Nijmegen School of Management
Master’s Thesis
The Relation between Ownership Structure and Capital Structure:
Empirical study of DAX firms
Abstract
This paper investigates the effects of managerial ownership and ownership block holders on the capital structure of the firm by using an agency framework. The research extends the literature on the subject by examining the relation between equity ownership and debt-ratios in DAX firms. Furthermore, possible interaction effects between management ownership and external block holder ownership are analyzed. The findings of the empirical study suggest that debt-ratio is negatively related to managerial ownership and positively related to external block holder ownership. Moreover, the negative relationship between managerial ownership and debt-ratio is negated by large external block holders. The observed negative correlation becomes insignificant in the presence of large external shareholders. The findings of this study suggest that large external block holders influence the agency cost of equity and debt.
JEL classification: G30, G32
Keywords: Capital structure; Management ownership; External shareholders; Agency cost
Author: Lambert Tran
Student ID: s4827775
Specialization: Corporate Finance and Control Supervisor: Dr. S. Nolte
Date: November 15, 2020
3
Table of Contents
1. Introduction 4
2. Literature Review 6
2.1 Management ownership and capital structure 6
2.2 External block holder ownership and capital structure 9
2.3 Interaction between managerial ownership and external block ownership 11
3. Research Method 12
3.1. Data sample description 12
3.2 Variables 13 3.2.1 Dependent variables 14 3.2.2 Independent variables 14 3.2.3 Control variables 16 3.3 Models 18 4. Results 19 4.1 Descriptive statistics 19 4.2 Correlation matrix 20 4.3 Regression results 20
4.3.1 Analysis 1: Capital structure and management ownership 20
4.3.2 Analysis 2: External block holder ownership 21
4.3.3 Analysis 3: Interaction effect 22
4.4 Robustness check 23
5. Discussion & Conclusion 24
5.1 Discussion 24
5.2 Conclusion 26
6. References 28
4
1.
Introduction
Since Modigliani and Miller (1958) irrelevance propositions, capital structure has received considerable attention. A range of studies have shed their light on the determinants of capital structure. Recently, there has been an increased interest in the role of managers in the capital structure puzzle. Under the managerial perspective, capital structure is not only determined by external and internal contextual factors (Brailsford et al., 2002). This perspective also considers the managers characteristics in the discussion surrounding the determinants of capital structure. Managers adverse incentives can affect corporate finance decision-making. In turn, the incentives for management to act opportunistically are influence by the composition of equity ownership (Angrawal and Mandelker, 1990).
Jensen & Meckling (1976) agency framework elaborates on the effects of separation of ownership and control. Their work suggests that the conflict of interest between management and shareholder can lead managers to act opportunistically by consuming corporate resources for their own benefit. This might be at the expense of shareholders. Another problem is that managers will often choose to lower leverage to suboptimal levels. Reason for this is the inability of managers to diversify their human capital invested in the firm. To ensure their employment and thus wealth, they can choose to reduce firm risk by underleveraging (Friend & Lang, 1988). The agency costs which arise due to conflict between managers and shareholders can reduce the firm value as markets incorporates them.
Different mechanisms have been proposed by Jensen & Meckling (1976) to address the agency problem. First, increased managerial ownership can be used to mitigate the agency problem. This will align the interest of shareholders and managers and therefore reduce the possible agency costs. Second, increased financing by debt can reduces the equity base, increasing the percental equity ownership of managers. This again will align the interest of management and shareholders. Additional debt also increases the chance of financial distress or even bankruptcy. This risk might motivate managers to increase their efficiency and reduce their perquisites consumption. Furthermore, it lowers the amount of free cash flow under their discretion. The third disciplining mechanism is external block holder equity ownership. External block holders are argued to actively monitor and control management to protect their investments. Thus, force managers to make optimal capital structure decisions. How these mechanisms are interrelated remains unresolved. Especially, the effect of equity ownership on debt-ratios raises questions. Hence, this study will address the following research question: To what extent does equity ownership affect capital structure?
Prior research has reported contradictory findings with respect to the effect of different equity ownership on capital structure. Angrawal & Mandelker (1987), Mehran (1992), Kim & Sorensen (1986), Agrawal & Knoeber (1996), Berger et al. (1997) and Short et al. (2002) find a positive relation between managerial ownership and leverage. On the other hand, Friend & Lang (1988), Friend & Hasbrouck (1988), Jensen et al. (1992), Bathala et al. (1994), Firth (1995) and Moh’d et al (1998) report a negative correlation.
5
Furthermore, Mehran (1992), Berger et al. (1997), Short et al. (2002) and Brailsford et al. (2002) have provided contradicting results regarding the relationship between leverage and large external shareholders. Some studiers report that external block holder ownership negatively affects debt-ratios, other empirical research supports a positive relationship. Moreover, empirical evidence argues for an interrelation between managerial and external block holder ownership (Brailsford, Oliver & Pua, 2002; Firth, 1995; Short et al. (2002). These studies however also do not come to a consensus on the interaction effects between the different types of equity ownership.
Motivated by conflicting empirical findings, this study re-examines the effect of equity ownership on capital structure. Specifically, it examines the influence of managerial and external block holder equity ownership on debt-ratios in German firms. In addition to the separate effects of managerial and external block holder ownership, we examine possible interaction effects between both types of equity ownership. This is done to see whether it can explain capital structure variation in firms. Prior literature has mainly focused on Anglo-American countries. This research extends existing literature by examining German markets. Germany is noted as a more bank-based system, in which long-term and network relationships are more prevalent (Edwards & Nibler, 2000; Schleifer & Visnhy (1989). Furthermore, German firms are characterized by high concentrated ownership and large permanent shareholders. This could result in different agency relations between management and external shareholders than in Anglo-American countries. Re-examining the relation in German firms might give us insights whether the effect of equity ownership on capital structure differs from firms in other countries. Additionally, investigating the relation between ownership and capital structure with a dataset after the recent financial crisis can give an updated view on the issue. Especially since prior research on the topic is focused on periods before the financial crisis of 2008.
The results of this study show that debt-ratios are negatively related to managerial ownership and positively related to external block holder ownership in the case of German firms. Furthermore, large external block holder ownership negates the negative relation between managerial ownership and debt-ratios. The observed negative relationship between managerial ownership and debt ratios is only present in the absence of large external block holders. These findings support the argument that large external block holders affect the relative size of the agency cost of equity and debt. Thus, influencing debt ratios indirectly via managerial equity ownership.
This paper continues by providing an overview on the most relevant and noticeable literature on the subjects. Subsequently, we formulate four hypotheses. Section 3 will elaborate on the data collection procedure and methodological approach. Section 4 will outline the empirical analysis and the results. Lastly, section 5 will provide a conclusion and elaborate on the contributions, limitations and recommendations for future research.
6
2.
Literature Review
Following Berle & Means (1932) statement that a modern corporation is characterized by separation of ownership and control. A range of studies have explored the effects of ownership and control on a company's operations. The relation between ownership structure and capital structure is often analyzed, both theoretically and empirically. Prior research has particularly focused on the influence of either managerial or external block holder ownership on capital structure. We will discuss both in the following chapters.
2.1 Management ownership and capital structure
In their work on agency theory, Jensen & Meckling (1976) suggest that the separation of ownership and control gives managers, as the agents, the opportunity and incentive to use corporate resources in their self-interest. This can be at the expense of external shareholders, the principals. The conflicts of interest however can be reduced by adopting corporate governance mechanisms. These mechanisms could lower management discretion and incentivizes them to act in the interest of shareholders. The managerial incentives structure is especially influenced by their equity ownership, as managerial equity ownership can be used to align managers incentives with shareholder interest. The alignment of interest can reduce the incentives to consume perquisites. This mechanism is referred to the convergence-of-interest or incentive-alignment hypothesis (Brailsford et al., 2002). According to the model of Jensen & Meckling (1976), an optimal level of debt is derived by minimizing the agency costs. This is done by balancing the agency cost of debt and external equity. The correlation between management ownership and debt therefore depends on the relative agency cost of debt and external equity. These costs vary at different levels of managerial ownership. Based on the theoretical work of Jensen & Meckling (1976), Short et al. (2002) conclude that the relationship between management ownership and leverage is positive. They state that increased managerial ownership results in lower agency costs of debt which in turn increases the use of debt.
Other theoretical explanations for the positive relation between managerial ownership and leverage are provided by Brealey et al. (1977), Harris & Raviv (1988) and Stultz (1988). Brealey et al. (1977) predict, based on signaling considerations, that firms with higher managerial ownership have higher debt-ratios. Higher debt-ratios allows management to retain more equity. This increased equity ownership reduces their welfare due to their risk aversion, but this effect is smaller for firms of high quality. Managers of high-quality firms can choose to hold more debt to signal their firm high-quality to outside investors. Harris & Raviv (1988) and Stultz (1988) theoretical work involves managers incentives to maintain control over their firms. They suggest that managers with higher equity ownership may issue more debt to retire external equity. Thereby, decreasing the voting rights of outside shareholder and increases managerial discretion. Furthermore, higher leverage could make companies less attractive for raiders. This will reduce the chance of possible takeovers, again ensuring the control of managers. The positive relation between managerial
7
ownership debt-ratios has been documented by different empirical studies. These studies employed cross-sectional multivariate regression with datasets on US firms (Agrawal & Mandelker, 1987; Berger et al, 1997; Angrawal & Knoeber, 1996; Mehran, 1992; Kim & Sorenson, 1986) and UK firms (Short et al., 2002).
However, other theoretical models and empirical work argue for a negative relationship between managerial ownership and leverage. Agency theory implies that debt can also be considered a discipling mechanism which reduces opportunistic behavior of managers. Jensen & Meckling (1976) suggest that high levels of debt reduce the amount of free cash flow available. Moreover, managers can use debt to signal that they are committed to debt obligations (Grossman & Hart, 1982). Since both managerial equity ownership and debt can be used as disciplining measures to lower agency cost, they could substitute each other. Hence, higher managerial equity ownership should also reduce the need to use debt as a disciplining measure. This implies a negative relationship between managerial ownership and debt-ratio. Furthermore, corporate managers lack the capacity to diversify their investments since their wealth is largely derived from non-diversifiable human capital invested in the firm. This creates incentives to lower their risk by ensuring employment by the firm. One way to ensure employment is to decrease the company's debt-ratio. This will lower the chances the firm goes bankrupt or go through financial distress, in turn decreasing their risk of unemployment. The lowered debt-ratios however may be suboptimal for the firm. The management entrenchment hypothesis states that with increasing managerial ownership, managers incentives and ability to lower leverage to suit their own interest, increases (Schleifer & Vishny, 1989).
Different empirical studies support a negative relationship between managerial ownership and debt-ratios. Friend & Lang (1988) find a negative correlation between the market value of insider shareholdings and leverage. Mehran (1992) however criticized their datasets, which is mainly based on transactions, as it does not accurately reflect the equity holdings. Therefore, making their findings questionable. The study of Firth (1995) alleviates the criticism by Mehran (1992) by using similar explanatory variables as Friend & Lang (1988) on a dataset which better reflects ownership. Their study also documents a negative relationship. Bathala et al. (1994) use a single equation regression and find no significant correlation between managerial ownership and leverage. Together with Jensen et al (1992), they argue that the single equation regression might not take in account the endogeneity problem. To account for this problem, they employed simultaneous-equations frameworks with which they again document a negative correlation managerial ownership and debt-ratios. Moh et al. (1998) use a pooled time-series cross-sectional OLS regression to add to the cross-sectional data analyses. Although less significant, they also find the negative correlation.
The contradicting findings on the relationship between management ownership and debt ratios motivated the empirical study by Brailsford et al. (2002). The authors suggest that these conflicting findings may be caused by the assumption that the relationship between managerial ownership and debt-ratios is
8
linear. Brailsford et al. (2002) propose a non-linear inverted U-shape as function to explain the relation. This function could explain both the incentive-alignment effect and the management entrenchment effect. When managerial ownership is low the incentive alignment effect will reduce agency cost of debt and therefore lead to higher debt-ratios. At high levels of managerial ownership, the management entrenchment effect dominates which suggests that managers lower the debt-ratios for minimize their exposure to firm risk. This hypothesis is confirmed in their study on Australian firms. Their results suggest that the relation between managerial ownership and debt-ratios indeed has non-linear inverted U-shape, with a positive correlation at low levels of managerial ownership and a negative correlation at high levels of managerial ownership. However, the question remains whether these findings can be extended to firms based in other countries.
An overview of the empirical studies regarding the relation between managerial equity ownership and debt-ratios is provided in Table 1. The overview shows that it is difficult to draw a conclusion on the effect of managerial ownership on capital structure based on US data. Some studies support the arguments put forward by the incentive-alignment hypothesis, others show results supporting the management entrenchment hypothesis. Moreover, agency theory can also explain both incentive-alignment and management entrenchment. The effect of managerial ownership appears to be dependent on the assumption of the influence it has on the agency cost of debt and equity, whether it complements or substitutes debt. The incentive-alignment hypothesis however is more strictly based on the agency model of Jensen & Meckling (1976). Specifically, when we consider the relative agency cost of debt or equity at different levels of managerial ownership. This study therefore follows the incentive-alignments hypothesis which predicts a positive correlation between managerial ownership and debt-ratios. This results in the following hypothesis:
H1: There is a positive relationship between management ownership and debt ratios.
Due to the conflicting findings in earlier literature, we also must also take in to account the arguments of the entrenchment hypothesis. Therefore, this study follows the research of Brailsford et al. (2002) by examining the data for a possible non-linear relation between managerial ownership and debt ratios.
H2: The relationship between managerial ownership and debt ratio is non-linear
The two hypotheses considering managerial ownership will be empirically tested with a data set consisting of German firms. In addition to the effect of managerial ownership on capital structure, we also examine the influence of external shareholders on capital structure decisions.
9
Table 1: Overview empirical studies on managerial ownership and capital structure
Author(s) Country Period Effect managerial ownership
Kim & Sorenson (1986)
US 1970-1980 +
Agrawal & Mandelker (1987)
US 1974-1982 +
Friend & Lang (1988) US 1974-1983 -
Friend & Hasbrouck (1988) US 1983 - Angrawal & Nagarajan (1990) US 1979-1983 - Mehran (1992) US 1973-1983 + Bathala et al. (1994) US 1988 - Firth (1995) US 1995 -
Angrawal & Knoeber (1996)
US 1981-1987 +
Berger at al. (1997) US 1984-1991 +
Moh’d et al. (1998) US 1972-1989 -
Short et al. (2002) UK 1988-1992 +
Brailsford et al. (2002) Australia 1989-1995 + / -
2.2 External block holder ownership and capital structure
External block holders are argued to have significant influence on the capital structure. Since they have the incentive to actively monitor and control management decisions to protect their investments (Stiglitz, 1985). Contrary to small shareholders who do not have the incentives and means to monitor and control management. Block shareholders can reduce managerial opportunistic behavior and thus lower agency costs (Schleifer & Vishny, 1986). The discussion remains whether the presence of external block holder complements or substitutes the discipling effect of debt. Zeckhouser & Pound (1990) suggest that external block holder ownership is a substitute for debt. It signals to the market that managers are less able to act opportunistically, which reduces the need to use debt. This implies a negative relation between external block holders and leverage. Short et al. (2002) also argue for a negative correlation which they support with a study on UK firms. They state that external block holders incentivize managers into risk shifting activities which benefit shareholders. The higher agency cost of equity and lower agency costs of debt due to these risk-shifting activities result in a negative correlation.
However, other studies argue for a positive relationship between external block ownership and leverage (Friend & Lang, 1988; Mehran, 1992; Berger et al., 1997; Brailsford et al., 2002). Friend & Lang (1988) suggest that external block holders hold more diversified portfolios than managers. Due to this diversification, block holders are less risk averse and therefore prefer higher debt levels. As mentioned, they
10
also can incentive and force managers to hold more debt and pursue activities in the interest of shareholders instead of debtholders. Hence, agency cost of debt will decrease. This in turn will results in a higher debt-ratios. Furthermore, external block ownership and debt could complement each other. Debt may be attractive for these block holders since their cost of monitoring is high due to their diversified portfolio. Hence, debt can be used as a complementary mechanism to their other disciplinary devices. This implies a positive relation between external block ownership and debt ratios which is supported by empirical results. Friend & Lang (1988), Mehran (1992) and Berger et al. (1997) find a positive correlation based on datasets on US firms. A positive correlation was also found by Brailsford et al. (2002) on Australian companies. These findings contradict the empirical evidence indicating a negative relationship by Short et al. (2002) and Zeckhouser & Pound (1990).
The positive relation between external block holder ownership and debt ratios has considerably more empirical support with studies using datasets in different regions. Therefore, this study predicts positive correlation between external block holder’s ownership and debt ratios, resulting in the following hypothesis:
H3: There is a positive relationship between external block holder ownership and debt ratios
However, we must take in account that large external block holdings are more present in Germany (Edwards & Nibler, 2000). Increased equity ownership of external block holders might reverse their incentives as their exposure to risk also increases. Their increased risk aversion could lead them to prefer lower debt ratios as entrenched managers do. Moreover, block holders with significant equity ownership might already have sufficient control mechanisms in place such as voting on board members. This alleviates the need for debt as a disciplining measure. Therefore, the positive relationship between external block holder ownership and debt ratios becomes insignificant with large external block holders.
H4: The positive association between external block holder ownership and debt ratios becomes insignificant in the presence of a large external block holder.
11
Table 2: Overview empirical studies on external block holder ownership and capital structure
Author(s) Country Period Effect external blockholder ownership
Friend & Lang (1988) US 1974-1983 +
Mehran (1992) US 1973-1983 +
Berger at al. (1997) US 1984-1991 +
Short et al. (2002) UK 1988-1992 -
Brailsford et al. (2002) Australia 1989-1995 +
2.3 Interaction between managerial ownership and external block ownership
Prior research mostly examined the effect of either managerial ownership or external block ownership on capital structures. Despite the possible interaction effect between managerial and external block holder ownership. External block holder ownership may affect debt-ratios indirectly by influencing the relation between managerial ownership and debt-ratios. The block holder equity ownership and managerial equity ownership are argued to positively influence debt ratio. This could indicate that both could be used as measures to align the incentives of managers with the interest of shareholders. Subsequently, it can be argued that managerial ownership and external block holder ownership are substitutes. External block holders, actively monitoring and controlling management could alleviate the need for managerial ownership as a disciplining measure. The monitoring effect of external block holders is enough to ensure optimal debt levels which makes the incentive-alignment effect of managerial ownership irrelevant. This study therefore predicts that the positive relationship between managerial ownership and debt-ratios is not present in the presence of large external block holders:
H5: The positive relation between managerial ownership and debt ratios becomes insignificant in the presence of large extern block holders.
Brailsford et al. (2002) examine the influence of the distribution of equity ownership on leverage in their empirical study of Australian firms. Their study suggests that external block holders and managerial ownership interacts such that a low level of managerial ownership, external block holders can more effectively monitor and control management. This results in a positive relationship between managerial ownership and leverage. The monitoring effects of external block ownership however weakens at higher levels of managerial ownership. Managers will become more entrenched which enables them to act in their own interest. Thus, the effect of external block ownership becomes less significant. The results of Short et al. (2002) imply that external block holders not only weaken the positive effect of managerial ownership on leverage but that this correlation is completely negated by the presence of block holders. Their explanation
12
is based on the relative size of the agency cost of debt and equity. They argue that the presence of large external shareholders, with the ability to control managerial behavior, forces management to be less risk averse. This in turn increases the potential cost to debt holders which increases the agency cost of debt. Moreover, the presence of large external shareholders also might decrease the cost of equity, due to external shareholders perceiving that their presence may limit the opportunistic behavior of management. However, it must be mentioned that Short et al. (2002) document negative relation between external block holder ownership and debt ratios which conflicts with the positive correlation found by Brailsford et al. (2002). Firth (1995) who found that debt-ratios are negatively related to managerial ownership and positively related to external block ownership, report that external block holders reduce the negative relationship between debt and managerial ownership. Previous literature shows that external block holders indeed influence the relationship between managerial ownership and debt-ratios. The examination of German listed firms can add to the empirical evidence.
3.
Research Method
This chapter will explain the methodology of the research. Section 3.1 elaborates on the sample and data collection procedure. Section 3.2 will outline the dependent, independent and control variables. At last, section 3.3 will explain the used empirical analysis.
3.1. Data sample description
To answer the research question, we will use a quantitative research method. Data on ownership and capital structure will be retrieved from Eikon’s Datastream. This database combines economic, financial and business information from a range of sources worldwide. Eikon provides annual tables on share ownership. These tables contain: ranking of holder, name of holder, type and percentual share ownership.
In addition, data on firm financials are also retrieved from Eikon. Since this study focuses on German evidence on the relation between ownership and capital structure. The sample will consist of German listed companies. Specifically, we sample based on the three indices DAX, MDAX and SDAX which results in an initial sample of 160 firms over the period January 1, 2013 and December 31, 2018. This specific time period could give us insights whether the relation between ownership and capital structure has changed after the recent financial crisis. Subsequently, we eliminate financial services firms and insurance companies from the sample due to their fundamentally different capital structure (Short et al., 2002). In addition, foreign companies and pure holding companies are eliminated. Furthermore, companies with insufficient available data on ownership and capital structure are also removed which results in a remaining sample of 137 firms.
13
For each firm in the sample we retrieved the share ownership of top 2 directors, all the directors and the share ownership of top two, five and top 10 largest shareholders.
3.2 Variables
In this chapter, the independent, dependent and control variables are outlined. Table 1 presents an overview of the operationalization and source of variables.
Table 3: Overview variables
Variable Name Measurement Source
Dependent Variable
DEBT The variable for debt-ratios defined as the book value of debt divided by the market value of equity plus debt.
Eikon
Independent Variables
MSO The percentage equity shares owned by all directors. Eikon EBO The percentage owned by the five largest external
shareholders
Eikon
Control Variables
SIZE Natural log of total assets. Eikon
VOLTY The standard deviation of annual percentual change in EBITDA is used as a proxy for volatility
Eikon
GROWTH Annual percentage change in total assets Eikon
FCF Operating income before tax + depreciation expense + amortization – total tax – total dividend.
Eikon
PROF Operating income before interest and taxes divided by total assets
Eikon
INTA Total intangible assets divided by total assets Eikon NDTS Annual depreciation expenses divided by total assets Eikon
14 3.2.1 Dependent variables
Capital structure ratio
Academic literature does not provide us with a clear operationalization of the leverage variable. Most prior research on the relation between ownership and capital structure use long-term debt to total capital as the dependent variable (Agrawal & Nagarajan, 1990; Mehran, 1992; Kim & Sorenson, 1986; Jensen et al.,1992; Moh et al., 1998; Firth, 1995); Bathala et al. 1994). They suggest that the use of total debt which also includes short-term debt might raise issues due to short-term debt having different determinants than long-term debt. Moreover, the agency theory of Jensen & Meckling (1976) is based upon long-run external financing. As done in prior research on the subject, we will also use long-term debt in this study. This will enable us to compare the results with prior work on the relation between ownership structure and capital structure. Furthermore, the issue arises whether to use market values or book values.
Previous theoretical studies use market values for both debt and equity. However, due to the unavailability of market values of debt, market values of debt are commonly substituted by book values which are often considered as a close proxy for their market equivalent (Brailsford et al., 2002). Furthermore, Friend & Lang (1988) argue for the use of book values of equity, since managers’ decision-making on capital structure is mainly based on book values. However, book values of equity may be altered by management with accounting tools which makes the comparability of firms questionable. Therefore, in accordance with prior empirical studies, we use market values of equity for determining leverage (Titman & Wessels, 1988). To summarize, the dependent variable leverage is operationalized as the book values of long-term debt divided by total capital which consist of the long-term debt and the market value of equity as is often done by studies on the relation between ownership and capital structure (Mehran, 1992; Kim & Sorenson, 1989; Jensen et al., 1992; Moh et al., 1998; Firth, 1995; Bathala et al., 1994; Brailsford et al., 2002; Short et al., 2002)
3.2.2 Independent variables Managerial ownership
Managerial ownership (MSO) is one of the independent variables included to examine the relation between ownership and capital structure. This study uses equity ownership of corporate directors as a proxy for managerial ownership as is done by earlier research (Mehran, 1992; Kim & Sorenson, 1989; Jensen et al., 1992; Moh et al., 1998; Firth, 1995; Bathala et al., 1994; Brailsford et al., 2002; Short et al., 2002). Prior studies mostly use percentual ownership instead of market value of equity as in better reflects the agency framework. We therefore define managerial ownership (MSO) as the percentage equity shares owned by all directors. To test for robustness, we also include an alternative definition of managerial ownership. In addition to the equity ownership of all directors, the equity ownership of the top two directors is included
15
as a proxy. As suggested in Hypothesis 1, we predict the relation between managerial ownership and debt-ratios to be positive.
External block holder ownership
External block holder ownership (EBO) is the other main independent variable which is defined as the percentage equity owned by the five largest shareholders (Brailsford et al., 2002). Furthermore, we include the dummy variable BLOCK to measure the presence of a large external block holder. This study mirrors the definition of Friend & Lang (1988) and Short et al. (2002) by considering external block holders large when they own 10 percent or more share equity. Thus, the dummy variable BLOCK will be one with large external shareholders with 10 percent or more equity share ownership and 0 otherwise. The expected relation between external block holder ownership and debt-ratios is positive, as suggested by hypothesis 3.
To test for the possible difference in effect for firms with large external block holders and without, we formulate the variables: EBONOBLOCK and EBOBLOCK. These are comprised of the external block holder ownership variables (EBO) and the dummy variable (BLOCK). As hypothesized, we expect the coefficient on EBONOBLOCK to be positive and significant and insignificant on EBOBLOCK
EBONOBLOCK = percentage equity shares owned by the five largest external shareholders if a large external block holder is not present.
= 0 if a large external block holder is present
EBOBLOCK = percentage equity shares owned by the five largest external shareholders if a large external block holder is present.
= 0 if a large external block holder is not present
To test the possible interaction effect large external block holders, have on the relation between managerial ownership and debt-ratios, two interaction variables are constructed: MSONOBLOCK and MSOBLOCK. The variables are comprised of the variable managerial ownership (MSO) and the dummy variable (BLOCK) which leads to the following operationalization:
MSONOBLOCK = percentage equity shares owned by all directors if a large external block holder is not present.
16
MSOBLOCK = percentage equity shares owned by all directors if a large external block holder is present.
= 0 if a large external block holder is not present
These variables enable the coefficient on managerial ownership to vary for large and small external block holder ownership. Hypothesis 5 predicts that the external block holders will negate the positive correlation between managerial ownership and debt-ratio. Therefore, we expect the coefficient on MSONOBLOCK to be positive and significant and the coefficient on MSOBLOCK to be insignificant.
3.2.3 Control variables
In accordance with prior research, we included a range of control variables to account for other potential variables which could influence capital structure. Incorporation of these variables enables us to isolate the effect equity ownership has on capital structure. The control variables can be divided into 4 categories: (1) risk, (2) agency cost, (3) asset specificity, (4) tax. This chapter will discuss the different control variables and their influence on capital structure.
Risk controls
Size (SIZE): natural logarithm of the book value of total assets. Larger, more diversified firms are subjected to lower risk of bankruptcy which enables them to hold more debt (Friend & Lang, 1988; Agrawal & Nagarajan, 1990). In addition, larger firms might have better access to credit markets, have more bargaining power and are able to profit from economies of scale (Short et al., 2002). These
arguments lead us to predict that firm size is positively related to debt-ratios. This is empirically supported by Friend & Lang (1988), Firth (1995), Berger et al. (1997) and Short et al. (2002).
Volatility (VOLTY): standard deviation of annual percentage change in EBITDA of the previous three years (Brailsford et al., 2002). The volatility of future income of firms is argued to be the main factor determining whether firms can meet interest obligations (Mehran, 1992). Earnings volatility thus indicates business risk. Debt holders take in account firm’s future earnings as protection. An increase in volatility will therefore decrease the supply of debt. Thus, we expect that the coefficient for volatility will by negative. Different measures are used in prior research. Titman and Wessels (1988) use operating income in their operationalization of the volatility measure. This study will follow the study of Brailsford et al. (2002) which use cash flows instead of operating income.
17
Agency cost controls
Growth (GROWTH): Annual percentage change in total assets. Multiple studies suggest that the growth opportunities of a firm can be used as a proxy for the agency cost of debt (Titman & Wessels, 1988; Jensen et al.,1992; Mehran, 1992). They argue that management of firms in growing industries are more prone to expropriate wealth from their debtholders. However, growth may also be an indicator for profitability. Profitable firms may have more available internal firms which could alleviate the need for external debt financing. As suggested by the pecking order theory of Myers and Majluk (1984), available internal funds are preferred over external financing. We therefore expect a negative coefficient on the variable GROWTH.
Free cash flow (FCF): operating income before tax + depreciation expense + amortization – total tax – total dividend. Free cash flow is operationalized as is done by Brailsford et al. (2002). External debt can solve problems with cash flow (Zwiebel, 1996). However, sufficient free cash flows negate the need for debt. This makes the effect of free cash flow difficult to predict.
Profitability (PROF): operating income before interest and taxes divided by total assets. More profitable companies need less debt due to available internal funds as suggested by Myers and Majluf (1984). Furthermore, profitable firms use their retained earnings to build equity relative to their debt. Multiple empirical studies have examined the relation between profitability on leverage (Friend & Lang, 1988; Jensen et al., 1992). Prior studies documented a negative correlation between profitability and leverage.
Asset specificity controls
Intangible assets (INTA): total intangible assets divided by total assets, measured at book values at the end of the year. Balakrisnan & Fox (1993) suggest that intangible assets increase a firm's asset specificity. Asset specificity may negatively influence debt financing due to the non-redeploy ability of specific assets. Specifically, firms are less able to borrow. Examples of intangible firm-specific assets are: R&D expenditure, reputational investments and brands (Balakrisnan & Fox, 1993). Myers (1933) also suggest that the agency cost of intangible assets is higher than tangible goods. We expect based on earlier findings that the coefficient of INTA will be negative.
Effect of taxes controls
Non-debt tax shield (NDTS): annual depreciation expense divided by total assets measured at year-end. DeAngelo and Masulis (1980) argue that non-debt tax shield like depreciation deductions, negate the tax benefits of additional debt. Thus, companies with larger non-debt shields compared to their expected
18
earnings, will receive lower tax benefits. These firms therefore will use less external debt. Based on these arguments and the empirical work of DeAngelo and Masulis (1980), we expect that the correlation between non-debt tax shields and leverage is negative.
3.3 Models
The empirical analysis involves a series of OLS regressions of the dependent variable leverage on the explanatory variables of ownership and the control variables. The data sample consist of multi-year observations which could cause heteroskedasticity. To account for the possible heteroskedasticity, we use the White technique to construct robust standard errors. The studies of Brailsford et al. (2002) and Short et al. (2002) also use standard errors in their empirical studies.
In the first stage, we analyze the relationship between leverage and management ownership. Thus, omitting the influence of external block holders. The goal of this stage is to determine the direction of the correlation between management ownership and leverage. To establish the direction, we use the following regression equation:
DEBTit = α0 + β0MSOit+ β1SIZEit+ β2VOLTYit+ β3GROWTHit+ β4PROFit + β5FCFit
+ β6INTAit + β7NDTSit (1)
The next step adds a squared term of the management ownership variable MSO to test for the non-linear relationship between management ownership and leverage, resulting in the following regression equation. This study mirrors the method used by Brailsford et al. (2002)
DEBTit = α0 + β0MSOit + β1(MSO)2it + β2SIZEit+ β3VOLTYit+ β4GROWTHit+ β5PROFit
+ β6FCFit + β7INTAit + β8NDTSit (2)
The second stage of the empirical analysis examines the relation between external block holder ownership and debt ratio. The external block holder ownership variable is included to determine the direction of the effect of external block holder ownership on debt ratio. At this stage, we do not yet incorporate the possible interaction effects between management ownership and external block holder ownership. The following regression equation is formulated.
DEBTit = α0 + β0EBOit+ β1SIZEit+ β2VOLTYit+ β3GROWTHit+ β4PROFit + β5FCFit
19
Short et al. (2002) argue that the effect of external block holder equity ownership on debt ratio might differ for small and large external block holders. We have formulated the following regression equation to test for the possible different effect:
DEBTit = α0 + β0EBONOBLOCKit+ β1EBOBLOCKit + β2SIZEit+ β3VOLTYit+ β4GROWTHit + β5PROFit + β6FCFit + β7INTAit + β8NDTSit (4)
The last stage of the analysis takes in to account the possible interaction effects between management ownership and the presence of external block holders as implied in H5. We replace the managerial
ownership variable with MSOBLOCK and MSONOBLOCK which represent managerial ownership with an external block holder and without. This allows us to examine the relationship between managerial ownership and debt ratios with different levels of external block holder ownership.
DEBTit = α0 + β0MSONOBLOCKit + β1MSOBLOCK2it + β2SIZEit+ β3VOLTYit+ β4GROWTHit
+ β5PROFit + β6FCFit + β7INTAit + β8NDTSit (5)
4.
Results
This chapter presents the results of the regression models. Section 4.1 will present the descriptive statistics of the variables. Section 4.2 provides the correlation matrix between variables. Section 4.3 will elaborate on the results of the empirical analyses and the hypotheses. And lastly, we will perform robustness test on the results. The findings of the robustness checks are displayed in section 4.4
4.1 Descriptive statistics
Table 4 in Appendix A displays the descriptive statistics of the variables employed in the empirical analysis. The debt-equity ratios of the firms in the sample range from 0 percent to 395.45 percent with a mean of 82.63 percent. Managerial ownership has a minimum of 0.0003 percent and a maximum of 84.2555 percent. The average managerial ownership is 29.761 percent is higher than reported in the studies of Brailsford et al. 2002 with 10.65 percent and Short et al. (2002) with 18.64 percent. The average level of external block holder ownership is 39.90 percent and ranges from 1.8 to 79.9 percent. The average size of firms based on total assets is 18.9 billion. However, the median firm size is only 3.2 billion euro with a minimum of 55 million euro and a maximum of 448 billion euro. The spread is relatively large due to inclusion of firms listed on the SDAX, MDAX and DAX. For the variables VOLTY, GROWTH, FCF, PROF, INTA and NDTS no rare findings are observed.
20 4.2 Correlation matrix
Table 5 in Appendix B presents the correlation matrix for all variables used in the empirical analysis. High correlation could indicate multicollinearity which occurs when multiple dependent variables are significantly correlated. This could result in biased coefficients, since independent variables can be predicted by the model itself. Values lower than -0.4 and higher than +0.4 could indicate multicollinearity. The correlation coefficients of the variables range between -0.3 and 0.3. This suggests that the independent variables included our dataset show little to no correlation.
In addition to the correlation matrix, we test for multicollinearity by calculating the variance inflation factors (VIF) of the explanatory variables. The calculated VIF values are presented in table 6 of Appendix C. Multicollinearity can be present when VIF values is greater than 5, which could indicate that independent variables are possibly correlated (O’Brien, 2007). The VIF values for the variables in the dataset do not exceed the critical value of 5. Therefore, we conclude that no multicollinearity is present in our dataset. This enables us to include all defined variables
4.3 Regression results
In this chapter, we present the results of the regressions. The goal of the empirical analysis is to determine the effect of equity ownership on capital structure in German based firms. The variables used in this study are debt-ratio, management equity ownership, external block holder ownership and the control variables. First, the effects of different ownership are analyzed. Subsequently, we include interaction terms to examine possible interaction effects between managerial and external block holder equity ownership.
4.3.1 Analysis 1: Capital structure and management ownership
This paragraph outlines the results regarding the first two hypotheses. The first hypothesis predicts that managerial ownership is positively related to debt-ratios based on prior literature. Table 7 in Appendix D presents the output of equation (1) where debt-ratio is regressed against managerial share ownership together with the control variables. The estimated coefficient on managerial ownership (MSO) is negative and significant at =0.10. This implies that the percentage ordinary shares owned by executive and non- executive management is negatively and significantly correlated with debt ratio. Therefore, we must reject hypothesis 1. There is not a positive relationship between managerial ownership and debt-ratios. On the contrary, the correlation between managerial ownership and debt-ratios is negative, based on our dataset. These finds support the entrenchment effect, which states that managers behave opportunistically and therefore prefer lower debt-ratio.
21
The second hypothesis states that there is a non-linear relation between managerial ownership (MSO) and debt-ratios. Table 7 in Appendix D displays the output of regression equation (2) which test for the possible non-linear relation between managerial ownership and debt ratios. The estimated coefficients by the OLS regression are negative for MSO and positive for MSO2 which indicates a non-linear U-shaped
relation instead of the predicted inverted U-shape (Brailsford et al., 2002). However, both coefficients are insignificant. We therefore must reject hypothesis 2. This implies that the hypothesized non-linear relation between managerial ownership and debt-ratios is not present in our data sample.
Regarding the control variables for both regressions. The estimated coefficient on PROFIT, SIZE and GROWTH show expected signs and are significant at 5 percent level of confidence. PROFIT shows a significant and negative coefficient which is consistent with the studies of Friend & Lang (1988), Titman and al (1988) and Brailsford et al. (2002). This implies that more profitable firms have lower debt-ratios which is in accordance with the pecking order theory (Myers & Majluf, 1984). The theory suggests that these firms prefer to finance their projects with internal firms instead of external debt. The coefficient for SIZE is negative and significant which supports the hypothesis that larger companies have better access to credit markets, more able to profit from economies of scale in debt financing and have lower agency cost due to lower risk and cost of bankruptcy. This positive relationship is in line with results found by Friend & Lang (1988), Firth (1995), Berger et al. (1997) and Short et al. (2002). The GROWTH variable has a negative and significant coefficient which is consistent with the results of Titman & Wessels (1988) and Mehran (1992). This supports the argument of Myers & Majluf (1984) stating that firms with higher growth opportunities have larger agency cost of debt. As for the other variables, INTA and NDTS are correctly signed but the estimated coefficients of these variables are insignificant. This could indicate that the proxies used in this study are not able to detect asset-specificity and non-debt tax shield considerations. However, it must be noted that other studies on the subject also find insignificant coefficient for INTA and NDTS (Brailsford et al., 2002). The coefficients of VOLTY and FCF do not show their expected signs. Moreover, both are insignificant at 5 percent level of confidence. Overall the empirical model explains 30 percent of the variation in DEBT which is in line with previous studies of Brailsford et al. (2002)
4.3.2 Analysis 2: External block holder ownership
Based on existing literature, we hypothesize that external block holder ownership has a positive influence on debt-ratios due to the active-monitoring argument. We test this with regression equation (3) where debt ratio is regressed on external block holder ownership (EBO). Table 8 in Appendix E presents the results of this regression. The estimated coefficient on EBO is shows negative sign which implies that the relation between external block holder ownership is negative. This indicates that debt and large external shareholder ownership substitute each other instead of complementing. The increased external block holder ownership
22
alleviates the need for higher levels of debt to control managers. However, the estimated coefficient on EBO is insignificant. Thus, indicating that external block holder ownership does not affect debt ratios at all. This suggest that the neither complementing nor the substituting effect of external block holders on debt is present in this dataset. Pound (1988) also find no significant relation between external block holders and debt ratios. We must reject hypothesis 2 based on the results of equation (3).
Equation (4) is formulated to test the arguments of Short at al. (2002) which suggest that the influence of external block holders differs for small and large external block holder ship. This equation included ownership variables for large and small external block holders. The results of equation (4) is reported in table 8 in Appendix E. The estimated coefficient for EBONOBLOCK is positive and significant at =0.05. Furthermore, the coefficient for EBOBLOCK is positive but insignificant. These findings indicate that the positive effect of external block holders on debt ratios is indeed only present for small external block holdings at levels below 10 percent. Therefore, we cannot reject hypothesis 3 completely. The results of equation (4) imply that there is a positive relation between external block holder ownership and debt-ratios. However, only in the case of small external block holders. Hypothesis 4 cannot be rejected based on the results of equation (4)
With regards to the control variables, signs and significance on all control variables remain similar as in equation (1) and (2) except FCF. The estimated coefficient on FCF is positive and significant at =0.05 Furthermore, the explanatory power of regression (4) is less than the regression considering managerial ownership with a R-squared of 22%.
4.3.3 Analysis 3: Interaction effect
The previous chapters examined the relation between managerial ownership, external block holder ownership and debt ratios separately. In the last stage of the empirical analysis, we include interaction variables MSONOBLOCK and MSOBLOCK to analyze the possible interaction effect external block holders have on the relationship between managerial ownership and debt ratios. With equation (5), we test hypothesis 5 which states that the positive relationship between managerial ownership and debt-ratios becomes insignificant in the presence of external block holders. The results of regression equation are presented in table 9 in Appendix F. The estimated coefficient on MSOB is positive and insignificant. The coefficient on MSONOB is negative and significant at a significance level of 10 percent. These results indicate that the negative relationship between managerial ownership and debt-ratios observed in the results of equation (5) only exist for firms in the absence of large external block holders. The presence of external block holders seems to negate the negative effect of managerial ownership on debt-ratios. This means that we must reject hypothesis 4. Regarding the control variables, estimated coefficients are slightly different as in equation (2) and (4) However, significance and signs on the coefficients again remain similar.
23 4.4 Robustness check
The dataset used in our empirical model includes pooled multi-year observations from the period 2013-2018. Time-series correlation between individual year observations could cause observations to not be independent, which can result in biased regression estimates. We split the data sample in to two sub-periods, to examine the robustness of the regression results. One sub-sample consists of observations from 2013-2015 and the other sub-sample includes observations from 2013-2015-2018. All regression equations are run again with the sub-samples. The signs on the variables of interest in both sub-samples show similar signs as found in the regressions on the full sample. In the second sub-period, coefficients show similar signs and significance as reported in the results of the full sample regressions except for EBO. The estimated coefficient on EBO becomes significant at a confidence level of 10 percent which was not the case in the full sample regression. The significance on the variables in question varies in the first sub-period sample. The coefficients on EBO, EBONOBLOCK, EBOBLOCK and MSOBLOCK show predicted signs but become insignificant. All other variables stay significant and show consistent signs as the full sample. These results might imply that the association between capital structure and ownership structure varies over time. However, varying results may be caused by smaller sample size as the formed sub-samples are relatively small.
Another issue could be correlation between explanatory variables. Significant correlation may lead to a problem of multicollinearity and in turn biased coefficients. A method of testing the influence of multicollinearity is excluding significantly correlated variables. However, none of the explanatory variables are significantly correlated as observed in the correlation matrix and VIF values.
Subsequently, we examine the robustness of results regarding alternative variable definitions. Prior literature on the subject have different operationalizations of the ownership variables. For managerial ownership, this study mirrors the operationalization of Brailsford et al. (2002) with managerial ownership (MSO) measured as the share ownership of all directors. This is also done by Angrawal & Mandelker (1987). The studies of Friend and Lang (1988) and Mehran (1992), in addition use the share ownership of the top five directors as proxies for managerial ownership. We re-run the empirical models with the alternative proxies for managerial ownership. The estimated coefficient on the variables of interest becomes insignificant. This together with the low significance of the original definition of managerial equity ownership may be reasons to be careful when drawing conclusions on the effect of managerial equity ownership on capital structure. Furthermore, we also include alternative ownership proxies for external block holder ownership. Top two and top ten largest external shareholders are used to measure external block holder ownership (EBO). After re-estimating, signs and significance remains consistent.
24
5.
Discussion & Conclusion
This study re-examines the relation between ownership structure and capital structure to contribute to the existing literature. Since there is no consensus on the effect managerial ownership and external block holder ownership have on debt-ratios. This study analyzes the effect of ownership structure on capital structure for firms located in Germany. German data samples have not yet been examined which could lead to new insights on the relation between ownership structure and capital structure. Especially since German markets differ from their Anglo-American counterparts. Furthermore, the chosen time frame after the recent financial crisis could give us updated view on the relation between ownership structure and capital structure. In order to provide an answer to the research question, we formulated hypotheses to determine the effect of managerial ownership, external block holder ownership and possible interaction effects. This chapter will discuss the results found in the previous chapter and elaborate on the hypotheses. Moreover, findings will be compared to research results of prior studies on the subject. And lastly, we provide a concise conclusion and discuss the contributions, possible limitations and recommendations for further research.
5.1 Discussion
The results of the first analysis show a linear negative relationship between managerial ownership and debt ratios. The proposed non-linear relationship by hypothesis 2 is not present in our dataset. The robustness test on these results suggest that we should be cautious assuming that managerial share ownership negatively influences leverage for German listed firms. A reason for this may the limited data available on managerial ownership which lead to a relatively small sample size. The reported negative relation is not in line with the incentive-alignment hypothesis put forward by this study. This incentive-alignment argues for positive relationship. In the case of German firms, the estimated negative coefficient indicates that managerial equity ownership and debt are substitutes in the agency framework. Increased levels of managerial ownership alleviate the need for higher leverage which explains the negative relationship. An alternative explanation is provided by the management entrenchment hypothesis. This theory suggests that increasing managerial ownership changes the agency cost of debt and equity. Since managers have invested their non-diversifiable human capital in the firm, increased managerial equity ownership could therefore increase the agency cost of debt and lower the agency cost of equity. Their risk aversion may cause them to pursue lower debt-ratios to sub-optimal levels to ensure their employment and wealth. The observed negative relationship is in line with the results reported by Friend & Lang (1988), Firth (1995) and Bathala et al. (1994) which examined firms in the US, but contradicts the empirical evidence of Kim & Sorenson (1986), Agrawal & Knoeber (1996), Berger et al. (1997), Short et al. (2002) and Brailsford et al. (2002). The contradicting results may be accounted to varying research designs and time periods. Furthermore, the lack of earlier literature
25
examining the relation between managerial ownership and capital structure in German or other continental European counties makes the comparison of the results difficult.
Subsequently, in the second analysis the relationship between external block holder ownership and debt-ratios is examined. Hypothesis 3 predicted a positive relation between external block holder ownership and leverage. The estimated coefficient on EBO indicates that the relation between external block holder ownership and debt-ratios is negative. This is in line with the study of Short et al. (2002). They argue that debt and external block holder ownership are substitutes. Increasing external block holder ownership mitigates the need for debt as a disciplining measure. The insignificant results however imply that external block holder ownership does not influence debt-ratios at all. Therefore, questioning the existence of the active monitoring effect of external block holders. The only empirical work supporting an insignificant relation between external block holders and debt-ratios is the study by Pound (1998) which questions the efficiency of external shareholder oversight. This is confirmed by Zeckhauser & Pound (1990) who investigated US markets. Based on these findings, we must reject hypothesis 3.
The relationship between external block holder ownership on debt-ratios may differ for small external block holdings and large external block holdings (Short et al., 2002). Since external block holdings in Germany are higher than the Anglo-American countries, the difference between small and large block holders might be more prevalent in this dataset. The results show that the positive association between external block holder ownership and debt-ratios is only present with small external block holders. This positive association is in line with the prior studies (Friend & Lang, 1988; Mehran, 1992; Berger et al., 1997; Brailsford et al., 2002; Firth, 2002). The positive correlation however becomes insignificant with external block holder share ownership above 10 percent. Increasing equity ownership may reverse the incentives of block holders. Their increased risk exposure may cause them desire lower debt-ratios to avoid chances of bankruptcy. Moreover, increased ownership could mean more discretion over management via control mechanisms which reduces the need for additional debt as a discipling device. External block holder ownership in this case is assumed to be a substitute for debt. These effects may cause the relation between external block holder ownership and debt-ratios to become insignificant. The results of equation (4) may explain the insignificant relation found in equation (3) as the significant positive relationship is not present for large external block holders. Short at al. (2002) also find differences between the effect large and small external block holders have on debt-ratios. They, however, report a negative correlation between external block holder ownership and debt-ratios. Their results suggest that firms with large external shareholders have smaller debt-ratios than firms with no large external shareholders. This again could indicate that increased ownership may inverse the incentives of external block holders.
Lastly, the interaction effects between managerial and external block holder ownership are examined to determine whether external block holder ownership indirectly influences debt-ratios. The
26
results imply that external block holders moderate the financial decision by management. The negative relation between managerial ownership and debt-ratios is only present for firms without external block holders. These findings are in line with the arguments of Friend & Lang (1988), Friend & Hasbrouk (1988) and Firth (1995). They assume that the risk aversion of managers paired with increased managerial equity ownership leads to incentives for managers to decrease debt-ratios. Thus, resulting in a negative relationship. Managers discretion to set sub-optimal debt-levels is constrained when external shareholders hold substantial stock (Firth, 1995). When external block holders prefer high debt-ratios, large external block holder ownership seems to force management being less risk averse. Thus, aligning their incentives which in turn could decrease the agency cost of debt and increasing the agency cost of equity. This in turn may increase the use of external debt by management. Thus, aligning the incentives of external shareholders and management. However, this effect is not enough to completely align interest between both parties. The managerial ownership does not become positive and significant in the presence of large external shareholders. Large external shareholders seem to only negate the entrenchment effect of managerial ownership. Short et al. (2002) also find that large external block holder ownership negates the relation between managerial ownership and debt-ratios. They however report a positive correlation between managerial ownership and debt-ratios in UK firms.
5.2 Conclusion
This study contributes to the existing literature to better understand the relationship between ownership structures and capital structure. This conclusion will try to answer the research question: To what extent does equity ownership affect a firm’s capital structure? This study has examined two types of equity ownership: managerial ownership and external block holder ownership. The results of the empirical analysis indicate that, for German firms, debt-ratios are negatively correlated to managerial ownership which supports the entrenchment hypothesis. Furthermore, we found that external block holder ownership is positively related to debt-ratios based on the sample of German listed companies. This supports the argument that external block ownership and debt ratios are complementary measures. In addition to the two categories of ownership, this research analyzes the possible interaction effects between managerial and external block holder ownership. The results show that the negative relation between managerial ownership and leverage is negated by the presence of large external shareholders. These findings are in line with the studies of Short et al. (2002), Firth (2005) and Friend & Lang (1988) which hypothesize that large external block holders influence the relative size of agency cost of equity and debt.
Some limitations are involved when reviewing this study, these can provide a foundation for further research. Firstly, the sample size is relatively small. Especially when comparing to prior research on the subject. Small sample size could make the results more susceptible to missing data and changes in variables
27
definition. This could be seen in the robustness checks on the isolated effect of managerial ownership on debt-ratios. When slightly changing the operationalization of managerial ownership variables, the negative effect on debt-ratios becomes insignificant. Increasing the sample with multiple European countries could solve this. Furthermore, it may give us more insight on the relation between ownership and capital structure in European countries.
Secondly, the use of cross-sectional regression may not be optimal to determine the effect of ownership on capital structure. In addition, to relieving some of the econometric problems of cross-sectional analyses, panel data may better explain the dynamic relation between ownership and capital structure through time. As earlier studies on the subject also use cross-sectional analysis, further research could contribute by adopting panel data in the empirical models. This may add information on the relationship between ownership structure and capital structure.
Thirdly, some authors argue that single-equation models are not suitable to determine the effects ownership structure has on leverage. Several variables related to ownership structure and capital structure may be simultaneously determined which could imply an endogeneity problem. This study assumes that ownership structure influences debt-ratios exogenously. However, as debt-ratios is determined by other factors, debt-ratios may also play a role in the composition of equity ownership. This results in possible endogeneity of ownership structure. Thus, questioning the direction and causality of the relation between ownership and leverage. Simultaneous equations models such as 2SLS may be more appropriate to analyze the relationship between ownership and capital structure as these could account for endogeneity. Nevertheless, the examination of German listed firms has provided us with some new insights on the relationship between ownership structure and capital structure. The findings suggest that the equity ownership structure and capital structure are indeed interrelated. This paper contributes to solving capital structure puzzle by adding to the understanding of cross-sectional variation in capital structure caused by equity ownership structure. This could benefit the financial efficiency of firms and in turn firm value. However, the contradicting results with prior studies indicate that more investigation into the effects of ownership on capital structure is needed.
