hand refers to the issue of reverse causality. Ownership structure and firm performance may be simultaneously determined and the question arises what causes what? Despite of the fact that firm performance is influenced by the ownership structure, recent studies suggest that ownership structure can also be influenced by the firm performance. This, however, would have the consequences that the results of the OLS-regression are biased.
Though, Edwards and Nibler (2000) argue that the ownership structure of the majority of German firms remains unchanged over time. So, they suggest that the ownership structure is not affected by market developments, and thus firm performance. Therefore, they treat the ownership variable exogenously. Furthermore, Edwards and Weichenrieder (2004) and Gugler and Weigand (2003) confirm the assumption of an exogenous variable with the results of their respective studies. Both studies treat the variable as exogenously at one time and endogenously at another time. No differences could be identified which forces the point of an exogenous ownership variable. Following the results of these studies, the ownership variable is assumed to be exogenous.
26
Tables 4 and 5 subdivide the relationship over the four-year timespan. The results from the regression are in general inconclusive. Different performance indicators are used in order to shed light on the issue of potential expropriation effects during the years of the financial crisis. The dependent variables of table 4 are both market based performance indicators, whereby table 5 shows the results of dependent variables (ROA, ROE) which have an accounting background. Reason for this grouping is based on the assumption that the results might have similar outcomes for each group.
Table 4
The relationship between ownership concentration and firm performance.
The table presents the pooled Ordinary Least Squares analysis results by the use of following regression model: . The dependent variable,
either Tobin’s q or Annual Stock Return, is regressed on the independent variables OC and OC_2. Four control variables are also included in the regression. The sample consists of 138 companies per year over the period of 2007-2010. Each column represents two different market based firm performance indicators for each single year. The intercept is not reported and industry dummies are excluded in this regression. Figures represented in brackets are t-statistics values. ***, ** and * represent the 1%, 5% and 10 % significance levels, respectively.
Variables
2007 2008 2009 2010
Tobin’s q Stock Return Tobin’s q Stock
Return Tobin’s q Stock Return Tobin’s q Stock Return
OC 0.948 (1.374) 0.204 (0.462) -0.025 (-0.055) -0.173 (-0.635) 0.074 (0.129) -0.192 (-0.414) -0.681 (-1.241) -0.488 (-1.105) OC_2 -0.837 (-1.294) -0.132 (-0.318) 0.150 (0.355) 0.354 (1.424) 0.112 (0.206) 0.082 (0.188) 0.523 (0.985) 0.282 (0.660) LIQUIDITY 0.332*** (3.921) 0.032 (0.623) 0.242*** (4.048) 0.003 (0.934) 0.177** (2.412) -0.019 (-0.322) 0.158*** (1.988) -0.028 (-0.430) LEVERAGE -1.525*** (-3.774) -0.039 (-0.149) -0.540*** (-2.066) -0.351*** (-2.319) -1.311*** (-3.599) -0.595*** (-2.041) -1.467*** (-3.743) -0.010 (-0.305) FIRM_SIZE (Ln) 0.019 (0.627) 0.047** (2.364) -0.006 (-0.295) -0.008 (-0.720) 0.014 (0.547) 0.035 (1.647) -0.025 (-0968) -0.051** (-2.429) AGE (Ln) -0.034 (-0.461) -0.067 (-1.402) 0.045 (0.909) 0.027 (0.928) 0.050 (0.665) -0.069 (-1.160) 0.121 (0.144) 0.034 (0.502) Industry NO NO NO NO NO NO NO NO R² 0.359 0.051 0.273 0.159 0.263 0.065 0.256 0.061 Adj. R². 0.319 0.007 0.240 0.120 0.229 0.023 0.222 0.018 N 138 138 138 138 138 138 138 138
27
Table 5
The relationship between ownership concentration and firm performance.
The table shows the pooled Ordinary Least Squares analysis results by the use of following regression model: . The dependent variable,
either ROA or ROE, is regressed on the independent variables OC and OC_2. Four control variables are included in the regression. The sample consists of 138 companies per year over the period of 2007-2010. Each column represents two different accounting based firm performance indicators for each single year. The intercept is not reported and industry dummies are excluded in this regression. Figures represented in brackets are t-statistics values. ***, ** and * represent the 1%, 5% and 10 % significance levels, respectively.
Variables
2007 2008 2009 2010
ROA ROE ROA ROE ROA ROE ROA ROE
OC -0.055 (-0.746) -0.162 (-0.900) 0.019 (0.220) -0.041 (-0.144) 0.060 (0.812) 0.100 (0.386) -0.047 (-1.092) -0.311** (-2.069) OC_2 0.047 (0.681) 0.151 (0.893) -0.011 (-0.145) 0.113 (0.435) -0.040 (-0.462) -0.020 (-0.081) 0.062 (1.483) 0.359** (2.463) LIQUIDITY 0.013 (1.501) 0.033 (1.583) 0.020* (1.826) 0.030 (0.811) 0.006 (0.593) -0.002 (-0.060) 0.011* 1.802 0.034 (1.553) LEVERAGE -0.153*** (-3.539) 0.167 (1.583) -0.201*** (-4.150) -0.200 (-1.243) -0.217*** (-4.660) -0.359** (-2.187) -0.171*** (-5.527) -0.121 (-1.129) FIRM_SIZE (Ln) 0.002 (0.569) -0.005 (-0.620) 0.005 (1.262) 0.011 (0.923) 0.003 (0.931) 0.007 (0.589) 0.002 (0.829) 0.007 (0.980) AGE (Ln) 0.015* (1.836) 0.014 (0.699) -0.002 (-0.271) -0.023 (-0.763) -0.002 (-0.231) -0.012 (-0.372) 0.001 (-0.179) -0.002 (-0.086) Industry NO NO NO NO NO NO NO NO R² 0.194 0.322 0.242 0.048 0.242 0.063 0.363 0.102 Adj. R² 0.157 0.304 0.208 0.003 0.208 0.020 0.334 0.061 N 138 138 138 138 138 138 138 138
As stated before, must be greater than zero and must be less than zero in order to meet the condition of a normal bell-shaped relation between ownership structure and firm performance. By studying all given results, both market-based performance indicators in the year 2007 and both accounting-based performance indicators in the year 2009 fulfill each, the above condition as well as ROA in the year 2008. For the calculation of the maximum of the curvilinear relation, the following formula is used: max. OC:
28
concentration variable. Thus, it is difficult to draw any conclusions about a possible left shift of the maximum point. Moreover, other applied performance proxies do not show similar outcomes and indicate a non-constant relation between ownership concentration and firm performance over the crisis periods. This fact might hamper the analysis and can be seen as a source for further research.
Table 6 below shows the results of the full period and the previous stated condition of a bell-shaped curve is again not met. On the contrary, the results indicate that the relation might be a U-shaped curve for all given dependent variables since the dependent variable OC is smaller than zero in each case (Tobin’s q: = 0.006; Stock return: = 0.386; ROA: = 0.292 and ROE: = -0.117) and OC_2 is greater than zero (Tobin’s q: = 0.068; Stock return: = -0.386; ROA: = 0.022 and ROE: = 0.167). As these results are insignificant and are not supported by the theory, this topic is not discussed further here.
29
Table 6
The relationship between ownership concentration and firm performance -full period.
The table displays the pooled Ordinary Least Squares analysis results for the full period by the use of following regression model: . The
dependent variable, either Tobin’s q, Annual Stock Return, ROA or ROE, is regressed on the independent variables OC and OC_2. Four control variables are also included in the regression. The sample consists in total of 552 observations over the period 2007-2010. The intercept is not reported and industry dummies are again excluded in this regression. Figures represented in brackets are t-statistics values. ***, ** and * represent the 1%, 5% and 10 % significance levels, respectively.
Table A4 illustrates the outcomes of the regression model subdivided into three different time classes – pre crisis, crisis, post crisis. Again, specific industry dummies are not included, yet. In regard to each individual period, most of the coefficients are insignificant except of the coefficient ROE in the last column. Consequently, it complicates the issue of interpreting the variable and draw further conclusions. In some cases the sign for the coefficients changes from one period to another. This fact would confirm the study of Franks and Mayer (2001) suggesting that this relation is not static, but rather more dynamic. Firms may adjust their own ownership structure in response to the business environment where they currently operate. So, the crisis could have an impact on the degree of ownership concentration as well as on the relation between ownership concentration and firm performance.
Full period (2007-2010)
Variables Tobin’s q Stock Return ROA ROE
30
The control variable LEVERAGE is significantly (at 1%, 5% and 10% significance level) and negatively correlated with the dependent variable in most of the cases. LIQUIDITY is positively related to firm performance in the majority of the results. Especially, this control variable is significantly (at 1% significance level) positively correlated with Tobin’s q in all three different time periods. The variable AGE has a changing and insignificant relation either negatively or positively, indicating no fix correlation.
Several scholars suggest that the relation between ownership structure and firm performance is influenced by firm characteristics or by specific industrial features. Therefore, industry dummies are included in the regression model in order to control for special sector patterns. Table A5 provides an overview of the individual results. Again, it is difficult to draw a specific conclusion based on the outcomes regarding OC and OC_2 because of the insignificance and regular sign change of these variables. Thus, it seems hard to elaborate whether the expropriation or the monitoring effect outweighs during the financial crisis, due to the complexity of the issue.
Regarding the control variables, the empirical outcomes provide a more conclusive and complete picture. FIRM SIZE is mainly positively related to firm performance and even in some cases significant at a 10% significance level. This result is consistent with other studies (Himmelberg, Hubbard, and Palia, 1999) stating that during times of crisis a bigger corporation is better able to devise strategies in order to overcome any market risks or arising uncertainties, and thus reduce potential future losses. Another fact is that bigger firms have a better bargaining power compared to smaller firms. This improves the firm performance of each single firm. LEVERAGE is negatively and significantly (either at 10% or at 1% significant level) correlated with firm performance most of the time. During the crisis-period this variable correlates significantly negative in respect to all given performance proxies. The main reason is already explained in the above section. AGE is predominantly positively but insignificantly related to firm performance.
6.2. Robustness Test
31
of a dependent variable. Thus, in case of analyzing panel data, either fixed effect or random effect models can be applied. The main difference between both models is based on the question whether the unique and time constant characteristics result from random variations or not. The Hausman-Test helps to determine which of both models should be taken. Considering that, cross-sectional as well as period data has both fixed and random effects.
In respect to the sample, only a random effect model can be applied because the regression includes industry dummies which remain constant during the time period. Therefore, the fixed effect model cannot be used. The outcomes for the whole sample are reported in table A6. Generally speaking, the outcomes are in majority similar to the OLS-regression results and even in case of deviations, they are relatively small.
Despite of using a different regression method, another robustness check is done. As already discussed in the previous part, a large number of firms (65% of the entire sample) operate in one of the following three industries: Machinery, equipment, furniture, recycling, Chemicals, rubber, plastics, non-metallic products and Wholesale & retail trade. Hence, there might be the problem that these industries and their individual characteristics regarding ownership structure distort the overall results. Hence, another OLS regression is applied excluding these firms from specific industries to check whether the outcomes differ significantly. Table A7 demonstrates both OLS results for the full-period; once applying the whole sample size (552 observations) and once excluding the three main industries (188 observations). Again, no ownership concentration turns to be significant and in general only small changes in in the results are obtained. All signs remain identical in both groups and in the majority of the cases the significance level does not change.
To sum up, the generated results from the “core” regression are robust and consistent even by changing the sample size or employing a different regression method.
7. Conclusion
32
studies. Based on the data, I am not able to draw a comprehensive conclusion whether the expropriation effects are more dominant during the crisis. Main reason for this problem is the frequent change of sign of the coefficients OC and OC_2 suggesting a non-static relation between ownership concentration and firm performance over the years of observations. Further, the results of the OLS regression show an insignificant relationship between the dependent variable (firm performance) and the independent variable (ownership concentration). The robustness test indicates that the results of the dependent variables behave robust even when another examination method is applied. Another pattern, which should be considered in this respect, is the problem of endogeneity. As other scholars have already argued before, it is difficult to analyze whether the firm performance is indeed influenced by the ownership structure or vice-versa – meaning the issue of reverse causality. Although I follow the statement of Edwards and Nibler (2000), stating that the ownership variable shall be treated exogenously in the German corporate governance context, it is possible that this does not refer to the given sample. Hereby, a different model, treating the variable endogenously might generate better and more plausible results.
In addition, the inclusion of industry dummies leads to slight changes of variables. In other words, individual sectorial features have an impact on the relation between the dependent and independent variables. Especially, the ownership concentration changes partially from a negative to a positive sign, indicating that firms operating in one sector have own characteristically ownership structures compared to other sectors. This pattern should be considered in case of future research in the German business environment.
7.1. Implications
33 7.2. Limitations
The research at hand is subject to six main limitations. First, the definition of ownership concentration is in some extent arbitrary. Although I used a common academic procedure which has been used by other scholars, other ways for measuring ownership concentration still exists. Second, the focus is set on voting rights instead of cash flow rights or the divergence of both rights because of data reasons. According to Edwards and Weichenrieder (2009), this distinction should be considered and not be seen as the same in cases of measuring the ownership structure in German corporations. Third, there is a problem concerning data access. Since many German corporations are built through pyramidal ownership structure or cross-shareholdings, finding the ultimate owner is difficult and sometimes even faulty. Despite the fact that a specific database is used to remedy this issue, still small variations might occur due to diverse reporting styles. Fourth, from an empirical point of view the sample size can be seen as too small. Fifth, I concentrate especially on listed and non-financial firms. So, in a certain way the German ownership structure of non-listed firms as well as firms operating in the financial sectors is ignored. Finally, the problem of endogeneity remains open which can be seen as the main limitation aspect of this study. More precisely, the ownership structure might be also determined by the firm performance which then would create bias in regard to the estimators of the given study.
7.3. Further Research
This study leaves some questions unanswered, which can be seen as a source for further research. The topic of expropriation effects on minority shareholders can be investigated based on other countries which have a low level of investor’s protection; for instance in developing countries. Since, the global crisis also had an impact on those nations, the probability to find such value-decreasing outcome is more likely. The investigation of other explanatory variables which might interact with the relation between ownership concentration, the effect of exploitation by larger shareholders and firm performance, remains as an open question.
34
