Influence of a corporate governance reform on the leverage level : German Governance Code case study

(1)

Influence of a corporate governance

reform on the leverage level

German Governance Code case study

Abstract

This study investigates the influence of a corporate governance reform on the leverage level. Evidence comes from the passage of the German Corporate Governance Code (2002). Arping and Sautner (2009) argue that corporate governance can be used as a device to mitigate agency problems. Prior studies from Zwiebel (1996) and Grossman and Hart (1982) state that high leverage ratios can also reduce the agency problems of firms. This paper argues that debt financing and a corporate governance improvement are substitutes when it comes to mitigating agency problems. Hence, leverage level would decrease after a corporate governance reform. A fixed effect model based on firm-specific characteristics and a variable that represents a system-wide corporate governance reform is used to distinguish this relationship. Given the German firm characteristic of being bank-orientated, this study additionally investigates if the German Governance Code had an influence especially on the proportion of bank debt of German firms. Although some statistical significance is found to argue a negative relation between the leverage level and the passage of the German Governance Code, no final conclusion can be drawn due to the fragility of the results.. Besides, there is no significant effect found on the bank debt level of the German firms used in the sample.

Edited by: B. Takacs, 10595783 BSc Finance & Organization January 17, 2017

(2)

2

Statement of originality

This document is written by Student Bianca Takacs who declares to take full responsibility for the contents of this document.

I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.

The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.

(3)

3

1 Introduction

Armour and McCahery (2006) state that the Enron scandal caused not only discussion about the effectiveness of the current corporate governance in the U.S., but also raised doubts about the corporate governance systems worldwide. Researchers blame the agency problem for the bankruptcy scandal of Enron. This created a public perception of firm managers acting in favour of their benefits rather than the benefits of the stockholders (Jiraporn et al. 2008). Resulting in policy and regulation questions to mitigate these problems. Regulators introduced the Sarbanes-Oxley Act to align interests of shareholders and managers. In the same year as the Sarbanes-Oxley Act, Germany formulated the German Governance Code.

Prior studies from Berger and Di Patti (2006) and John and Senbet (1998) found that corporate governance improvements could mitigate the agency problems of firms. A relevant study from Zwiebel (1996) looked into the effect of leverage on agency problems and found that high leverage ratios often lead to less agency problems. Hence, this paper argues that leverage level and corporate governance substitutes in mitigating agency problems. Based on the agency theory this paper will investigate the association between corporate governance and leverage level. This leads to the central research question: What is the influence of a corporate governance reform on leverage level of firms?

Evidence comes from the effect of the passage of the German Governance Code (2002) on the corporate leverage of German listed companies. The Governance Code Commission (2002) formulates the purpose of the German Corporate Code as follows:

“The Deutscher Corporate Governance Kodex (German Corporate Governance Code) presents essential statutory regulations for the management and supervision of German listed companies and contains, in the form of recommendations and suggestions, internationally and nationally acknowledged standards for good and responsible corporate governance."

Source: German Corporate Governance Code (Governance Code Commission, 2002)

In addition to a similar study of Arping and Sautner (2009), who found that Dutch firms significantly reduced their leverage after the passage of the corporate governance reform in the Netherlands, this study focuses on evidence from Germany.

(5)

5

This study highlights some German firm characteristics and provides evidence to answer the question if the reform has the expected negative effect on leverage level, even though the corporate governance system of Germany is already seen as one of the best worldwide (Schleifer and Vishny, 1997).

John and Kedia (2006) and Antoniou et al. (2008) classify Germany as a bank-oriented country. Not only because of the relatively large amounts of bank debt the firms hold but mainly based on the corporate control the banks have within German firms. This paper investigates whether the German Corporate Governance has a significant effect on the level of bank debt of German firms.

This paper contributes to the existing literature by providing practical evidence to show the significant economic relation between leverage level and corporate governance by an econometrical analysis. Second, instead of just investigating the leverage level of German firms this research also focuses on the portion of bank debt of German firms. Third, this study provides an analysis of the influence of the corporate governance code on other firm characteristics than just the primary variable of interest; leverage. Finally, next to the traditional theories from Jensen and Meckling (1976), Modigliani and Miller (1958) and Schleifer and Vishny (1997) on agency theory, capital structure, and corporate governance, this study adds new insights from more recent studies from Jiraporn et al. (2011) and Arrawal et al. (2010)

The primary objective is to fill the gap between theory, predictions and the practical effect of the passage of the German Corporate Governance Code. A fixed effects model is used to investigate the effect of a corporate governance reform on leverage level of German firms. This model derives from Arping and Sautner (2008) by including the same explanatory variables and the variable of interest that represents the passage of the German Corporate Code. By using panel data empirical support is found for a negative influence of the German Corporate Governance Code on leverage level. However, this is strongly depending on how leverage is measured. The main dependent variable in this study, the debt to capital ratio, is significantly affected by the passage of the corporate governance code. However, from the regressions run to check for robustness only two regressions provide statistically significant evidence. Hence, we cannot draw a final conclusion about the relationship between a corporate governance reform and leverage level. The results of this study partly follow the findings of Arping and Sautner (2008). However, the robustness checks show how fragile the outcome of the main regression is.

(6)

6

Besides, no statistically significant evidence is found for the effect of the German Corporate Governance Code of the level of bank debt of German firms.

This study is organized as follows. Section 2 reviews the background literature, which leads to the hypotheses. Section 3 describes the data and includes the methodology that is used to test whether the German Governance Code has a significant impact on leverage level. Section 4 contains the data analysis and presents the results. Section 5 concludes and provides some suggestions for further research.

2 Literature review

An extensive amount of literature has been written on the agency problem, capital structure, and corporate governance. The aim of this section is to provide a theoretical background and explain the insights given by the prior literature, which leads to the formulation of the hypotheses. First, the agency problem will be described followed by a theoretical background of capital structure and corporate governance, which are argued to be ways to mitigate the agency problems. Additionally, German institutional factors will be discussed as well as the German Corporate Governance Code.

2.1 Agency problem

The essence of the agency theory is the problem that arises when ownership and control are separated. John and Senbet (1998) state that, given the presence of well-functioning capital markets, managers can be entrusted to make decisions regarding investments or production to maximise firm value. In practice, the economy is characterized by imperfect information, agency conflicts, and imperfect markets. La Porta et al. (2000) state that the agency problem occurs when insiders, who control the corporate assets, use corporate assets for purposes other than beneficial to the interests of outsiders1. When these agency problems occur, the agency costs of outside ownership equal the lost value from inside managers maximizing their utility, rather than the firm value (Berger and Di Patti, 2006). Agency costs can also be defined as the sum of the monitoring expenditures by the principal, the bonding expenditures by the agent and the residual loss (Jensen and Meckling, 1976).

1_{Examples are self-diverting of the assets, dilution of outside investors through share issues to insiders, excessive} compensations, asset sales to themselves, favourable transfer pricing to self-controlled corporations (La Porta et al. 2000)

(7)

7

John and Senbet (1998) outline two classes of agency problems. Firstly managerialism, which refers to self-serving behavior, can lead to different conflicts. It can lead to ‘empire building' at the expense of the capital investments of the owners and to undertaking conservative investments. These investments are in the form of seeking safe, but inferior projects to maintain the safety of wage compensation and are not at risk of the manager’s tenancy. Secondly, agency costs of debt arise when managers can make sub-optimal decisions, which differ from the value maximizing decisions, given by the debt contract.

Another problem that might occur is the free cash flow problem. The essence of the free cash flow problem is that managers must trade off the private benefits of current spending against accumulating excess cash reserves that benefit the firm. The conflict derives when managers are motivated to waste excessive cash, cash flow in excess of that required, on organization inefficiencies (Jensen, 1986). The following sections argue that corporate governance and debt financing can both serve as a device to reduce these problems and the costs that come with it.

2.2 Capital structure

Capital structure theories attempt to explain the mix of equity and debt used by corporations to finance real investments. Since the work of Modigliani and Miller (1958) capital structure has been a topic of interest. There is no universal theory of the debt-equity choice, however there are several relevant theories provided by studies after the work of Modigliani and Miller (Myers, 2001). Examples are the trade-off theory, pecking order theory and the agency theory. This paper will mainly focus on the relationship between agency theory and capital structure and specifically on how the leverage level of a firm can reduce agency problems.

Berger and Di Patti (2006) state that the choice of capital structure may help mitigating agency costs. They conclude that high leverage ratios reduce the agency costs of outside equity and increase firm value by motivation managers to act more in the interests of the stakeholders. Also, Jensen and Meckling (1976) state that agency theory postulates that capital structure is determined by the amount of agency costs.

There are multiple theories on how a higher leverage ratio can reduce agency problems. Firstly, the increasingly use of debt financing often increases the probability of bankruptcy. The risk of financial distress, in case a firm is unable to fulfill its interest obligation, combined with a potential loss of the managers’ benefits may motivate managers to limit their consumption and can lower the agency costs (Grossman and Hart, 1982).

(8)

8

Secondly, debt financing can be beneficial in reducing agency conflicts in the free cash flow problem. Debt financing comes with obligated interest payments that reduce the free cash flow, hence might reduces the agency costs of a firm. Zwiebel (1996) developed a model in which debt is used as a commitment device to reduce agency problem. In this model managers voluntarily chose debt financing to constrain their future empire building.

Also, Jiraporn et al. (2011) argue that leverage and corporate governance can both be used as a device to alleviate agency costs. The relationship between corporate governance and agency costs will be discussed in the next section.

2.3 Corporate governance

2.3.1 Defining corporate governance

Schleifer and Vishny (1997) describe corporate governance as an appliance for suppliers of finance to corporations to assure themselves of a return on their investment. Cromme (2005) defines corporate governance as a description of good, efficient management and supervision of firm, which is based on internationally recognized standards in the interests of the firm's owners and its social environment. John and Senbet (1998) state that corporate governance deals with mechanisms by which stakeholders are able to exercise control over the corporate insiders so that their interests are protected. Stakeholders fulfil as equity holders, creditors and other stakeholders such as employees, consumers, suppliers, and the government. The powers given to stakeholders by corporate governance can range from the right to receive the same per share dividends as the insiders, to the voting right regarding important corporate matters, to the right to take a damage caused by insiders to court (La Porta et al. 2000). Arrarwal et al. (2010) contribute by stating that governance depends on both country-level as well as firm-level mechanisms. Country-level mechanisms include culture and norms, law, and the institutions that enforce the law, whereas firm-level mechanisms are those that operate within the firms. The existing corporate governance mechanisms are best described as economic and legal institutions that can be adjusted through the political process.

Schleifer and Vishny (1997) argue against a traditional evolutionary theory that product market competition would result in firms automatically adapting to rules, including corporate governance mechanisms, as part of forced cost minimization. Hence, competition would enhance corporate governance, and there is no need for reforms.

Schliefer and Vishny (1997) conclude that solving the agency problem requires more than competition and improving the corporate governance mechanisms could be a solution.

(9)

9

Additionally, John and Senbet (1998) conclude that effective corporate governance and effective contracting among various parties with a diverse interest fosters economic development and enhances the effective allocation of resources promoted, therefore reduces the agency problems of firms. Moreover, Berger and Di Patti (2006) outline the use of corporate governance as a mitigating device for agency problems by stating that better corporate governance reduces agency costs in both financial and nonfinancial firms.

2.3.2 Influence of corporate governance on leverage

The association of corporate governance and agency theory is frequently discussed in previous research. A substantial number of studies used corporate governance measures to measure agency problems.2 Jiraporn et al. (2011) state that the level of sub-optimal leverage managers can take are depending on the strength of corporate governance, hence corporate governance quality influences capital structure decisions. Jiraporn et al. (2011) conclude that firms with poor governance, where governance quality is measured by broad-based governance metrics, are significantly more leveraged.

La Porta et al. (2000) first document the substitution hypothesis, which is further implemented by Jiraporn et al. (2011). The essence of the substitution hypothesis is that firms with weak governance have a stronger need to establish a reputation for not expropriating wealth from shareholders. A way to create this reputation is to keep more debt as fixed interest payments reduce what is left for expropriation. This conclusion is based on empirical results that show that an improvement in the governance index from the 25th to 75th percentile, leads to a decrease of leverage by 12,88% (Jiraporn et al. 2011).

Additionally, Liu and Miao (2006) found that stronger corporate governance was associated with higher equity value, less concentration of ownership and lower leverage levels. They also found that in the case of imperfect corporate governance leverage increases. This conclusion is based on the assumption that in the case of imperfect corporate governance the market price of risk increases and managers tend to consume more. Resulting in a manager’s preference towards debt when it comes to investment decisions. The outcome of this research of Liu and Miao (2006) is that in the case of imperfect corporate governance the leverage level of firms increases.

2_{Gompers, Ishii and Metrick (2003) used the Governance index, and so did Jiraporn and Gleason (2007). Harford et al.} (2006) used multiple governance measures such as ISS governance metrics and GI.

(10)

10

2.4 German institutional factors

2.4.1 Corporate governance in Germany

Schleifer and Vishny (1997) state that the United States, Germany, Japan, and the United Kingdom do not only contain the essential elements of good governance systems, but they are also even described as the best corporate governance systems worldwide. They argue that the differences in corporate governance systems partly result from the differences in the legal obligations that managers have to financiers and partly on how courts interpret and enforce these obligations.

Board effectiveness is an essential component of corporate governance (John and Senbet, 1997). Evidence shows that the board characteristics within these four best corporate governance systems differ. For example, Germany and Japan seem to have passive boards, except in extreme circumstances, compared to the others (Kaplan, 1994). Another characteristic for Germany is the voting arrangements with large commercial banks. These banks often control over a quarter of the votes in major companies (Franks and Mayer, 1994). In Germany and Japan, the influence of banks is significant, caused by the dominant role the banks play in lending and the direct role through board membership. Schleifer and Vishny (1997) conclude that, regarding the differences, these best corporate governance systems have the combination of large investors and a legal system that protects investor rights in common.

2.4.2 German leverage level

Rajan and Zingales (1995) provide some international empirical data regarding capital structure. According to this study at an accumulated level, firm leverage is considerably similar to the G-7 countries3_{. The side note to this finding is that the UK and Germany are}

relatively less levered. Besides, Rajan and Zingales (1995) conclude that German firms depend substantially more on debt than equity issuance than the other G-7 countries.

A more recent study form Antoniou et al. (2008) investigates the differences between capital market-oriented and bank-oriented economies. This classification is based on the financial and institutional traditions, whereas Germany is classified as a bank-oriented economy.

3_{The IMF reports the Group of 7 (G7) as the seven major advanced economies. This group consists of Canada, France,} Germany, Italy, Japan, UK, and the US.

(11)

11

According to Antoniou et al. (2008) capital, market-oriented economies contain high transparency and investor protection, for example, the U.S. and the U.K, whereas bank-oriented economies include countries with lower transparency and investor protection. Table 1 from Antoniou et al. (2008) summarises the major institutional factors from the G-5 countries. 4

Figure 1: Table 1 from Antoniou et al. (2008)

Source: Antouniou et al. (2008)

According to Agarwal and Elston (2000) close bank-firm relationships developed in universal banking systems like Germany, result in easier access to bank debt.5_{Agarwal and Elston}

(2000) outline the three primarily ways that German banks affect firm behaviour. Firstly, as a creditor, through loans to the firms the bank monitors the process of capital allocation within the firm. Secondly, as representation on the supervisory board (Aufsichtsrat), which provides a strong channel of information in both directions. Finally, banks are allowed to exert control through voting rights. Hence, theoretically close bank-firm relationships might lead to a higher level of leverage.

Cromme (2005) contributes to the existing literature by outlining the two main sources of funding from German companies: retained earnings and loans.

4_{The G-5 countries consist of France, Germany, Japan, UK and US.} 5_{See Gerschenkron (1966) and Hoski et al. (1994).}

(12)

12

This started changing in the 1990s together with the globalization of the economy, the international competition and the liberalization of the financial markets. From this point, German firms started to rely more on the international capital market as a source of funding. Figure 2 shows the evolution of leverage over the period 2000-2006. Leverage is measured as the average debt-to-capital ratios for the firms used in the first sample of this research. This figure shows a decrease in leverage after the year of the passage of the German Governance Code. This research will investigate if this decrease is attributable to the passage of the German Governance Code.

Figure 2: Evolution of leverage over time. Leverage is measured as the average of the debt-to-capital ratios of the firms used in the first sample of this research.

2.4.3 The German Corporate Governance Code

Shortly after high regarded firms like Enron Corporation and Wordcom filed for bankruptcy the Sarbanes-Oxley Act was formulated. In The Sarbanes-Oxley Act (SOX) the Congress introduced a series of corporate governance initiatives into the federal securities laws (Romano, 2004). The main incentive of The Sarbanes-Oxley Act is to prevent balance sheet manipulation and restore investor confidence and is mainly focused on the oversight of auditors and corporate governance at listed firms (Cromme, 2005).

Cromme (2005) argues that German governance is strongly influenced by developments in the USA and, as the debate over corporate governance continued, Germany formulated the German governance code in the same year as The Sarbanes-Oxley Act.

0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 2000 2001 2003 2004 2005 2006 D eb t-to-c ap ital r ati o Year Evolution of leverage

(13)

13

Cromme (2005) states that the German understanding of corporate governance was underdeveloped and lacking in transparency for a significant amount of time, which was the starting point of a corporate governance debate in Germany. In 2001 the Government Commission of the German Governance Code was appointed to communicate the code to the financial community.

The Government Commission outlined the key objective of the Code as a boost of confidence in the management of German companies. The German Governance Code should serve as a guideline on both German and international investors, by setting out the particularities of German business in line with the capital market and match them to international standards.

Cromme (2005) states that the key function of the German Governance Code is to create transparency, as there is no better form of control than transparency. More transparency could lead to the mitigation of the agency problem, therefore could reduce the level of leverage when leverage is used as a device to alleviate agency problems.

The German Governance Code should not be seen as legislation but is based on recommendations. German firms are required to declare whether they comply with the recommendations of the Code. Appropriate Management Board compensation is one of the key content areas of the Code and is set out in criteria for appropriateness of compensation. The main incentive is that executives must be compensated in line with their individual performance. Another key component of the Code is the role of Supervisory Boards. The function of the German Supervisory Boards is to oversee the firm’s management. The Code states that there should be an increase in committee work of the members of German Supervisory Boards. Additionally, the Code states the need of an increase in preliminary discussions about issues facing both the stockholders and the employee representatives within the Supervisory Boards (Cromme, 2005).

2.5 Hypotheses

This paper follows the argument of Gillian (2006) that debt financing can be seen as a self-enforcing governance mechanism and can be substituted by external governance mechanisms. Debt financing and the improvement of corporate governance can both mitigate the agency problems of a firm. The aim of the passage of the German Governance Code is to increase the corporate governance quality of German firms.

(14)

14

Based on the expectation that firms with high-quality corporate governance have lower levels of debt this paper argues there ought to be a significant negative relation between leverage and the passage of a corporate governance code. This leads to the first hypothesis of this paper:

H0: The German Governance Code had no significant effect on the leverage level of German firms.

H1: The German Governance Code had a significant negative effect on the leverage level of German firms.

As explained in the prior section Germany is classified as a bank-oriented country. Antoniou et al. (2008) conclude that bank loans are the main financing device for German firms. Therefore it is interesting to investigate if the passage of the German Governance Code had an influence on the proportion of bank debt of German firms. This leads to the second hypothesis of this paper:

H0: The German Governance Code had no significant effect on the amount of bank debt of German firms.

H1: The German Governance Code had a significant effect on the amount of bank debt of German firms.

3 Methodology

To analyse the effect of the German Governance Code on leverage level and the proportion of bank debt of German firms, a fixed effect model is used. Firstly, the data used in the samples will be described. Secondly, the model will be specified and discussed. Finally, this section provides expectations about the relation between the explanatory variables used in the model and leverage level based on existing theory.

3.1 Data

To analyse the effect of the German Governance Code on the leverage level panel data will be used. This data is obtained from DATASTREAM and ORBIS. The sample exists of German DAX and MDAX listed firms.

(15)

15

Following previous research, firms whose capital decision may reflect special factors, like financial and regulated firms, are excluded. Eliminating firms whose accounting information is not sufficient to estimate the variables used in the model reduces the sample. The sample is also reduced by excluding the firms that do not have data available for the used timeframe. There are two final samples. The first sample consists of 40 firms, both from DAX and from MDAX indices and is used for the first model. The second sample is used to estimate the influence of the German Governance Code on the proportion of bank debt. This final sample consists of 33 firms where information about the amount of bank loans the firms hold is available. Each variable consists of six yearly observations from 2000 up to 2006. The year 2002 is excluded to serve as a reference year. The descriptive statistics of the first and second sample after eliminating the missing values and outliers are summarised in table 3 and 4 respectively.

Here we can see that the number of valid observations is equal to 240 and 198 for sample one and two respectively. Notable is the big variation between the minimum and maximum value of the variable MB, which stands for growth opportunities. This is caused by the values of the firm SAP. Considering the industry of this firm and high book to market value throughout all the years there is no need to exclude it from the sample.

Table 1. Descriptive statistics of sample one

Statistics Mean Maximum Minimum Std. Dev. Observations

BOOK 0.212 0.542 0.001 0.135 240 MARKET 0.981 1.000 0.351 0.076 240 LT/ASSETS 0.143 0.413 0 0.100 240 DEBT/CAP 0.354 0.778 0.17 0.198 240 LT/CAP 0.238 0.599 0 0.151 240 SIZE 15.948 19.148 12.820 1.651 240 MB 1.939 18 0.38 1.885 240 LIQ 1.589 4.91 0.67 0.689 240 TANG 0.281 0.584 0.060 0.111 240 PROFIT 0.122 0.370 -0.057 0.057 240 POST 0.667 1 0 0.472 240

Table 2. Descriptive statistics of sample two

Statistics Mean Maximum Minimum Std. Dev. Observations

BANK1 0.136 2.155 0 0.343 198 BANK2 23.489 292.438 0 48.011 198 SIZE 16.043 19.138 12.821 1.723 198 MB 1.790 18 0.38 1.792 198 LIQ 1.563 4.91 0.67 0.689 198 TANG 0.285 0.584 0.060 0.114 198 PROFIT 0.118 0.370 -0.057 0.058 198 POST 0.667 1 0 0.114 198

(16)

16

This study looks into variables from multiple firms over multiple years, therefore based on panel data. Antoniou et al. (2008) outline that panel data increases the degrees of freedom, reduces the possibility of multicollinearity among the explanatory variables and consequently leads to more efficient estimates.

In appendix 1 the firms remaining in the final two samples are listed.

3.2 Fixed Effects Models

When analysing panel data either fixed effects or random effects regressions are used. With fixed effects, time-invariant factors are controlled for. With random effects, the regression assumes the variation across companies is random and uncorrelated with the dependent and independent variables. This study focuses on the variables that vary over time. Hence, to control for the time-invariant factors the fixed effects regression model is used (Stock and Watson, 2012, p.402). In this way, unobserved variables that can influence the estimated effect of variables in the model, such as firm-specific preferences, are eliminated. This reduces the threat of omitted variable bias.

Two models are used in this research and both derive from Arping and Sautner (2009). The first model to estimate the leverage level is specified as follows:

𝐿𝐸𝑉𝐸𝑅𝐴𝐺𝐸_!" = 𝛼 + 𝛽_! ×𝑃𝑂𝑆𝑇_!" + 𝛽_! × 𝑀𝐵_!" + 𝛽_! × 𝐿𝐼𝑄_!" + 𝛽_! × 𝑃𝑅𝑂𝐹𝐼𝑇_!" + 𝛽_! × 𝑆𝐼𝑍𝐸_!"+ 𝛽_! × 𝑇𝐴𝑁𝐺_!" + 𝜀_!"

In this equation the subscript i is the index for the firms and t is the time index. Leverage is modeled as a function of a variable that represents the passage of the German Governance Code combined with various firm-specific factors that appear regularly in prior literature and studies. These variables should account for possible determinants of capital structure choices. Rajan and Zingales (1995) found that the best way to measure the dependent variable leverage for Germany is to use the debt-to-capital ratio. Pension liabilities are often treated differently in accounting terms in Germany compared to other countries; therefore the debt-to-equity ratio is the most accurate leverage measurement.

(17)

17

To check for robustness different measures of leverage are regressed separately on the group of variables defined in the model. The different measures of leverage used in this research are book leverage, market leverage, the long term assets ratio, the long term debt-to-capital ratio and the main dependent variable; the debt-to-debt-to-capital ratio.

The second model, which is used to estimate the proportion of bank debt, consists of the same group of explanatory variables as the first model holds.

Bank debt is defined as the proportion of bank loans the German firms hold. Bank debt is measured in two ways; bank loans over assets and bank loans over capital. In this research the main dependent variable bank debt is measured as bank loans over assets. Bank debt over capital is used to provide evidence for the validity of the model. The second model is specified as follows:

𝐵𝐴𝑁𝐾𝐷𝐸𝐵𝑇_!" = 𝛼 + 𝛽_! ×𝑃𝑂𝑆𝑇_!" + 𝛽_! × 𝑀𝐵_!" + 𝛽_! × 𝐿𝐼𝑄_!" + 𝛽_! × 𝑃𝑅𝑂𝐹𝐼𝑇_!" + 𝛽_! × 𝑆𝐼𝑍𝐸_!" + 𝛽_! × 𝑇𝐴𝑁𝐺_!" + 𝜀_!"

In appendix 2 the definition of the depended variables and the sources can be found.

3.3 Independent Variables

Post

Post is the main variable of interest. This dummy variable represents the passage of the German Corporate Governance Code in 2002. The dummy variable is equal to 1 for the period 2003-2006 and 0 for the years before 2002.

MB

The variable MB represents the growth opportunities of the firms and is measured by the market-to-book ratio. Often used indicators for growth opportunities such as price/earnings are partly determined by a firm’s leverage level, however to limit biases due to reverse causality these indicators are excluded. Titman & Wessels (1988) mention that firms with more growth opportunities tend to have more flexibility in their choice of future investment. According to the pecking order theory, that states that firms prefer internal finance to external finance, firms with distinguished growth opportunities might have to raise external capital when the internal resources are not sufficient to finance the investment opportunities.

(18)

18

In the presence of information asymmetries firms prefer to issue equity instead of debt, which would imply a negative relation between growth opportunities and leverage level. Antoniou et al. (2008) state there are differences in disclosure practices and lender-borrower relations in firms operating in bank-oriented and market-oriented countries. Hence, German firms should have a less strong negative coefficient than firms within capital-market-oriented economies due to fewer agency conflicts. This study follows the expectation of a negative relation between growth opportunities and leverage level.

PROFIT

Profitability is measured by the EBITA over total assets. Myers (1984) argues that the pecking order theory implies a negative relationship between leverage and profitability. When internal cash flow is not sufficient the firm will borrow rather than issue equity (Myers, 2001). Less profitable firms have less opportunities to retain earnings for the financing of their activities, thus are expected to attract more external capital (Titman & Wessels, 1988). Especially in closely held firms there is a strong preference not to raise external equity to avoid dilution of ownership structure. Atoniou et al. (2008) argue that bank-based countries like Germany, are closely held compared to market-based nations. Therefore Atoniou et al. (2008) expect an even stronger effect from profitability on leverage levels in Germany.

TANG

Tangibility is measured as net PPE over assets. Arping and Sautner (2009) state that a higher level of tangibility makes it easier for firms to pledge collateral. This often results in an enlargement of debt capacity. Besides, Campello and Hackbarth (2008) argue that an investment in tangible assets helps to relax financing constraints and enhances a firm’s credit capacity. These arguments imply a positive relationship between tangibility and leverage.

SIZE

The logarithm of assets is used as an indicator of size. Rajan and Zingales (1995) provide an explanation for the expectation of a positive relation between leverage increases and size. According to Rajan and Zingales (1995), larger firms are better diversified and have a lower probability of being in financial distress. Hence, lower expected bankruptcy costs enable larger firms to take on more leverage. Also, White (1993) and Kaiser (1994) state that when German firms are entering bankruptcy this usually results in liquidation.

(19)

19

Since liquidation values are generally lower than going concern values, bankruptcy costs are virtually higher in Germany. This implies an even stronger positive relation between size and leverage level in Germany. A more recent study from Beck et al. (2006) finds empirical evidence for the expectation that larger firms report less financing obstacles and have higher leverage levels.

LIQ

The final explanatory variable of the models is liquidity, measured by a firm’s current ratio. Arping and Sautner (2009) argue that a higher current ratio resolves in lower costs of financial distress. This in turn results in an easier attraction of additional leverage, which implies a positive relationship between liquidity and leverage.

Table 3: Expected signs of the estimated independent variable coefficients

In appendix 3 the independent variables and the sources are defined.

4 Results

In this section the results of the analysis, the correlation matrices and some univariate results will be summarised and discussed. First, the regressions to estimate leverage and bank debt will be discussed and in the second section the regressions used as robustness checks will be analysed.

4.1 Empirical results

To check for the possibility of multicollinearity the variable correlation matrices for the first and second sample are presented in table 4 and 5 respectively. The high correlations can be found between the different dependent variables. Hence, it can be assumed that the model will not suffer from multicollinearity and there is no need to exclude one of the explanatory variables from the model.

Variable Expected relation to leverage

Growth opportunities Negative

Profitability Negative

Tangibility Postive

Size Positive

(20)

20 Table 4. Correlation matrix of sample one

Table 4. continued

Table 5. Correlation matrix of sample two

Table 6 shows univariate results to clarify the stasticical significance of the difference between the varaibles before and after the passage of the German Corporate Governance Code. The mean of the variables Debttocap, MB, LIQ, PROFIT, SIZE and TANG for the period pre 2002 and post 2002 are reported in the table with a sample period from 2000-2006. The results show a significant difference of the debt-to-capital ratios of the firms after 2002. This suggests that German firms reduced their leverage subsequent to the passage of the German Governance Code. Table 6 also reports the changes of the explanatory variables after 2002. Size significantly increased after the passage in 2002, which suggest a potential increase in leverage level based on the expected positive relation between size and leverage. Tangibility and growth opportunities significantly decreased after 2002.

Debttocap Booklev Mrktlev Ltcap Ltass SIZE

Debttocap 1.000 Booklev 0.8797 1.000 Mrktlev 0.3859 0.3405 1.0000 Ltcap 0.8845 0.7734 0.3531 1.000 Ltass 0.7858 0.9023 0.3194 0.8846 1.0000 SIZE 0.4249 0.3316 0.0957 0.3769 0.2857 1.000 MB -0.2376 -0.2013 -0.5006 -0.2448 -0.2186 -0.0702 LIQ -0.4993 -0.3952 -0.2784 -0.4010 -0.3035 -0.4504 TANG 0.1658 0.0783 0.1691 0.1652 0.0729 0.2084 PROFIT -0.3064 -0.2151 -0.3853 -0.3268 -0.2522 -0.2543 POST -0.1311 -0.0681 -0.0643 -0.0038 0.0398 0.0472

MB LIQ TANG PROFIT POST

MB 1.0000

LIQ 0.1393 1.0000

TANG -0.1143 -0.1791 1.0000

PROFIT 0.4707 0.2990 0.1959 1.0000

POST -0.2040 -0.0267 -0.0877 -0.0278 1.0000

Bankdebt1 Bankdebt2 SIZE MB LIQ TANG PROFIT POST

Bankdebt1 1.0000 Bankdebt2 0.9759 1.000 SIZE -0.1929 -0.1929 1.0000 MB 0.0825 0.0451 -0.0131 1.0000 LIQ 0.3168 0.2662 -0.4728 0.0943 1.0000 TANG -0.2077 -0.1921 0.2687 -0.1296 -0.2346 1.0000 PROFIT -0.0260 -0.0434 -0.2182 0.4154 0.2909 0.2340 1.0000 POST -0.0938 -0.11038 0.0434 -0.1778 0.0661 -0.0950 -0.0253 1.0000

(21)

21

Based on the expected signs of tangibility and growth opportunies this implies a potential decrease and increase in leveral level respectively. Finally, the results show no statistical significant for a change in liquidity and profitability following the passage of the German Governance Code.

Table 6 . Univariate results of sample one

Debttocap MB LIQ PROFIT SIZE TANG

Pre 0.3906 2.4818 1.5631 0.1241 15.837 0.2954

Post 0.3357 1.6679 1.6021 0.1208 16.0022 0.2748

Diff -0.0549* -0.8139*** 0.039 -0.0033 0.1656*** -0.0206**

*p-value < 0.10, ** p-value < 0.05, *** p-value < 0.01

For the first sample five regressions are run, all containing a different measure of leverage as the dependent variable. The first model represents the regression with the main dependent variable; the debt-to-equity ratio. Table 7 shows the results from the five regression controlled for firm fixed effects. For all regressions, robust standard errors are used to control for possible heteroskedasticity and adjusted for 40 firm clusters. To test the significance of the model an F-test is used. The F-statistic, F(6,39), provided by the first regression has a critical value of 3.305 at a significance level of 0.01. The value of the F-statistic for this regression is 5.43. It can be concluded that at a 1% significance level the explanatory variables used in the model have a significant relation with the leverage level.

To indicate the degree to which the dependent variable is explained by the first regression model the R-squared value is used. The firm fixed effects regression conducts three R-squared values; the within, between and overall R-squared. In this study, the within R-squared is the value of interest. With this value, the goodness of fit measure of the individual data is provided and all the between information in the data is ignored. The first regression has a within R-squared value of 0.2478, which can be interpreted as a 24.78% explanatory power of the model.

The intraclass correlation, rho, is also provided by the fixed effect regression.6 𝜌 = _(𝜎 (𝜎!)!

!)!+ (𝜎!)!

In this regression 𝜌 is 0.8856, which means that 88.56% of the variance can be explained by the differences across the firms.

6_{Where 𝜎}

! is the standard deviation of residuals within groups 𝑢! and 𝜎! is the standard deviation of residuals 𝑒!

(22)

22

In the first firm fixed effects model the variables PROFIT and SIZE are significant at a 1% significance level. In addition, the variable TANG and the dummy POST are significant at a 5% level. The variable PROFIT has a significant negative relation with leverage level measured as the debt-to-capital ratio. When PROFIT increases, the leverage level declines. This result aligns with the predicted coefficient sign of profit in relation to leverage based on theory.

According to these results, the firm size seems to have a statistically significant positive effect on the leverage level. When the size of a firm increases, so does the debt-to-capital ratio, which is the expected relation. This is the same for the variable TANG, however this relation is statistically less strong. The variables MB and LIQ do not show any statistical significance; hence the expected relation cannot be confirmed.

This regression provides statistical evidence for a negative relation between leverage level, measured as debt-to-capital ratio, and the dummy variable POST with a significance level of 5%. Hence, the null hypothesis that the German Governance Code had no significant influence on the leverage level of firms can be rejected.

Table 7. Regression coefficients of sample one with corresponding robust standard errors in parentheses. The dependent variables are the debt-to-equity ratio, book leverage, market leverage, long term debt-to-capital ratio and the long term debt-to-assets ratio in regresstion 1,2,3,4 and 5 respectively.

1 2 3 4 5 Constant -2.136** (0.889) -1.210*** (0.446) 1.367*** (0.383) -1.948 (0.747) -1.101** (0.417) POST -0.7831** (0.024) -0.035** (0.014) -0.001 (0.004) -0.022 (0.002) -0.005 (0.011) MB -0.004 (0.009) -0.003 (0.006) 0.006 (0.004) -0502 (0.664) -0.003 (0.005) LIQ -2.902 (1.988) -0.021 (0.013) -0.019 (0.153) 2.766 (1.787) 0.019 (0.012) PROFIT -0.581*** (0.194) -0.491)*** (0.140) -0.208 (0.145) -0.331** (0.158) -0.304*** (0.104) SIZE 0.160*** (0.052) 0.094*** (0.027) -0.022 (0.022) 0.131*** (0.043) 0.076*** (0.025) TANG 0.381** (0.282) 0.154 (0.160) 0.049 (0.056) 0.359* (0.224) 0.162 (0.136) F-stat 5.43*** 4.84*** 0.70 3.41*** 4.03*** 𝑅! _0.2478 _0.2257 _0.1110 _0.1159 _0.1481 𝜌 0.8856 0.8964 0.7519 0.8718 0.8715

4.1.1. Bank debt

To find an answer to the question if the passage of the German Corporate Governance Code had an influence on the level of bank debt of firms the same explanatory variables are modeled, against the dependent variable bank debt.

(23)

23

Table 8 shows univariate results from the second sample. The mean of the variables Bankdebt1, MB, LIQ, PROFIT, SIZE and TANG of sample two for the period pre 2002 and post 2002 are reported in the table. The results show a significant difference in the proportion of bank debt after the passage of the German Governance Code. This implies a decrease of the proportion of bank debt German firms hold after 2002. The difference of the explanatory variables show the same significance and sign as in the first sample discussed in the previous section.

Table 8. Univariate results of sample two

Bankdebt1 MB LIQ PROFIT SIZE TANG

Pre 0.1913 2.2392 1.4988 0.1199 15.9376 0.3007

Post 0.1232 1.5648 1.5951 0.1168 16.0960 0.2778

Diff -0.0681** -0.6744*** 0.0963 -0.0031 0.1584*** -0.0229**

The critical F-value for this model is 3.43 at 1% significance level and 2.40 at a 5% significance level. The value of the F-statistic provided by the first regression is 2.18. Hence, it cannot be assumed that the independent variables used in the model are jointly significant in relation to the dependent variable bank debt. Also, the within R-squared of this model is relatively low with 11,89% explanatory power of the model. The 𝜌 of this model is 0.8743, which is similar to the first model.

The estimated regression coefficients for sample two are summarised in table 9. The second regression provides a robustness check and will be discussed in the next section. The only significant variable at a 5% significance level is SIZE. The statistical evidence for this variable implies a negative relation between firm size and the proportion of bank debt that German firms hold. Based on theory the expected relation between firm size and leverage is positive. Based on the obtained results an increase in firm size implies a increase in leverage measured as the debt-to-capital ratio and a decrease in the proportion of bank debt the German firms hold.

The estimated sign for the dummy variable POST is negative, however insignificant. Therefore we cannot reject the null hypothesis and no conclusion can be drawn from this regression about the influence of the German Corporate Governance Code on the proportion of bank debt.

(24)

24 Table 9. Regression coefficients for sample two with corresponding robust standard errors in parentheses. The dependent variables are bank loans over assets and bank loans over capital in regression 1 and 2 respectively.

1 2 Constant 3.691** (1.508) 5.77*** (1.861) POST -0.043 (0.033) -0.067 (0.055) MB -0.014 (0.012) -0.024 (0.017) LIQ -0.018 (0.025) -0.035 (0.036) PROFIT -0.414 (0.506) -0.599 (0.715) SIZE -0.213** (0.889) -0.333*** (0.109) TANG 0.003 (0.325) 0.049 (0.439) F-stat 2.18* 2.51** R! _0.1189 _0.1264 ρ 0.8743 0.8694

* p-value < 0.10, ** p-value < 0.05, *** p-value < 0.01

4.2 Robustness

To check for robustness each dependent variable is regressed for both samples separately on the group of variables defined in the measurement model. As mentioned before, in the first model leverage can be measured by the debt-to-capital ratio, book leverage, market leverage, the long-term debt-to-capital ratio and the long-term debt-to-assets ratio. Regression 1 and 2 provide a significant outcome at a 1% and 5% level respectively for a negative relation between leverage and the passage of the German Governance Code. However, the robustness check shows how fragile these results are. To outline the fragility of the results the p-values of the variable POST of the 5 regressions are summarised in table 10. The p-values show that the significance deviates from highly significant to highly insignificant. Further research is needed to explain the difference in significance between the regressions.

Table 10 also includes the p-values for the variable POST from the second model. Two regressions are run with both a different measure of the proportion of bank debt. As discussed in the previous section the variable POST has no statistical significance in relation to the proportion of bank debt that the German firms hold. The results of the second regression, which serves as a robustness check, align with the first findings. The fact that it was not mandatory for German firms to comply with the rules provided by the German Governance Code might be an explanation for the insignificant results.

(25)

25 Table 10. P-values of the regressions from the first and the second model. POST1 comes from the first model and POST2 from the second.

1 2 3 4 5

POST1 0.003*** 0.015** 0.784 0.252 0.678

POST2 0.209 0.233

5 Conclusion

This section summarises the main findings of this research and provides suggestions for further research.

5.1 Summary and concluding remarks

Some statistical significant results are found to distinguish a relationship between leverage level and the passage of the German Governance Code. When leverage is defined as the debt-to-capital ratio or as book leverage significant results are obtained that provide evidence for this negative relation. However, with market leverage, long-term debt-to-assets ratio and the long-term debt-to-capital ratio as leverage measures the results were insignificant. The model used to estimate the proportion of bank debt of the German firms used in the sample provides no significant results in relation to the passage of the German Governance Code in 2002. The results contribute to the understanding of the relationship between corporate governance and leverage level; however the question if a corporate governance reform has an influence on leverage level cannot be answered based on the performed analysis.

5.2 Further research

As the timeframe of the data was only 2 years before and 4 years after the passage of the German Governance Code the data of the variables was slightly limited. The use of more observations might provide better insights into the influence of the corporate governance reform on leverage level. The difference in outcome when other leverage levels are used as dependent variables requires further research. Complying with the rules provided by the German Governance Code was not mandatory for the German firms. This might be a cause of the insignificant results. Hence, to limit this threat the firms that declare not to comply with The Code could be excluded from the sample in further research.

(26)

26

In this research a fixed effects model is used based on the belief that sufficient explanatory variables are included in the model and that the unobservable factors are time-invariant. However, this might not be the case and if the unobservable factors are not time-invariant there might still be an omitted variable bias. Therefore, another suggestion for further research is to include other variables that could influence the leverage level of firms and have a likelihood of being time-variant into the model. The interest rate possibly is such a variable. It often chances over time and can be seen as the price of debt. Hence, the interest rate might influence the leverage level that firms hold when it changes. Therefore, future research might take the interest rate into consideration.

The fixed effects model also does not control for the effect of omitted time-invariant variables that have time-varying effects. For example, to estimate the time-varying effects of more or less stable firm-characteristics, interactions with time could be included in the models of further research.

(27)

27

6 Appendix

6.1 Table appendix

Appendix.1 Firms used in the samples

DAX MDAX

BASF SE Infineon Airbus Group SE Jungheinrich

Bayer AG Linde AG Aurubis AG Krones

Beiersdorf AG* Lufthansa AG Axel Springer SE* Leoni*

BMW AG Merck KGaA Bilfinger SE Metro AG

Continetal AG RWE ST DMG Mori Seiki AG* Rheinmetall

Daimler AG SAP AG Duerr AG Salzgitter AG

Deutsche Telekom AG Siemens AG Fielmann Stada Arzeimittel

E.ON AG ThyssenKrupp AG* Fuchs Petrolub SE K & S AG

Fresenius Medial Care & Co KGaA*

Volkswagen GEA Group Suedzucker

Fresenius SE* ProSiebenSat.1Media Hochtief HeidelbergCement

*Firms excluded from second sample

Appendix 2. Dependent variables used for robustness check

Variable Description

Debttocap The debt –to-capital ratio is defined as the total debt

divided by the book value of equity plus the book value of total debt.

Source: Datastream Worldscope

Booklev Book leverage is defined as the total debt divided by

the total of assets.

Mrktlev Market leverage is computed by dividing the total debt

market value of equity plus the book value of total debt.

Ltcap The long-term debt-to-capital ratio is computed by

dividing the long-term debt by the book value of equity plus the book value of total debt.

Ltass Long-term debt-to-assets is defined by the long-term

debt over total assets.

Bank1 Bank debt measured as the total of bank loans divided

by total assets Source: Orbis

Bank2 Bank debt measured as the total of bank loans divided

by capital Source: Orbis

(28)

28 Appendix 3. Independent variables included in the models

Variable Description

SIZE Firm size is measured as the logarithm of total assets.

MB Growth opportunities for the firms are measured as the

market to book ratio.

Source: Datastream Worldscope

LIQ Liquidity is measured as the current ratio: current

assets divided by current liabilities. Source: Datastream Worldscope

TANG Tangibility is measured by the net property, plant and

equipment divided by the total of assets. Source: Datastream Worldscope

PROFIT Profitability is measured by the earnings before

interest, taxes, depreciation and amortization divided by the total of assets.

Source: Datastream Worldscope

6.2 References

Akkermans, D., Van Ees, H., Hermes, N., Hooghiemstra, R., Van der Laan, G., Postma, T., & Van Witteloostuijn, A. (2007). Corporate Governance in the Netherlands: an overview of the application of the Tabaksblat Code in 2004. Corporate Governance: An International Review, 15(6), 1106-1118. Agarwal, R., & Elston, J. A. (2001). Bank–firm relationships, financing and firm performance in Germany.

Economics Letters, 72(2), 225-232.

Aggarwal, R., Erel, I., Stulz, R., & Williamson, R. (2010). Differences in governance practices between US and foreign firms: Measurement, causes, and consequences.Review of Financial Studies,23(3), 3131-3169 Aguilera, R. V., & Jackson, G. (2003). The cross-national diversity of corporate governance: Dimensions and

determinants. Academy of management Review, 28(3), 447-465.

Antoniou, A., Guney, Y., & Paudyal, K. (2008). The determinants of capital structure: capital market-oriented versus bank-market-oriented institutions. Journal of financial and quantitative analysis, 43(01), 59-92.

Armour, J., & McCahery, J. A. (Eds.). (2006).After Enron: improving corporate law and modernising securities regulation in Europe and the US. Bloomsbury Publishing.

Arping, S., & Sautner, Z. (2010). Corporate governance and leverage: Evidence from a natural experiment. Finance Research Letters, 7(2), 127-134.

Beck, T., Demirgüç-Kunt, A., Laeven, L., & Maksimovic, V. (2006). The determinants of financing obstacles. Journal of International Money and Finance, 25(6), 932-952.

Berger, A. N., & Di Patti, E. B. (2006). Capital structure and firm performance: A new approach to testing agency theory and an application to the banking industry.Journal of Banking & Finance,30(4), 1065-1102.

Byoun, S. (2011). Financial flexibility and capital structure decision. Available at SSRN 1108850. Code, G. C. G. (2002). German Corporate Governance Code. Berlin, Government Commission.

Cromme, G. (2005). Corporate governance in Germany and the German corporate governance code. Corporate Governance: An International Review, 13(3), 362-367.

(29)

29 Fama, E. F., & Miller, M. H. (1972). The theory of finance (Vol. 3). Hinsdale, IL: Dryden Press.

Flannery, M. J., & Rangan, K. P. (2006). Partial adjustment toward target capital structures. Journal of financial economics, 79(3), 469-506.

Frank, E., & Daniel, F. (1991). The economic structure of corporate law.

Gerschenkron, A. (1962). Economic backwardness in historical perspective: a book of essays (No.HC335 G386). Cambridge, MA: Belknap Press of Harvard University Press

Grossman, S. J., & Hart, O. D. (1982). Corporate financial structure and managerial incentives. InThe economics of information and uncertainty(pp. 107-140). University of Chicago Press.

Harford, J., Mansi, S. A., & Maxwell, W. F. (2012). Corporate governance and firm cash holdings in the US. InCorporate Governance(pp. 107-138). Springer Berlin Heidelberg.

Hoshi, Takeo, Anil Kashyap and Gary Loveman (1994) “Lessons from the Japanese Main Bank System for Financial Reform in Poland” in Masahiko Aoki and Hugh Patric eds.: The Japanese Main Bank System Oxford U.K.: Oxford University Press

Jensen, M. C., & Meckling, W. H. (1976). Theory of the firm: Managerial behavior, agency costs and ownership structure.Journal of financial economics,3(4), 305-360.

Jensen, M. C. (1986). Agency cost of free cash flow, corporate finance, and takeovers.Corporate Finance, and Takeovers. American Economic Review,76(2).

Jensen, M. C. (1993). The modern industrial revolution, exit, and the failure of internal control systems.the Journal of Finance,48(3), 831-880.

Jensen, M. C. (1997). Eclipse of the public corporation.Harvard Business Review (Sept.-Oct. 1989), revised. John, K., & Senbet, L. W. (1998). Corporate governance and board effectiveness.Journal of Banking &

Finance,22(4), 371-403.

Jiraporn, P., & Gleason, K. C. (2007). Capital structure, shareholder rights, and corporate governance.Journal of Financial Research,30(1), 21-33.

Jiraporn, P., Kim, J. C., & Kim, Y. S. (2011). Dividend payouts and corporate governance quality: An empirical investigation. Financial Review, 46(2), 251-279.

Kaiser, K. (1994). Corporate restructuring and financial distress: an international view of bankruptcy laws and implications for corporations facing financial distress. INSEAD, mimeo.

Kaplan, S. N. (1997). Corporate governance and corporate performance: A comparison of Germany, Japan, and the US.Journal of Applied Corporate Finance,9(4), 86-93. Franks, J., & Mayer, C. (1994). Corporate Control: A Comparison of insider and outsider systems.London Business School.

La Porta, R., Lopez‐de‐Silanes, F., Shleifer, A., & Vishny, R. W. (2000). Agency problems and dividend policies around the world.The journal of finance,55(1), 1-33.

Litvak, K. (2007). The effect of the Sarbanes-Oxley Act on non-US companies cross-listed in the US. Journal of corporate Finance, 13(2), 195-228.

Myers, S. C. (2001). Capital structure. The journal of economic perspectives,15(2), 81-102.

Parsons, C. A., & Titman, S. (2007). Capital structure and corporate strategy.Available at SSRN 983553. Parsons, C., & Titman, S. (2009). Empirical capital structure: A review. Now Publishers Inc.

(30)

30 Rajan, R. G., & Zingales, L. (1995). What do we know about capital structure? Some evidence from

international data. The journal of Finance, 50(5), 1421-1460.

Romano, R. (1993).The genius of American corporate law. American Enterprise Institute.

Shleifer, A., & Vishny, R. W. (1997). A survey of corporate governance.The journal of finance,52(2), 737-783.

Stock, J. H., & Watson , M. M..’’ Introduction to Econometrics’’. Pearson Education Limited: Harlow, England (2012).

Romano, R. (2004). The Sarbanes-Oxley Act and the making of quack corporate governance.

Torres-Reyna, O. (2007). Panel data analysis fixed and random effects using Stata (v. 4.2). Data & Statistical Services, Princeton University.

Werder, A., Talaulicar, T., & Kolat, G. L. (2005). Compliance with the German corporate governance code: An empirical analysis of the compliance statements by German listed companies. Corporate Governance: An International Review, 13(2), 178-187.

White, M. J. (1993). The Costs of Corporate Bankruptcy: A US. In European Comparison.” University of Michigan Research Forum on International Economics, Discussion Paper 346.

Zwiebel, J. (1996). Dynamic capital structure under managerial entrenchment.The American EconomicReview, 1197-1215.

Influence of a corporate governance reform on the leverage level : German Governance Code case study